Effect of Three Bean Species on the Development and Reproduction of a Population of the Parasltold, Encarsia bimaculata, on the Whitefly, Bemisia tabaci

Developmental time, parasitism, emergence, longevity, fecundity and demographic parameters of population of Encarsia bimaculata Heraty and Polaszek (Hymenoptera: Aphelinidae), a parasitoid attacking Bemisia tabaci (biotype B) (Gennadius) (Homoptera: Aleyrodidae) infesting soybean, Glyine max L. (Merr), cowpea, Vigna unguiulata L. and garden bean, Phaseolus vulgaris L. (Fabeles: Fabaceae) were quantified and compared. Encarsia bimaculata was able to complete its life cycle independent of the B. tabaci instar parasitized. However, parasitoid development was significantly slower when first (19 d), second (15 d) instars or pharate adults (14 d) were parasitized compared to the third (13 d) or fourth (13 d) instars. Consequently, percent parasitism was higher when the third (51 %) or fourth (46 %) instars were parasitized compared to the first (22 %), second (25 %) instars or pharate adults (36 %) of B. tabaci. Similarly, percent parasitoid emergence was significantly higher when third (83 %) or fourth (76 %) instars were parasitized compared to when the first (34 %), second (64 %) or pharate adults (54 %) were parasitized. Host plant species significantly influenced egg to adult developmental time, percent parasitism and the day on which E. bimaculata nymphs hatching from eggs was first observed. More nymphs were parasitized on cowpea (40 %) followed by garden bean (36 %) and soybean (32 %), while percent hatching was significantly higher on soybean (76 %) followed by cowpea (68 %) and garden bean (42 %). Adult parasitoid females lived an average of 6.7 d on soybean, 7.6 d on cowpea and 7.2 d on garden bean and laid a lifetime average of 27 eggs on soybean, 31 eggs on cowpea and 30 eggs on garden bean. The daily mean fecundity of E. bimaculata was not significantly different on the three bean species. Life table parameters showed that the net reproductive rate (Ro) was 14.50, generation time (Tc) was 17.16, intrinsic rate of natural increase (rm) was 0.16, finite rate of growth (λ) was 1.17 and doubling time (Td) was 4.44 for parasitoids on soybean. On cowpea, Ro was 15.32, Tc was 18.59, rm was 0.15, λ was 1.16 and Td was 4.72, while, on garden bean, Ro was 8.95, Tc was 19.28, rm was 0.11, λ was 1.12 and Td was 6.08. Given these life table parameters, higher population build up of the parasitoid will be expected on cowpea and soybean, respectively, compared to garden bean. Thus, for an effective augmentative release program involving E. bimaculata for the control of B. tabaci, it is important to take into consideration both the host stage of B. tabaci and the nature of the host plant on which it is developing.


Introduction
The sweet potato whitefly Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae) biotype B, is a major pest and plant virus vector attacking a wide variety of food crops worldwide (Perring 2001). The expolsive increase in B. tabaci population has been attributed in part, to the heavy application of pesticides to control it that in turn has resulted in the development of pesticide resistance and in the decline of the whitefly's natural enemies (Van Lenteren 2000). Therefore, it became important to promote the use of effective biological control agents that can efficiently control the pest, are safe to the environment and are acceptable to the farmers and greenhouse growers (Heinz 1995).
Parasitoids of the genera, Amitus, Eretmocerus and Encarsia are among the most important natural enemies of the B. tabaci complex (Cock 1996, Hoelmer 1995. Heinz and Parella (1998) showed that the percentage of parasitoids successfully developing to adults was greater for Encarsia formosa than for Eretmocerus sp., independent of the host plant. Collection records and descriptions of parasitism of Bemisia spp. worldwide indicate differences in guild structure and parasitism intensity depending on the species of host plant sampled and host plant age and size (Hu and Vinson 2000;Hu et al. 2002).
A study on the effects of two different host plant species, each infested with B. tabaci, on the fitness of two populations of Eretmocerus sp. was conducted by Powell and Bellows (1992). The study showed significant differences in preimaginal (development during the egg stage) developmental rate, survival and fertility between host plant species. Also, Liu and Stansly (1996) described the effect of B. tabaci age at the time of parasitization on the number of egg oviposited, egg to adult development and survivorship of Encarsia pargandiella.
The objectives of this study were to compare reproduction of Encarsia bimacalata attacking various stages of B. tabaci infesting soybean, Glycine max L. (Merr), cowpea, Vigna unguiculata L. and garden bean, Phaseolus vulgaris L. (Fabeles: Fabaceae) and to see how best host plant resistance and biological control can be integrated to suppress whitefly population.

