SrprnN4eun 1989 B.r.r. rN A MIcHIcAN RIvER 397 A BROADEVALUATION OF B.T.I. FOR BLACK FLY (DIPTERA: SIMULIIDAE) CONTROLIN A MICHIGAN RIVER:EFFICACY, CARRY AND NONTARGETEFFECTS ON INVERTEBRATESAND FISH RICHARD W, MERRITT,I EDWARD D. WALKER,I MARGARET A. WILZBACH,, KENNETH W. CUMMINS2 eno WILLIAM T. MORGAN1 ABSTRACT. Efficacy for black fly control, carry and nontarget effects of B.f.i. (Teknar@HP-D^-): applied in the Betsie River, Michigan, were studied in June 1988.Black fly mortality was high (-100%) f6i a2,200 m stretch downstream from the application site, declined to 30% at 3,200m, and was_nilat 4,500m. Drift of black flies greatly increasedafter application at a downstreamsite, but did not change ai an upstream site. There *ere no detectablenontarget effects of B.t.i. application on: 1) invertebrate rnu".o--o. micro-driftt 2) numbers of invertebrates in benthic Surber samples;3) mortality or feeding of drifting and nondrifting insects; 4) growth or mortality of cagedStenomena sp. Iarvae; 5) invertebrate functional group composition; 6) mbrtality or weight change of caged rock bass; or 7) fish numbers, speciescomposilion, length-weight (rock bass only) relationships or rock bass diet. Sampling of Rheo' tinytarsus sp. midgesonnatural substratesindicated low (27%) mortality owing to B.t.i. at only 100 m downstreapfrom ihe application site, with negligible mortality at all other downstream and upstream sites. This information, iombined with no pronounced changes in numbers of midges in macro-drift after application, indicated that midge populations were not adverselyaffected by B.t.i. in the study. INTRODUCTION have prompted the developmentof more ecolog- ically sound management strategies employing Judged by the degreeof annoyance and irri- nonpersistent agents presumedto have little or tation they causein North America, black flies no toxicity to nontargetorganisms. rank with mosquitoesas major pests of people The most successfulbiological control agent (Newson in recreationareas 1977,Kim and Mer- developedto date for black flies is B.t.i. fBacillus ritt 1987).Annually, black flies account for large thuringiensis var. israelensisde Barjac (serotype fiscal losses in Canada and the north central H-14)1,first isolatedby Goldbergand Margalit and northeastern United States. They greatly (1977) from samplestaken in the Negev Desert discouragetourism and outdoor recreation, and of Israel. It has proven to be very effective in a interfere with lumbering, mining and building multinational black fly control program con- activities during early spring and summer ducted by the World Health Organization in (Jamnback 1969, Fredeen 1977, Merritt and West Africa (Laceyet al. 1982),and at several Newson 1978). In addition to humans, other locations in North America (Molloy and Jamn- animals also are affected by black fly attacks back 1981, Molloy 1989, Colbo and O'Brien (Steelman 1976).Fredeen (1985)estimated that 1984,Back et al. 1985,Gibbs et al. 1986,Pistrang lossesto beef and dairy producersin Sasketche- and Burger 1984). Lacey and Undeen (1986, wan in one year exceededUS $2.9million owing 1987),and most recently Molloy (1989) have to mass outbreaks of a single species.Besides revieweduses of B.r.l. for black fly control. economic effects, the bites of these insects can To be effective,B.t.i. must be eatenby black create a variety of pathological conditions in fly larvae. The principal mortality agentof B.t.i. humans requiring hospitalization of sensitized is a parasporal particle which contains a pro- individuals (Newson 1977). teinaceous protoxin (Dubois and Lewis 1981, In recreation areas,spraying to control adult Aronson et al. 1986).The toxin becomesactive black flies with insecticideshas sometimesbeen after solubilization of protoxin in the presence effective, but the benefits are of short duration of proteases and alkaline pH in the black fly and this method presents risk of environmental Iarval midgut. The toxic polypeptides bind to contamination. Broad spectrum chemical insec- surface receptors on the midgut epithelial cells ticides applied to streams for larval control re- and causethe cells to lyse and disintegrate, and sult in death of many nontarget organisms (Fre- the larva dies. deen 1975,1983; Mohsen and Mulla 1982).In To date, there have been no published studies addition, biomagnification of insecticidesin the in the use or effect of B.t.i. against black flies in food chain has undesirable consequences Michigan waterways. In June 1988, the Michi- (Woodwell et al. 1967). These adverse effects gan Water ResourcesCommission and Depart- ment of Natural Resourcesissued a permit al- I Iowing B.t.i. to be experimentally applied to a Department of Entomology, Michigan State Uni- Betsie trout stream versity, East Lansing, MI 48824. sectionofthe River, a brown 2Pymatuning Laboratory of Ecology, Department Iocated in the lower peninsula. This paper