SEPTEMBER1989 TRAPPTNG Cuttcones SABIIPENNIS

COMPARISONOF SEVERAL DIFFERENT METHODS FOR CUZrcOIDES VARIIPENNIS (DIPTERA: CERATOPOGONIDAE)

JOHN R. ANDERSON AND ARICIO X. LINHARES'

Departm.entof Entomological Sciences,Uniuersity of California, Berheley,Berhelcy, CA 94720

(with.and ABSTRACT. When battery operated CDC miniature incandescentand black light- traps *ii[o"t figftt nulbs) were oper"t.a *ittt and without CO:, the rank oftrap_effectivenessfor total numbers of female"Culicoid.esuariiiinnrs caughtwas: black lighi plus COz;COz-baitedtrap without light bulb; ;i".il'l'gita;iiit;ui cor; l-*""a.r."it light plus co, a".l incaldescentlight without Cor. In 1983,the any-other traps, and the ;i;;i; ffiiii ttip pt". C'O"caught signifriantty more males and females than incande"scentlight trap 1vitftoritCOr"caught sigrrifrcantly fewer femalesthan, any other traps. The-rcwere go-notrophiccondition of G;iit.;;;-;itii;."".". "t.o in times gnit. *ere collecied, as well as in the i"-"""1". .""gftt in different traps. Co"mparativecatches with a vehicle-mounted net revealed that few -"i". ** Jaught in traps and ihat different trnps caught differelt physiological/reproductivesegments of the populatidn. All caichesrevealed that gravid-femalesbegan fl-ying only a.few min betbre sunsetancl larger numbers of females than other ;;;i;;,ii; th; lst h after sunset. Becauseii caught significantly paious females,the black-light trap.would i."p.-th;;;t;i u1d becauseit was seiectivelyattractive tdgravid and ;; cftol.u *tt"tr t"rge numbers of paroul femalesaie desiredin arbovirus isolation studies.

INTRODUCTION 1965,Nelson and Bellamy 1971,Reuben 1963, Rowley,2Service 1971,1973), gnat catchesfrom Light traps commonly have been used for different traps were compared with vehicle- collecting speciesof Ceratopogonidaesince first mounted net catches to assesspotential trap usedby James (1943),but they also have been bias. criticized for not providing an unbiasedestimate ofinsect flight activity andthe gonotrophic state MATERIALS AND METIIODS of the population sampled. When estimates of population agestructure were the objective,var- Study site and trapping of adults: A dairy iousworkers (Barnard 1979,Barnard and Jones located 15 km northwest of Santa Rosa in west- 1980, Bidlingmayer 1961, Nelson and Bellamy ern Sonoma County (ca. 80 km north of San 1971,Rowley2) used such other samplingmeth- Francisco)was selectedfor trapping adults. This ods as suction traps and vehicle-mounted net dairy had a wastewaterholding pond (ca. 100 x catches. 40 x 3 m) positive for larvae of.Culicoides uari- Although it is known that certain kinds of ipennis (Linhares and Anderson, unpublished light are more attractive than others to several data). Culicoidesspecies (e.g., Belton and Pucat 1967, From August 26 through September30, 1982, Rowley and Jorgenssen1967), there is little the following 7 types of traps (Table 1) were information on whether the difference is only exposed 10 m apart along an unobstructed numerical or whether light sourcesof different north-south pasture fencelocated 1 km from the a selective attraction wavelengths might exert lagoon-Type breeding site: to females in different gonotrophic states or at 1: A 6-V CDC miniature light trap3 with different times in a diel. The primary objectives incandescentIight bulb. of this study, therefore' were to determine the Type 2: Same as type 1 and baited with COz. relative effectiveness(: attractiveness) of dif- Type 3: Same as type 2, but without a light ferent kinds of light traps for female Culicoides bulb (: suctiontrap). uariipennis(Coq.) and to determinethe gono- Type 4: A 12-V CDC miniature light trap with trophic condition of femalescaught in different a 4-W black light bulb (ca.320-420 nm). traps and at different times. Since vehicle- Type 5: Same as type 4 and baited with COz. mounted sweepnets and nonbaited suction traps Type 6: A 25 x 38 cm white plywood panel are thought to provide an unbiased sample of coveredwith clear Mylar,@a transparent plastic the flying population of many Culicoidesspecies material. The Mylar sheet was coated with (e.g.,Barnard and Jones1980, Lewis and Taylor TangleTrap,@a sticky substanceused to catch t . P"*"rrt address:Departamento de Parasitologia, Type 7: Same as type 6 and baited with CO2. Campinas,Caixa I.B., UniversidadeEstadual de Dry ice@was used as the sourceof COz' In 6109,13081 Campinas-SP, Brasil. Postal ca.l kg was placed on the ground beneath 2Rowley, W. A. 1965.The occurrenceand bionom' 1982, in the CentralColumbia ics of bloodsuckingmidges t Basin.Ph.D. Dissertation, Washington State Univer- eff figfrttraps were manufactured by the JohnW. sity,92 pp. HockCo., P.O. Box 12582,Gainesville, FL 32604. 326 Jounrel oF THE AtrleRrceuMosqurro CoNtnor, Assocrlrror,t VoL. 5,

Table 1. Number of Culicoidesuariipennis collectedin 1982by ? different trap types, from 2 h before to t h after sunset. Trap typer Black Incandes- Black Suction Sticky Incandes- Black light light cent Iight cent trap * CO2 trap' trap trap * light + + coz + co, trap light trap CO, CO" only only Total Date