Plant materials
Seeds of G. max, P. vulgaris and V. unguiulata were obtained from Guangdong Agricultural Institute in Guangzhou, South China. The seeds were placed in Petri dishes with water to initiate germination. The partially germinated seeds were grown individually in 12 cm diameter plastic pots and used in the experiment at the 4-6 leaf stage. These pots were placed into cages (60 x 60 x 60 cm).

Host and parasitoid culture
Adult B. tabaci were collected from cucumber plants in the field and released into cages (60 cm x 60 cm x 60 cm) containing cucumber as the source host. These cages were placed in a temperature controlled room maintained at 25 o C with a relative humidity of 70 % under a photoperiod of 14:10 (L:D).  (sedentary first, second, third, fourth and pharate adults) were independently exposed to 3 mated E. bimaculata parasitoids held in a clip cage. Bemisia tabaci nymphal instars were identified by measuring the body length of the whitefly, and young instars were selected based on their relatively flat appearance (Gelman et al. 2002b). All whitefly nymphs that were not of the desired instar were removed from the leaf before the parasitoids were released. This was done to ensure that all the exposed instars were of the same age. Clip cages with parasitoids were removed after 24 hours and the soybean leaf was left in the cage to trap emerging adult flies. On each subsequent day, several randomly selected B. tabaci nymphs were dissected until the egg, first, second and third instars of E. bimaculata were obtained. Parasitoid developmental time was counted from the day the female parasitoid was released. Egg to first instar developmental time was determined when the first instar was observed during dissection. Developmental times for second instar, third instar, prepupa (yellowish-white instar preceding the appearance of black wasp pupa, with two large meconia on either side of the host puparia), and pupa were determined in the same way. For each instar parasitized, at least 10 parasitized whitefly nymphs were dissected each day after parasitization until parasitoid adults emerged. When a wasp pupa was formed, it was removed and placed in a petri dish. The developmental time from egg to pre-pupa, pre-pupa to pupa and pupa to adult E. bimaculata eclosion was recorded separately on all whitefly nymphs (sedentary first, second, third, fourth and pharate adults). Parasitoid total developmental time (egg to adult) for each instar was also recorded from the same observations. Percent parasitism was estimated for all B. tabaci instars based on the number of nymphs in which parasitoid eggs were found following dissection. Percentage of successful adult parasitoid eclosion were also estimated based on the number of adult parasitoids that eclosed and the total number of whitefly nymphs that were parasitized (which was only an estimate of total parasitism) separately for each host stage of B. tabaci. All experiments were repeated six times for each B. tabaci host stage.

Developmental time and parasitism with respect to host plant
Developmental time and parasitism of E. bimaculata on the three test plants was achieved using the same procedure as in the age specific and developmental time experiment except that B. tabaci third instar was used as host for the parasitoid on the three test plants. Approximately 40-50 unsexed B. tabaci adults collected from the culture were released into each clip cage placed on the leaves of the test plants, soybean, cowpea and garden bean. After 24 hours, all adult whiteflies were aspirated from the clip cages and the area of the leaf covered by the clip cage was marked with indelible ink. The egg infested plants were confined in cages (60 cm x 60 cm x 60 cm) and immatures were allowed to develop to the third instar stage which is the most preferred stage for parasitization by Encarsia sp. (Donnell and Hunter 2002). When the desired third nymphal stage was reached, 40-50 -nymphs were exposed to 3 mated E. bimaculata parasitoids held in a clip cage. Clip cages with parasitoids were removed after 24 hours and the plants left in cages. Developmental time, percent parasitism and percentage emergence on the three host plants were recorded as described in the age specific experiment. All experiments were repeated 6 times for each host plant.

Female adult longevity and fecundity
To determine female fecundity and longevity, a newly emerged E. bimaculata female from the same colony was mated and confined by means of a leaf cage (2.5 cm diameter) to the abaxial side of a fully expanded host plant leaf containing between 15-20 whitefly nymphs (third instar). The female was transferred to a fresh leaf every 24 hours until death to determine the fecundity (number of eggs laid by female parasitoids over her lifetime). After each day, the leaves of the test plants with B. tabaci nymphs that were exposed to the parasitoid were examined under a dissecting microscope for oviposition. Mortality was registered daily. All the host plants with nymphs that were exposed to the parasitoid were kept in cages until the parasitized nymphs contained wasp pupae. The number of eggs laid in the life time of each female was determined based on the number of wasp pupae formed and from nymphs in which the parasitoid immature died before reaching the wasp pupal stage. Leaves with wasp pupae from each host plant were enclosed in Petri dishes with the top covered with plastic for adult emergence and subsequent sex determination. The parasitoids that emerged from the pupae were counted and sexed as described by Heraty and Polaszek (2000). Survivorship and number of eggs laid each day by the F1 generation were recorded. Fecundity and longevity data were used to calculate daily and lifetime fecundity of E. bimaculata. Twenty parasitoids were tested for each host plant with each host plant replicated 6 times.

