Biological Control of Mosquitoes with Mermithids ~ E. G. Platzer" Abstract: Mermithid nematodes parasitizing mosquitoes have substantial potential for vector control. Studies on the physiological ecology of Romanomermis culicivorax have defined some of the general requirements of mermithid nematodes and produced general guidelines for the ex- perimental release of mermithids in biological control. Experimental field studies have established the biological control potential of R. culicivo~ax, but further development and ulilization of this parasite will require a substantial commitlnent of scientific man-years and ftmds. Key words: Integrated Pest Management, physiological ecology, Roma*wmermis culicivorax.

The Phylum Nematoda has five orders maturing in adult mosquitoes. There, of 14 with 14 families of obligate parasites, host species 86% are Aedes sp. but only the have been found in natural populations of mosquitoes (29, HOST SPECIFICITY 30). Mosquitoes are also utilized by filarioids as intermediate hosts, but these nematodes In most cases, we have little information have no promise for biological control and on host specificity other than that the hosts will not be considered further. There are are generally mosquitoes. Some exceptions many reports of mermithids in are Hydromermis churchilliensis, which populations but, because the nematode naturally inlects chaoborid midges as well identifications were based on larval stages, as Aedes sp., and Romanomermis culici- they are often incomplete or inaccurate. vorax, which also can infect blackfly larvae However, there appear to be 15 acceptable ttnder abnormal conditions (30). These ex- species in eight genera (29,30). These mer- atnples appear to be extraordinary excep- mithids can be divided into two groups: tions to the specificity for mosquito larvae. those that complete the parasitic phase of Specificity within mosquitoes has received life primarily in tim larval stages of the mos- little attention. Perutilimermis culicis dem- quito host and those that complete most of onstrates species specificity and parasitized the parasitic growth phase in the adult mos- only Aedes sollicilans; StreIhovimermis quito (Table 1). In the first group there are peterseni demonstrates generic specificity fonr genera and ten species. Of the 133 and parasitized only sp. (13). natural and laboratory infected host species The most thorough examination of host listed, 53% are Aedes sp., 20% are Ano- specificity llas been carried out mainly by pheles sp., 19% are Culex sp., and the re- Dr. Petersen with R. culicivorax (17,30). mainder of tim hosts are species of Armi- More than 82 species of mosquitoes in 13 geres, Culisela, Deinocerites, Mansonia, genera have been infected in the laboratory Orthopodom3,ia, Psorophora, Toxorhyn- or found infected under natural conditions. chiles, Tripteroides, Uranotaenia, or Wy- R. culicivorax was unable to complete its eomyia (17,29,30). It may appear obvious development in five mosquito species: Aedes from this list that mermithids have had triseriatus, Anopheles sinensis, Culex ter- their greatest success as parasites of culicine ritans, Mansonia uniformis, and Psorophora mosquitoes, particularly those in tlm genus ]erox. Aedes. However, the list probably more ac- curately reflects the geographic distribution ENVIRONMENTAL LIMITATIONS of entomologists interested in biological Most investigations of environmental control organisms and, therefore, such an limitations of mermithids as biological con- assumption is probably premature. A sim- trol agents have used R. culicivorax, but the ilar assumption, and subsequent explana- observations are probably valid for other tion can be made for hosts of mermithids temperate zone mermithids. Temperature: Petersen (13) reported Recei~.-ed for publication 12 December 1980. that R. culicivorax was infective when air 1Symposium paper presented at the annual meeting of the Society of Nematologists, New Orleans, Louisiana, Au- temperatures were above 15 C. Subse- gust 1980. aDepartment of Nematology, Universilv of California, quently, Brown and Platzer (1) and Gallo- Riw'rside, CA 92521. way and Brnst (8) found that some infec-

257 258 Journal of Nematology, Volume 13, No. 3, July 1981

