Field Efficacy and Nontarget Effects of the Mosquito Larvicides Temephos, Methoprene, and Bacillus Thuringiensis Var

Home , Aedes taeniorhynchus

Journal of the American Mosquito Control Association, 15(4):446_452,1999 Copyright O 1999 by the American Mosquito Control Association, Inc.

FIELD EFFICACY AND NONTARGET EFFECTS OF THE MOSQUITO LARVICIDES TEMEPHOS, METHOPRENE, AND BACILLUS THURINGIENSIS VAR. ISRAELENSISIN FLORIDA MANGROVE SWAMPS

SHARON P LAWLER,' TRULS JENSEN,,., DEBORAH A. DRITZ, ,qNo GEoRGE WICHTERMANI

ABSTRACT. We compared the efficacy and nontarget effects of temephos, Bacillus thuringiensis var. israe- (B.t'i.), lensis and methoprene applied by helicopter to control mosquito larvae in mangrove ,*u-p, on Sanibel Island, FL, in May 1997. Three sites per treatment and 3 untreated sites were used. Temephos (Abate@) was applied at 37 mllha (43Vo active ingredient tAIl), B.t.i. granules (Vectobac G@.; were uppii.d ut 5.606 kg/ha (200 International Toxic Units/mg), and methoprene (Altosid@ ALL) was applied at213 ni/ha (5vo AI). EffiJacy was quantified by monitoring the survival of caged and uncaged larval Aides taeniorhynchus. We quantified mortality of sentinel nontarget amphipods (Talitridae) at all sites, monitored the effect bf temephos on flying arthropods using light traps, and collected dead insects in tarps suspended under mangroves in areas treated with either temephos or methoprene. Each pesticide showed good overall efficacy but occasional failures occurred. No detectable mortality of amphipods or flying insects attributable to pesticides was found. The inconsistent field efficacies of the pesticides indicate a need for reinspection of treated sites in this habitat.

KEY WORDS B.r.i., Abate@, Altosid@, mosquito control, Aedes taeniorhynchus

INTRODUCTION safe for vertebrates at levels used in mosquito control, but vary in risk to invertebrates. Temephos is Coastal marshes and mangrove swamps are often an organophosphate pesticide that acts by inhibiting managed for mosquito control because salt-marsh cholinesterase, and it is toxic to insects and some mosquitoes can emerge in numbers that threaten the other nontarget invertebrates (Smith 1987, Brown health of humans and livestock. The salt-marsh et al. 1996).The bacterium B.t.i. is a microbial in- mosquito Aedes taeniorhynchus (Wiedemann) can secticide. It controls mosquitoes with toxins whose make our study area, Sanibel Island, FL, nearly un- action is specific to nematoceran dipterans (e.g., inhabitable, and it can harm livestock (Addison and mosquitoes and black flies), and is expected to have Ritchie 1993). Aedes taeniorhynchus c^n fly long little effect on other macroinvertebrates (Back et al. distances to obtain blood meals (Ritchie and Mon- 1985, Federici 1995). Methoprene is similar in tague 1995), so one of the best methods of con- structure to insect juvenile hormone, and it causes trolling this species is to kill the larvae in local mortality in mosquitoes by interfering with meta- breeding sites before they disperse as adults. Aedes morphosis. Although the action of methoprene is taeniorhynchzs larvae develop in temporary pools not specific to mosquitoes, many insects are insen- that form in depressions above the upper intertidal sitive to the levels of methoprene used in mosquito zone in mangrove swamps and salt marshes. They control (Hershey et al. 1995, but see Gelbic et al. often occur in such large numbers that emergence 1994). Methoprene is not expected to affect adult of even a small percentage of the population can insects. Methoprene and B.t.i. were recently re- create problems for residents. Because salt marshes ported to have negative indirect effects on preda- tory insects (Hershey and mangrove swamps are productive habitats that et al. 1998). However, declines of predators occurred only after 2-3 years of sustain a variety of wildlife (review, Mitsch and frequent application of the materials at maximum Gosselink 1993), environmentally sound mosquito label rates, and no such effects have been docu- management is a conservation priority. We con- mented during normal use of these materials. ducted a large-scale field study on 3 larvicides, te- This study offers a side-by-side comparison of mephos, Bacillus thuringiensis var. israelensis (de the field efflcacy of the pesticides when applied by Barjac) (B.t.i.), and methoprene, to assess whether aircraft to swamps composed of mangroves and these larvicides could control mosquitoes in man- grasses. To compare the nontarget effects of the grove areas without causing substantial mortality of pesticides, we monitored the survival of a common nontarget amphipods and canopy insects. amphipod (Talitridae) in treated and control sites. The 3 pesticides have different modes of action We also tested whether temephos would kill inver- and expected nontarget effects. All are relatively tebrates inhabiting the mangrove canopy. Nontarget insects in the mangrove canopy could be affected by aerial applications of temephos because it is a I Department of Entomology, University of Califomia contact poison. at Davis, One Shields Avenue, Davis, CA 95616. 2 Present address: 3721 E Vermilion Avenue, Kanab, uT 84741. MATERIALS AND METHODS r Lee County Mosquito Control District, PO 60005, Study sites: The study was conducted in May Fort Mvers. FL 33906. 1997, during the 1st larviciding operations of that

