Egg parasitoids of noctuid stemborers on maize in Kenya

Abstract

The West African scelionid egg parasitoid Telenomus isis was released into Kenya in 2005 for control of the noctuid stemborer Busseola fusca. Pre- and post-release surveys were carried out between 2004 and 2006 to assess parasitism of eggs of noctuid stemborers in ecozones where B. fusca is the dominant pest, namely in the mid-altitude areas of Taveta and Kiboko, in the highlands along the Eldoret–Kitale transect, and in Wundanyi, which stretches from the mid-altitudes to the highlands. Apart from Eldoret, where surveys were carried out weekly during the cropping seasons of 2004–2005, the sites were visited once during the short and/or long rainy seasons. In addition, after release of T. isis in Wundanyi in October 2005, two post-release surveys were conducted in 2006. Busseola fusca was the dominant stemborer species at Eldoret and Wundanyi. At Taveta and Kiboko only eggs of the noctuid Sesamia calamistis were recovered. There was generally a high variation in the numbers of egg batches and eggs per batch among the sites surveyed. Most of the infested plants had one egg mass. Egg parasitism varied between 10% and 32% across the study sites. The predominant parasitoid species was the scelionid Telenomus busseolae while the trichogrammatids Trichogramma bournieri, Trichogrammatoidea nr. lutea. Trichogrammatoidea nr. cryptophlebiae and Trichogramma mwanzai were minor species. In 2006, T. isis was recovered in Wundanyi. This study showed that although egg parasitoid species diversity was higher, egg parasitism was considerably lower in Kenya than in western Africa.

Keywords:

• egg parasitism
• Noctuidae
• Telenomus
• Trichogramma
• T. busseolae
• T. isis
• establishment
• Kenya

Introduction

Maize is a major food crop in East and Southern Africa (ESA) (FAO 2003). Among the most important biotic constraints to maize production are lepidopteran stemborers such as the noctuid Busseola fusca (Fuller) and the invasive crambid Chilo partellus (Swinhoe) (Polaszek and Khan 1998; Seshu Reddy 1998). Various stemborer control techniques, such as chemical control, botanicals, host plant resistance and cultural control, have been researched and tested (Sétamou et al. 1995; Ndemah and Schulthess 2002; Bruce et al. 2004; Ali et al. 2006; Matama-Kauma et al. 2006; Wale et al. 2006a; Songa et al. 2007). Since the 1990s, both the International Institute of Tropical Agriculture (IITA) in West Africa, and the International Centre of Insect Physiology and Ecology (ICIPE), Nairobi, Kenya, have given increased emphasis on biological control. While ICIPE focused on classical biological control of the exotic C. partellus using the Asian braconid larval parasitoid Cotesia flavipes Cameron (Zhou et al. 2001; Omwega et al. 2006), IITA focused on the ‘redistribution’ approach, which aims at exchanging natural enemy species and strains between African regions (Schulthess et al. 1997). Thus, in the early 1990s, IITA introduced a strain of Cotesia sesamiae (Cameron) from the coastal region of Kenya into Benin, where it established successfully on the noctuid Sesamia calamistis Hampson (Schulthess et al. 1997). In addition, several strains of Co. sesamiae are considered for introduction from Kenya into Cameroon for the control of B. fusca (Ndemah et al. 2007).

There are several studies on the relative importance of natural enemies of B. fusca in ESA (Mohyuddin and Greathead 1970; Kfir 1995; Phiri 1995; Bonhof et al. 1997; Mochiah et al. 2002; Matama-Kauma et al. 2007) but little attention has been given to egg parasitoids (Mathez 1972; Bruce et al. 2006; Okoth et al. 2006). Egg parasitoids can be an important source of mortality because the pest is killed before reaching the destructive stage (Temerak 1981). For instance in Benin, parasitism by the scelionids Telenomus busseola Gahan and T. isis Polaszek has been reported to reach 95%, and several studies yielded a positive relationship between egg parasitism during the vegetative stages of growth and maize yields at harvest (Sétamou and Schulthess 1995; Schulthess et al. 2001; Ndemah et al. 2003).

