Running title: Natural mortality of the diamondback moth

Abstract

The impact of natural enemies on immature stages of the diamondback moth was quantitatively assessed in commercial cabbage crops in southeast Queensland, Australia, in 2000 and 2001. The survivorship of P. xylostella eggs, larvae and pupae on cabbage plants caged to exclude predators and larval and pupal parasitoids was compared with survivorship on cabbage plants which were caged to simulate the ambient conditions created by the exclusion cages, but which allowed natural enemy access. In 2000, six cohorts of P. xylostella were studied on three commercial Brassica farms which adopted an integrated approach to pest management and six cohorts were studied on three farms which adopted a calendar spray approach to pest management; a further two cohorts were followed in cabbage plots at Gatton Research Station. In 2001, two cohorts were followed on a farm adopting IPM and two cohorts were followed on a farm practising calendar spraying. Estimated losses due to predation ranged from 2-85% in 2000 and from 22-77% in 2001. When losses and mortality due to parasitism were combined, total estimated losses ranged from 2-98% in 2000 and from 22-90% in 2001. Larval and pupal parasitism rates were low in both years of the study but, in order of abundance, the hymenopteran parasitoids detected could be ranked Diadegma semiclausum > Diadromus collaris > Apanteles ippeus > Oomyzus sokolowskii. This study represents the first record of O. sokolowskii attacking P. xylostella in Australia. Pitfall trapping indicated that Araneae (Lycosidae and Oxyopidae) were the most abundant insectivorous predators but that Coleoptera (Carabidae, Staphylinidae and Coccinellidae) and Hemiptera were also relatively abundant on commercial Brassica farms in southeast Queensland.

Key words

Diamondback moth, predators, parasitoids, exclusion.

Introduction

Naturally occurring arthropod predators and parasitoids are known to significantly affect populations of pest Lepidoptera in various Brassica crops (Jones, 1987; Schmaedick and Shelton, 1999). In order to maximise the contribution which natural enemies can make to integrated pest management programs it is imperative that their potential value is recognised and that the factors underpinning their activity are elucidated. Although the parasitoid fauna associated with the diamondback moth (Plutella xylostella L.) is well documented (Talekar and Griggs, 1986; Talekar, 1992; Sivapragasam et al., 1997), quantitative assessment of its impact, and that of generalist predators, on populations of the pest is much less well understood. In order to assess the impact of predators as well as parasitoids, an exclusion cage approach was chosen over the more simple life table studies which identify agents which cause mortality, but do not allow its quantification (Luck et al., 1988). This study summarises an area wide approach that was taken towards understanding the importance of arthropod parasitoids and predators of the diamondback moth on commercial farms in the Lockyer Valley, southeast Queensland, Australia.

Materials and Methods

Experimental sites.

Following a series of interviews and farm visits conducted in the Lockyer Valley in February 2000, six commercial Brassica farms were identified as suitable experimental sites. Consideration of the types of insecticide used, the frequency of application, scouting procedures, initial planting and final harvest dates were used to broadly categorise farm management practice. Three farms that were considered to adopt an integrated approach to insect pest management (“IPM” sites) and three farms that were considered to practise an approach tending towards calendar insecticide application (“CAL” sites) were selected. Although it was not possible to select adjacent farms practising different methods of insect pest management for inclusion in the study, suitable sites were selected on the basis of location as well as proposed management practice. The first experimental site (IPM1) was located in the southeast of the Lockyer Valley and the second site (CAL1) was located in the north of the valley, thereafter experimental sites were selected so that site categories alternated with each other (Figure 1). The second set of experiments was performed in the same order as the first set, these experiments are denoted by the suffix “a” in site labels in all subsequent tables and figures. Experiments performed in 2001 are followed by the suffix “(01)”.

Figure 1. The location of experimental sites within the Lockyer Valley, Queensland.

Plants and insects.