Table I. List of mermithids, their hosts, and sites of collection.*

Mermithid Mosquito host Locality

Mermithids maturing primarily in larval stages of the host: Hydromermis churchilliensis Aedes--3 sp. Manitoba Octornyomermis muspratti Aedes--8 sp. Zambia Anopheles--2 sp. Culex--5 sp. Octomyomerrnis troglodytis Aedes sierrensis California Romanomermis culicivorax Aeries--29 sp. Florida, Anopheles--3 sp. Louisiana Armigeres subalbatus Culex--18 sp. Guliseta--4 sp. Deinocerites pseudes Mansonia--2 sp. Orthopodomyia signifera Psorophora--7 sp. Toxorhynchites rutilus Tripteroides bambusa Uranotaenia--3 sp. Wyeomyia smithii R omanomermis corn munensis A edes--4 sp. Manitoba Romanomermis hermaphrodita Aedes--2 sp. Manitoba R omanomermis iyengari Anopheles subpictus India Rornanomermis kiktoreak Aedes--5 sp. Northwest Territories Romanomermis nielseni Aedes--5 sp. Wyoming Culex--2 sp. Strelkovimermis peterseni Anopheles--9 sp. Louisiana Mermithids maturing primarily in adult stages of the host: Culicimermis schahhovii Aedes--6 sp. USSR Culicimermis sp. Aedes--4 sp. Manitoba Ernpidomermis cozii Anopheles [unestus West Africa Paramermis canadensis Aedes--2 sp. British Columbia Perutilimermis culicis Aedes sollicitans Louisiana, New Jersey

*Derived from Petersen and Chapman (17) and Poinar (29,30) tions occurred at water temperatures of 12 pH: Petersen (15) found that R. culici- and 10 C, respectively. However, the op- vorax was fully infective from pH 5.4 to 7.9 timum temperature was in the 21-33 C (4.8 to 8.5 were transitory pH exposures). range. In addition, the optimal temperature Hence, pH of most natural waters should for development of the parasitic stage of not be a limiting factor. R. culicivorax was in the 20-32 C range. Salts: Petersen and Willis (20) reported Petersen (14) found that Romanomermis that mild salinity (0.04M NaC1) inhibited nielseni postparasites developed and com- the infectivity of R. culicivorax. This was pleted oviposition within 7 wk at 17 C. In- confirmed by Brown and Platzer (2) who creasing the temperature to 23 C suppressed also reported the hierarchy of ion toxicity oviposition but not postparasite develop- on a molar basis for R. culicivorax as fol- ment. In contrast, R. culicivorax required lows: cations, sodium < potassium < cal- 23 wk to complete oviposition at 17 C, but cium: and anions, chloride < carbonate = only 3-7 wk at 23 C. This information pro- sulfate < nitrate < nitrite < phosphate. vided a unique physiological comparison Therefore, it appeared that R. culicivorax between the two species. Temperature is an would not be an effective biocontrol agent important consideration in the use of R. in feedlot runoffs, fertilizer plant waste- culicivorax as a biological control agent; water, and brackish water situations. Re- its use in colder temperature zones is pre- cently, Petersen (16) has found that Octo- cluded (8). myomermis muspratti is tolerant of diluted Mermithids of Mosquitoes: Platzer 259 seawater (3,000-4,000 ~mhos/cm) and water aquatic ecosystems and combination treat- from tree holes (10,000 t~mhos/cm; high in ments are possible. organic matter). These findings suggest that Invertebrate predators: Most research O. muspratti has great potential for mos- on environmental limitations has been di- quito biocontrol in polluted waters. rected towards evaluation of abiotic factors. Oxygen: In investigations to determine Mitchell et al. (11) were the first to recog- why polluted water compromises the infec- nize that ostracods preyed on mermithid tivity of R. culicivorax, Brown and Platzer preparasites. Similarly, copepods and young (3) reported tile effects of lowered oxygen ganlmarids will attack preparasites (27), availability on the infective nematode. and the dynamics of copepod predation on Transient exposure to low 02 tension in- mermithids has been reported in detail (28). creased the survival, and thereby the infec- Both satiated and starved adult copepods tivity, of the preparasites of R. culicivorax. (Cyclops vernalis) attack preparasites rap- This effect was explained as resulting from idly in small volumes of water (less than 2 induced quiescence under low O2 tensions ml). This attack or predation rate