Aug.26 7 310 1830210 25000 0 10 539 Sept.2 0 31 0 50190 00102058 Sept.9 3 46 0 009002500 05105 Sept.16 0 3 0 00150270000045 Sept.23 2l ror 6 151 4 173 2 5911500 34 565 l RO A ,250 Sept.30 9 18 3 0 11300 17 424 Oct.7 13 4l 43 324 0 220390 30110 0 56 440 Oct. 14 6 118 21 158 0 11 12 t20 2 69020 0 41 478 Oct.2l 23 726 11 57 1000210 70000 37 311 o 1qa Oct. 28 8 69 LO 113 33 161 0 00000 66 470 Nov.4 14 134 35 193 t) 195 4 L79 0 00000 59 70r Nov. 11 0 z 070 580300 00000097 Total 104 1,199 119 866 38 856 56 1,037 4 191 4 820 2 325 4,233 (a)' (a) (a.) (a) (a) (a) (a) (a) (a) (a) (a) (b) (a) (b,c) t For all trap types except black light trap plus COz above, 1 kg of dry ice was exposedin a perforated container set on the ground beneath each baited trap. The unbaited sticky trap caught ' no C. uariipennis. A CDC miniature light trap operatedwithout light bulb. o Results of Kruskal-Wallis' one way ANOVA test applied for female catcheswere significant at p < 0.001 QZ: a7.59) and for male catchesat P < 0.001 (x3 = 80.99).For each sex,trap catchesfollowed by at least one similar letter did not differ significantly when Bonferroni's inequality multiple comparisonstest was used for 21 comparisons(overall errol rate a : 0.0b), for pairwise comparisonsofdifferent trap types. each baited trap at the start of trapping. Each that each trap had at least 0.25 kg at the end of block of dry ice was exposedin a 1.gliter waxed the last period. cardboard container with eight 1-2 cm holes Each light trap had an identical motor and near the bottom. Each sticky panel trap was fan mounted in the samesize cylinder. Voltages staked with the bottom of the panel about 0.Sm at the motor terminals (with or without the from the ground, and light traps were hung from lights operating) are 6.1 for the CDC incandes- a fencewith entrancesca. 0.5 m from the ground. cent trap and 6.3 for the UV trap (John W. PrevailingW, SW or NW winds blew acrossthe Hock, personal communication). On each night north-south oriented trap line so that COz- of exposure all light traps were operated with baited traps were not exposedupwind of non- fully charged batteries. Each light trap was baited traps. On each collection date, the traps' equippedwith a removablenylon mesh collect- relative positions were randomly assignedalong ing bag into which insects were blown afber the fenceline. The 10-m spacingwas selectedso passing through the fan located beneath the that insects attracted into the vicinitv of the light. Collection bags for all traps were replaced traps being compared could perceive and re- hourly, as were gnat-positive Mylar sheets. spond to any of several traps. Most traps were Temperature data for the Santa Rosa mete- nearly equally visible from most orientations. orological station was obtained from the Cli- Trapping was conductedweekly from 2 h before matological Data for California, Vol. 82. Na- to t h after sunset, from August 26 through tional Climatic Data Center, Asheville, NC. November 11, for a total of 12 collection dates. Times of sunset on collection dates were ob- On October 7, a CDC miniature black light trap tained from annual issues of the American baited with CO2was addedto the design (Table Ephemerisand Nautical Almanac(U.S. Govern- 1). With this trap, 1 kg of dry ice was placed in ment Printing Office, Washington, DC). Wind a perforated plastic bag on top of the trap. velocity and direction, temperature, RH and Collection bags for all traps were replaced Iight intensity were also recorded at half-hour hourly. The dry ice supply for each trap was intervals at the sampling site. monitored and supplemented as necessary so In 1982,trapping was terminated on Novem- SEPTEMBER1989 TRAPPING CULrcOIDES VANIIPENNIS

ljd. ber 11 becauseregular sampling of immatures kg of dry ice and coveredwith a styrofoam holes at the lagoon breeding site revealedthat pupa- The COz escapedthrough ten 1-1.5 cm tion had ceasedabout 3 weekspreviously. Pre- punchedin the bottom ofthe pails.In 1983tr?ps liminary sampling in previous years also had were operated from June 9 through November shown that. at the onset of winter, all larvae 3. Trapping stopped after November 3 because, that disappearedfrom the shoreline and pupation as in i982, samplesof immatures revealed previously. In ceased. pupation had ceaseda few weeks at In 1983,we modifiedthe experimentaldesign Loih y.utt, the collecting bags were checked pre- by: 1) operating 2 replicates of trap typeg 1-5 hourly intervals and changedif gnats were (Table 2) and eliminating trap types 6 andT;2) sent. The dry ice supply was monitored and placing the dry ice above all traps; and 3) oper- supplementedas in 1982' of light ating traps from 2 h before to 2 h after sunset ln addition to the regular operation (115 (Table 3). traps in 1983,a vehicle-mountedsweep net of Trap types 2,3 and 5 eachhad the bottom of cm diam x 150 cm long), similar to that to the a styrofoaminsulatedpail (20 cm diam x 15 cm Nelsonand Bellamy (1971),was attached insects deep) bolted to the topframe located 8-10 cm top of a station wagon and used to catch 18 and above the trap entrance. At the beginning of on 8 collecting dates between August each trapping night each pail was baited with 1 September 29. On each date multiple samples

types, from Table 2. Number of male and female Culicoid.esuariipennis collected in 1983by 5 different trap 2 h before to 2 h after sunset.r TraP tYPe Incan- Incandes- Black light Suction trap2 Black light descent cent Total trap * CO2 + CO2 traP onlY Iight trap light + COz only