446 DeceNanen1999 Epp'ecrs op Tnngn LARVICIDESrN MeNcnovgs 447

Fisherman's Key qt/ 4

1fl-\^L r

Sanibel Island

- 1 Kilometer

Fig., 1. Map of a study area, Sanibel Island, FL, showing the location of mangrove areas treated with mosquito control pesticides or used as controls. Sites ranged in size from 3.2 to 2i.6 ha. Treitments correspond to symbois as follows: r?r temephos, ) Bacillus thuringiensis var. israelensis, I methoprene. I control.

year. We established 11 study sites on Sanibel Is- quitoes in treated sites using dippers. We evaluated land plus one on nearby Fisherman's Key (Fig. l). whether temephos would affect terrestrial insects Sites ranged in size from 3.2 to 2i.6 ha. The dom- by comparing insect collections from sites treated inant species in all areas were red mangrove (Rftl- with temephos to sites that were either untreated or lophora mangle (L.)) and black mangrove (Avicen- treated with methoprene. Details are given below. nia germinans (L.)), with a few areas of srasses Pesticide application: Lee County Mosquito or white mangroves (Laguncularia raJemosa Control District personnel applied all pesticides by (Gaertn)) in some sites. These sites typically flood helicopter, and cleaned application gear between during high-tide cycles that coincide with onshore applications of different materials. The B.t.i. srarl- winds, and the water usually disappears within sev- ules (Vectobac G@, Abbott Laboratories, Iiorth eral days (G. Wichterman, personal communica- Chicago, IL) were applied on May 14, 1997, at a tion). All sites flooded with rain and tide warers rate of 5.606 kg/tra (200 International Toxic Units during a storm on May 12, 1997, and lst-stage Ae. per mg; 5 lb/acre). Temephos and methoprene were taeniorhynchus appeared the next day. Each larvi- applied on May 15, 1997, except for I application cide was applied to 3 sites and 3 control sites were of temephos to Fisherman's Key on May 16, 1997. not treated. Control sites flooded simultaneouslv Temephos (Abate@, Clarke Mosquito Control prod- with treated sites but produced few mosquitoes. WL ucts, Roselle, IL) was applied at 37 mllha (43Vo were unable to assign control sites at random with active ingredient [AI]; O.5 fl oz./acre). Methoprene respect to the presence of mosquitoes, because even (Altosido ALL, Wellmark International, Benson- small untreated areas can produce enorrnous num- ville, IL) was applied at 213 mllha (5Vo AI, 3 oz./ bers of mosquitoes that attack nearby residents. acre). One B.t.i. site was retreated with temephos We sampled study sites on the day before pesti- after 2 days to prevent emergence of mosquitoes cide application and on the subsequent 6 days. At that survived the B.t.i. application. all sites, we monitored survival of caged mosquito Sentinels: We used field-collected Ae. taenior- larvae and amphipods, and sampled uncaged mos- hynchus larvae and amphipods in the family Tali- 448 JounNll or rHe AlarntcaN Mosquno CoNrnol AssocnrroN VoL. 15,No.4 tridae as sentinels to monitor insecticide activity. that insects were too sparse in the mangrove canopy Aedes taeniorhynchus and the amphipods were the for effective sampling by sweep net (15 canopy only aquatic macroinvertebrates consistently pres- sweeps typically yielded fewer than 4 insects). ent in study sites after flooding, probably because Therefore, we tested whether temephos affected fly- the habitat is ephemeral. Other aquatic macroinver- ing insect abundance by collecting insects with tebrates occasionally observed during field work Centers for Disease Control light traps (CDC traps) were fiddler crabs, water boatmen (Corixidae), and and ultraviolet light rraps (UV traps) in 2 sites treat- dytiscid beetles. Sentinels were held in floating ed with temephos (Wulfert's Point and West Im- predator-exclusion cages. Cages were 1.9-liter plas- poundment Swale), and in 2 sites that are not larv- tic buckets that were suspended in the water by a icided (Pole Line and Tarpon Bay). We placed 2 ring of styrofoam and attached to a stake by string. CDC traps and 1 UV trap in each site, at least 20 Cages had 2 screened windows measuring 5 X 10.5 m apart. The vegetation was thick in these areas cm, and had screened lids that were removed dur- and so it is unlikely that the traps attracted insects ing insecticide applications. Before pesticide appli- from outside the designated sites. We could