BIOLOGICAL CONTROL OF CITRUS WOOLLY WHITE FLY ALEUROTHRIXUS FLOCCOSUS MASK (HOMOPTERA: ALEYRODIDAE) BY CALES NOACKI HOW (HYMENOPTERA: APHELINIDAE) IN SOME TANZANIAN SMALL-SCALE CITRUS ORCHARDS

Cales noacki How (Hymenoptera: Aphelinidae) wwas released in citrus groves in Muheza (05°10' 0S, 38°46' 0E) and Morogoro (06° 49' 0" S, 37° 40' 0 E) Townships to control Aleurothrixus floccosus Mask (Homoptera: Alleyrodidae). Sleeve cage and “free release” methods were used to introduce the parasitoid onto selected citrus trees. C. noacki adults and mummies were released at selected sites between 08:00 and 09:30 am on 14th September 1999. Adults were first observed 90 days after introduction. They were also recovered from 2nd and 3rd CWWF instars. At Morogoro, CWWF infestations by adults, eggs and nymphs were reduced by between 6 and 12 fold 90 days after release and between 30 and 300 fold 60 months after. At Muheza decreases were between 140 and 700 and 12 and 17 fold, respectively compared. Parasitoid recovery indicated its establishment. Recoveries were also on trees up to two kilometres away from release sites. Parasitoid release was also accompanied by vigour improvement of previously heavily infested citrus trees. This is the first report of successful use of C. noacki for control of A. floccosus in Tanzania, which was prevented from spreading from the original small infestation foci.


INTRODUCTION
The citrus woolly white fly, Aleurothrixus floccosus Mask (CWWF) was accidentally introduced into Eastern Africa, probably from the Mediterranean region and North Africa, where citrus is an important cash crop. In this region, the pest had earlier caused severe crop losses in the 1970s in Morocco (Abbasi, 1975), Egypt (Vulic and Betran, 1977) and more recently in Ethiopia (Alemu et al, 2014).
The presence of A. floccosus in Tanzania was first reported near Tengeru in Arusha, by Brigitte Nyambo in 1993, and later at Morogoro around Sokoine University of Agricultural Sciences (Lohr, personal communication). Specimens collected by B. Nyambo were later identified by CIE as the citrus woolly white fly. Since then, the pest has been reported in some citrus-growing areas of Morogoro Township (Mwaiko and Kyamba, 1995). A new isolated infestation was also discovered in Muheza Township in Tanga region (Seguni, 1998). Sporadic pest presence was also reported in small areas of Kilimanjaro region (Lohr, 1997) and Shinyanga and Dodoma (Mfugale, personal communication). However, the pest was not detected at Kyela, which neighbours Malawi where CWWF has been reported (Lohr, 1997). The pest was also discovered in small numbers on orange and lemon trees at Wanging'ombe in Njombe region (Seguni own observation, 2003).
The presence of CWWF in Morogoro and Muheza districts in particular posed a potential threat to the Tanzanian citrus industry if the new pest would not be immediately controlled in the few and small infestation foci. Moreover, the pest had become a new element in the indigenous homopteran pest complex, for which no sustainable control measures are undertaken. The presence of the pest in Morogoro and Muheza Townships, both surrounded by important citrus growing areas producing more than 250,000 metric tons annually (DALDOs Office Muheza District Council, 2016), was threatening the heart of the citrus industry which is a dependable source of district revenues and rural economies of these areas. CWWF affects citrus trees by the adults' and nymphs' feeding effects on tender leaves thus depriving the plant of vital manufactured nutrients. However, more significant is the sooty mould that develops on leaves on residues of honeydew that is produced copiously by the pest, which reduces significantly the photosynthetic ability of infested plants, leading to leaf fall, senescence and eventual death of severely infested plants. Control of CWWF by various insecticides has been attempted (David 2002) but is largely impractical (Lohr, 1997) because the infestation occurs wholly on the underside of leaves where accessibility to insecticides is limited. Moreover, the frequency of sprays required for effective control of the pest would make the method expensive and unacceptable environmentally. In addition, most smallholder farmers growing most of the citrus are unable to purchase insecticides and the spraying gear.
The use of the oligophagous parasitic wasp, Cales noacki How (Hymenoptera, Aphelinidae) (Carl, 1995) is an effective and sustainable biological control method against A. floccosus. The parasitoid was earlier used successfully against the pest in the Mediterranean region, saving the industry from virtual collapse (Carrero, 1975;Magalhaes, 1980;Liotta and Maniglia, 1984). In East Africa the parasitoid was used successfully in Uganda and Kenya in 1996 (Lohr, Seif and Nang'ayo, personal communication). In both countries, the parasitoid established and spread in citrus groves very quickly after only a couple of months, effecting good control over CWWF. To curb spread of the pest from spreading from the relatively small foci a research project to introduce the parasitoid was executed between 1999 and 2004.