Date 2i 406 June 9 2l 204 / 0 81 0 7 0 0 14 366 June 16 13 239 0 13 I 111 0 0 0 8 0 0 I 333 June 23 4 I II) 0 82 0 t27 0 6 77 June 30 0 13 0 8 0 7 b 49 0 0 I I 143 July 7 0 57 0 13 0 Ib 0 0 0 0 1 54 July 14 0 JI 0 2 0 L2 1 9 24 252 July 21 94 0 T2 0 15 2l 131 0 0 L7 aln July 28 3 L45 0 to 0 10 l4 r46 0 0 en 369 Aug.5 34 256 0 28 0 40 43 0 2 13 13? Aug. 12 7 85 0 10 0 2 6 40 0 0 ,.4 156 1,668 Aug. 18 149 1,085 0 90 6 468 1 0 1 254 Aug. 25 13 68 2 126 0 39 1 2l 0 0 16 2t 449 Aug. 29 2l 0 J+ 0 40 0 24 0 0 I 76 OU.J Sept. 2 n1 386 0 10 0 72 0 /7 28 332 Sept.9 2t 129 0 20 0 31 r52 0 0 1n 1 143 Sept. 16 0 84 0 2T 0 1 2T 0 0 81 1,093 Sept. 19 17 273 1 138 0 32 69 650 0 0 0 62 916 Sept. 22 31 318 0 15 0 I6 ot 565 0 8 r97 Sept. 29 6 135 0 0 t2 2 19 0 0 2 250 Oct. 5 138 0 48 0 46 0 15 0 29 1,068 Oct. 13 22 659 0 170 7 202 0 o{ 0 0 106 t,476 Oct. 20 81 812 6 316 13 276 6 70 0 2 0 98 2,250 Oct.27 o, r,404 1 7l t7 583 13 r92 0 I I 20 586 Nov. 3 l4 362 0 DO 2 118 50 0 13,639 Total 600 7,444 t4 r,443 46 2,378 195 2,364 0 10 855 (a)' (a) (b) (b) (b, c) ft) (a, c) (b) (b) (c) in a {\r" ""plt."t"" "f """h t.^p ere operated each date. For each trap 1 kg of dry ice was exposed perforated container suspendedover the entrance to the trap' ' A CDC miniature light trap operatedwithout light bulb. (1! : 3The results of Friedman's'two way ANOVA test were significant for female catches at P < 0'001 letter 66.g9),and for male catchesat ]? < 0.001 (x|:53.82). For each sex, trap catches followed by the same aia "oL aifi"" *ignifrcantly, when Bonferroni's inequality multiple comparisonstest was usedfor 10 comparisons (overall errot .ate a = 0.05),for pairwise comparisonsof different trap types. 328 JouRNer,oF THEArr,rpRrcnr.i Moseurro CoNrnor,Assocrarron VoL.5,No.3

Table 3. The number and gonotrophic/physiologicstatus of female Culicoidesuariipennis collectedin 1983by 4 different trap types. Gonotrophic/physiologicstatus