Telenomus isis is commonly recovered from eggs of noctuid stemborers in West Africa and in Cameroon in Central Africa (Sétamou and Schulthess 1995; Moyal 1998; Ndemah et al. 2001; Schulthess et al. 2001; Chabi-Olaye et al. 2006). It has been proposed for release in ESA (Schulthess et al. 1997; Chabi-Olaye et al. 2001a) from where it has never been reported (Polaszek 1998). The parasitoid was thus introduced into the containment facilities of ICIPE in 2003.

Before release of a natural enemy into a new area, the Kenya Plant Health Inspectorate Service (KEPHIS) requires information on its possible impact on target and non-target species as well as on the indigenous egg parasitoids. In a first step, a catalogue and the relative importance of egg parasitoids of noctuid stemborers were established in those ecozones of Kenya, where B. fusca and the alternative host S. calamistis are common (Zhou et al. 2002). The present results are part of the documentation submitted to KEPHIS to obtain an application for a release permit, which was granted in 2005. Telenomus isis was then released in one location in late 2005. The present work also presents some data from post-release surveys.

Materials and methods

Survey locations

The study was carried out in the mid-altitude areas of Taveta and Kiboko (700–1300 m asl), and in the highlands (1500–2400 m asl) along the Eldoret–Kitale transect, and in Wundanyi, which stretches from the mid-altitudes to the highlands (Figure 1). All sites, with the exception of Eldoret, which has only one long cropping season from April to September, have a bimodal rainfall distribution pattern, which allows for two cropping seasons – one from April to July (long rainy season, LR) and one from September to January (short rainy season, SR).

Figure 1. Map of Kenya with the study locations.

Sampling scheme

Fields visited were divided into five subplots – one at each corner and one at the centre of the field. The size of subplots varied from 42 to 56 m² reflecting differences in farm size. Egg sampling was carried out at the pre-tasseling stage of maize. Surveys were conducted in the following regions (Figure 1):

a.	Kiboko: 22 and 23 fields surveyed in December 2004 and 2005, respectively. Each subplot was 56 m2.
b.	Taveta: seven and five fields surveyed in September 2004 and 2005, respectively. The subplots were 42 m2.
c.	Eldoret: 76 and 74 fields were visited in May to October 2004 and 2005, respectively. The subplots were 42 m2.
d.	Wundanyi: during the pre-release phase, 22 and 24 fields were sampled in September and November 2004, and 16 fields in August 2005. After the release of T. isis, 16 and 15 fields, respectively, were sampled during March and December 2006. The subplots were around 54 m2.

For each subplot, all plants were checked for noctuid stemborer eggs, which are oviposited between the leaf sheath and the stem (Kaufmann 1983). The number of egg batches per plant and number of eggs per batch was recorded. Egg batches were individually placed in glass vials, plugged with cotton wool, labelled and held in environmental chambers at ICIPE at 25 ± 1°C and 60%–70% relative humidity (RH) until borer larvae or parasitoids emerged. Larvae were reared to adult stage according to the method described by Onyango and Ochieng-Odero (1994) for species identification. For each species, three kinds of parasitization were calculated: (a) mean egg parasitism per field calculated as percentage of eggs parasitized within an individual egg batch, averaged over all egg batches found in the field; (b) percentage of egg batches with parasitoids per field – referred to as ‘discovery’ efficiency by Bin and Vinson (1991) – which provides information about the searching ability of a parasitoid; and (c) percentage of eggs parasitized within discovered egg batches, averaged over all egg batches per field, termed ‘parasitism efficiency’ by Bin and Vinson (1991). Parasitism efficiency is determined by parasitoid fecundity, host acceptance and host patch exploitation strategy. Identification of parasitoids was done using keys developed by Polaszek (1998), Douth and Viggiani (1968), and by J. M. Baya (ICIPE, Nairobi, Kenya) and Gerald Delvare (CIRAD, Montpellier, France).

