Abstract

A population of the diamondback moth, Plutella xylostella (L.) (Lepidoptera: Plutellidae) (DBM), was detected in 1999 on sugar snap peas (Pisum sativum) in the export vegetable growing area south of Lake Naivasha in the Rift Valley Province of Kenya. The performance of this new pea strain (DBM-P) was compared with the normal crucifer strain (DBM-C) in laboratory studies. DBM-P survived equally well on kale and on pea, whereas DBM-C survival on pea was only 2.4%. Larval development of DBM-C took five days longer on pea than on kale and larval growth was severely reduced. Pupal weights of DBM-C on pea (3.8 mg) were significantly lower than those of DBM-P (4.6 mg) and those of both strains on kale (5.7 and 5.3 mg, respectively). Crosses between the strains in both directions produced fertile offspring.

The influence of the host shift on Diadegma mollipla (Holmgren), the most abundant indigenous parasitoid of DBM in the Kenyan highlands, was also investigated. The number of larvae parasitised in a no-choice situation by two strains of D. mollipla, one collected on pea and the other from cabbage, was higher on pea than on cabbage. In a choice situation with entire plants, parasitism on pea offered alone was four times higher than on cabbage. When both host plants were offered simultaneously, the level of parasitism dropped to the level of cabbage offered alone. In conclusion, it is suggested that D. mollipla is a non-specialist parasitoid in the crucifer/ DBM system and the significance of this observation for biological control of DBM is discussed.

Key words: Plutella xylostella, crucifers, host shift, Pisum sativum, Diadegma mollipla, tritrophic interactions

Introduction

Most herbivorous insect species are specialised feeders on just one plant family (Bernays & Chapman 1994). Still, a few abrupt host shifts to new plant families have been reported (Strong 1979, Bush 1994). The most serious pest of crucifer crops worldwide, the diamondback moth (Plutella xylostella L.) was thought to be restricted to the family Cruciferae (Talekar & Shelton 1993). However, in 1999, DBM was first collected from a heavily infested sugar snap pea (Pisum sativum L.) field in Naivasha in central Kenya (Löhr 2001). This localised population seemed to have overcome two major hurdles for its adoption of peas as host plants: females are attracted to peas for egg laying and larval survival is sufficient to allow a viable population to develop (Löhr & Gathu, in preparation).

Apart from the obvious relationship with their primary pest species, plants are known to strongly influence the evolutionary and behavioural ecology of host-parasitoid associations (Godfray 1994). Therefore, the effect of this host shift on Diadegma mollipla (Holmgren), the most abundant indigenous parasitoid of DBM in the east African highlands, was considered worth studying. Very little information is available about the biology of D. mollipla in association with DBM. The species was first described as Limneria mollipla (Holmgren, 1868), a parasitoid of the potato tuber moth (PTM), Phthorimaea operculella (Zeller) (Gelechiidae). It is reported to be indigenous to southern and eastern Africa, but the original host is unknown (Broodryk 1971, Gupta 1974). Because of its effectiveness on PTM, this species was introduced for biological control into various countries, e.g. Peru (Smit et al. 1998) and Yemen (Kroeschel 1993). In a recent revision of the genus Diadegma as DBM parasitoids, Azidah et al. (2000) grouped all African specimens under D. mollipla. In Kenya, D. mollipla is frequently found on DBM on Brassica crops, but parasitism rates are not particularly high. Field parasitism in Kenya, with D. mollipla being the most abundant species, was less than 20% (Oduor et al. 1996).

In this paper, the performance of the novel DBM pea strain is compared with the common crucifer strain of DBM on both the original and acquired host. In addition, the effect of the host shift by DBM on its most important indigenous parasitoid is assessed.

Materials and methods

DBM laboratory cultures

The cabbage strain (DBM-C) was originally obtained from a cabbage field in Limuru, a peri-urban vegetable growing area about 25 km north-west of Nairobi. The culture had been maintained for about one year, first in the laboratory and then in a purpose-built insectary, as described by Löhr and Gathu (in preparation).