Statistical Analysis
Data for developmental time (egg deposition to adult emergence), percentage parasitism, and percentage emergence on different B. tabaci instars and the three host plants were analysed using one-way analysis of variance (ANOVA); lifetime fucundity (eggs laid and hatched over the female's lifetime), daily fecundity (eggs laid and hatched per day), and longevity on the three host plants were also analysed using one-way analysis of variance (ANOVA). Means were compared using Fisher's Least Significant (LSD) test at 0.05 level of significance (SAS Institute 2001).

Results and Discussion
Developmental time of E. bimaulata with respect to B. tabaci host age In a non-choice arena, E. bimaculata was capable of parasitizing all nymphal stages of B. tabaci and was able to complete development through to the adult stage.
The developmental rate of E. bimaculata varied with host age (instar parasitized) ( Table 1). The duration of parasitoid development from oviposition to adult emergence was significantly different (F 29,60 = 139.28, P = 0.0001) longer when the first instars were parasitized than when the second, third, fourth and pharate adults were parasitized (Table 1). There was no significant difference in developmental times when third or fourth instars were parasitised. The durations of individual parasitoid stages were also influenced by host age at the time of parasitization (Table 1). Egg to pre-pupa (F 29,60 = 290, df = 4 , P = 0.0001), pre-pupa to pupa (F 29,60 = 140, df = 4 , P = 0.0001) and pupa to adult (F 29,60 = 55.92 , df = 4, P = 0.0001) developmental times were significantly different depending upon host age. Developmental times for these three stages were always significantly longer when first instar whiteflies were provided for oviposition (Table 1). The lengthening of the egg and larvae developmental times in younger B. tabaci instars has also been reported for other heteronomous aphelenides (Walter 1983 However, these stages developed significantly faster when the third or fourth instars were parasitised than when first, second or pharate adults were parasitized (Table 1). Similar findings were reported by Donnell and Hunter (2002); who observed significantly shorter developmental time and more synchronous adult emergence of E. formosa when third and fourth instars were parasitized. For egg to adult column (Table 1) values do not appear to be significantly different from those reported by Hu et al. (2002) (standard error overlap) which is surprising considering the reported influence of host plant species and the fact that different wasps were used.
The general preference of E. bimaculata for older instars may be related to certain advantages these instars confer to both the ovipositing females and the would-be developing immature. Firstly, the female can assess the host resources exactly with regard to size, quality and suitability at the moment of oviposition. Secondly, eggs laid in the earlier instars will suffer the same age-related mortality risks such as predation and disease. Thirdly, shorter developmental time in the late instars reduces the period for which the offspring is susceptible to hyperparasitism. Nechols and Tauber (1977) reported that wasps that begin development in third and fourth instars nymphs exhibited higher survivorship and developmental times were shortened by approximately 38 %.
Percentage parasitism and emergence of E. bimaculata with respect to B.tabaci host age. Percent parasitism was also influenced by host age (Table 2). Percent parasitism was significantly (F 29,60 =3097 df = 4, P = 0.0001) higher when the third or fourth instars of B. tabaci were parasitized than when the pharate adults, second or first instars were parasitized ( Table 2). The mean percent parasitism across the various instars were not much different from the values reported by Antony et al. (2003) for E. transvena parasitizing B. tabaci instars. These authors observed that even though E. formosa oviposited in all B. tabaci instars, the third and fourth instars were mostly preferred for oviposition compared to the pharate adults or younger instars. Lopez-Avilla (1988) also reported that E. formosa, E. luteola, E. adrinae and E. cibcenses parasitized all instars of B. tabaci but the third  Nell et al. (1976) reported that wasp may use their antennae to obtain olfactory and resonance information about hosts, and this information coupled with stimuli received while making 180 o turns on the dorsum of the nymph may be used to determine host size.
When E. bimaculata parasitized third and fourth instars of B. tabaci, successful adult parasitoid emergence was significantly (F 29,60 = 124.20, df = 4 , P = 0.0001) higher when third or fourth instars were parasitised as compared to when first, second or pharate adults were parasitised (Table 2). Similar observations were reported by Antony et al. (2003) for E. transvena parasitizing B. tabaci instars. Nell et al. (1976) also reported higher emergence when E. formosa parasitized third and fourth instars of T. vaparariorum. The low emergence of parasitoids in the younger instars of B. tabaci could be related to the high percentage of age-related immature mortality.