Statistical analyses

The dispersal of egg batches was analysed using Taylor's (1961) power law, which describes the relationship between the variance (s²) and the mean, i.e. s² = am^b, where the slope b is a measure of dispersion of a species with b > 1, b = 1 and b < 1, indicating aggregated, random and regular distribution, respectively. These coefficients were computed by regressing the logarithm of the between plant variance (logs²) against the logarithm of mean density (logm), for each field at each sampling occasion. Student's t-test was used to compare b with unity.

Correlation analysis was computed to assess the relationship between different components of the system. Data collected were analysed using the generalised linear model (PROC GENMODE, SAS 1997). Count data (egg batch and number of eggs per batch) were analysed using a Poisson error distribution with a logarithmic link function (McCullagh and Nelder 1989). While proportion data (discovery efficiency, egg parasitism, parasitism efficiency and sex ratio) were analysed using a binomial error distribution with a logistic link function (Collett 1991). Significance level was set at P ≤ 0.05. In addition, LSM were used to compare between locations and egg parasitoid species.

Results

Dispersion pattern of egg batches

Dispersal of egg batches varied significantly with locality (Table 1). Egg batches were randomly distributed at Kiboko, Taveta and Wundanyi, and regularly at Eldoret. The slope (b) was positively and significantly correlated with egg batch density across all locations with r-values varying from 0.61 (P = 0.0023) at Wundanyi after release of T. isis to 0.73 (P < 0.0001) at Kiboko.

Table 1. Taylor's coefficients (± SE) for number of noctuid egg batches per maize plant

Year	Locations	b	a	F	P	R^2
2004/2005^a	Kiboko	1.17 ± 0.079a**	2.90	222.8	<0.0001	0.87
2004/2005^a	Taveta	1.01 ± 0.004ab**	1.01	1111.9	<0.0001	0.99
2004/2005^a	Eldoret-Kitale	0.98 ± 0.001b*	0.92	10,178.81	<0.0001	0.99
2004/2005^a	Wundanyi	1.00 ± 0.00ab**	0.50	1.5	<0.0001	0.99
2006^b	Wundanyi	1.02 ± 0.017ab**	1.35	3903.8	<0.0001	0.99
F		9.93
df		4248
P		<0.0001
Note: ^a pre-release survey, ^b post-release survey; *b < 1, **b = 1, t-test, P ≤ 0.001; slopes (b) followed by the same lower case letter(s) are not significantly different at P ≤ 0.05 (SNK test).

Borer species and egg batch density

In Eldoret and Wundanyi, more than 80% of the batches collected were from B. fusca and the remainder from S. calamistis. At Taveta and Kiboko, only S. calamistis eggs were recovered. The relative importance (in % of all egg batches) of B. fusca increased while that of S. calamistis decreased with altitude (r = 0.64, P < 0.0001 and r = −0.64, P < 0.0001, respectively).

There was a high variation in the number of egg batches and number of eggs per batch among the sites (Table 2). Most of the infested plants (97%) yielded a single egg mass with mean batch size ranging between 21 and 34. At Wundanyi, there was a significant difference in number of batches per plant between pre- and post-release surveys (χ² = 76.9, df = 1, P < 0.0001), while the number of eggs per batch did not vary (χ² = 1.61, df = 1, P = 0.2046).

Table 2. Average number (±SE) of egg batches per 100 plants and of eggs per batch, discovery efficiency (%), egg parasitism (%) and parasitism efficiency (%) at various sites in Kenya between 2004–2006

