Joumal of the American Mosquito Control Association, l3(l):71J5' L99'l Copyright @ 1997 by the American Mosquito Control Association' Inc.

HOST PREFERENCE OF MOSQUITOES IN BERNALILLO COUNTY NEW MEXICO'''

K. M. LOF[IN,3j R. L. BYFORD,3 M. J. LOF[IN,3 M. E. CRAIG,T t'ro R. L. STEINERs

ABSTRACT. Host preference of mosquitoes was determined using animal-baited traps. Hosts used in the study were cattle, chickens, dogs, and horses. Ten mosquito species representing 4 genera were collected from the animal-baited traps. Aedes vexans, Aedes dorsalis, Culex quinquefasciatus, Culex tarsalis, and Culiseta inomata were used as indicator species for data analysis. Greater numbers of Ae. vexans, Ae. dorsalis, and C.s. inornata were collected from cattle and horses than from chickens or dogs. In addition, engorgement rates were higher on mammals than on chickens. Engorgement and attraction data for Cx. quinquefasciatrs suggested a preference for chickens and dogs over cattle and horses. A slight preference for chickens and dogs was seen with Cx. tarsalis, but the degree of host preference of C.r. tarsalis was less than that in either Ae. vexans or Cx. quinquefasciatus.

INTRODUCTION used cattle-baited traps to determine the biting flies and mosquitoes attacking cattle in Canada. Host preference is an important aspect of arthro- Studiesin the USA comparing relative attraction pod-borne diseases. Determining the host prefer- of multiple host species in traps are limited. In a ence of mosquitoes can aid in understanding the Texas rice-growing area, Kuntz et al. (1982) used transmissionof diseaseswithin a geographicalarea multiple, paired host speciesto determine that cattle (Defoliart et al. 1987). Mosquito host preference and horseswere the preferredhosts for Psorophora varies among geographical strains of the same mos- colurnbiae (Dyar and Knab). quito species.In addition, host preference studies The only previous study concerning host pref- potential have led to mosquito control strategies erence in New Mexico was conducted by Jones et (Kuntz et d. 1982). al. (1977), who identified the biting flies and mos- Bertsch and (1983), Norment using the capillary quitoes that attack horses. The presence of mos- precipitin method, determined the sources of blood quito-borne diseasessuch as canine heartworm dis- meals from field-collected Culex quinquefasciatus ease(Pfaffenberger and Fawcett 1983) and western Say in Mississippi. Their study showed prefer- a equine encephalitis (WEE) (Clark et al. 1986) sug- ence for avian hosts. Other researchers(Jones et al. gests that host preference research is needed in 1977, Kuntz et ^1. 1982, McCreadie et al. 1984) New Mexico. have used animal-baited traps to study mosquito ln our study, traps baited with chickens, a cow, host preference. a dog, and a horse were used to determine the host Most studies using animal-baited traps have de- preference of mosquitoes in north-central New termined the mosquito species attracted to a single Mexico. The study design allowed mosquitoes to host specieswithin a region. Dog-baited traps were choosefrom 4 host speciesand evaluatedpreferred used to find species composition and abundance in domestic hosts, potential vectors of diseasesof do- Indiana (Pinger 1985), Tennessee(Hribar and Ger- mestic animals, and relative host utilization of sev- hardt 1986), and Louisiana (Villavaso and Steelman eral mosquito species. I970a,1970b). Joneset al. (L977')used horse-bait- ed traps to identify potential vectors of Venezuelan equine encephalitis (VEE) in the southwestern MATERHLS AND METHODS United States. Horses also were used to identify potential vectors of Potomac horse fever in Mary- Study sites: Mosquito trapping was conducted at land (Fletcher et d. 1988). McCreadie et al. (1984) 2 sites in the Rio Grande Valley in Bernalillo Co., New Mexico. The south valley site (site l) was an agricultwal area consisting of small, irrigated farms rThis research was supported in part by the New Mex- and pastures. In addition, a dairy operation was lo- ico Agricultural Experiment Station, New Mexico State cated 1.2 km north and a small dog kennel was University, Las Cruces, NM, and the Albuquerque Envi- located 0.4 km west of the site. No domestic ani- ronmental Health Department, Albuquerque, NM. mals were present in the pasture during trapping. 'Animals used in this study were handled with the ap- The north valley site (site 2) was 24.2 km north proval and under the guidelines of the New Mexico State of site 1. Trapping was conducted on fenced prop- University Institutional Animal Care and Use Committee. 3 erty between an irigation canal and a pasture. The Department of Entomology, Plant Pathology and area directly east consisted of mixed hardwood Weed Science, New Mexico State University, Las Cruces, tree NM 88003. species and other vegetation. In comparison to site a Present address: 4823 Old 12 Crossroad. Chester. AR l, fewer domestic livestock were present in the sur- 72934. rounding area and no domestic animals were pres- 5 University Statistics Center, New Mexico State Uni- ent in the immediate trapping area. versity, Las Cruces, NM 88OO3. Iraps; Stable traps, 1.22 x 2.2O X 1.68 m, were