Trap type Hourt Nullioarous Parous Gravid Engorged2 Total

Black light I 387 269 o 0 659 * COz II r,782 763 0 2,548 III 7,2T3 1,458 639 a 3,312 NIV 263 387 274 I ot< Total3 3,645(a) 2,877(a) ele (b) 7,444(a) Incandescentlight I48 20 I 0 69 + co, II 600 236 I 0 837 III 2II 160 z 0 d l.l IV66 89 q 0 164 Total 925 (b) 505 (b) 13(c) 0 1,443(b) Suction trap I 155 88 0 0 243 + COz II 969 409 2 0 1,380 III 463 236 2 1 702 NIV23 30 0 0 53 Total 1,610(a, b) 763 (b) a (c) 1 2,378(b) Black light IO0 0 0 0 IIO 0 0 0 0 uIII22 2t4 t,209 2 1,447 ry29 118 766 ^ 917 Total 51 (c) 332 (b) 1,975(a) 6 2,364(b) Grand totals 6,23r 4,477 2,91r 10 13,629 I There were two l-h trapping periods before, and two after, sunset. 'All femalesappeared nearly fully engorgedwith blood. 3 The two way ANOVA (SAS GLM procedure)for differencesin numbers of nulliparous, parous, gravid and engorgedfemales caught in the different traps was highly significant (F = 15.55,P < 0.001).Total trap catches in vertical columns followed by the sameletter did not differ significantly when Bonferroni's inequality multiple comparisonstest was usedfor 6 comparisons(overall error rate: a : 0.05),for pairwise comparisonsof different trap types. were taken as the vehicle was driven over a 2.9 trophic state. For several of the collections, km route. The collecting route followed an off- hourly catches of females were very large (sev- road track leading through a hayfield and an eral hundred/trap) for some of the traps. For unused field around 3 sides of a cornfield. The practical reasons,therefore, a maximum of 50 cornfield was adjacent to and N and ENE of females/trap/h was randomly selected and ex- wherethe light traps were operated.During each amined to determine their gonotrophic state. hour that light traps were operated,vehicle net The results were then used to estimate the pro- collections were started 15 min into the hour portion of flies in different gonotrophic states and 15 min before a trap hour ended.These two for those specific traps and hourly catches. collections were combined to represent hourly Determination of gonotrophic conditian: Col- catches for comparison with the Iight trap lected femaleswere examined with a dissecting catches.During the collectingtrips the vehicle microscopeand classifiedinto 4 categories:a) was driven at 16-25 km/h. It took about 8 min nulliparous,b) parous,c) gravid and d) blood to completethe route, after which the collecting engorged. If a female was neither gravid nor bag was replaced. Each trip started and ended engorged,the parity status was determined by at a point about 40 m from the fence supporting the absenceof abdominal pigment as described the light traps. Driving lights on the vehicle by Dyce (1969)and Potter and Akey (1978),and were operatedon low beam when it becametoo by the pattern of abdominal tergites II and III dark to follow the route safely without their use. (Akey and Potter 1979). Flies collected in the light traps and vehicle- Since flies had been frozen at -30'C for sev- mounted net were held in ice-cooled insulated eral weeksbefore they were examinedand dis- containers until taken to the laboratory. Here, sected, it was not possible to check for the they were anesthesizedwith COr, placed in snap possible occurrence of multiple oviposition cap glassvials, sealedin polyethylene bags and cyclesby using the method of Polovodova (De- frozen at -30'C until they could be dissected tinova 1962).Loss of elasticity in the ovarian and examined for determination of their gono- tissues of frozen flies made it too difficult to 329 SEPTEMBER1989 TnepprNc Cut'tconns rARIIPENNIS stretch the ovarioles to check for degenerative light traps, and none were caught in nonbaited follicles or oviposition scars in the ovaries as incandescentlight traps. However, in southern (1971) describedby Mullens and Schmidtmann(1982). California, Nelson and Bellamy caught We were not able to determine the parity much greater numbers of males in COz-baited status of gnats collectedby sticky traps in 1982 incandescent light traps than were caught in becausethe sticky material permeatedthe cuti- this study. Since male swarms have a nonran- cle and obscuredcolor differences. These flies, dom distribution (Zimmerman et al. 1982)' it is therefore, were not included in the following possible that some of the Nelson and Bellamy analysis. Iraps *e.e located near swarming sites or that Data analysis:The Kruskal-Wallis one-way malesmay have been attracted to femalesin the ANOVA test (Zar 1974) was used to determine traps. if differences existed in the numbers of gnats For females,statistical analysesshowed a sig- caught among the several trap types used in nificant differencein effectiveness(: attractive- 1982.The Friedman two-way ANOVA test (Zar ness) among the severaltypes of traps used. As 1974)was used for the 1983 data. In both years, the identical motors and fans on all traps pro- Bonferroni's inequality multiple comparisons vided equal suction we conclude,in all analyses, test (Miller 1981) also was used to Iook for that the differences in numbers of females specific pair-wise differences among the traps. caughtwere directly related to the attractiveness The Kruskal-Wallis test was used instead of the ofthe light sourceand to the presenceor absence Friedman test for the 1982 data becauseof the of COz. For 1982, multiple comparison tests smaller number of replicate nights for the black revealed that Coz-baited traps were signifi- Iight trap baited with COz above. The unequal cantly more attractive to femalesof C- uariipen' number of replicates in 1982 precluded the use nis regardlessof the kind of light source used- of the Friedman test. However, no significant difference was found A Log linear analysis(Fienberg 1977) using a among the COz-baitedIight traps even though, three dimensional table (trap type vs. trapping on average,the black light trap with COz above time vs. gonotrophic status) was used for simul- collected many more gnats/trapping night than taneoustesting for complete independence,par- the other traps (Table 1). tial independenceand conditional independence The more extensiveoperation of traps in 1983 amongthe 3 variablesfor both 1982and 1983. revealed clear differences in attractiveness The analysiswas performed on a computer using among the different types of traps tested. The either the SPSSx Loglinear Procedure(SPSS Coz-baited black light trap was significantly Inc. 1983) or the SAS Catmod Procedure (SAS more effective than all other traps in catching Institute 1985). females of C. uariipennzs(Table 2). About 55% The ANOVA was conducted with the SAS of all trapped females in 1983were collected by GLM program (SAS Institute 1985)to test for this type of trap. The other 2 types of COz- significant differences in numbers of nullipa- baited traps (incandescentlight trap and suction rous, parous, gravid and engorged females trap ) collected,respectivel y, t07o and L77o of the caught in the different traps and to determine if flies caught. Of the 2 types of nonbaited traps, there were signifrcant differencesin numbers of the one with a black light bulb was as effective nulliparous, parous, gravid and engorged fe- as the Coz-baited light trap without a light bulb' males caught in each trap during different trap- at least in numerical terms. It caught aboutl7% ping periods. The same procedure was used to of the total flies collected.The nonbaited incan- test for differences in numbers of nulliparous, descentlight trap was by far the least attractive; parous and gravid femalescaught in the vehicle it caught only 10 of 13,639female gnats in 1983 mounted sweep net during different trapping and only 2 of 4,248 females in 1982. Incandes- periods. In all analyses,the dependentvariable cent light, in fact, appearedto have a repellent was Log (n + 1). effect on females (compare catches in suction trap plus COz with incandescentlight plus CO2 in Tables 1-3). AND DISCUSSION RESULTS Although the black light trap was very effec- Analysisof trap catches:Total trap catchesfor tive in catching females of C. uariipennls, its 1982 and 1983 are shown in Tables 1 and 2, attractivenesswas enhancedwhen COzwas used respectively.In 1982, few males were caught in as an additional attractant. The positive attrac- any traps. In the more extensive 1983 study, tive effect of COz is well-known for mosquitoes black light traps caught significantly more males (Gilliesand Wilkes 1969;Reeves 1951, 1953), as than any other traps. Unlike females,males were well as for other blood sucking flies (Anderson not attracted to CO2; only small numbers of and Hoy 1972,DeFoliart and Morris 1967),in- males were caught in Coz-baited incandescent cluding Ceratopogonidae(Birley and Boorman 330 Jounulr, oF THEAunucan Mosquno Coxtnol AssocretroN Vor,.5, No. 3