The pea strain (DBM-P) originated from an export vegetable farm along South Lake Road, Naivasha. After initial difficulties, a self-sustaining culture was established in August 2000 and repeatedly replenished with field-collected material from the original collection site. Snowpea cultivar “Snowgreen” and sugar snap pea cultivar “Sugar Pod” were used to sustain the culture.

Comparison of performance of DBM strains on peas and cabbage

Small tissue paper strips (approx. 15 x 30 mm) were placed inside transparent acrylic 30 ml vials and moistened with a drop of tap water. A piece of kale leaf, (Brassica oleracea var. acephala L. cv. Thousandheaded), of approximately the same size was placed on top of the tissue paper. In the pea treatment, an entire pea leaflet (Pisum sativum cv. Oregon sugar pod) was used. Neonate larvae were transferred with a fine brush into the prepared vials. Because of their delicate nature, the larvae were not touched, but lifted on their own thread and gently lowered onto the leaf surface. The procedure eliminated larval mortality due to handling. Pea leaflets were changed whenever the leaflet showed signs of wilting, while kale was changed when it started to turn yellowish. Care was taken to have enough food at any time. The vials were closed with a plastic cap without ventilation holes. Balancing the amount of water to avoid condensation inside the vial was delicate, but unventilated vials had proven the best option in preliminary trials.

The treatments consisted of a) DBM-C larva on kale; b) DBM-C on pea; c) DBM-P on kale and d) DBM-P on pea, all were evaluated concurrently in any one replication of 50 vials of each treatment. Five replications were run between July and November 2000. The vials of each treatment were placed on a metal grid and the whole set kept on a laboratory bench at room temperature (23 ± 3ºC during the day, 19± 2ºC at night).

The vials were checked every morning between 09:00 and 11:00. Parameters recorded were larval position during feeding (mining or on the leaf surface), survival, day of pupation and day of adult emergence and sex. Pupae were removed with the cocoon from the vial, weighed on a Mettler analytical balance and returned for adult emergence.

Rearing of parasitoids

(a) Cabbage strain of Diadegma mollipla

Cultures were established from cocoons collected from cabbage fields at Kapsabet in Nandi District of western Kenya and Limuru, Kiambu District of central Kenya. Parasitoids were reared on diamondback moth larvae on excised cabbage leaves in small Perspex cages. DBM larvae were renewed every two to three days until the parasitoids died. Parasitised DBM larvae were maintained separately. Pupae were collected into a vial and newly emerged parasitoid adults of both sexes were then kept together for at least one day to ensure mating.

(b) Pea strain of D. mollipla

The culture was established with a single pair collected with DBM larvae from pea fields at Naivasha. The culture was maintained on DBM-P larvae as previously described for the DBM-C culture.

Parasitism

All experiments were conducted under laboratory conditions (23 ± 2°C). To compare parasitism of the two D. mollipla strains on DBM-C and DBM-P, single mated 2–3 day old female D. mollipla were used. Preliminary tests with D. mollipla showed peak searching activity beyond this period. Females were kept for 24 h in small plastic containers (5 x 8 x 17 cm) with 25 II instar DBM larvae, 4 days old, on leaves. Fully expanded leaves from 4 to 6 week old plants of both species were used. Parasitism experiments were carried out with both c-D. mollipla on DBM-C and on DBM-P, p-D. mollipla on DBM-C and DBM-P.

After removing the parasitoid, the DBM larvae were reared individually in vials until reaching adult stage on leaves or leaf discs of their respective food plants. Dead DBM larvae were dissected in order to search for parasitoid eggs.

In this experimental setup, naïve and experienced parasitoids were tested. To gain experience, females were allowed to parasitise larvae of the DBM strain they emerged from 24 h before the trial.

Effect of host plants on parasitism

In an exploratory test, single excised cabbage and pea leaves of similar size were kept in vials with water. A day before exposure, the leaves were infested with ten II instar DBM larvae each (DBM-C on cabbage and DBM-P on pea). Both vials were placed at a distance of approximately 30 cm in a Perspex cage (43 x 23 x 22 cm). Three 3–4 day old mated and experienced (on DBM-C) females of c-D. mollipla were then released into the cage. Larvae were removed after 24 h and reared in separate containers. The number of parasitoid pupae on both plants was recorded. The treatment was replicated five times.