is de- and hence greater viability under the test pendent on the volume of water, thus in- conditions. More recently, I have found dicating that searching or hunting behavior that the preparasites of R. culicivorax stop controls predation rate. In laboratory stud- moving within 8 h in water high in organic ies with larger volumes of water (one liter), content but low in O2 (6). Therefore, it was shown that mernlithid preparasite lowered oxygen tensions in polluted waters populations were reduced significantly by may be responsible for the inability of R. copepod densities of 20-100/liter. Although culicivorax to infect mosquitoes under such field studies have not been carried out, I conditions. suspect that copepod density might play a significant role in the success or failure of INTEGRATED PEST MANAGEMENT field applications. Chemicals: Integrated Pest Manage- Invertebrate predators aIso may reduce ment, the multipronged approach to pest the potential of establishment of aquatic control, encourages assessment of pesticide mermithids (27). Diving beetles and gam- effects on mermithid infections. The first marids, dragonfly and damselfly naiads, and assessment was that of Mitchell et al. (11) small crayfish attack mermithid postpara- in Taiwan, who found that the usual levels sites rapidly (27). lsopods, however, will of Abate, Dieldrin, and Gama HCH did attack postparasites only if other food is not adversely affect R. culicivorax infec- limited. Such predation may account for tions. Finney et al. (7) reported that Altosid poor or no establishment of the mermithid 5E, an insect growth regulator, at concentra- after field application of R. culicivorax. tions typically used for control of mosquito FIELD APPLICATIONS larvae did not interfere with any phase of the parasite's development. However, host Essentially all field releases of mermithid mortality was considerably increased when parasites of mosquitoes have been con- Altosid 5E and R. culicivorax were used in ducted with R. culicivorax, although in one combination on in labora- study, Petersen and Willis (22) released S. tory experiments. This finding suggests a peterseni and after 5 yr obtained more than unique chemical-biological control ap- 80% parasitization. In 1971 Petersen and proach in mosquito control: "If the chem- Willis (21) treated 10 natural sites 20 times ical don't get 'em, the nemas will!" Platzer with preparasites of R. culicivorax and ob- and Brown (25) reported that a variety of tained 65, 58, and 33% parasitism of sec- copper-based organic algicides and copper ond, third, and fourth instars, respectively, sulfate didn't compromise the infectivity of of the Anopheles sp.; 94% of second-instar R. culicivorax as long as the concentrations anophelines and 64% of all Anopheles were in the ranges usually used for algae were infected at rates of 1,000 prepara- and weed control. In summary, it appears sites/m ~- of surface area application. Appli- that R. culicivorax is tolerant of a variety cation rates less than 1,000 preparasites/m 2 of chemical pest control measures used in of surface area were less effective, but higher 260 Journal of Nematology, Volume 13, No. 3, July 19s1 application rates did not increase the num- Aedes taeniorhynchus, Psorophora colum- ber of infected mosquito larvae. Establish- biae, and Psorophora ciliata. In some cases ment of the nematode was observed in 7 of all larvae were killed within 24 h due to the 10 sites. multiple parasitism. Petersen et al. (19) treated fallow rice Levy and Miller (9) released prepara- fields in Louisiana infested with Psorophora sites into two unused sewage settling tanks confinnis and Anopheles quadrimaculatus. at 64,000 and 110,000 preparasites/m = of Increasing the application rate from 180 to surface area and obtained 37 and 54% in- 1,450 preparasites/yd ~ of surface area in- fection of C. pipiens quinquefasciatus, re- creased the parasitism of P. confinnis from spectively. Although these treatment levels 10 to 38%. They estimated that 3,900 pre- seem high, large numbers of C. p. quin- parasites/yd'-' would give 95% infection. quefasciatus larvae were present and host: Sixteen percent of A. quadrimaeulatus were parasite treatment ratios were