Cabbage was chosen as the study crop as it is grown over a wide area and a long period (multiple sequential plantings February-August) and it is relatively easy to sample. Three cabbage varieties, “Warrior”, “Neptune” and “Sugarloaf” were grown on the experimental sites. The variety to be grown at a particular site was determined by consultation with the relevant grower approximately 1 week prior to field transplantation. One-two weeks after field transplantation, 50 seedlings of that variety were transplanted in 20 cm diameter pots containing a 1:1 mixture of Gatton Research Station soil and potting mix and grown in a glasshouse for two weeks.

Plants were exposed to diamondback moth for oviposition by placing 40 potted seedlings on the floor of a nylon net cage (2 m x 1.5 m x 0.4 m) in a controlled environment room (22 ±2°C; 12: 12 (L: D) h; 60% RH) and releasing approximately 1500 recently emerged (1-2 days post eclosion) and mated diamondback moth adults into the cage. After 12-18h exposure, plants were removed from the cage and examined. Excess eggs were removed from leaves so that all plants supported 25 eggs that were distributed as evenly as possible between the plant's six leaves. The location of each egg on each plant was marked with a permanent marker. Eggs that were laid on the plant pot were removed by washing the pot with a scourer and warm soapy water and eggs that were laid on the soil surface were destroyed by turning over the soil with a fork.

Exclusion of predators and larval and pupal parasitoids

Survivorship of P. xylostella on plants caged to exclude predators and larval and pupal parasitoids was compared with survivorship on plants caged to allow total or restricted predator access. All cages consisted of a central cylindrical frame (45 cm high x 45 cm diameter) made from wire (2 cm diameter mesh). Modifications to the fine nylon-netting sleeve (mesh 0.5 mm) covering the central frame allowed construction of cages to effect:

a) Total natural enemy exclusion. The nylon mesh sleeve was buried under the plant pot and completely covered the central wire frame.

b) Total natural enemy access. The cage consisted of only the central wire frame.

c) Natural enemy access and ambient environmental conditions similar to those within total exclusion cages. One half of the central wire frame and the top of the cage were covered with the nylon mesh.

d) Natural enemy access and ambient environmental conditions similar to those within total exclusion cages, but ground predators excluded. As in c) above but the top 4 cm of the plant pot was treated with Tanglefoot.

The bottom of all cages, except those providing total exclusion of natural enemies, was buried approximately 10 cm into the ground and a single egg-laden potted plant was placed into each cage. In all treatments, the pots were buried so that the top of the pot, or the bottom of the sticky barrier, was flush with the soil and the cages anchored to the ground by 3 bamboo canes. The tops of the nylon sleeves of the total exclusion cages were tied and the cages fixed in place by 4 bamboo canes placed around the edge and fastened with string (Figure 2).

At each field site, cages were arranged in four blocks within two adjacent beds of cabbage plants; the two blocks in the same bed were placed 20 m apart. Cages within a block were spaced 5 m apart and their position within the block randomly assigned. Each treatment was replicated twice within a block to give a total of 8 replicates per treatment.

On the day that the experimental plants were transferred to the field, 30 plants within the field were randomly selected and sampled for diamondback moth, other cabbage pests and natural enemies. Sites were visited regularly during the course of the experiments (2-3 times each week); all experimental plants were examined and field-laid P. xylostella eggs removed. Regular liaison with growers determined when insecticides were to be applied to the crops and all experimental cages were covered with large garbage bags before insecticide application and removed approximately 1 h later. The experiment was terminated when the majority of insects on the plants in the cages had reached the pupal stage.

Plants were cut at their base and transferred individually to labelled plastic bags. Any diamondback moth larvae or pupae on the cage or rim of the pot were transferred to the same bag as the plant. In the laboratory, individual plants were carefully examined and all larvae and pupae recorded and then reared (22 ±2^oC; 12: 12 (L: D) h; 60% RH) until moths or parasitoids emerged or it was certain that individuals had died.

In 2000, the experiment was performed twice within cabbage crops at each field site between March and August and in experimental plots at Gatton Research Station from August-September. In 2001, the experiment was performed twice within cabbage crops at two of the field sites between June and August (Figure 1; Table 1). Insecticide input at each field site prior to and during each experiment was recorded (Table 1).

Table 1: Insecticide input at experimental sites prior to and during natural enemy impact assessment trials.