METHODS
Colonies of CWWF for raising C. nocki were reared on caged young potted orange seedlings at the Kibaha Biological Control Station (KBCS) in April 1999 from cultures of adult insects collected from citrus groves in Morogoro between April and May 1999. By June 1999, wellestablished CWWF colonies in the screen house were ready for inoculation with C. noacki. Between May and June 1999, baseline surveys of CWWF infestation levels were conducted in the Morogoro and Muheza Townships and surrounding areas. The surveys were also aimed at selecting suitable sites for initial release of the parasitoid. A shipment of C. noacki was made in Mid September 1999, part of which was reared on CWWF in the screen house at KBCS and part introduced directly in the field at Morogoro and Muheza (Seguni, 1999).

2.1.FIELD EVALUATION OF CWWF INFESTATION AND INCIDENCE OF OTHER CITRUS PESTS AND NATURAL ENEMIES
Evaluation of CWWF population was according to Lohr, (1997); Mwaiko and Kyamba, (1995) and Seguni, (1998). Samples of 10 trees were taken randomly per orchard, but in an area where citrus was scattered, individual trees were considered along roadside transects. The same trees were used for the survey throughout the period of the experiment. From each of the four poles of a tree i.e. north, south, east and west, four leaves: two tender and two older, were inspected for about 5 minutes for different life stages of CWWF. Adult white flies were quickly counted using a counter while eggs and larval instars were estimated according to the area on leaf undersides each of the sedentary larval stages covered as follows: from zero indicating absence to 100 indicating entire leaf surface covered. Old CWWF attack was recorded as present (1) or absent (0). Other pests and natural enemies on the trees were noted down and where possible, classified to genus level or simply by their common names.

2.2.FIELD INTRODUCTION OF C. NOACKI
Both the sleeve cage (Plates1and 2) and the open or 'free release' methods were used in releasing adult parasitoids and mummies respectively. Grease was smeared on both ends of the branch on which the cage was mounted to prevent noxious insects, particularly ants, from getting into the cage. Cages were untied two weeks after parasitoid release. A total of 120 and 180 C. noacki adults in glass vials from icipe were transported in cool boxes to release sites in Muheza (05°10' 0S, 38°46' 0E) and Morogoro (06° 49' 0" S, 37° 40' 0" E), respectively. In addition, the release at Muheza also included mummies on three infested leaves. The glass vials were supplied with thinly honey-smeared strips of paper as food for the adult wasps. The cool boxes stocked with blocks of ice were aimed at reducing the parasitoids' temperature and thus their metabolic activity in order to increase their survival prior to release.
Parasitoids at Muheza were released by one team of researchers on two orange trees at the district headquarters and by a second team of researchers on five orange trees at the Prison's the parasitoid. Direct quick counts of adult parasitoids were done on CWWF-infested leaves in the crown of ten sample trees by two observers using counters each working on one half of the crown. The numbers of observed adults in a period of five minutes were added up for each tree and then pooled for all observed trees and expressed as adults per tree. In addition, mummified CWWF were collected from infested trees and placed in paper bags for incubation in the laboratory at ambient temperatures for three to four days, during which time emerging parasitoids were counted under a stereomicroscope. Most reliable information, however, was by direct tree crown observations due to scarcity of mummified CWWF at sampling time resulting from depletion of CWWF by C. noacki. Data was analysed by the chi square test at the 95% probability level.

3.1.DEVELOPMENT OF CALES NOACKI POPULATION IN CITRUS ORCHARDS WHERE IT WAS INTRODUCED IN MID SEPTEMBER 1999 IN MUHEZA AND MOROGORO TOWNSHIPS
Adults of C. noacki were first sampled in detail at both release sites at the end of December 1999, nearly 90 days after its initial release ( Table 1). The population increased most rapidly between May 2000 and March 2001, especially at Muheza. Thereafter, the population of adult parasitoids decreased at both sites whereby in April 2004, sixty months after initial release of the parasitoid, none could be found on citrus. The decrease in the number of parasitoids corresponded with a marked decrease in all stages of CWWF (Table 3). If the two sites are compared, the population of C. noacki increased faster at Muheza than at Morogoro during the same observation period (Table 1). This could possibly have been attributed due to higher temperatures at Muheza (150-200 masl) favouring faster parasitoid development compared to Morogoro (525 masl). CWWF and C. noacki population fluctuations were in line with pestparasite rhythms (Southwood, 1976). There must have been higher numbers of adult parasitoids in the tree crowns which were possibly naturally missed by the observers due to their minute size.