approxi- infected at the rate of 181 preparasites/yd 2 mately 5:1 amt 3: 1. High concentrations of and 61% at 724/yd'-' of surface area. They phosphates, chlorides, and carbonates prob- estimated that 1,300 preparasites/yd 2 of sur- ably lowered the infectivity of tile prepara- face area would give 95% infection. Cau- sites, as much higher infection levels would tion should be used in making this type of I)e expected in this species at the dosages prediction. Brown et al. (4), for example, applied. found that increasing treatment rates from Petersen et al. (18) successfully con- 1,000 to 2£000 preparasites/m 2 of surface trolled Anopheles albimanus and Anopheles area did not significantly increase the num- punctipennis in Lake Apostepeque, E1 ber of Culex tarsalis larvae infected. Salvador, with R. culicivorax. The mos- Petersen and Willis (22) made a total of quito breeding areas, a 2-5 meter band 30 releases in 21 sites (15 at rates of 1,000 (total area = 10,700 m 2) around the lake and 15 at rates of 2,000 mermithid prepara- was treated 11 times during 7 wk with sites/yd 2 of surface area) to control Ano- 2,400-4,800 preparasites/m 2 at each appli- pheles larvae. The primary species present cation. The Anopheles population declined was Anopheles crucians. At the lower rate, from 10 per sample dip prior to the first an average of 76% of the hosts were in- preparasite application to 0.6 per dip by the fected and parasitism averaged 60, 80, 86 end of the release period; a 17-fold reduc- and 77% in first through fourth instars, re- tion in the number of mosquito larvae. This spectively. At 2,000 preparasites/yd 2, an was the first successful large-scale attempt average of 85% of the larvae were infected, to control mosquitoes with a parasite. Poor correlation was found between infec- Another technique for application of R. tion rates and vegetation and water depth. culicivorax was reported by Petersen and Brown et al. (4) conducted field tests Willis (23) who distributed the eggs and against four species of mosquitoes in three postparasitic stages of R. culieivorax in 13 natural (rice fields and ponds) and two habitats known to breed pasture mos- artificial (1- and 6-m 2 ponds) habitats at quitoes. They reported 52% of Aedes at- treatment levels ranging from 706 to 25,000 lanticus, 59% of Aedes tormentor, 38% of preparasites/m 2 of surface area. Anopheline P. columbiae, and 51% of Psorophora larvae were more susceptible than culicine. howardii were parasitized after the pastures Although both C. tarsalis and CuIiseta were flooded. Brown-Westerdahl et al. (5) inornata were reported in laboratory stud- found that early season application of R. ies to be good hosts for R. culicivorax, field culicivorax postparasites was effective in studies with these two species were disap- providing continuous partial control pointing. (weekly mean infection -- 60%) of A. I,evy and Miller (10) applied R. culici- freeborni and C. tarsalis throughout the vorax at the rate of 3,600 preparasites/m e rice growing season in Northern California. of surface area to control mosquitoes breed- Infections were observed up to 12 m from ing in a grassy field. In 10 potholes and point of application of 1,500 postparasites. ditches sampled in the field, the 5' found The nematodes overwintered, and infected 88-100% infection in Culex nigripalpus, sentinel and native mosquitoes were found Mermithids of Mosquitoes: Platzer 261 the following season. Platzer and Eby (26) we are to fully utilize the biocontrol po- also showed that postparasites can produce tential of these nematodes. A greater com- sufficient eggs to persist for two seasons of mitment of scientific and economic resources mosquito control. are required. Biocontrol with R. culicivorax In field releases of mermithid nema- has been under development for 10 yr; at todes, it is important that the application tile current level of interest and funding it be made by trained personnel who thor- will probably require 10 more years to ob- oughly understand the environmental lim- tain effective use of this or alternate or- itations of the organism. Under such con- ganisms in biological control of vector pop- ditions, the full potential for biological ulations. In addition, we need a strong com- control will be achieved with mermithid mitment from agencies to train personnel in nematodes. tile handling and use of such biological con- trol organisms. ESTABLISHMENT LITERATURE CITED Petersen and Willis (26) have conducted 1. Brown, B. J., and E. G. Platzer. 