Insecticides applied2 Site1YearDuration1

Bt	S’sad	P’zole	E’ctin	P’roid	OP	OC	C’ate	Total
IPM1	2000	15/3 - 30/3	3	-	-	-	-	-	-	-	3
CAL1		8/4 - 28/4	2	-	-	1	3	1	-	1	8
IPM2		19/4 - 15/5	3	-	-	-	-	1	1	-	5
CAL2		8/5 - 4/6	1	-	4	-	-	-	-	-	5
IPM3		15/5 - 18/6	6	-	1	1	-	-	-	-	8
CAL3		1/6 - 11/7	4	-	4	-	-	-	-	-	8
IPM1a		8/6 - 16/7	4	-	-	-	-	-	-	-	4
CAL1a		18/6 - 22/7	6	-	1	1	5	-	-	-	13
IPM2a		8/7 - 9/8	1	-	-	-	-	-	1	-	2
CAL2a		21/7 - 25/8	1	1	2	-	-	-	-	-	4
IPM3a		23/8 - 15/9	6	-	1	-	-	-	-	-	7
GRSA		27/8 - 18/9	-	-	-	-	-	-	-	-	0
GRSB		27/8 - 18/9	-	-	3	-	-	1	-	-	4
IPM2	2001	19/6 - 24/7	3	-	-	-	-	1	1	-	5
CAL1		23/6 - 24/7	2	-	-	1	3	1	-	1	8
IPM2a		28/7 - 28/8	1	-	-	-	-	-	1	-	2

¹The date on which the experiment commenced to the date on which P. xylostella were recovered from the cages in the field.

² Abbreviations for insecticides applied during the course of the experiments; Bt= Bacillus thuringiensis, S’sad= Spinosad, P’zole= Pyrazoles, E’ctin= Emamectin, P’roid= Pyrethroids, OP= Organophosphates, OC= Organochlorines, C’ate= Carbamates.

Sampling for ground dwelling predators

In the third week of each experiment set up after June 7, 2000 (Sites IPM1a, IPM2a, IPM3a, CAL1a, CAL2a, GRSA and GRSB) and in all experiments performed in 2001, pitfall traps were used to sample ground dwelling predators within the immediate vicinity of the exclusion cages. Traps were constructed by burying plastic cups (200 ml; 6.5 cm diameter) so that soil was level with the rim, adding approximately 100 ml of detergent solution (1% (vol./ vol.)) to each and then covering each cup with plastic disc (15 cm diameter) supported 3 cm above ground level by 3 steel nails (10 cm). Sixteen pitfall traps were placed within each experimental site and each trap was positioned a minimum distance of 10 m from its nearest neighbour. Traps remained in the field for 7 days before they were collected and returned to the laboratory for analysis.

Statistical Analysis

The number of P. xylostella recovered from each of the cage treatments at each of the experimental sites was compared by ANOVA. Similarly, the number of parasitised P. xylostella recovered from each of the cage treatments at each of the experimental sites was also compared by ANOVA. In order to estimate the impact of predators on each P. xylostella cohort at each of the experimental sites, the number of pupae and final instar larvae recovered from the open and half cage treatments was expressed as a proportion of the original number of eggs on each plant and corrected by the proportion of individuals recovered from the total exclusion treatments. To increase the reliability of corrected estimates of predation and to ensure that the confidence intervals calculated were centred around the appropriate mean, a modified version of Abbott’s formula (Abbott, 1925) incorporating an estimate of the variance of the control mean (total exclusion recovery rate) was used (Rosenheim and Hoy, 1989).

Within years, total predator numbers sampled at each experimental site were compared by ANOVA.

Results

Exclusion of predators and larval and pupal parasitoids

Application of insecticide to the cabbage crop before experimental cages could be covered rendered data from sites CAL1 and CAL3a useless (Table 2). Recovery of P. xylostella varied enormously between the remaining experimental sites (Table 2). Initial analysis of variance showed that Tanglefoot around the top of the plant pot in the nylon mesh covered cages to which predators and parasitoids had access had no significant impact on the number of P. xylostella recovered (LSD P> 0.05) at any of the experimental sites. These data were combined for further analysis and the Tanglefoot treatment was subsequently excluded from experiments performed at Gatton Research Station in 2000 and all the experiments performed in 2001.