Further tests described in this chapter were conducted with whole plants in a screened metal-framed cage measuring 60 x 45 x 45 cm. Only the experienced c-D. mollipla strain was used. To reduce the influence of variability of performance for individual females, three parasitoids were released in the cage. All treatments were replicated three times.

Single host plant and mixed host plant exposure

Four cabbage plants (4–6 weeks after transplanting; 6–8 leaves) were placed in the cage at a distance of approximately 20 cm. Each plant was infested a day before exposure with ten II instar DBM-C larvae. They were then exposed for 48 h to three 3–4 day old parasitoids. The larvae were subsequently reared in plastic containers on cabbage leaves. Larvae of the same plant were kept together. The number of parasitised larvae was recorded. A similar experiment was conducted with DBM-P larvae on four pea plants offered as single host and in a mixed host plant situation with two pea and two cabbage plants with their respective DBM larvae.

Statistical analysis

Analysis of variance was performed on the larval development of DBM in order to determine significant differences among treatments. A mixed model was assumed with the trial effect as random. Treatment means were separated using the Tukey’s HSD test at 5% significance level. For the performance of the parasitoid, the Student-Newman-Keuls test at 5% significance level was used additionally.

RESULTS

Effect of host plant and DBM strain on development and survival

Larval development

Neonate larvae of DBM usually pass the first larval stage mining within the leaf parenchyma. DBM-C larvae mined for an average of 2.1 days in kale leaves, but only one of the 250 tested larvae managed to mine a pea leaflet for one day. In contrast, mining time of DBM-P larvae on kale was even longer than that of DBM-C (3 days). Pea strain larvae also managed to mine pea leaves, though for a shorter period than they mined on kale (Table 1). Differences in mining time were significant between both strains and on both crops.

Table 1 Performance of a cabbage and a pea strain of diamondback moth on kale and pea.

DBM strain	Host plant	Mining days1	Larval period	Pupal period	Pupal weight	% survival	Sex ratio2
cabbage
	kale	2.1 ± 0.05 b	8.5 ± 0.07 d	5.6 ± 0.07 ab	5.7 ± 0.06 a	87.6	1.33
	pea	0.01 ± 0.01 d	13.6 ± 0.55 a	5.8 ± 0.17 a	3.8 ± 0.13 d	2.4	2.00
pea
	kale	3.0 ± 0.06 a	9.0 ± 0.10 c	5.3 ± 0.06 b	5.3 ± 0.06 b	85.2	0.94
	pea	0.6 ± 0.04 c	10.6 ± 0.13 b	5.5 ± 0.07 ab	4.6 ± 0.06 c	82.8	0.93

¹ average ± SE

² Females:males

Means in the same column having same letter are not significantly different at 5% level using Tukey’s HSD test

Larval development showed a similar pattern for both DBM strains: faster development on kale and slower development on peas. However, whereas on kale DBM-C developed significantly faster (8.5 days) than DBM-P (9.0 days), its development on peas (13.6 days) was greatly retarded, also in comparison to DBM-P (10.6 days, Table 1). Development of both strains on kale and of DBM-P on pea was also more homogeneous than for DBM-C on peas, where some larvae remained in the larval stage for 22 days.

Heavy larval mortality occurred only among DBM-C on peas and was concentrated during the first three days. After the second day, 55.6% of the larvae had died (Figure 1). In total, 92% of DBM-C completed the larval stage on kale compared with only 7.2% on peas, while 88.4 and 90.0% of the pea strain completed larval development on kale and peas, respectively.

Figure 1: Survival of a cabbage (C) and a pea (P) strain of diamondback moth on kale and pea

Pupal development

Differences in the duration of the pupal stage were small and only significant between those of DBM-C on peas (5.8 days) and DBM-P on kale (5.3 days, Table 1). There were differences, however, in the rates of survival during pupal stage: 95.6% of pupated DBM-C raised on kale emerged as adults as compared with 33.3% on pea. Survival of the DBM-P was similar on both plants, kale and pea (97.3% and 92% respectively, Figure 1).