1977. The the only long-term studies on the ability of effects of temperature on the infectivity of Romano- R. culicivorax to become established after mermis culicivorax. J. Nematol. 9:166-172. application. They showed that this nema- 2. Brown, B. J., and E. G. Platzer. 1978. Salts and tode can become established in many semi- the infectivity of Romanomermis culicivorax. J. Nematol. 10:53-64. permanent and permanent mosquito breed- 3. Brown, B. J., and E. G. Platzer. 1978. Oxygen ing sites, but that certain problems existed: and the infectivity of Romanomermis culicivora×. lack of host populations, periodic lack of J. Nematol. 10:110-113. breeding water in the semipermanent sites, 4. Brown, B. J., E. G. Platzer, and D. S. Hughes. and inaccessibility due to changes in site 1977. Field trials with the mermithid nematode, Romanomermis culicivorax, in California. Mosq. access. Periods of little or no water tended News. 37:603-608. to reduce the levels of infection in suscep- 5. Brown-Westerdahl, B., R. K. Washino, and tible hosts, and flushing of the sites by E. G. Platzer. 1979. Early season application of heavy rains reduced nematode populations. Rmnanomermis culicivorax provides continuous partial control of rice field mosquitoes. Proc. Calif. It also appeared that sites which produced Mosq. Vector Cont. Assoc. 47:55 (Abstr.). most mosquitoes also produced most R. 6. Dhillon, M. S., M. S. Mulla, and E. G. Platzer. culicivorax infections. Petersen and Willis 1980. Evaluation of nematode Romanomermis (24) noted that recycling occurred in 10 culicivorax against cemetery mosquitoes. Mosq. News. 40:531-535. sites 6-29 wk after treatment with prepara- 7. Finney, J. R., R. Gordon, W. J. Condon, and sites. In 1974 three of the sites treated in T. N. Rusted. 1977. Laboratory studies on the 1971 were still producing infected hosts (7- feasibility of integrated mosquito control using an 25%). Five of six sites treated in 1971 and insect growth regulator and a mermithid nematode. 1973 and 7 of 12 sites treated only in 1973 Mosq. News 37:6-11. 8. Galloway, T. D., and R. A. Brust. 1977. Effects were producing parasitized hosts in 1974. of temperature aml photoperiod on the infection of Petersen and Willis (23) released S. two mosquito species by the mermithid, Romano- peterseni in Louisiana in 1971. Two years mermis cnlicivorax. J. Nematol. 9:218-221. later 88% of the Anopheles sp. were para- 9. Levy, R., and T. W. Miller, Jr. 1977. Experi- mental release of a mcrmithid to control mosquitoes sitized. Nickle (12) reported that R. culict- breeding in sewage settling tanks. Mosq. News 37: vorax was established and overwintered in 410-414. Maryland where winter temperatures 10. Levy, R., and T. W. Miller, Jr. 1977. Experi- dropped as low as -17 C. mental release of Romanomermis culicivorax (Mermithidae: Nematoda) to control mosquitoes breeding in Southwest Florida. Mosq. News 37:483- SUMMARY 486. 11. Mitchell, C. J., P. Chen, and H. C. Chapman. Although substantial potential has been 1974. Exploratory trials utilizing a mermithid nema- attributed to the effectiveness of mermithids tode as a control agent for Culex mosquitoes in for vector control, more time and research Taiwan. J. Formosan Med. Assoc. 73:241-254. is required to realize this potential. Addi- 12. Nickle, W. R. 1979. Probable establishment and overwintering of a mermithid nematode para- tiotm~ studies on the biology of R. culici- site of mosquitoes in Maryland. Proc. Helminthol. vorax and other mermithids are needed if Soc. Wash. 46:21-27. 262 Journal of Nematology, Volume 13, No. 3, July 1981 13. Petersen, J. J. 1973. Role of mermithid nema- quito larvae. Mosq. News 32:312-316. todes in biological control of mosquitoes. Exp. 22. Petersen, J. J., and O. R. Willis. 1974. Ex- Parasitol. 33:239-247. perimental release of a mermithid nematode to 14. Petersen, J. J. 1976. Comparative biology of control Anopheles mosquitoes in Louisiana. Mosq. the Wyoming and Louisiana populations of Reesi- News 34:316-319. mermis nielseni, parasitic nematode of mosquitoes. 