7l 72 JounNer- oF THE AMERTCAN Moseurro Cournol AssoctATroN Vol. 13, No. I constructed of standard aluminum screen (7 X 6 indicator species for data analysis. Site differences mesh/cm) and square steel rubing (1.9 cm). The ba_ within year for number of mosquitoes collected sic-design of the traps was similar to that used by were not statistically significant (X, : 6.48, p : Roberts (1965); however, rhe sides were modified, 0.09), but there were differences between years (1, with openings(1.3-1.9 : cm) running the length of 336, P < 0.0001). This year difference was de- the traps at heights of 22.9, 77.j, and 121.9 cm tected because of the large number (and therefore from ground level. These openingscould be closed high power of the test) of mosquitoes in the study. easily to prevent mosquitoesfrom entering or leav- Although the year difference was detected, trends ing traps. Nylon netting was placed behind each for host use were the same over years, so data were door to minimize mosquito loss when animals were pooled for final analysis. removed. Traps were placed 8 m apart and appro- Data summaries (Thbles 1-5) showed similar priate animal restraints were used to restrict animal seasonal distributions for Aedes and Culex indicator movement within traps. species, with low numbers of mosquitoes collected Design: Tfapping was conducted monthly from in June and September, and higher abundance in Junethrough Septemberat both sites.Each trapping July and August. Culiseta inornata catches were period consisted of 4 consecutive nights at each lowest in July, but due to the small number col- site. Using a random number table, relative loca- lected, this may not have shown the actual seasonal tions of animals were assignedfor each trapping distribution. Peak catches occurred in August for night. Because of the individualized animal re- all 5 species. straining modifications and possible residual host Total numbers collected, engorgement percent- odor, traps were rotated daily to assigned animal age, and proportion of engorged mosquitoes locations. Also, traps were washed out daily with showed preferences for either large or small animal tap water to minimize residual odor. hosts. Signiflcant differences were observed in host Animals used in the study were a cow (270 kg), preference parameters for attraction (X' : 7,514, P a horse (290 kq-),a dog (20 kg), and 6 chickens (3 < 0.0001) and engorgement (X' : 4,470, P < kg). Animals were placed in their respectivecages 0.001). at 1900 h; once all animals were in place, side slots Aedes vexans.' Collection numbers, percentage were opened,allowing host-seekingmosquitoes ac- of engorgement, and the proportion of engorged Ae. cessto bait animals. Side slots were closed at 070O vexans collected from each host showed that this h the following morning and traps were left undis- species preferred cattle and horses over chickens turbed for an additional 