Period	Site	# egg batches	# eggs/batch	Discovery efficiency	Egg parasitism	Parasitism efficiency
Pre release	Kiboko	0.42 ± 0.09c	20.6 ± 1.18b	37.3 ± 1.74b	32.6 ± 1.27a	86.8 ± 1.26a
(2004–2005)	Taveta	1.58 ± 0.61ab	18.3 ± 1.02b	40.3 ± 1.01a	31.7 ± 1.78a	74.9 ± 1.10c
	Eldoret	2.09 ± 0.19a	34.2 ± 1.36a	14.2 ± 1.86d	10.0 ± 0.45c	63.6 ± 1.53d
	Wundanyi	1.01 ± 0.12b	25.7 ± 1.41b	32.9 ± 1.66c	25.9 ± 1.96b	79.9 ± 1.15b
	χ^2	91.4	147.4	1749.3	1791.9	122.2
	df	3	3	3	3	3
	P	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001
Post release (2006)	Wundanyi	3.85 ± 0.85	20.3 ± 1.31	48.1 ± 1.35	37.9 ± 1.71	73.2 ± 0.62
Note: Within column, means followed by the same letter(s) are not significantly different at P ≤ 0.05 (Chi-square χ^2 test).

Relative importance of parasitoid species

Multiple-species parasitism of egg batches was not observed during the surveys. The egg parasitoids identified were the scelionids T. busseolae (from 2521 batches), T. isis (293) and an undescribed Telenomus spp. (596 batches). In addition, there were the trichogrammatids Trichogramma bournieri Pintureau & Babault (151), Trichogramma nr. lutea Girault (15), Trichogramma mwanzai Schulten & Feijen (11) and Trichogramma nr. cryptophlebiae Nagaraja (29) (Table 3).

Table 3. Relative importance of egg parasitoids at different sites in Kenya between 2004–2006

Period	Pre-release (2004–2005)							Post-release (2006)
Site	Kiboko	Taveta	Eldoret	Wundanyi	χ ^2	df	P	Wundanyi
Parasitoid species	Egg parasitism (%)
T. bussoelae	38.95 ± 4.49b	40.67 ± 0.07a	11.33 ± 1.70d	27.06 ± 3.10c	634.7	3	<0.0001	38.64 ± 4.34
T. isis	0.00 ± 0.00	0.00 ± 0.00	0.00 ± 0.00	0.00 ± 0.00				6.34 ± 0.68
Telenomus spp.	2.93 ± 0.92a	3.63 ± 0.54a	0.49 ± 0.03b	1.02 ± 0.75a	31.1	3	<0.0001	3.99 ± 0.88
Tr. bournieri	0.64 ± 0.03a	1.41 ± 0.09a	0.69 ± 0.32a	0.41 ± 0.03b	6.8	3	0.0778	0.99 ± 0.02
Tr. nr lutea	1.36 ± 0.09a	0.83 ± 0.05b	0.00 ± 0.00c	0.00 ± 0.00c	163.9	3	<0.0001	0.00 ± 0.00
Tr. mwanzai	0.00 ± 0.00b	0.35 ± 0.02a	0.00 ± 0.00b	0.00 ± 0.00b	17.6	3	0.0005	0.00 ± 0.00
Tr. nr cryptophlebiae	0.27 ± 0.02b	1.17 ± 0.07a	0.00 ± 0.00c	0.00 ± 0.00c	65.7	3	<0.0001	0.00 ± 0.00
	Discovery efficiency (%)
T. bussoelae	45.29 ± 5.91b	49.53 ± 8.77a	15.11 ± 1.98d	32.83 ± 3.68c	567.3	3	<0.0001	45.38 ± 5.69
T. isis	0.00 ± 0.00	0.00 ± 0.00	0.00 ± 0.00	0.00 ± 0.00				8.36 ± 1.85
Telenomus spp.	2.95 ± 0.09a	5.49 ± 0.06a	0.21 ± 0.01c	2.53 ± 0.75b	61.2	3	<0.0001	5.47 ± 0.01
Tr. bournieri	0.79 ± 0.03b	2.40 ± 0.01a	1.23 ± 0.04ab	0.64 ± 0.05c	12.7	3	0.0053	2.08 ± 0.05
Tr. nr lutea	0.96 ± 0.06a	1.16 ± 0.07a	0.00 ± 0.00b	0.00 ± 0.00b	127.9	3	<0.0001	0.00 ± 0.00
Tr. mwanzai	0.00 ± 0.00b	0.54 ± 0.03a	0.00 ± 0.00b	0.00 ± 0.00b	27.1	3	<0.0001	0.00 ± 0.00
Tr. nr cryptophlebiae	0.32 ± 0.02b	1.94 ± 0.01a	0.00 ± 0.00c	0.00 ± 0.00c	100.3	3	<0.0001	0.00 ± 0.00
	Parasitism efficiency
T. bussoelae	85.35 ± 2.63a	73.08 ± 4.83b	62.02 ± 2.86c	77.09 ± 2.56b	181.5	3	<0.0001	61.38 ± 3.58
T. isis	–	–	–	–				65.39 ± 4.33
Telenomus spp.	93.57 ± 2.07a	66.78 ± 5.85b	53.00 ± 0.23d	54.28 ± 9.48c	192.3	3	<0.0001	76.72 ± 3.58
Tr. bournieri	83.79 ± 9.23a	55.31 ± 0.15b	51.50 ± 7.33b	52.59 ± 4.87b	79.3	3	<0.0001	55.49 ± 7.24
Tr. nr lutea	92.69 ± 7.30a	76.02 ± 1.91b	–	–	234.8	1	<0.0001	–
Tr. mwanzai	–	66.61 ± 1.57b	–	–				–
Tr. nr cryptophlebiae	100.00 ± 0.00a	86.94 ± 3.78a	–	–	2.2	1	0.1367	–
Note: – No observation; within row, means followed by the same letter(s) are not significantly different at P ≤ 0.05 (Chi-square χ^2 test).