Developmental time of E. bimaulata with respect to host plant species
Parasitoid development can be influenced by the nature of plant secondary compounds eaten by their herbivorous hosts (Awmack and Leather 2000). Thus, variation in secondary compounds among individual plants may cause variation in development and successful parasitism of herbivores by specialist parasitoids such as E. bimaculata. With respect to the bean species, the duration of parasitoid development on the third instar of B. tabaci from oviposition to adult emergence was significantly (F 29,60 = 45.80, df = 2, P = 0.0002) longer for parasitoids that developed on garden bean compared to cowpea or soybean ( Table 3).
The mean developmental time of the parasitoid on cowpea and garden bean were not much different from the range (15-15.80 d) reported by Shishehber (1996) for E. formosa at the same temperature. On the other hand, the mean value recorded on soybean was slightly different from the range (16-18  Barro et al. (2000) for E. bimaculata on soybean under a temperature regime of between 22-30 o C. Similarly, Arakawa (1982) reported that E. formosa raised on T. vaporariorum showed differences in developmental time of between 15 days on tobacco, egg plant, and tomato compared with 24.5 days on poinsettia at a mean temperature of between 22.5-25 o C.

Percentage parasitism and Emergence of E. bimaculata with respect to host plant
The role of plants in the interaction between parasitoid and hosts has been known for some time (Price et al. 1980;Turling et al. 1995). Thus, the host plant species on which the parasitoid develops plays a significant role in determining the level of parasitism in subsequent generations.
Comparatively, more nymphs were parasitized on cowpea followed by garden bean and soybean (Table 4); but, the differences were not significant. Similar observation was reported by Goolsby et al. (1998) in which Encarsia sp. were found to parasitized more nymphs on rockmelon, cotton and hibiscus than tomato and soybean. The reason for the differential parasitization rates on the three bean species may be related to slight differences in leaf morphology among the bean species particularly between soybean and the other two bean species.The leaves of the soybean species used in the study were covered with hairs (trichomes) which prevent the underside of the whitefly nymphs from fitting level to the surface of the leaf thus making it difficult for E. bimaculata to oviposit on the nymphs (Headrick et al. 1996). Also, these hairs may interfere with the movement of the parasitoids thereby reducing parasitism (Headrick et al. 1996). This may account for the low parasitism rate on soybean compared to the other species. On the other hand, the leaves of the other two species were relatively smooth, hence the margin of the whitefly nymph's body can fit level with the leaf surface making it easier for the parasitoid to oviposit. The slight difference in percentage parasitism between cowpea and garden bean may be attributed to the comparatively larger veins on garden bean that might interfere with the foraging behaviour of the parasitoid ).
When third instars of B. tabaci were presented to E. bimaculata for oviposition, percent adult parasitoid emergence was significantly (F 29,60 = 3.96, df = 2, P = 0.0480) different (Table 4). The highest emergence rate was recorded on soybean followed by cowpea and garden bean. This observation is in accordance with studies carried out by Reed et al. (1991), which showed that plant species or cultivars identity may influence the proportion of emerging parasitoids. The comparatively higher emergence rate of the parasitoid on soybean and cowpea, respectively, compared to garden bean may be attributed to the high immature mortality of the parasitoid on garden bean. The negative effect of a cultivar exhibiting antibiosis on the pre-imaginal mortality of parasitoids has also been reported by Van Emden (1991) and Farrar et al. (1994). It is well-known that secondary substances ingested by phytophagous insects may have negative effects on the biological characteristics (survival, fecundity and development) of parasitoids (Godfray 1994). It has been demonstrated that parasitoids are generally more sensitive to toxic compounds than their polyphagous hosts, as they appear to be incapable of metabolizing the different concentrations of plant secondary compounds present in their host (Turling et al. 1995).