not see cation on May 14, 25 2nd-stage mosquito larvae from 1 trap to the next and had to follow flagged were placed in each of 2 cages per site, and 25 trails to find them. However, insects were free to amphipods were placed in each of 2 cages per site. fly in and out of sites during the study. We collected We recorded sentinel organism survival each day. insects on the night before temephos was applied, Dip samples.' We monitored pesticide efficacy and on the subsequent 4 nights. We used ANOVA using 3 transects of 15 mosquito-dipper samples to determine whether treated areas showed greater per site (45 dips of 350 ml each). We used a ran- differences in abundance than controls after the dom dipping pattern at all sites except for I site pesticide application date. Data points were trans- treated with methoprene. This site contained large formed as Un(pretreatment abundance) numbers of mosquitoes that formed dense aggre- ln(posttreatment abundance)1, where the pretreat- gations, and random sampling underestimated their ment data were collected May 14 p.m. to May 15 abundance. Beginning on May 18, we targeted ag- a.m, and the posttreatment data were collected May gregations in our dip transects to check for mos- 16 p.m. to May l7 a.m.. The posttreatment collec- quitoes dying as mature 4th instars or pupae, as is tion occurred on the night after the highest mortal- typical of methoprene. On this date we also put 1O ity was seen in larval mosquitoes. 4th-stage larvae into each of 3 additional sentinel Not all insects are attracted to light traps and cages, to see what proportion of previously uncaged some cannot fly, so we used a 2nd method to test mosquitoes would metamorphose. Aggregations whether temephos caused mortality of canopy in- did not pose sampling problems at other sites be- sects. We suspended 3 3.8-m' tarps under the man- cause mosquitoes either were dead by the time ag- grove canopy in all sites treated with methoprene gregations formed or were too sparse to form ag- and temephos to collect any insects killed by in- gregations. secticide application. We used methoprene sites as We used analysis of variance (ANOVA) to com- controls for the effects of temephos, because both pare the abundance of mosquitoes before vs. after insecticides are applied as spray from helicopters. treatment in treated sites. Abundances were trans- Methoprene should not kill canopy insects imme- formed as ln(n + l) to normalize their distribution. diately because few of the insects would be likely Data were the mean of the number per dip averaged to metamorphose during the brief study. The center over 2 pretreatment dates for temephos (May 14, of each tarp was weighted to form a funneled shape 15), compared to the mean abundance of 2 post- so that insects would not be swept out of the tarps treatment dates (May 18, l9). Only I pretreatment by breezes. We collected all dead or moribund in- date was available for the B.r.i. sites (May l4), and sects from these tarps on the 2 days following in- we used May 15 as the posttreatment date because secticide applications. 1 site had to be retreated with temephos after May 15. Natural mortality cannot be factored out of the data set, but the data allow a comparison of the RESULTS proportion relative of mosquitoes emerging from Sentinel organisms each treatment. Analysis of the sentinel data from controls showed whether mortality could be attri- All 3 materials killed sentinel mosquitoes. Heavy buted to pesticides. Unfortunately, we could not mortality occurred in treated sites but survival was perform a statistical test for uncaged mosquitoes high in controls (Fig. 2A). Ninety-five percent of exposed to methoprene because we had suitable sentinel mosquitoes died in 2 of 3 sites treated with data from only I site. The 2nd site dried, and larval temephos, but all survived in the 3rd site, Fisher- aggregation at the 3rd caused an apparent rise in man's Key. This site was treated I day later, and sample numbers after treatment. However, sentinel contained older 4th-stage larvae that may have been data were available for statistical tests of the effect less susceptible to temephos (G. Wichterman, per- of methoprene. sonal communication). The effect of temephos was Canopy insects: Preliminary sampling showed significant when we eliminated this site from sta- Dscerr,rssn1999 Errecrs op TgRen LARVTcTDESrN MANcRovEs