3Omin, allowing any unfed and dogs (Table 1). For all monthly trapping peri- mosquitoesthe opportunity to engorge.Mosquitoes ods, the cow and horse attracted almost the same then were collected with a battery-powered aspi- numbers of mosquitoes, and both hosts consistently rator (Meek et al. 1985) and identified according to attracted more than 3-fold as many mosquitoes as Darsie and Ward (1981). Engorgement was deter- the dog or chickens. Ninety-seven percent of allAe. mined by the expansionand coloration of engorged vexans collected from both the horse and cow were mosquitoes (Edman and Kale 1971). Partially en- engorged, versus engorgement rates of 78Vo for the gorged individuals were classif,edas engorged. dog and 34Vo for the chickens. The proportion of Data analysis was based on the number of mos- engorged mosquitoes collected from the large ani- quitoes attracted and number of engorged mosqui- mal hosts was similar, with 0.43 collected from the toes collected from the various hosts by site and horse and 0.46 collected from the cow. Small ani- year. The chi-square likelihood ratio statistic was mal hosts were less utilized, witln 8Vo of engorged used to compare utilization of each animal by se- mosquitoes collected from the dog and 3Vo from lected mosquito species. The chi-square test for chickens. These results agree with studies on wild- equal attraction or engorgementwas rejected at P collected Ae. vexans in Florida (Edman 1971) and < 0.01. Indiana (Nasci 1984). Both authors found that larg- er mammals served as primary blood sources for Ae. vexans. However, Edman (197I) suggested that RESULTS AND DISCUSSION blood meal sources are closely related to the avail- Ten mosquito speciesrepresenting 4 generawere ability of hosts within an area. Although less than collected:Aedes vexans(Meigen) (n = 22,886),Ae- lOVo of Ae. vexans were collected on the dog, the des dorsalis (Meigen) (n : 1,326) Aedes nigro- high engorgement rate suggests that although dogs maculis (Ludlow) (71 : 597), Aedes melanimon may not be the preferred host, they are utilized. Dyar (n : 244), Cx. quinquefasciatus(n : 2,896), Although chickens attracted marry Ae. vexans, per- Culex tarsalis Coq. (n = 1,051), Culex erythro- cent engorgement was greatly reduced. thorax Dyar (n : 1), Culiseta inornata Williston (n Aedes d.orsalis: Although this species was less : 144), Anophelesfreeborni Aitkin (n = 6), and abundant, host preference data (Table 2) were very Anophelesfranciscanus McCracken (n : 1). Aedes similar to data for Ae. vexans. Again, large animal vexans,Ae, dorsalis, Cx. quinquefasciatus,Cx. tar- hosts were preferred, and nearly all Ae. dorsalis salis, and Cs. inom.atawere the most abundantspe- collected from large animals were engorged (97 cies in their respective genera and were used as and 94Vo from the horse and cow, respectively), 73 M,c.ncFr1997 Hosr PnereneNcEoF MosQUITopsIN New Mrxtco