Egg parasitism, discovery efficiency and parasitism efficiency

Parasitism varied with site (Table 2) and was similar in Kiboko, Taveta and Wundanyi but significantly lower in Eldoret. Discovery efficiency followed the same pattern as parasitism in all the localities.

T. busseolae was the most important species, with parasitism ranging between 11.3 in Eldoret to 40.7% in Taveta (Table 3). For any of the other species, parasitism did not exceed 4%. In Wundanyi, parasitism was higher in 2006 than 2004–2005. Discovery efficiency followed the same pattern as parasitism, ranging from 32.8% to 45.4% and was higher in 2006 than 2004–2005. Discovery efficiency and parasitism were negatively correlated with egg batch size (r = −0.63, P = 0.03043 and r = −0.31, P < 0.0001, respectively) and egg density (r = −0.37, P < 0.0001 and r = −0.30, P ≤ 0.0001, respectively).

At Wundanyi, discovery efficiency and egg parasitism pooled across species varied significantly with season (χ² = 175.3, df = 1, P < 0.0001 for discovery efficiency and χ² = 161.4, df = 1, P < 0.0001 for egg parasitism). The same pattern was observed for the predominant parasitoid T. busseolae alone (χ² = 17.5, df = 1, P < 0.0001 and χ² = 21.1, df = 1, P < 0.0001, for discovery efficiency and egg parasitism, respectively). Discovery efficiency was 3.76% for Telenomus spp. and 0.47% only for Trichogramma spp., and varied significantly (χ² = 1559.0, df = 1, P < 0.0001).

Parasitism efficiency was significantly higher for Telenomus spp. (74%) than Trichogramma spp. (59%) (χ² = 1401.4, df = 1, P < 0.0001); it varied significantly among parasitoid species, and ranged from 48% for Tr. bournieri to 93% for Telenomus spp. At Wundanyi, parasitism efficiency pooled across species varied significantly between pre- and post-release periods (79% for 2004–2005 and 73% in 2006, χ² = 131.6, df = 1, P < 0.0001).

For all species sex ratio was female-biased and varied from 0.51 for T. busseolae to 0.90 for Tr. bournieri (χ² = 0.50, df = 6, P = 0.99).

Discussion

The dominance of B. fusca increased with altitude and it became the major cereal pest above 1000 m asl, while S. calamistis was mainly found in the mid-altitude zones. This trend is only observed in ESA (Kfir 1997; Zhou et al. 2001; Ong'amo et al. 2006; Wale et al. 2006b). In Cameroon, Central Africa, B. fusca is the major pest across all altitudes (Ndemah et al. 2002), while in West Africa it is only common in the Northern Guinea and Sudan savannah, where it attacks sorghum (Schulthess et al. 1997). Recent phylogeographic analyses revealed distinct populations of B. fusca occurring in Africa, which vary in their climatic requirements (Sézolin et al. 2006).