100 analysis. Methoprene significantly decreased sentinel mosquito survival (ANOVA, df 1,3, F : 29.9, P: 0.01). No effects were detected of any of the 3 insec- (Fie. J ticides on amphipod survival 2B; ANOVAs 5uo on: temephos vs. control, df 1,3, F = 1.4, P : O.32; iso B.t.i. vs. control, dt 1,4, F: 3.6, P: 0.13; meth- f oprene vs. control, df 1,3, F : O.22, P = 0.67). We |,40 s used data from day 4 for all but 1 site because all 30 materials should cause mortality within a few days of treatment, and later data would contain more noise from natural mortality. For 1 of the B.t i. sites we used survival data from the 2nd day posttreatment because the site was subsequently retreated CONTROL with temephos. Mosquito dip transects.' 100 Dip transects for mosquito larvae showed similar mortality to that ob- 90 tained from the sentinel data (Fig. 3), but no effects were statistically significant at the P : 0.05 level. For temephos, numbers of uncaged mosquito larvae J dropped dramatically at the 2 sites where sentinel 5uo mosquitoes showed poor survival. However, at I of iso these sites approximately 3OVo of mosquitoes sur- (,40= vived to emergence even though the sentinel data ;s indicated survival of less than 5Vo. The temephos 30 application completely failed to control mosquitoes 20 at Fisherman's Key, where the larvae were treated 't0 as late 4th instars. All sites treated with B.t.i. showed decreases in 0- mosquito populations (Fig. CONTROL TEM. BTI METH. 3B), although I site could not exhibit much of a decrease because very Fig. 2. Survival of Aede.s taeniorhynchus larvae (A) few larvae were present initially. An ANOVA in- and Talitridae amphipods (B) held in cages in Florida dicated a strong trend toward an effect of B.r.i. on mangrove areas that were treated with temephos, Bacillus mosquito abundance whether or not this site was thuringiensis var. israelensis (.8.t i.), or methoprene, or un- included (all sites, df 1,4, F : p < treated. Bars represent the proportion of 50 sentinel 4.9, 0.09; I site or- : ganisms surviving in I site until 4 days posttreatment, excluded, df 1,3, F 13.0, P < O.O7). except for the lst B./.i. site, where data are from day 2 Methoprene killed nearly all mosquitoes at meta- posttreatment; this site was later retreated with temephos. morphosis in I of the 2 sites that remained wet until Two cages of 25 of each type of organism were used per mosquitoes emerged (Fig. 3C). At the 2nd site, the site, with 3 sites per treatment. CONTROL : untreated, final dip sample consisted of 35Vo shed pupal skins TEM. : : : temephos, BTI B.t.i., METH. methoprene. and 65Vo dead pupae or larvae and many biting 0 : no survival, X : loss of site through drying- adults were present, showing that methoprene did not yield adequate control. This was the same site that showed emergence of 5Vo of the original caged : : tistical analysis (df 1,3, F 34.5, P 0.01). The sentinels and lOVo of field-exposed mosquitoes that B.r.i. also killed sentinel mosquitoes (ANOVA, df were caged on day 5. Dip samples at both sites : : I,4, F 8.8, P 0.04). AX sentinel mosquitoes contained many dead pupae, 4th-stage larvae, and died in 2 treated, sites but 607o survived in the 3rd partially emerged adults, as is typicat of the action site until the 2nd day posttreatment, when they of methoprene. were killed with temephos. Survival in the cages Canopy insects: Analysis of light trap data did was higher than in the uncaged population (iee not indicate any consistent loss of insects from Mosquito dip transects below). This could result treated sites (Fig. 4). No significant decreases oc- from poor distribution of B.t.i. granules into the curred in the abundance of flying insects captured cages. Methoprene prevented the maturation of all in CDC traps in sites treated with temephos (AN- sentinel mosquitoes at 1 site, but approximately 5Vo OVA, df 1,2, F : 4.3, P : O.l7). Results from UV matured at a znd site where larvae were abundant light traps were similar. Unfortunately, 2 of the UV and formed aggregations. We had also caged 30 traps failed during the pretreatment night, 1 from a additional larvae at this site that had been free- treated site and the other in a control. Therefore, swimming for 5 days postapplication, and lOVo of instead of analyzing pre- vs. posttreatment differ- these emerged. The 3rd site dried shortly after pes- ences for UV traps, we compared the posttreatment ticides were applied, and was eliminated from the abundance of insects in controls vs. treated areas JounNlr- oF THE AMERTCANMoseurro CoNTRoL Assocrerrolr Vor-.15, No.4