1982,Mullens and Rutz 1984,Nelson 196b.Nel- and parous femalesboth beforeand after sunset, son and Bellamy 1971). and to the attraction of gravid femalesto black Host-seeking , females of C. uariipennis werc light after sunset. As the non-COz-baitedblack first caught in large numbers in Co2-[ai1sdtraps light trap caught no gnats during the first two (1965), by Nelson who suggestedthat CO, exposureperiods (Table 3), it was surprising to played a role in directing host-seeking female seethat the CO2-baitedblack Iight trap caught gnats to hosts. Nelson basedthis interpretation so many more host-seeking females than the on the fact that nearly all femalescolleited were other 2 types of Co2-baited traps. neither gravid nor engorged.The results ofthis The Coz-baited incandescentlight trap and study support that interpretation. However, our Coz-baitedsuction trop: Catches in these traps results also revealedthat the type (color) of light for both 1982 and 1983 were similar (Tables 1- used with the traps also influenced the nature 3). As the non-Coz-baitedincandescent light of the catch. The greater effectiveness of the trap rarely caught C. uariipennis (Tables 1 and black light traps in this study confirms the find- 2) we concludethat the COz-baitedincandescent ings of Belton and Pucat (1967)and Rowleyand trap and the Coz-baited suction trap (1967), trgh! Jorgenssen who found that black light (Table 3) both functioned essentially as COr- traps were much more efficient than incandes- baited traps in this study. Although there is an cent light traps for catching females of several indication that incandescentlight had a repel- Culicoides species, including C. uariipennis. Ient effect on somefemales, there was no signif- However, these researchersdid not use COr as icant difference in numbers of femalescaught in bait, and they did not determine the gonotrophic these 2 types of traps (Table 3). Except for state of the femalescollected. Significant biolog- nullipars caught in the suction trap, both types ical differences related to these factors are dis- of traps caught significantly fewer femalesthan cussedbelow. the Coz-baited black light trap (Table 3). Re- The results of the Log linear analysis were sults for the 1982 catches were similar highly (P significant < 0.001)for both the 1982 (Linharesa).The most interestingresult wasthe and 1983 data, revealing that the type of trap, poor attractivenessof both trap types for gravid time gnat of collection and gonotrophic status females.Hence, in contrast to black light, incan- were n-otpartially, conditionally nor totally in- descentlight provided little attraction, if any, to dependent. females of C. uariipennis, regardlessof their The nonbaitedblack light trap: This trap was gonotrophic state. highly selective for gravid females and, to a Gonotrophicstate in relation to trapping time: certain degree,for parousfemales (Table B). Ir The results for all 1983 traps combined (Table also waseffective only afbersunset with no gnats 3) showed that significantly fewer nulliparous, being collected during the first 2 h of trapping. parous and gravid females were caught during Results for 1982were similar (Linharesu).Ovei- the first hour of trapping, and that sigaificantly all, this trap caught significantly fewer nullipa- greater numbers of gravid females were caught rous femalesthan any other tlpe of trap (Table during period III (Table 3). All engorged,and dr. most gravid females,were collectedafter sunset, The Coz-baited btack light trap: Combining COzwith this trap resulted in a highly attractive synergistic effect (Tables 2 and B: see Table 6). Table4. The numberand gonotrophic status of This trap caught significantly more nulliparous Culi.coi.desuariipennis caught in 19g3in vehicle- and parous femalesthan the other types of traps mounted sweep net collections. (Table 3), and it caught significantly more Gonotrophic status of gravid femalesthan the other 2 types of COz- femalesl baited traps (incandescent light plus CO2 and Nullip- suction plus COr). However, it caught signifi- Date Males arous Parous Gravid Total cantly fewer gravid femalesthan the black light trap without COz (Table 3), indicating that CO, Aug. 18 2,150 3061450 Aug. 25 had a general repellent effect on gravid 3,632 6061480 females Sept.2 2,99I (Table 3). Nevertheless, general 3945396 the effective- Sept.9 4,860 1633352 ness of this trap appeared to be due to release Sept.16 2,147 26 14 13 53 of COz,which attracted host-seekingnulliparous Sept.19 9,563 95 27 23 t45 Sept.22 3,799 75 20 19 tt4 a Sept.29 342 153220 Linhares, A. X. 1984. The biology of Culicoi.d,es uariipennis,(Coquillett)(Diptera: Ceratopogonidae)in Total 29,484 356 83 I7t 610 Northern California. Ph.D. Dissertation. Universitv Percentage (58.36) (13.61) (23.03) of I California, Berkeley, 153 pp. No blood engorged females were caught. SnprnNrenn1.989 TnepprNc Cuttcotoos qARIIPENNIS JOt and the vast majority of both were caught in Of the 610 females collected in the vehicle- black light traps (Table 3). The 10 gravid fe- mounted net catches, 28Vo were gravid, 58Vo malescaught prior to sunset (Table 3) may have were nulliparous and only 147o were parous been prematurely stimulated to undertake a (Tables5 and 6). These resultsdiffer consider- host-seekingflight when COz-baitedtraps were ably from those of the different trap types op- placed near their resting sites. Since it is not erated during the same period (August 18 unusual to collect small numbers of gravid and through September 29) (Table 6). During this engorgedfemales of haemotophagousspecies in time gravid females comprised only L4% of the host biting collections,it doesnot seemunusual total catch in COz-baited black light traps that a few such C. uariipennis femaleswould be (Table 6). The two other types of COz-baited attracted to COz. traps wereeven less effective in attracting gravid Cornparison of trap catches with uehicle- females,although they collected large numbers rnountednet catches:Unlikethe Co:-baited light of nulliparous gnats. All 3 types of traps oper- traps, the vehicle-mounted sweep net collected ated with COzalso attracted a higher percentage large numbers of males, greatly outnumbering of parous femalesthan was caught in the vehi- females(Table 4). While making net collections cle-mountednet (Table 6). with the vehicle, we noticed that, at times, the During the 8-day comparisonof trap and ve- vehicle was driven through severallarge swarms hicle net catches,85% of the females caught in of males. This occurred mainly around sunset, the non-COz-baitedblack light trap were gravid and we believe it accountsfor the large catches (Table 6), a percentagevery close to that for of males. this trap for the total trapping periods in 1982 In the case of females: 1) significantly fewer (Linharesa)and 1983 (Table 3) and verv differ- nulliparous individuals were caught during the ent from the 28% caughtby the vehicle-mounted first and last hoursofcollection,2) significantly net. Although gravid females comprised nearly fewer parous individuals were caught during the 15% (24 out of 1?1) of the total catch in the first hour and 3) sigrrificantly more gravid indi- vehicle-mountednet before sunset (Table 6), the viduals were caught during the third hour (Table data in Table 5 show that 23 of the 24 gtavid Dl. femalesnetted during the secondhour of collec-