In the present study, 97% of the infested plants had only one egg batch, and the distribution of egg batches was regular to random, corroborating findings by Sétamou and Schulthess (1995), Schulthess et al. (2001), Ndemah et al. (2002) and Chabi-Olaye et al. (2005). As also reported by these studies, the aggregation index increased with egg batch density. Female noctuid stemborers rarely lay more than one egg batch per plant. It was suggested that female moths are able to recognize and avoid plants with egg batches and oviposit at a distance from an occupied plant leading to a regular distribution of egg batches (Sétamou and Schulthess 1995; Chabi-Olaye et al. 2005). More than one egg batch per plant would lead to early destruction of the food source, causing forced emigration and exposure to predators, eventually resulting in increased larval mortality. The distribution of egg batches might have implications for the host searching efficiency of the parasitoid, whereby a more aggregated distribution of egg batches should enhance its host finding.

In West Africa and Cameroon, the only parasitoids obtained from noctuid stemborer eggs were the scelionids T. busseolae and T. isis, and the trichogrammatids Lathromeris ovicidae Risbec and Paracentrobia sp. (Bosque-Pérez et al. 1995; Ndemah et al. 2001; Schulthess et al. 2001). However, while parasitism by L. ovicidae never exceeded 3%, only a few specimens of the latter were recovered. In contrast, four trichogrammatid and two scelionid egg parasitoids species were obtained from eggs of noctuid stemborers in Kenya alone.

In spite of the higher parasitoid diversity, egg parasitism in Kenya during the 2 years of study was mostly below 33%. In Côte d'Ivoire, Moyal (1998) reported 72% parasitism of B. fusca eggs by T. busseolae and T. isis. In Benin, noctuid borer attacks never reach the outbreak levels common in other West African countries as a result of high Telenomus spp. egg parasitism of up to 95% (i.e. regional mean) at the beginning of the late season (Schulthess et al. 2001). In the humid forest zone of Cameroon, B. fusca egg batch densities increased up to the beginning of the second season, but larval densities during the late season were lower than during the first season (Ndemah et al. 2002). This was attributed to high egg parasitism by T. isis and T. busseolae averaging 66% (Ndemah et al. 2003). Schulthess et al. (2001) and Ndemah et al. (2003) suggested that high S. calamistis egg parasitism as encountered in the Dahomey Gap, encompassing Benin, Togo and the southeastern part of Ghana, was due to a high density of alternative host plants in the vicinity of maize fields, which maintain stable parasitoid populations during the off-season, when the crop was not available. Unlike B. fusca (Kfir 1991), S. calamistis does not diapause and depends on alternative host plants during the off-season. Thus, the catalogue of host plants of S. calamistis includes more than 20 species, while B. fusca, with exception of wild sorghum species, is rarely found on wild host plants (Gounou and Schulthess 2004; Le Rü et al. 2006). Therefore, the relatively high egg parasitism rates in Taveta and Kiboko could be explained by the prevalence of the non-diapausing S. calamistis. This suggests that the presence of significant populations of suitable, non-diapausing non-target species would increase the chance of establishment of T. isis in East Africa.