re- of Biological Sciences,University of Pittsburgh, Pitts- ports results of this study, the major objectives burgh,PA 15260. of which were to: 1) determine the extent of Seprnrr,rsnn1989 B.r.r. rN I Mrcnrcelt Rrvnn sporesin the river water were done to estimate axis of an aluminum frame (45 x 15 cm rectan- the carry of the B.t i. downstream from the gular aperature) holding the net. AII drift sam- release site. Water samples were taken at sta- ples werepreserved in 70% ethanol and returned tions 100,300, 600 and 1,800m belowthe treat- to the lab for later analysis. ment site. For each station, sampleswere drawn Drift was sampledfrom 2000 h to 0800 h (for 5 min before B.t.i. application and at intervals 12 h) during each of 6 successivedays or sam- of 5, 10, 15, 30 and 60 min after application. pling periods, as natural drift of stream orga- Samples were taken by submerging and com- nismsis greatestduring this time (Waters1972). pletely filling 500 ml sterile tissue culture flasks. The first two sampling periods started 48 and Sampleswere put on ice immediately after col- 24 h before B.t.i. application. The beginning of lection and kept on ice until they were returned the third period was concurrent with the appli- to the lab, where subsampleswere drawn and cation, while the remaining three periods began preservedin 10% formalin (final concentration, 24, 48 and 72 h after application. The drift net v/v). Dilutions (1:5 or greater) of the samples positions were the same for each of the 6 sam- were prepared using filter-sterilized, deionized pling periods. Stream depth and current flow in water and then stained with DAPI fluorochro- front of each net were measuredjust prior to matic stain (Porter and Feig 1980,Walker et al. removal of the sample to allow for corrections 1988)at a concentration of -6 pg/mI for 25 min in calculating actual drift densities.Drift collec- at 4'C in the dark. Slides were prepared from tions per net per day were then expresseduni- each sample accordingto the method of Hobbie formly as the number of individuals of each et al. (1977),using lrgalan-blackstained filters taxon per total volume of water that passed for a neutral, nonfluorescingbackground. Direct through each net during the sampling period counts of total bacteria (including B.t i. spores) (Waters 1972). were done for 4 replicates of each sample,from The sampling design (6 replications at each at Ieast 15 fields to a total count of at least 200 site with 6 sampling periods, 2 before the treat- bacteria for each filter, using a Leitz Laborlux ment and 4 after the treatment) provided infor- 11 microscopewith the appropriate epifluores- mation on pretreatment and posttreatment drift cent light fittings and excitation and barrier above and below the B.t i. releasepoint. Use of filters. Counts were converted to numbers of an experimental design incorporating both spa- bacteria per ml of river water sample, using a tial (above-and-below)and temporal (before- standard formula (APHA 1980). and-after) sampling (cf., Green 1979)provided Two attempts were made to obtain indirect controls to the treatment effect in space and counts of bacterial spore numbers using agar time. plating proceduresof Gibbs et al. (1986).Sub- Large numbers of chironomids were captured samples(i0 ml) drawn from the original samples in drift nets: therefore. collections of these or- were heat-shockedin a water bath at 60'C for ganismswere subsampledwith a devicemodified 30 min to kill vegetativegrowth, serially diluted, from Waters (1969). An evaluation of the sub- and 1 ml of each dilution plated onto agar in sampler with drift samples revealed no signifi- petri plates, which were then incubated for 24 h cant differencesin the number of invertebrates before inspection for the presenceof bacterial distributed among subunits after subsampling colonies. In the first trial, dilutions were in (12tests, P > 0.05).Total countswere made for powersof 10 from 1:1 to 1:10,000,and plating other invertebrates,including black fly larvae. was directly onto the top of solidified blood agar. Micro-drift sarnples:ln order to determine the In the secondtrial, dilutions were in powers of effects of the B.t.i. application on smaller-sized 5 from 1:1to 1:625,and the sampleswere mixed insects and to determine the suitability of using into unsolidified tryptose agarduring the plating the coarser mesh-sized,Iarger volume nets for procedure. sampling drift, micro-drift sampling was con- Macro-drift samples: Drift sampling has ductedas follows. The nets usedto collect micro- proved an effective method for evaluating the drift were of the "windsock" type, 0.8 meters effects of B.t.i. on black flies and nontarget Iong with a mesh size of 250 pm and a collection invertebrates(Gibbs et al. 1986,Molloy 1989), cup, with a removablescrew cap, at the terminus and was used in this study. Drift nets were of the net (Wilzbach et al.