3.2.STATUS OF CWWF ON CITRUS TREES BEFORE AND AFTER RELEASE OF CALES NOACKI IN MUHEZA AND MOROGORO TOWNSHIPS
The status of CWWF prior to establishment of C. noacki is presented in Table 2, which shows that the pest was abundant both at Muheza and Morogoro. Eight months after initial and subsequent parasitoid releases, significant reduction in the population of all growth stages of CWWF occurred at both sites (

DISCUSSION
The trend of decreasing numbers of CWWF at both release sites corresponded with increasing numbers of C. noacki adults in citrus tree crowns, which suggested impact of the parasitoid on CWWF population. The changes in parasitoid and CWWF numbers were in general agreement with the pest-parasitoid dynamics, whereby to a certain degree, CWWF abundance favoured rapid parasitoid population build-up until at such level that low numbers of CWWF became a limiting factor for parasitoid multiplication, thus lowering its population. It is expected that the dynamic pest-parasitoid relationship will remain favourable unless some adverse factors are imposed such as farmers using insecticides unwisely thus upsetting the parasitoid numbers. It was not easy to relate directly the effect of reduced citrus infestation by CWWF due to establishment of C. noacki and fruit production because CWWF is not a direct pest. Moreover, in experimental areas citrus is most often subjected to a wide range of similarly damaging homopteran pests such as black flies, scales and aphids which were abundant on the citrus trees. However, at Muheza 7-10 years old orange trees that were prior to release of C. noacki heavily infested by CWWF and which had ceased fruit bearing, had resumed growth vigour and orange production (Plate 3).  Presence in the citrus cropping systems of abundant indigenous and exotic fauna of parasitoids and predators was an indication of a balanced agro-ecosystem, which should be maintained through periodic updating of the inventory of the economic pests and their natural enemies with a view to develop augmentative or classical biological control strategies, which in combination with cultural practices should sustain environment-friendly production of citrus in Tanzania.
It is not clear from our observations whether introduction of CWWF and its subsequent reduction by C. noacki has had an impact on the population and distribution of indigenous pests and natural enemies. However, it would be expected that a competition must have occurred between pests that similarly infested citrus such as between the black fly and CWWF. Occurrence in several citrus areas of E. serius was interesting and showed that biological control programs should be undertaken as regional initiatives because the wasp was initially released around Mombasa Kenya in the 1950s (Hill, 1975) and must have spread naturally into Tanzania from there.
There were some indigenous predators notably ladybird beetles that had become adapted to CWWF, but these have not been identified nor their effect on the pest evaluated. They are likely to be generalists. The major indigenous foraging ant species P. magacephala, Camponotus sp complex, A. custodiens and in a few cases Oecophylla longinoda, had become very well adapted to collecting honeydew from CWWF in addition to that produced by indigenous Homoptera such as aphids, scale insects, mealy bugs and black flies. These are likely to foster non beneficial associations with injurious homoptera.

CONCLUSION AND RECOMMENDATIONS
In conclusion, this research provided proof of successful establishment of C. noacki in experimental citrus groves and that the parasitoid had profound impact on CWWF which became severely reduced. Reduction in CWWF infestation pressure on heavily infested citrus trees had resulted in rejuvenation of the trees, which were brought back to vigour and production. The biological control effected by the parasitoid benefited small-holder citrus groves in release areas and beyond as the parasitoid spread spontaneously without further human intervention. However, to safeguard the parasitoid farmers have been urged to refrain from unwise use of contact insecticides.

ACKNOWLEDGEMENTS
Many thanks are to the GTZ IMP Project ICIPE Nairobi Kenya and the Ministry of Agriculture Livestock and Fisheries for financing the research activities. Thanks are also to different farmers and extension staff in project areas for cooperation. Last but not least to my colleagues in Pest Control Unit: Valentina Nyange, C. Materu, K. Lema and G. Mwingira and collaborators at KBCS in particular Joyce Kiboya for dedicated work. Last but not least I would like to regretfully m oan the untimely death of the co author who passed away sometime in 2014 when this paper was being prepared.