Table 5. The number and gonotrophic status of female Culicoidesuariipennis collectedin 1983with a vehicle- mounted sweepnet in relation to hour of collection. GonotroPhicstatusr Nulliparous Parous Gravid Total Collection 12 0 0 0 0 Collection 2 0 0 d Total, Hour I3 3 (b) 0 (b) o (c) 3 (c) Percentage (0.49)4 (0) (0) (0.4e) Collection 3 t4 1 16 Collection 4 24 23 224 Total, Hour II 191(a) 25(a) 24 (b) 240 (a) Percentage (31.31) (4.10) (3.e3) (3e.34) Collection 5 r45 26 116 287 Collection 6 10 10 2l 44 Total, Hour III 155(a) 36 (a) 137(a) 328 (a) Percentage (25.4r) (5.e0) (22.46) (53.77) Collection 7 o tl tb Collection 8 D to 23 Total, Hour IV 7 (b) 22 (a) 10 (b, c) 3e (b) Percentage (1.15) (3.61) (1.64) (6.40) TotaI 356 83 t7l 610 Percentage (58.36) (13.61) (28.03) 1No blood engorgedfemales were caught. 2Collections 1-4 (first 2 h) were before sunset and collections 5-8 (last 2 h) were after sunset. 3The two way ANOVA (SAS GLM Procedure)for differencesin numbers of nulliparous, parous and gravid femalescaught in the different trapping periods was significant (F: 9.50,P < 0.01). Total catchesin vertical columns foll,owedby the sameletter did not differ significantly when Bonferroni's inequality multiple compar- isons test was used for 6 comparisons(overall error rate: a : 0.05), for pairwise comparisonsof different trap types. aThe numbers in parenthesesreflect the percentagesof the total number collected (610)' 332 JounNar,oF THEAlrnRrceN Mosqurro Cournol Assocrlrrox VoL.5,No.3

Table 6. The number of female Culicoidesuariipennis caught in relation to trap type and gonotrophic status on 8 dateswhen all traps and the vehicle-mountednet were operated.r Trap type (B) Incan- (c) (D) Traps Vehicle- (A) descent Suction Black (A), (B), (C) Gonotrophic mounted Black light light trap Iight and (D) status sweep net * co: + co2 + COz only combined