collected from Table 1. Seasonal abundance, number engorged (Vo), and proportion of engorged Aedes vexans animal-baited traps in Bernalillo Co., New Mexico.

Horse Cow Dog Chicken Total No. No. No. No. No. No. No' No. No' No' collec- engorged collec- engorged collec-engorged collec- engorged collec- engorged (Vo) ted lVoj ted (Vo) ted (Vo) ted (Vo) ted June 300 292 584 565 7l 59 47 17 I,OO2 933 (e7) (97) (83) (36) (93) July 3,667 3,604 3,83s 3,762 1,019 794 983 349 9,504 8,509 (e8) (e8) (78) (35) (e0) Aug. 4,t72 4,114 4,336 4,259 670 564 437 130 9,615 9,067 (e8) (e8) (84) (30) (85) Sepr 1,202 1,088 r,126 1,009 323 209 ll4 44 2,765 2,350 (91) (e0) (65) (38) (85) Total 9,34t 9,098 9,881 9,595 2,O83 r,626 1,581 540 22,886 20,859 (97) (e7) (78) (34) (el) Proportion engorged 0.08 0.03 1.00

whereaslower engorgementrates were observedon that Cx. quinquefasciatuswere general feeders in the dog (78Vo)and chickens (39Vo).The proportion Florida. He again suggested that blood meal of engorged mosquitoes showed a greater utiliza- sources varied with available hosts within an area. tion of the horse (0.52) over the cow (0.37). However, if this were the case in Bernalillo Co. one Culex quinquefasciatus: More Cx. quinquefas- would expect more C.r. quinquefasciaruscollected ciatus were consistently attracted to small animal from the cow or horse than was observed. than to large animal hosts (Table 3). The dog at- Culex tarsalis.' Host preference of Cx. tarsalis tracted more mosquitoes than the chickens, but the varied more than with Aedes spe*ies or Cx. quin- engorgementrate was highest on chickens (88Vo). quefasciatus(Table 4). In general, Cx. tarsalis ap- The proportion of engorged, Cx. quinquefasciatus peared to prefer the small animals over the large was practically the reverseof results recorded with animals, but the degree of preference was smaller Aedes species, with most of the engorged mosqui- than wittt Cx. quinquefasciatus. Although greater toes collected from the small animal hosts. The numbers were attracted to the dog (348) and chick- largest proportion engorged on chickens (0.49), fol- ens (346) than to the cow (226) and horse (l3l), lowed by the dog (0.38), cow (0.10), and horse engorgement rates of Cx. tarsalis on the different (0.02). This preference for chickens agrees with a hosts were similar, ranging from 76Vo on both the study by Bertsch and Norment (1983) with Cx. horse and cow to 88Voon the chickens.Proportions quinquefasciarzs in Mississippi. They found that of engorged mosquitoes were greatest with the 6O-:7OVoof wild, bloodfed Cx. quinquefaEciatus chickens and dog. It appearsthat although this spe- had fed on birds. However, Edman (1974), using cies shows differences in attraction to various hosts. blood meal analysisof wild mosquitoes,concluded host preferenceis not sufflcient to prevent engorge-

Thble 2. Seasonal abundance, number engorged (Vo), and proportion of engorged Aedes dorsalis collected from animal-baited traps in Bernalillo Co., New Mexico.

Horse Dog Chicken No. No. No. No. No. No. No. No. No. No. collec- engorged collec- engorged collec-engorged collec- engorged collec- engorged ted (Vo) ted (Vo) ted (Vo) ted (Vo) ted (vo\

June l3 t2 36 JJ 2 6 I 62 48 (e2) (e2) (2e) (17) (77) July l7 17 ll n 4 4 0 0 32 32 (100) (100) (100) (l0o) Aug' 57r 556 383 365 r27 t02 84 t7 1,165 I,O40 (e7) (es) (80) (20) (8e) Sept. 3l 26 29 2l 5 3 4 I 67 51 (84) (72) (100) (2s) (76) Total 632 6ll 459 430 lll t9 r,326 l,t7l (97) (e4) (78) (3e) (88) Proportion engorged o.52 o.37 0.09 o.o2 1.00 74 Jounull or tHe AurnrcAN MoseulTo CoNrnol AssocranoN Vol. 13, No. t

Table 3' Seasonal abundance, number engorged (vo), ylproponion of engorged culex quinquefasciarus collected