Table 2: The recovery and parasitism rates of P. xylostella from each of the cage treatments at each of the field sites

	Mean No. P. xylostella recovered (±SE)						Mean No. P. xylostella parasitised (±SE) 2
Field Site	Total exclusion	Half cage	Open cage	F1	P1	LSD0.051	Total exclusion	Half cage	Open cage
IPM1	21.4 (±4.4)	21.8 (±2.5)	7.9 (±1.8)	6.15	0.006	4.75	0	2.4 (±1.5)	2.1 (±1.3)
CAL1	-	-	-	-	-	-	-	-	-
IPM2	19.1 (±2.2)	19.8 (±1.4)	12.9 (±2.3)	4.27	0.024	2.83	0	0.4 (±0.3)	1.9 (±0.9)
CAL2	12.0 (±1.4)	11.8 (±1.0)	8.0 (±1.0)	2.87	0.073	1.98	0	0	0
IPM3	9.8 (±1.9)	7.7 (±1.2)	8.9 (±1.5)	0.57	0.574	2.30	0	0	0
CAL3	10.6 (±1.9)	5.3 (±1.0)	2.6 (±0.8)	9.38	<0.001	1.89	0	0	0
IPM1a	22.1 (±1.8)	13.3 (±1.7)	12.9 (±2.6)	4.52	0.020	3.60	0	2.3 (±1.0)	2.8 (±0.8)
CAL1a	20.8 (±1.4)	20.4 (±1.3)	13.6 (±1.4)	4.93	0.015	2.87	0	0.1 (±0.1)	0
IPM2a	21.3 (±2.3)	12.9 (±1.3)	10.6 (±1.9)	9.23	<0.001	2.67	0	3.1 (±1.0)	3.8 (±1.3)
CAL2a	20.1 (±2.1)	11.3 (±1.2)	7.4 (±1.4)	13.91	<0.001	2.48	0	4.5 (±1.2)	3.9 (±1.5)
IPM3a	16.3 (±1.8)	12.4 (±1.3)	11.3 (±2.2)	1.82	0.181	2.77	0	0.1 (±0.1)	0.3 (±0.2)
CAL3a	-	-	-	-	-	-	-	-	-
GRSA	23.8 (±2.7)	8.4 (±0.8)	4.1 (±0.8)	37.36	<0.001	5.02	0	3.5 (±1.4)	3.6 (±1.1)
GRSB	21.3 (±2.8)	10.0 (±1.7)	3.4 (±0.7)	21.27	<0.001	5.83	0	2.9 (±0.9)	2.8 (±1.1)
IPM2(01)	17.3 (±1.9)	8.6 (±1.7)	4.0 (±0.6)	16.66	<0.001	4.80	0	3.6 (±1.2)	2.3 (±0.9)
CAL2(01)	7.4 (±0.9)	5.8 (±2.2)	3.4 (±0.6)	1.76	0.196	4.45	0	0	0
IPM2a(01)	19.0 (±1.5)	7.9 (±1.6)	7.3 (±1.3)	17.89	<0.001	5.60	0	6.6 (±1.9)	5.4 (±1.2)

¹The number of P. xylostella recovered from each type of cage was analysed by ANOVA. For field sites IPM1, CAL1, IPM2, CAL2, IPM3, CAL3, IPM1a, CAL1a, IPM2a, CAL2a, IPM3a degrees of freedom =2, 30 and for field sites GRSA, GRSB, IPM2(01), CAL2(01) and PM2a(01) degrees of freedom = 2, 22.

² The number of P. xylostella recovered from each type of cage subsequently found to be parasitised. Within an experiment, there was no significant difference between the numbers of parasitised P. xylostella recovered from either type of cage to which parasitoids had access (LSD, P>0.05). Parasitoids found attacking P. xylostella larvae or pupae were Diadegma semiclausum, Diadromus collaris, Apanteles ippeus and Oomyzus sokolowskii (Table 3).