1,000 TEMEPHOS 500 L20 t,u 200 u 100 - tc z o I 5U u ""zil, U A. v EMERGENCE z l 20 5ru "ei o a" '\;...... ^...""t. .t0 o s ...... : z >5 J \. a 1 ,uf 16 17 DATE 30

t25 B.T.t. 1,000 ffzo o 300 B. u !< o o'- O F e 1oo 3ro o o RE.TREATED 3so E5 o .-*- .-O zru J 0 ooNTROL --+ -! 15 J OATE 1 METHOPRENE 15 16 17 18 19 LB ..1 MAY TE 71 U Fig. 4. Light trap catches of flying insects in Florida o-6 (t mangrove areas that were treated with temephos for mos- c. U quito control (dashed lines), or left untreated (solid lines). v =4 Two Centers for Disease Control light traps (A) and 1 l ultraviolet light trap (B) were placed at each site. The zrrow on the x axis indicates the date of pesticide appli- >z ar '!-- i\ \ cation. \I \ EMERGENoE 12 14 tu tt 20 | 22 Tarps strung under mangroves treated with te- oora mephos and methoprene collected similar numbers Fig. 3. Numbers of Aedes taeniorhynchus mosquito of dead arthropods in each treatment over the 48 h larvae per dip in 3 transects of 15 mosquito-dipper sam- after pesticide application (ANOVA, df 1,4, F = ples per site, per date, in Florida mangrove areas treated 0.206, P = O.67:Table 1). with (A) temephos, (B) Bacillus thuringiensis var. israelensis, or (C) methoprene. Each line represents a separate site. The arrow on the x axis indicates the date of oesticide DISCUSSION application. All 3 pesticides caused significant mortality of sentinel mosquito larvae in treated sites compared on May 17, the lst night when all traps rnn prop- to controls, demonstrating that the pesticides (rather erly. No effect of temephos (ANOVA, dt L,2,F : than predation or other natural sources of mortality) 2.1, P : 0.28) was detected. Most of the insects were responsible for reducing mosquito popula- captured in light traps were small flies in the fam- tions. The mortality of sentinel mosquitoes paral- ilies Psychodidae and Ceratopogonidae. Other com- leled decreases in natural populations in treated mon taxa included mosquitoes, crane flies, moths, sites. However, sentinel mosquitoes sometimes beetles, and Hymenoptera. showed greater or lesser declines after pesticide ap-

Table 1. Numbers of insects caught in 3.8-m'tarps hung below mangrove canopy sprayed with either methoprene or temephos, during 2 days after pesticide application.