Horse Cow Dog Chicken Total No. No. No. No. No. No. No. No. No. No. collec- engorged collec- engorged collec- engorged collec- engorged collec- engorged ted (Vo) (Vo) ted ted (Vo) ted (Va\ ted (Vo) June 3 I 17 6196 t5 14 54 27 (33) (35) (32) (93) (50) July 47 t4 r23 62 296 t74 309 285 /t> 535 (30) (50) (5e) (e2) (6e) Aug. 94 30 226 ll3 665 472 604 520 1,589 1,135 (32) (s0) (7r) (86) (71) Sept. '70 20 2 15 225 80 160 143 4'75 240 (10) (2r) (36) (8e) (51) Total r64 47 436 196 r,205 732 r,088 962 2,893 r,937 (2e) (45) (61) (88) (67) Proportion engorged 0.o2 0.10 0.38 o.49 r.00 ment. Our results agree with a study by Hayes et or Cx. tarsalis, but a lower proportion of the Ae. al. (1973) on wild-collected C.r. tarsalis in Texas. vexans population fed on the dog. Haddock (1987) They found that 3O-4OVo of Cx. tarsalis had fed on suggests that heartworm transmission is facilitated chickens but feeding preference was highly vari- by high mosquito populations. However, the ob- able. served host preference coupled with Edman's Culiseta inornata: This species showed a dis- (1971, L974) theory that mosquito host preference tinct preference for the large animal hosts, with 91 is influenced by the available hosts suggests that and 44 of the total 144 mosquitoes collected from the relative availability of large domestic hosts the cow and horse, respectively (Table 5). In ad- within an area may influence Ae. vexans feeding on dition, all Cs. inornata taken from large animal dogs. hosts had engorged. We consistently saw reduced attraction and en- This study showed that dogs were fed upon by gorgement of selected Aedes species on chickens. Ae. vexans, Ae. dorsalis, Cx. quinquefasciatus, and Ttvo possible explanations exist. Edman and Kale Cx. tarsalis. A previous study on potential mos- (1971) and Edman et al. (1974) suggest that host quito vectors of canine heartworm, Dirofilaria im- defensive behavior reduces engorgement. However, mitis (Leidy), in Bernalillo Co. showed that Ae. this does not seem to be the case in our study be- vexans has the greatest vector efficiency, Ae. dor- cause Culex species had high engorgement rates on salis can support development of infective D. im- chickens. Specific cues such as host odor may help mitis, and Cx. quinquefasciatus and Cx. tarsalis are mosquitoes distinguish mammalian from avian generally of minor importance (Loftin et d. 1995). hosts. Gillies (1980) suggested that host odors and In the present study, dogs were fed upon more fre- other factors are more powerful than carbon diox- quently by Ae. vexans than by Cx. quinquefasciatus ide at close and moderate ranges. Probably the best

Table 4. Seasonal abundance, number engorged (Vo), and, proportion of engorged Culex tarsalis collected from animal-baited traps in Bernalillo Co., New Mexico.

Cow Dog Chicken

No. No. No. No. No. No. No. No. No. No. collec- engorged collec- engorged collec- engorged collec- engorged collec- engorged ted (Vo) ted (Vo) ted (Vo) ted (Vo) ted (Vo)

June 20 I2 JZ 2T 49 37 52 48 153 ll8 (60) (65) (76) (e2) (77) July 5l +J 88 69 r4l t26 r2a 115 408 353 (84) (78) (8e) (e0) (86) Aug. 59 44 96 75 139 t25 155 130 449 374 (75) (78) (e0) (84) (83) '7 Sept. I 0 lo l9 t4 II n 41 32 (70) (74) (100) (78) Total 99 172 302 304 r,051 877 (76) (76) (87) (88) (83) Proportion engorged 0.11 o.20 o.34 0.35 1.00 75 MARCH 1997 Hosr PnsFEneNce oF MosQunoEs IN NEw Mextco

collected from Table 5. Seasonal abundance, number engorged (7o), and proportion of engorged Culiseta inornata animal-baited traps in Bernalillo Co., New Mexico.

Horse Cow Dog Chicken Total No. No. No. No. No. No. No' No. No' No' collec- engorged collec- engorged collec-engorged collec- engorged collec- engorged ted 9d ted (Vo) ted (Vo) ted (Vo) ted (7o)