Pupal weight

Significant differences were also recorded in pupal weight between all treatments and strains (Table 1). Cabbage strain pupae on kale were heaviest (5.7 mg), followed by DBM-P on kale (5.3 mg) and peas (4.6 mg). The lightest (3.8 mg) were the pupae of the DBM-C that had survived on peas. Their weight was so low that many of them failed to emerge as adults.

The sex ratio of emerging adults was female-biased for DBM-C, irrespective of the host plant, while the ratio for DBM-P was slightly male-biased.

Effect of host shift on the parasitoid

Parasitism by Diadegma mollipla

In Table 2, the number of parasitised larvae is shown for the two D. mollipla strains, c-D. mollipla and p-D. mollipla on both DBM strains. Generally parasitism was low, but parasitoid performance was better on peas than on cabbage. The number of parasitised larvae per naïve female ranged from 0 to 11 for c-D. mollipla on DBM-C, 0 to 17 on DBM-P and from 0 to 20 for p-D. mollipla on DBM-P.

With experienced D. mollipla, the number of parasitised larvae was higher and ranged between 1 to 15 for p-D. mollipla on DBM-P and 10 to 14 for c-D. mollipla on DBM-P. In both experiments, parasitism on DBM-P was higher than on DBM-C. The range of parasitised larvae per female given in Table 2 illustrates the high variation in performance of individual females. Therefore, these figures are not statistically different, but can be taken as tendencies. Missing data are due to microsporidian infection problems in the DBM culture and the collapse of the p-D. mollipla culture during the experiments.

Table 2: Parasitism of naïve and experienced Diadegma mollipla on diamondback moth larvae reared on cabbage and on pea

Parasitoid strain	Host strain	No. parasitoids tested	No. larvae exposed	Range of larvae parasitised	Mean no. larvae parasitised1
Naïve D. mollipla
cabbage	cabbage	12	23	0 - 11	3.0 ± 1.15 ns
	pea	12	23	0 - 17	3.8 ± 1.62 ns
pea	cabbage	-	-	-	-
	pea	7	22	0 - 20	5.8 ± 2.19 ns
Experienced D. mollipla
cabbage	cabbage	3	25	6 - 11	8.3 ± 1.44 ns
	pea	3	23	10 - 14	12.6 ± 1.33 ns
pea	cabbage	5	23	1 - 15	7.3 ± 3.49 ns
	pea	-	-	-	-

Effect of host plants on parasitism

When larvae were exposed on excised leaves kept in vials, parasitism was low with 24 DBM larvae out of 100 parasitised. However, with 16 larvae parasitised on peas, preference was more biased towards peas than towards cabbage (eight parasitised larvae).

When DBM was exposed on cabbage plants only, the parasitism rate of c-D. mollipla was 6.1%, thus being significantly lower than on pea plants alone (26.5%) (Table 3). When both host plants were offered simultaneously, parasitism was comparable to cabbage offered alone (3.5%). However, a higher proportion of larvae was parasitised on peas (2.6%) than on cabbage (0.9%).

Table 3: Influence of host plant on parasitism of diamondback moth by Diadegma mollipla reared on cabbage

Host plant	No. DBM larvae exposed	Average no. of larvae recovered	Average no. of larvae parasitised1	Parasitism rate [%]
cabbage alone	40	38	2.3 ± 0.88 a	6.1 a
pea alone	40	38	10.0 ± 2.52 b	26.5 b
pea/cabbage	40	39	1.3 ± 0.33 a2	3.5 a [*) 2.6, **) 0.9]

¹ average ± SE

² average number parasitised on pea =1.0, on cabbage = 0.3

*⁾ parasitised on peas, **⁾ parasitised on cabbage. Means from 3 replicates.