Wasps Beetles Moths Spiders Flies Bugs Other

Temephos 91 2 J 0 J 6 o 3 lo8 Methoprene t73 19 I 6 2 9 461 Decervrsrn1999 Epppcrs or Tunrs Lanvrcroes w MlNcnoves 451

plication than natural populations. These differenc- lished of the nontarget effects of these pesticides es could result from uneven distribution of pesti- on salt marshes or mangrove swamp fauna with cides over the sites or different levels of exposure which to compare our results, but we review those due to the cages. available below. We found few aquatic nontarget Larvicides were often effective in killing un- organisms in our ephemeralsites; however, we re- caged Ae. taeniorhynchas larvae underneath the port studies of macroinvertebrates or fish that may mangrove canopy, yielding average control levels enter the high marsh during more extended periods of 8O-9OVo. This is a good level of control for this of flooding. As a caveat in interpreting these stud- habitat, which is difficult to treat because the can- ies, all but the lst 2 presentedwere conducted in opy can interfere with delivery of the materials and laboratories. Laboratory studies often result in the water is often hidden from view. However. in greater exposureof organisms to toxins becauseof at least 1 site per material, the larvicides failed to clean water conditions and the inability of organ- yield the high level of control necessary for salt- isms to behaviorally avoid contaminants. marsh mosquitoes. In a field study, the granular formulation of te- Temephos failed to control mosquitoes adequate- mephos was reported to cause behavioral changes ly at 2 sites, including 1 complete failure. This may in fiddler crabs. The flddler crabs actively collected have been due to a l-day delay in treating the site, and ingested the granules, which were applied at 63 because older Ae. taeniorhynchas larvae that are g AI/ha (Ward and Busch 1976, Ward et al. L976). approaching pupation are less susceptible to teme- Crabs may have been attracted by the celatom in phos (G. Wichterman, personal communication). 't\e the granules. This effect is not expected with the B.t.i. reduced dip-counts of mosquitoes to 0 liquid temephos used in our study area because the at 2 sites but a 3rd site had to be retreated with liquid is applied at a lower rate and the crabs cannot temephos to further reduce emergence from l4Vo to collect it. The 2nd field study was performed by nearly 0. Mosquitoes were more abundant at this Pierce et al. (1989), who quantifled the effects of site than at any of the others (Fig. 3), and B.t.i. liquid temephos on 2 species of shrimp, juvenile sometimes fails to control dense populations of snook, and sheepsheadminnows. Temephos was mosquitoes if they deplete the toxin from the water applied at twice the application rate used in our before all obtain a lethal dose (e.g., Mulla et al. study. Although temephos was safe for the flsh, re- 1990, Becker et al. 1992). sults for the shrimp were unclear because32Vo mor- Finally, methoprene yielded I success and I par- tality occured at 1 of their 3 treated sites. However. tial failure. Dame et al. (1998) recenrly demonstrat- this mortality was not correlated with the amount ed resistance to methoprene for a population of Ae. of temephos in the water. taeniorhynchas on Captiva Island that had been ex- Roberts (1995) found that B.t.i. bad no effecr on posed to extended-release, 150-day Altosid briquets eitler a salt-marsh gammaroid amphipod or a over a period of 6 years. Captiva and Sanibel is- shrimp. McKenney and Celestial (L996) found that lands are part of the same land mass, and it is pos- methoprenekilled a marine mysid shrimp exposed sible that the population of Ae. taeniorhynchus is to L25 pglliter over 4 days, with sublethal effects also the same. This could explain the emergence of occurring at levels as low as 2 p"gfliter.In a com- mosquitoes from our methoprene-treated sites, al- parative study, Lee and Scott (1989) used the mum- though uneven application