Adult longevity and fecundity
Adult longevity was recorded as part of the fecundity experiment. Although most female parasitoids died by day 9, 11 and 10 on soybean, cowpea and garden bean, respectively, some females were long lived, surviving to 11 d on soybean and 12 d on cowpea and garden bean (Figure 1). For each experiment, 20 female parasitoids were utilized. A one-way ANOVA followed by the Fisher's Least Significant (LSD) post hoc test was used to determine if there were significant differences in fecundity and longevity. Means in column with the same letter are not significantly different at P > 0.05 (LSD). The day on which adult parasitoid emergence was first observed was designated day 1 ("onset") and the corresponding survival was 1 ("100 percent survival"). Percent adult survival for day 1 and for each succeeding day = number of adults that survived on a given day/total number of original cohort that survived. Standard errors have not been indicated, but this value was about 10% of the point's value. High quality figures are available online.
There were significant (F 29,60 = 4.15, df = 2, P = 0.0439) differences in the mean longevity of E. bimaculata on the three bean species (Table 5). Encarsia bimaculata lived longer on cowpea and garden bean as compared to soybean (Table 5).
The mean longevity on the three bean species were lower than the value (10.4 d) reported by Lopez and Botto, (1995) for local populations of E. Formosa on tomato. These means were however, higher than the value (5.4 d) reported by Powell and Bellows (1992) at 29 o C for an Eretmocerus species (likely E. eremicus), on B. tabaci on cotton. The host plant species we used (soybean, cowpea and garden bean) were different from those used by Lopez and Botto (tomato) and Powell and Bellows (cotton). Also, Powell and Bellows used a different wasp species (Eretmocerus sp). A combination of the above factors might have resulted in the discrepancies between our results and those of the other authors. The relatively higher survivorship of the parasitoid on cowpea and garden bean respectively compared to soybean could be attributed to differences in both surface structure and chemistry among the three bean species.
Female E. bimaculata laid an average 26.95 eggs during their lifetime on soybean, 31.10 eggs on cowpea and 30.05 eggs on garden bean. These means were within the reported range for other aphelinids (Viggiani 1984). The means were significantly different (F 2,57 = 45.80, df = 2, P = 0.0002) on the three bean species (Table 5). The daily mean fecundity were however not significantly different (F 4,57 = 0.3907, df = 2, P = 0.6937) on the three bean species (Table 5).

Life table
Results from development and fecundity experiments were used to develop lx-mx life tables for E. bimaculata on the three bean species. These tables were used were used to calculate the demographic parameters shown in table 6. The sex ratio of E. bimaculata was 0.7087.
The net reproductive rates (R o ) of E. bimaculata on soybean, cowpea and garden bean are shown in Table 6. The values recorded on soybean and cowpea were higher than that on garden bean. These values for soybean and cowpea were higher than that reported by Powell and Bellows (1992) for wasps attacking B. tabaci on cotton (11.74) but, all recorded values were lower than the value (26.50) reported by the same authors for wasps reproducing on cucumber. The low reproductive rate of the parasitod on garden bean indicates that less parasitod females were produced on it compared to soybean and cowpea, respectively.
Generation times (T c ), the finite rate of increase 45!"$-7"71<?@ , -2"., A /"(T d ) on soybean, cowpea and garden bean are shown in Table 6. B1<?@ , -2"., A /% were not much different among the bean species but, were lower on soybean followed by cowpea and garden bean. The recorded values for 5"1soybean and cowpea were slightly different from garden bean The most important parameter calculated from the life table is the intrinsic rate of increase, r m , which is the rate of increase per individual in an environment where fecundity and survivorship are maximal in the absence of external mortality factors. The larger the r m value, the greater the potential of a species in reproducing and increasing in number within a given environment. This parameter allows for the comparison of r m among species, and it also facilitates the evaluation of a parasitoid concerning its use in strategies where biological control is an option (Jansen and Sabelis 1992). The r m value for soybean, cowpea and garden bean are shown in Table  6. These values are lower than the value (0.28) reported by Lopez and Botto (1995) for E. formosa. Small differences in r m values can make remarkable differences in expected population growth over time, therefore, to compare the population growth of the parasitoid on the three bean species over time, the exponential equation for population growth N t = N o e rt was used, where N o is the initial number of parasitoid on the plant, N t is the number of parasitoid at time t, r m the intrinsic rate of increase and t the time in days. Given a stable age distribution, the parasitoid population on soybean, cowpea and garden bean were calculated as 211.43, 236.38 and 81.11, respectively, within two generations.
Given these life table parameters, parasitoid populations should build up relatively slowly on garden bean compared to the other two bean species. Thus, the parasitoid is comparatively a poor biological control agent on garden bean compared to the other bean species. This conclusion however, cannot be applied to all populations of parasitoids on the three bean species used as the effectiveness of a parasitoid as a biological control agent is dependent, among other factors, on the morphological and chemical composition of the host plant and on the identity the parasitoids population.