Figure 2. Design of the various field cages used in the experiments to permit natural enemy access or to effect total natural enemy exclusion.

The different cage designs significantly affected the number of P. xylostella recovered at all but four (CAL2, IPM3, IPM3a and CAL2(01)) of the experimental sites (Table 2). At all remaining sites the number of P. xylostella recovered from completely open cages was significantly lower than the number recovered from cages that totally excluded natural enemies (LSD, P<0.05; Table 2). At sites IPM1, IPM2 and CAL1a there was no significant difference between the number of P. xylostella recovered from cages which totally excluded all natural enemies and cages which were partially covered with nylon netting but which permitted natural enemy access (LSD, P>0.05). At all other experimental sites (CAL3, IPM1a, CAL1a, IPM2a, CAL2a, IPM3a, GRSA, GRSB, IPM2(01) and IPM2a(01)) the number of P. xylostella recovered from cages which totally excluded all natural enemies was significantly greater than the number recovered from cages which were partially covered with nylon netting, but which permitted natural enemy access (LSD, P<0.05; Table 2).

Table 3: Species of hymenopteran parasitoids attacking immature stages of P. xylostella at field sites in the Lockyer Valley, March-September 2000 and June-August 2001

		Total number of P. xylostella parasitised by each parasitoid at each site
Field site1	Year	D. semiclausum	D. collaris	A. ippeus	O. sokolowskii
IPM1	2000	2	34	0	17
CAL1		-	-	-	*
IPM2		1	20	0	0
CAL2		0	0	0	0
IPM3		0	0	0	0
CAL3		0	0	0	0
IPM1a		47	11	0	1
CAL1a		1	0	0	0
IPM2a		77	1	3	0
CAL2a		103	0	0	0
IPM3a		2	0	0	0
GRSA		30	0	27	1
GRSB		26	0	18	0
IPM2	2001	47	0	0	0
CAL1		0	0	0	0
IPM2a		93	0	0	0

¹ Quantitative assessment of data from field sites CAL1and CAL3a was impossible. However, some larvae recovered from CAL1 were parasitised by O. sokolowskii.

No parasitised larvae were recovered from cages that totally excluded P. xylostella natural enemies (Table 2). There was no significant difference between the overall parasitism rates in larvae recovered from cages which were partially covered with nylon netting but which permitted natural enemy access and completely open cages at any of the experimental sites (P>0.05; Table 2). The major parasitoids attacking the immature stages of P. xylostella were Diadegma semiclausum, Diadromus collaris, Apanteles ippeus and Oomyzus sokolowskii and their relative abundance varied both between and within experimental sites over time (Table 3). In 2000, the rates of parasitism were generally low and no parasitism was detected in experiments that were initiated between the first week of May and the first week of June. Rates of parasitism increased in experiments performed subsequently, but never exceeded 17% of available larvae at a given experimental site (Table 2). In 2001, only parasitism by D. semiclausum was detected and, although parasitism rates were slightly higher than in 2000, they never exceeded 26.5% of available larvae (Table 2).

Figure 3. The corrected proportion of P. xylostella “missing” from totally open cages and cages which were partially covered with nylon netting, but which permitted natural enemy access.

Figure 4. The corrected proportion of P. xylostella “missing” or parasitised in totally open cages and cages which were partially covered with nylon netting, but which permitted natural enemy access.

When P. xylostella recovery rates from cages that were partially covered with nylon netting but which permitted natural enemy access and completely open cages were corrected by recovery rates from cages from which natural enemies were totally excluded (Rosenheim and Hoy, 1989) estimated rates of predation ranged from 2-85% in 2000 and from 22-77% in 2001 (Figure 3). Similar statistical treatment of data that incorporated P. xylostella mortality due to parasitism into overall recovery data estimated that the overall combined mortality due to predation and parasitism ranged from 2-98% in 2000 and from 22-90% in 2001 (Figure 4).