of materials is also a michog fislr (Fundulus heteroclitus) to compare the possibility. acute toxicities of B.r.i.,methoprene,and temephos. The occasional failures of the 3 larvicides dem- Temephoswas more toxic than the other larvicides. onstrate that reinspection and retreatment of breed- but all 3 were safe for the flsh at the expected water ing sites are very important in ensuring an effective concentration for mosquito control. Brown et al. mosquito control program. However, a 2nd appli- (1996) compared the roxicity of B.t.i., methoprene, cation of larvicide is only feasible for sites treated and temephos to an Australian estuarine shrimp with B.t.i. early in larval development, or before (Leander tenuicornis). Methoprene and B.r.r. were the late 4th stage with temephos. Methoprene does not toxic at levels used in mosquito control. The not af,ford this opportunity because it kills mosqui- median lethal concentrationof temephos was O.0l toes during pupation and emergence, when it is too ppm, and concentrationsof temephosthis high may late to re-treat with a larvicide. occur under operational conditions in Florida man- Pesticides did not cause detectable mortalitv of groves: The expected concentration of temephos amphipods. The amphipods had been collected in applied at 14 g (0.5 oz) of 43VoAI temephoslacre aquatic samples from the sites and were sometimes (as in our study) is 0.026 ppm, with realized con- seen swimming in the water; however, they often centrationsof 0.12-{.0045 ppm temephosfound in climbed up the screens of the sentinel buckets and surface and midwater samples, respectively (pierce rested above the waterline. This could have reduced et al. 1996). However, whether laboratory studies their exposure to pesticides. Most Thlitridae are on an Australian shrimp can predict toxicity for semiaquatic species, which explains their abun- southeastern North America intertidal fauna under dance in the intermittently flooded high marsh. field conditions is not known. In conclusion. al- Comparatively few other studies have been pub_ though temephos had some documented toxic ef- 452 JounNll oF THE AMERIcIN Mosquno CoNTRoL AssocnnoN VoL. 15. No.4 fects, little f,eld evidence exists that it causes mor- capoda: Paleamonidae). J. Am. Mosq. Control Assoc. tality of nontarget macroinvertebrates or fish when l2:721-724. applied at rates used in mosquito control. Relatively Dame, D. A., G. J. Wichterman and J. A. Homby. 199g. low potential exists for mortality of aquatic nontar- Mosquito (Aedes taeniorynchus) resistance to methoprene in an isolated habitat. gets in our study area because few species utilize J. Am. Mosq. Control As- soc. I4:2OO-2O3. the high marsh during the brief period that it is Federici, B. A. 1995. The future of microbial insecticides flooded, and the more species-rich low marsh is as vector control agents. J. Am. Mosq. Control Assoc. rarely treated because regular tidal flooding reduces 1l:26O-268. or prevents mosquito breeding. Gelbic. 1.. M. Papacek and J. Pokuta. 1994. The effects This is the lst study of whether temephos affects of methoprene ,Son the aquatic bug llyocoris cimicoides mangrove canopy arthropods. We did not find sig- (Heteroptera, Naucoridae). Ecotoxicology 3:89-93. nificant decreases in nocturnal flying insects in sites Hershey, A. E'., A. R. Lima, G. J. Niemi and R. R. Regal. 1998. Effects of Bacillus thuringiensis treated with temephos compared with controls. We israelensis (B.r.i.) and methoprene on nontarget macroinvertebrates also compared the number of insects collected in in Minnesota wetlands. Ecol. Appl. 8:41-60. tarps under canopy treated with temephos vs. meth- Hershey, A. E'., L. Shannon, R. Axler, C. Ernst and P. oprene, and found a similarly low number of dead Mickelson. 1995. Effects of methoprene and, Bti (Ba- insects (an average of l0/tarp/day). In conclusion, cillus thuringiensis var. israelensis) on non-target in- temephos, B.t.i., and methoprene effectively con- sects. Hydrobiologia 3O8:219 --227. trolled Ae. taeniorynchus without observable mor- Lee, B. M. and G. I. Scott. 