Means in the same column having same letter are not significantly different at 5% level using Tukey’s HSD test

Discussion

Diamondback moth has been recorded in the field from host plants outside its “natural” host plant range before. Reichart (1919, in Talekar et al. 1985) reported DBM on chickpea and a chenopodiaceous vegetable in Russia. More recently, DBM was found on okra in Ghana (Anonymous 1971) and on faba beans in Egypt (Badr et al. 1986). These occurrences seem to have been either sporadic and were not given due attention, or were the result of misidentifications and, therefore, never heard of again. In the case reported here, crosses between DBM-C and DBM-P in both directions produced fertile offspring (Löhr, unpublished data), so the identity of the species is not in question.

To date, only few studies have looked into the host plant range of DBM, all under laboratory conditions. Gupta and Thorsteinson (1960a) observed that DBM fed on nine non-cruciferous species under no-choice situation in the laboratory. Pea was one of the six species in the Leguminosae family found suitable for DBM feeding in these studies. An additional 12 species were accepted only when sinigrin was added as phagostimulant on the leaf discs. The mentioned authors did not observe any differences between the survivors on peas and the ones raised on mustard. However, in our experiments, pupal weights and in consequence, adult size of the surviving normal strain DBM on pea was significantly lower than on kale. Gupta and Thorsteinson (1960b) attributed the non-host character of peas in the field to the egg-laying behaviour of the female moths, implying that those may not be attracted to the crop. In Kenya, profuse egg-laying has been observed in the field and Löhr and Gathu (in preparation) investigated the ability of DBM to adapt to new hosts. They found that in only four generations, a DBM strain could be selected that survived equally well on kale and pea.

The data presented here also show that D. mollipla, the major DBM parasitoid in the Kenyan highlands, has managed to follow its host onto the new host plant. Preference experiments revealed significantly higher parasitism of DBM larvae on peas compared with cabbage. Two conclusions can be derived from these observations: D. mollipla is only loosely associated with DBM and its original host plant range and there must be a factor that renders DBM on crucifers less attractive than DBM on peas. The first observation is supported by reports of D. mollipla as an important parasitoid of the potato tuber moth (PTM) on potato and tobacco in southern Africa (Broodryk 1971) and on potato in Yemen (Kroeschel 1993). However, as PTM is an introduced species to Africa and D. mollipla seems to be indigenous, PTM cannot be the original host of this species. It is therefore reasonable to assume that D. mollipla is a parasitoid with considerable host plasticity and that it might be found to parasitise more free-living or leaf mining species of microlepidoptera. This lack of specialisation may also explain the generally low parasitism rates of DBM observed in the laboratory (Akol 2001) and thus its irrelevance for the control of DBM field populations (Oduor et al. 1996, Löhr, unpublished survey data). As for the second observation, concerning factors for the higher attractivity of peas, a few published papers relate to the influence of DBM host plants on parasitism, but they all refer to DBM specialist parasitoids on crucifers (Talekar & Yang 1993, Beck & Cameron 1990, Idris & Grafius 1996). However, the preference of D. mollipla for DBM feeding on peas may be explained by cues used for host location. The use of infochemicals by hymenopterous parasitoids to locate their hosts is well documented (e.g. reviewed in Vet & Dicke 1992). Parasitoids associated with crucifer specialist herbivores were shown to be attracted by volatile isothiocyanates (mustard oils) typically released by crucifers when injured (Pivnick 1993, Murchie et al. 1997). For D. mollipla, crucifer volatiles are unlikely to be used for host location. We assume that these substances, known to be toxic for many herbivores, even reduce attractiveness of cruciferous plants for D. mollipla. We therefore assume that the crucifer growing system is not a preferred habitat for D. mollipla. However, it is commonly accepted because of its wide availability and therefore easy accessibility in Kenya.

Acknowledgements

Technical assistance during these studies was provided by J. Gatama and C. Momanyi, both are gratefully acknowledged. Thanks also to A. Hassanali for valuable suggestions on earlier versions of this paper and to D. Odulaja and M. Wabiri for help with the statistical analysis. This study was conducted within the DBM Biocontrol Project for eastern and southern Africa of the International Centre of Insect Physiology and Ecology, funded by the German Ministry of Technical Cooperation and Development. The work of the second author was partially supported by the German Academic Exchange Service (DAAD).

References