Sampling for ground dwelling predators

The number of potential predators of P. xylostella caught in pitfall traps varied significantly between experimental sites in both 2000 (F_6,97= 7.8, P< 0.001; Figure 5a) and 2001 (F_2,41= 12.2, P< 0.001; Figure 5b). In 2000 the total numbers of potential predators caught at CAL2a, IPM3a and GRSB were significantly lower than those caught at the remaining sites (LSD, P<0.05). In 2001 the total number of potential predators caught at IPM2a was significantly greater than at the other two sites (LSD, P<0.05).

Of the potential P. xylostella predators caught in pitfall traps, spiders (Lycosidae and Oxyopidae) were the most abundant, followed by Coleoptera (Carabidae, Staphylinidae and Coccinellidae) and Hemiptera.

Figure 5a. Predaceous arthropods sampled at experimental sites in 2000.
5b. Predaceous arthropods sampled at experimental sites in 2001.

Discussion

The great variability in the rates of recovery of insects throughout the series of experiments indicated that biotic and abiotic mortality factors varied both over time and between experimental sites. Comparison of survival rates of test insects between cages that totally excluded natural enemy access and those that allowed natural enemy access shows that both parasitoids and predators can cause significant mortality of P. xylostella infesting cabbage on commercial properties (Table 2; Figure 3).

The low levels of parasitism by hymenopteran parasitoids detected during the experiments were probably due to the low population pressure of P. xylostella (populations on commercial properties never exceeded 1.5 larvae per plant). Over both years, D. semiclausum was the most abundant parasitoid of P. xylostella followed by D. collaris, A. ippeus and O. sokolowskii. O. sokolowskii is an important parasitoid of P. xylostella throughout Asia and much of Africa and samples collected in these experiments represent the first record of O. sokolowskii in Australia (it has since been recorded in Tasmania and Western Australia, M. Keller, F. Berlandier, pers. comm.). It appears that this species of parasitoid is widespread throughout the Lockyer Valley with specimens being reared from P. xylostella collected at central, northern and south-eastern experimental sites as well as from larvae feeding on a wild host (Rapistrum sp.) in the western region of the valley.

Pitfall trapping indicated that spiders (Lycosidae and Oxyopidae) were the most abundant insectivorous predators present on commercial Brassica farms in the Lockyer Valley, but Coleoptera (Carabidae, Staphylinidae and Coccinellidae) and Hemiptera were also relatively abundant.

At the beginning of the study, experimental sites were chosen depending on the farmer’s planned approach to pest management. Farmers at three sites (IPM1, IPM2 and IPM3) planned to take an integrated approach to insect pest management (e.g. use of insecticides which are less harmful to parasitoids and predators, observation of a break in Brassica production, utilisation of a crop scout to collect information for informed decision making, application of techniques to encourage natural enemies) while farmers at the remaining three sites (CAL1, CAL2 and CAL3) planned a less flexible approach to insect pest management, intending to apply insecticides on a calendar basis. In the event, only site CAL1 was managed by prophylactic application of insecticides and frequent use of chemicals known to have a detrimental effect on predators and parasitoids (Table 1). The management practices adopted by farmers at sites CAL2 and CAL3 were more similar to those at designated IPM sites than predicted while application of B. thuringiensis to site IPM3 was far more frequent than expected. Consequently, direct comparison between management practices was not possible. However, significant natural enemy activity was recorded at sites IPM1, IPM2, CAL2, IPM3 and CAL3 with combined predator and parasitoid diversity being particularly great at sites IPM1 and IPM2; farms where an integrated approach to insect pest management has been taken over several years. Estimated rates of predation and monitored natural enemy activity were consistently low at CAL1 over both years of the study indicating that pest management practices at this field site may have negatively impacted on natural enemy populations.

The great variation in the estimated rates of predation and the measured rates of parasitism illustrates the unpredictability of the impact of natural enemies on P. xylostella populations. However, at times the effect of natural enemies was extremely significant and further research is required to understand the processes underlying natural enemy activity so that they can be effectively incorporated into integrated pest management programs for the diamondback moth.

Acknowledgements

This work was funded by the Australian Centre for International Agricultural Research. Sue Scull, Vilma Amante and Sandra Dennien are thanked for excellent technical assistance.

References