1989. Acute toxicity of te- tality of the nontarget amphipods, and temephos did mephos, fenoxycarb, diflubenzuron, and methoprene and, Bacillus thuringiensis var. israelensis to the mum- not cause detectable changes in the abundance of michog (Fundulus heteroclitus). Bull. Environ. Contam. flying insects. Toxicol. 43:827*832. McKenney, C. L., Jr. and D. M. Celestial. 1996. Modified survival, growth and reproduction in an estuarine mysid ACKNOWLEDGMENTS (Mysidopsis bahia) exposed to a juvenile hormone an- alogue through a complete life cycle. Aquat. Toxicol. We thank S. Stenquist, J. O'Neal, L. Hinds, L. 35:.ll-2O. Hamilton, and J. Coppen of the U.S. Fish and Wild- Mitsch, W. J. and J. G. Gosselink. 1993. Wetlands. Van life Service for their assistance in planning and fa- Nostrand Reinhold, New York. cilitating this project, and we also thank J. Coppen, Mulla, M. S., H. A. Darwazeh and M. Zgomba. 1990. E. Jensen, and several Youth Conservation Corps Effect of some environmental factors on the efficacy of Bacillus sphaericus 2362 and,Bacillus thuringiensis (H- volunteers for help with fleld work. We are grateful 14) against mosquitoes. Bull. Soc. Vector Ecol. l5:166- to the Lee County Mosquito Control District for 775. their generous help with many aspects of the work, Pierce, R. H., R. C. Brown, K. R. Hardman, M. S. Henry, with particular thanks to W. Opp and T, Stewart. C. L. P Palme! T. W. Miller and G. Wichterman. 1989. We are indebted to R. K. Washino for help in or- Fate and toxicity of temephos applied to an intertidal ganizing the study. The comments of 2 anonymous mangrove community. J. Am. Mosq. Control Assoc. 5: reviewers improved the paper. This project was 569-578. Pierce, R., M. Henry, D. Kelly, P funded by the U.S. Department of the Interior and Sherblom, W. Kozlow- sky, G. Wichterman and T, W. Miller. 1996. Temephos the U.S. Fish and Wildlife Service Cooperative distribution and persistence in a southwest Florida salt Agreement 14-48-0001-94582, Mosquito Manage- marsh community. J. Am. Mosq. Control Assoc. 12: ment on National Wildlife Refuges. 637-646. Ritchie, S. A. and C. L. Montague. 1995. Simulated populations of the black salt marsh mosquito (Aedes tae- REFERENCES CITED niorhynchus) in a Florida mangrove forest. Ecol. Mod- Addison. D. S. and S. A. Ritchie. 1993.Cattle fatalities el.77:123-141. Roberts, G. M. 1995. Salt-marsh crustaceans, Gammarus from prolonged exposure to Aedes taeniorhynchus in duebeni and Palaemonete.r varians as predators of mos- southwestFlorida. Fla. Sci. 56:65-69. quito larvae and their reaction to Bacillus thuringiensis Back, C., J. Boisvert, J. O. Lacoursiereand G. Charpen- subsp. israelensis. Biocontrol Sci. Technol. 5:379-385. tier. 1985. High-dosagetreatment of a Quebec stream Smith, G. J. 1987. Pesticide use and toxicology in relation with Bacillus thuringiensis serovar.israelensis: efficacy to wildlife: organophosphorus and carbamate com- fly (Diptera: against black larvae Simuliidae) and im- pounds. Resource Publication 170. U.S. Department of pact on non-target insects. Can. Entomol. ll7:-1523- the Interior, Fish and Wildlife Service, Washington, DC. t534. Ward, D. V. and D. A. Busch. 1976. Effects of temefos, Becket N., M. Zgomba, M. Ludwig, D. Petric and E Ret- an organophosphorus insecticide, on survival and es- tich. 7992. Factors influencing the activity of Bacillus cape behaviour of the marsh fiddler crab Uca pugnax. thuringiensis vaf. israelensistreatments. J. Am. Mosq. Oikos 27:331-335. Control Assoc.8:285-289. Ward. D. V.. B. L. Howes and D. E Ludwid. 1976. Inter- Brown, M. D., D. Thomas, K. Watson, J. G. Greenwood active effects of predation pressure and insecticide (Te- and B. H. Kay. 1996. Acute toxicity of selectedpesti- mefos) toxicity on populations of the marsh fiddler crab cides to the estuarine shrimp Leander tenuicornis (De- Uca pugnax. Mar. Biol. 35:.119-126.