Hour I D 47 11 21079 Hour II 191 653 278 427 0 1,358 Nulliparous Hour III 155 482 44 42 3 571 Hour IV 7 235 15 3 14 267 Total 356 t,4t7 (a)' 348 (b) 4e3 (b) 17 (c) 2,275 (58.36)3 (57.23) (77.16) (76.55) (1.12) (44.65) Hour I 0 297 13049 Hour II 25 269 64 96 0 429 Parous Hour III 36 347 20 39 134 540 Hour IV 22 76 11 3 74 t64 Total 83 72L (a) 102(b) 151 (b) 208 (b) 1,182 (13.61) (2e.L2) (22.62) (23.45) (13.65) (23.20) Hour I 0 00000 Hour II 24 10001 Gravid Hour III IJI 258 0 0 931 1,189 Hour IV 10 79 1 0 368 448 Total t7l 338 (b) 1 (c) 0 (c) 1,299(a) 1,638 (28.03) (13.65) (0.22) (0) (85.24) (32.15) Totala 610 2,476 457 644 t,524 5,095 l Trapping dates are the same as those in Table 4. The two-way ANOVA (SAS GLM Procedure) for differencesin numbers of femalesin each gonotrophic state caught in trap types A-D was highly significant (F :13.47.P<0.001). 2Total trap catches in horizontal rows followed by the same letter did not differ sigrrificantly when Bonferroni's inequality multiple comparisonstest was used for 6 comparisons(overall error rate: a : 0.05), for pairwise comparisonsof different trap types. 3The numbers in parenthesesreflect the percentagesof the total collected by each kind of trap (column only). a In addition to the flies accountedfor in this table, 2 engorgedfemales were caught in the black light trap plus COz;one was caught during hour III and one during hour IV. tion were caught in the last of the two 8-min baited traps during the first hour. Becausethe vehicle runs (completedonly a coupleof minutes vehicle net collections (Tables 5 and 6) and non- before sunset).This confirms the results of trap COz-baitedtraps (Tables 3 and 6) revealedthat catches that revealed that few gravid females females in all gonotrophic states did not begin were flying prior to sunset (Tables 3 and 6). In flying until a few minutes before sunset (e.g., northeastern Colorado, Akey and Barnard Table 5, coll. 4), the small catchesof femalesin (1983) also reported a large increase in gravid Cor-baited traps during the first hour may have femalescaught in a vehicle net just after sunset. been due to the fact that the COz prematurely A comparisonof trapping resultswith vehicle- stimulated a host-seekingflight by femalesrest- mounted net catches(Table 6) showsthe nature ing in the near vicinity of the traps. Overall, of the bias for each trap type. Overall, the COz- proportionately more nulliparous females were baited black light traps overestimatedthe pop- caught before sunset by all trap types with CO2 ulation of parous flies but underestimated the and by the vehiclenet. Akey and Barnard (1983) population of gravids. The COz-baited suction obtained similar results for vehicle net catches. trap and COr-baited incandescent light trap When catch results for all light and CO2- overestimatedboth nulliparous and parous pop- baited traps are combined, the bias associated ulations but greatly underestimated the gravid with individual trap types decreasesconsider- segmentof the population. The nonbaited black ably (Table 6). Hence, the percentagesof gravid light traps greatly overestimatedthe gravid seg- femalescaught in all trap types and the vehicle- ment of the population but exerted little attrac- mountednet are very similar (32% vs.28%), as tion to host-seekingfemales. The flying activity are the catchesfor the host-seekingsegment of of C. uariipennis was overestimatedby all COz- the population (68Vovs.72Vo). Parous females SEPTEMBER 1989 TupprNG Cunrconas SARIIPENNIS 333 are still somewhat overrepresentedwhen trap iipennis (Diptera: Ceratopogonidae)in northeastern catches are combined, and nulliparous females Colorado.Environ. Entomol. 12:91-95' 1979.Pigmentation somewhat underrepresented. Nevertheless, as Akey, D. H. and H. W. Potter. with oogenesisin the biting fly Culicoides judged by the nonattractive samplesobtained in associated uariipennis (Diptera: Ceratopogonidae):determina- for all the vehicle-mounted net, combined data tion of parity. J. Med. Entomol. 16:67-70. types of traps provides a better measureof the Anderson,J. R. and J. B. Hoy. 1972.Relationship parity profile and the reproductive status ofthe between host attack rates and COz-baited population than the catch from any single type flight trap catchesof certain Symphoromyiospecies. of trap. The biased nature of female catches in J. Med. Entomol. 9:373-393. relation to their gonotrophic state limits the Barnard, D. R. 1979. A vehicle mounted insect trap. usefulnessof certain types of light traps in stud- Can.Entomol. 111:851-854. DieI and sea- ies of the population age structure of female C. Barnard. D. R. and R. H. Jones. 1980. patterns of flight activity of Ceratopogonidae Bidlingmayer (1974) reported simi- sonal uariipennis. in northeastern Colorado: Culicoides.Environ. En- lar results for mosquitoes. tomol. 9:446-451. Belton, P. and A. Pucat. 1967.A comparisonof differ- CONCLUSIONS ent lights in traps for Culicoides(Diptera: Cerato- pogonidae).Can. Entomol. 99:267-27 2. This is the first time that black light traps Bidlingmayer, W. L. 1961.Field activity of adult Cul- have been found to be selectively attractive to icoi.desfurens. Ann. Entomol. Soc. Am. 54:149-156. gravid and parous females of C. uariipennis. Bidlingmayer,W. L. 19?4.The influenceof environ- Zimmerman and Turner (1983)previously re- mental factors and physiological stage on flight in the vehicle aspirator ported high parous rates for C. uariipennis patterns of mosquitoestaken suction, bait and New Jersey Iight traps. traps; but since only black and truck, caught in black light J. Med. Entomol. 11:119-146. possible light traps were used,it was not to make Birley, M. H. and J. P. T. Boorman.1982. Estimating comparisons.Since our black light traps baited the survival and biting rates of hematophagousin- with COz also caught significantly higher num- sects, with particular reference tn the Culicoi'des bers of gnats than did other traps, this kind of obsolctus group (Diptera: Ceratopogonidae). J. trap should be the trap of choice when the Anim. Ecol.51:135-148. objective is to catch large numbers of parous De Foliart, G. R. and C. D. Morris. 1967.A dry ice- females,as in arbovirus isolation studies. baited trap for the collection and field storage of Entomol. 4:360- The ineffectiveness of black light traps for hematophagousDiptera. J. Med' gravid before sunset seemed related 362. females Detinova,T. S. 1962.Age-grouping methods in Dip- more to the fact that such femalesusually began tera of medical importance. World Health Organ. flying only a few minutes before sunset (Tables Monogr.Ser. 47, 266pp. 3 and 5) than to thesetraps not being visible to Dyce, A. L. 1969.The recognition of nulliparous and the flies prior to sunset.In 1983,the black light parots Culicoideswithout dissection. J. Aust. En- trap plus COz, in fact, caught 6 gravid females tomol. Soc.11:91-96. before sunset (Table 3), as well as catching more Fienberg, S. E. 1977.The analysis of Cross-Classified host-seeking females before sunset than the categoricaldata. The Cambridge,MA: M.I.T. Press, other types of COz-baitedtraps (Tables 3 and. 151pp. Gillies, M. T. and T. J. Wilkes. 1969.A comparisonof 6). the rangeofattraction ofanimal baits and ofcarbon dioxide for some west African mosquitoes. Bull. ACKNOWLEDGMENTS Entomol.Res. 59:441-456. James,M. T. 1943.The genus Culicoidesin northern We thank Manweiler for assistance Stephen Colorado (Diptera: Ceratopogonidae).Pan. Pacific in operating traps on several nights and Harri- Entomol. 19:148-153. son Stubbs for advice on statistical analysis of Lewis, T. and L. R. Taylor. 1965.Diurnal periodicity data. We also thank Darold Batzer, Stephen of flight by insects.Trans. R. Entomol. Soc.Lond. Manweiler and Bradley Mullens for reviewing 116:393-469. the manuscript. This study was supported, in Miller. R. G. 1981.Simultaneous statistical inference. part, by funds from a University of California Seconded. New York: Springer-Verlag,299 pp. Cooperative Extension Management Mullens, B. A. and D. A. Rutz. 1984. Age structure Grant. A. X. Linhares receivedpartial financial and survivorship of Culicoidesuariipennis (Diptera: York state. Ann. support from the Brasil National Council for Ceratopogonidae)in central New Entomol. Soc.Am. 76:913-917. Scientific and Technologic Development and Mullens, B. A. and E. T. Schimidtmann. 1982. The the State University of Campinas. gonotrophiccycle of Culi.coidesuariipennis (Diptera: in age-grad- REFERENCES CITED Ceratopogonidae)and its implications ing field populations in New York State, USA. J. Akey,D. H. and D. R. Barnard.1983. Parity in air- Med. Entomol. 19:340-349. bornepopulations of the biting gnatCulicoid.es uar- Nelson, R. L. 1965. Carbon dioxide as an attractant 334 JounNer,oF THEAnrpnrclN Moseurro Conrnol AssocrATroN Vor,.5, No. 3