23. Petersen, J. J., and O. R. Willis. 1974. Dixi- J. Nematol. 8:273-275. merinis peterseni (Nematoda: Mermithidae): A po- 15. Petersen, J. J. 1979. pH as a factor in para- tential biocontrol agent of Anopheles mosquito sitism of mosquito larvae by the mermithid larvae. J. Invert. Pathol. 24:20-23. Romanomermis culicivorax. J. Nematol. 11:I05-106. 24. Petersen, J. J., and O. R. Willis. 1975. Es- 16. Petersen, J. J. 1981. Observations on the biol- tablishnrent and recycling of a mermithid nematode ogy of Octomyomermis musprani, a nematode para- for the control of mosquito larvae. Mosq. News 35: site of mosquitoes. J. Invert. Pathol., in press. 526-532. 17. Petersen, J. J., and H. C. Chapman. 1979. 25. Platzer, E. G., and B. J. Brown. 1976. Physi- Checklist of mosquito species tested against the ological ecology of Reesimermis nielseni. Proc. Int. nematode parasite Romanomermis culicivorax. J. Colloq. Invert. Pathol., Kingston, Canada, pp. 263- Med. Entomol. 15:408-471. 267. 18. Petersen, J. J., H. C. Chapman, O. R. Willis, 26. Platzer, E. G., and J. E. Eby. 1980. Mer- and T. Fukuda. 1978. Release of Romanomermis mithid nematodes in mosquito control: persistence culicivorax for the control of Anopheles albimanus in small experimental ponds. Proc. Calif. Mosq. in E1 Salvador. II. Application of the nematode. Vector Cont. Assoc. 48:40-41. Am. J. Trop. Med. Hyg. 27:1268-1273. 27. Platzer, E. G., and L. L. MacKenzie-Graham. 19. Petersen, J. J., c. D. Steehnan, and O. R. 1978. Predators of Romanomermis culicivorax. Proc. Willis. 1973. Field parasitism of two species of Calif. Mosq. Vector Cont. Assoc. 46:93 (Abstr.). Louisiana rice field mosquitoes by a mermithid 28. Platzer, E. G., and L. L. MacKenzie-Graham. nematode. Mosq. News 33:573-575. 1980. Cyclops vernalis as a predator of the prepara- 20. Petersen, J. J., and O. R. Willis. 1970. Some sitic stages of Romanomermis culicivorax. Mosq. factors affecting parasitism by mermithid nematodes News 40:252-257. in southern house mosquito larvae. J. Econ. 29. l'oinar, G. O., Jr. 1975. Entomogenous Nema- Entomol. 63:175-178. lodes. Leiden: E. J. Brill. 21. Petersen, J. J., and O. R. Willis. 1972. Re- 30. l'oinar, G. O., Jr. 1979. Nematodes for Bio- suits of preliminary field application of Reesimermis logical Control of . Boca Raton, Florida: nielseni (Mermithidae: Nematoda) to control mos- CRC Press.

Mermithid Parasites of Agricultural Pest Insects 1 W. R. Nickl&

Plant and insect nematologists generally mithids, called preparasites, in soil samples agree that nematodes are well adapted to from agricultural fields. They are about 1 the soil environment. And since over 90% nnn in length, narrow, and resemble long of all agricultural pest insects spend at least Tylenchus sp. larvae in their shape and a part of their lives in the soil, a situation moving habits. After infecting their host, exists which favors an encounter between the preparasites continue their development the two organisms. When the encounter in- through the several larval stages and egress volves an insect susceptible to a mermithid from the host insect as fully grown post- nematode, a parasitic relationship may en- parasites which no longer feed. Egress from sue. Additionally, infective stages of several the host insect causes a large hole from species of mermithids climb low-growing which the insect loses its essential fluids and plants to infect susceptible host insects dies. The postparasites, 2-25 cm in length, which feed on the leaves or in stems. Nema- re-enter the soil and move to a depth of tologists can find the infective stage of mer- 15-50 cm, often well beneath the plow layer where the soil is more sterile and free Received for publication 12 December 1980. from fungal parasites and predators. Here aSymposium paper presented at the annual meeting" of the Society of Nematologists, New Orleans, Loulsiana, Au- they molt to adults and mate. In the fall, gust 1980. they often form balls (Fig. 1) consisting of 2Nematology Laboratory, Plant Protection Institute. USDA SEA AR, BARC.West, Beltsville, MD 20705. a female and one or more males. The eggs