In western Africa and Kenya, T. busseolae was the predominant parasitoid species and produced considerably higher parasitism rates than any of the other species. However, this could not be explained by the biotic potential of the parasitoid alone. For example, at 28°C with S. calamistis as the host, intrinsic rates of increase (r _m) were 0.359, 0.278, 0.266 and 0.205, respectively, for Tr. bournieri, T. isis, T. busseolae and L. ovicidae (Chabi-Olaye et al. 1997, 2001a, 2004; Bruce et al. 2006). In the present study, parasitism efficiency was significantly higher in Telenomus than Trichogramma spp. High parasitism efficiency was also reported for Telenomus spp. by Schulthess et al. (2001) and Agkoba et al. (2002), and for the trichogrammatids L. ovicidae and Tr. bournieri by Chabi-Olaye et al. (2004) and Bruce et al. (2006), respectively. It was suggested that these parasitoid species tended to fully exploit discovered egg batches, but the results from the current study showed that egg batches with more than 50 eggs were in most cases only partially parasitized by Trichogramma spp. Thus, the low parasitism by the two trichogrammatids in the field might be the result of egg limitations, which allowed them to parasitize fewer egg batches, as also indicated by the low discovery efficiency when compared with T. busseolae.

Furthermore, differences in parasitism between parasitoid species could be explained by the host specificity and host finding ability of a species. With the exception of the crambid Coniesta ignefusalis (Hampson), which is attacked by T. busseolae, T. busseolae and T. isis only parasitize eggs of noctuids, which are laid hidden between the leaf sheath and the stem. By contrast, the other parasitoid species are also reported from eggs of species of various lepidopteran families that oviposit on the lower or upper surfaces of leaves (Polaszek 1998). In general, the Telenomus spp. appeared to be more specialized than the trichogrammatids; T. isis was only obtained from noctuid eggs, while the hosts of trichogrammatids included species of up to four families (Polaszek 1998). Fiaboé et al. (2003) showed that the sex pheromones of S. calamistis serve as chemical cues that help T. busseolae and T. isis in restricting the search area and thereby direct the parasitoid to areas where host mating is in progress or where oviposition has taken place. Once the oviposition site is discovered, contact kairomones, produced by the female moth, are directly associated with the attack on host eggs (Chabi-Olaye et al. 2001b). Furthermore, T. busseolae and T. isis were significantly more attracted to calling virgin female S. calamistis than L. ovicidae. According to Kaiser et al. (1989), the higher the level of specialization of a parasitoid the more it should be able to detect the appropriate host.

Based on comparisons of egg retention capacity under host deprivation situations and the ability to parasitize older eggs, Chabi-Olaye et al. (1997, 2001a) proposed that T. busseolae and T. isis both may be adapted to areas with strong environmental fluctuations, but that T. busseolae was much better adapted to long dry seasons than T. isis. Telenomus busseolae was also reported from C. ignefusalis in the Sahel (Youm 1990), whereas T. isis does not attack this borer species (Chabi-Olaye et al. 2001a). It was suggested that while T. busseolae was originally a savanna species, adapted to long dry seasons and situations of host scarcity, T. isis had evolved in humid forest habitats where the dry season lasts only 2–3 months. This was corroborated by Chabi-Olaye et al. (2006) who found that in the inland valleys in the forest zone of Cameroon, where both alternative host plants and maize maintain noctuid populations all year round, T. isis outcompeted T. busseolae. Chabi-Olaye et al. (2001a), based on the reproductive strategy of T. isis and the availability of host eggs during the off-season, predicted that the parasitoid would establish in mid-altitude areas in ESA. Wundanyi, T. isis was released in Kenya in 2005, is situated between the mid-altitude and highlands, and where the rainfall pattern is bimodal. Recovery of T. isis from Wundanyi appears to validate the prediction by Chabi-Olaye et al. (2001a). However there exists a network of rivers in Wundanyi, which enables farmers to cultivate maize throughout the year. It is suspected that, due to the continuous cropping of maize, a part of the B. fusca population does not diapause, as shown by Chabi-Olaye et al. (2006) for the inland valleys in Cameroon, which should facilitate establishment of T. isis.

In the laboratory, T. isis accepted and parasitized eggs of 12 noctuid borer species (i.e. seven Sesamia, two Busseola, two Manga and one Sciomesa) (Bruce, 2008) found in cultivated and wild habitats (Le Rü et al. 2006), which could serve as alternative hosts in areas where B. fusca is the main borer species and maize is not grown during the off-season, such as the high production zones in the highlands. However, further studies are needed to assess if T. isis attacks eggs of those borer species in their natural habitats.