for Culi.coi.des.J. Med. Entomol. 1:56-57. 1479. Nelson, R. L. and R. E. Bellamy. 1971.Patterns of SAS Institute Inc. 1985. SAS users guide: statistics, flight activity of Culicoidesuariipennis (Coquillet) version 5thed. Cary, NC. (Diptera: Ceratopogonidae). J. Med. Entomol. Service,M. W. 1971.Adult flight activities of some 8:283-291. British Culicoidesspecies. J. Med. Entomol. 8:605- Potter, H. W., and D. H. Akey. 1978.Pigmentation 609. associatedwith oogenesisin the biting midge,Culi- Service, M. W. 1973. Further results of catches of coidesuariipennls: changes in abdominal tergite pat- Culicoides(Diptera: Ceratopogonidae)and mosqui- terns. Mosq. News 38:499-504. toes from suction traps. J. Med. Entomol.ll:471- Reeves,W. C. 1951. Field studies on carbon dioxide 479. as a possible host stimulant to mosquitoes. Proc. SPSS Inc. 1983.SPSSx usersguide. A completeguide Soc.Exp. Biol. Med. 77:64-66. to SPSSx languageand operations.Chicago, IL. Reeves, W. C. 1953. Quantitative field studies on Zar, J. H. 1974. Biostatistical analysis. Englewood carbon dioxide chemotropismof mosquitoes.Am. J. Cliffs, NJ: Prentice-Hall,620 pp. Trop. Med. Hyg. 2:325-331. Zimmerman,R. H., S. J. Barker and E. C. Turner, Jr. Reuben, R. 1963. A comparison of trap catches of 1982. Swarming and mating behavior of a natural Culicoidesinpunctatus (Diptera: Ceratopogonidae) population of Culicoi.desuariipennis (Diptera: Cer- with meteorological data. Proc. R. Entomol. Soc. atopogonidae).J. Med. Entomol. 19:151-156. Lond. (A) 38:181-193. Zimmerman,R. H. and E. C. Turner, Jr. 1983.Sea- R.owley,W. A. and N. M. Jorgenssen.1967. Relative sonal abundanceand parity of common Culicoides effectivenessof three types of light traps in collect- collected in traps in Virginia pastures. ing adult Culicoides.J. Econ. Entomol. 60:1478- Mosq.News 43:63-69.