June 8 8 28 28 0 0 0 0 38 JI (10o) (l0o) (e7) July I I I I 0 0 0 0 2 z (l0o) (100) (100) Aug. az )z 56 56 3 I J 0 94 89 (100) (l0o) (33) (e4) Sept. .J 5 6 6 0 0 1 I 10 l0 (l0o) (100) (r00) (100) Total 44 9t 9l 2 1 t44 138 (1oo) (100) (4) rr5\ (e6) Proportion engorged o.32 0.65 0.0r 0.01 1.00

explanation is that chickens are simply not attrac- finding by mosquitoes (Diptera: Culicidae). Bull. En- tive to the selected Aedes species. tomol. Res. 7O:525-532. Feeding of Cx. tarsahs and, to a lesser extent, Haddock, K. C. 1987. Canine heartworm disease: a re- view and pilot study. Soc. Sci. Med.24:225-246. Cx. quinquefasciatus, on both birds and mammals Hayes, R. O., C. H. Tempelis, A. D. Hess and W. C. was observed. Results suggest that both Cx. quin- Reeves. 1973. Mosquito host preference studies in quefasciatus and Cx. tarsalis display host prefer- Hale County, Texas. Am. J. Trop. Med. Hyg. 22:270- ence characteristics that could favor transmission of 277. endemic WEE, as suggested by Hayes et al. (1973). Hribar, L. J. and R. R. Gerhardt. 1986. Mosquitoes at- tacking dogs in Knox County, Tennessee. J. Am. Mosq. Control Assoc. 2:552-553. REFERENCES CITED Jones, R. H., R. O. Hayes, H. W. Potter, Jr. and B. D. Bertsch,M. L. and B. R. Norment. 1983. Host-feeding Francy. 1977. A survey of biting flies attacking patterns of Culex quinquefasciarzsin Mississippi. equines in three states of the southwestern United Mosq. News 43:203--206. States, 1972. J. Med. Entomol. 14:441-447. Kuntz, K. J., J. K. Olson and B. J. Rade. 1982. Role Clark, G. G., C. L. Crabs,C. L. Bailey, C. H. Calisher of domestic animals as hosts for blood-seeking females of and G. B. Craig, Jr. 1986. Identificationof Aedescam- Psorophora columbiae and other mosquito species pestris from New Mexico: with in notes on the isolation Texas ricelands. Mosq. News 42:202-210. of westernequine encephalitisand other arboviruses.J. Loftin, K. M., R. L. Byford, M. J. Loftin and M. E. Craig. Am. Mosq. Control Assoc.2:529-534. 1995. Potential rnosquito vectors of Dirofilaria immitis Darsie, R. E and R. A. Ward. 1981. Identification and in Bernalillo County, New Mexico. J. Am. Mosq. Con- geographical distribution of the mosquitoes of North trol Assoc. ll:90-93. America, North of Mexico. Mosq. Syst. I (Suppl.):l- McCreadie, J. W., M. H. Colbo and G. F Bennett. 1984. 3r3. A trap designed for the collection of hematophagous Defoliart, R. M., P R. Grimstad and D. M. Watts. 1987. Diptera from cattle. Mosq. News 44:212-216. Advances in mosquito-borne arboviruVvector research. Meek, C. L., M. V. Miesh and T W. Walker. 1985. Por- Annu. Rev. Entomol. 32:479-505. table, battery-powered aspirators for collecting adult Edman, J. D. 197 l. Host-feeding patterns of Florida mos- mosquitoes. J. Am. Mosq. Control Assoc. 1:102-105. quitoes L Aedes, Anopheles, Coquilletidia, Mansonta Nasci, R. S. 1984. Variations in the blood-feeding pat- and, Psorophora. J. Med. Entomol. 8:687-695. terns of Aedes vexans and Aedes trivittatus (Diptera: Edman, J. D. 1974. Host feeding patterns of Florida mos- Culicidae). J. Med. Entomol. 2l:95-99. quitoes III. Culex (Culex) and Culex (Neoculex). J. Pfaffenberger, G. S. and R. F Fawcett- 1983. Canine Med. Entomol. 11:95-104. heartworm disease in New Mexico. Stud. Nat. Sci. (East. N.M. Univ.'t 2:l-12. Edman, J. D. and H. W. Kale lI. 1971. Host behavior: its Pinger, R. E. 1985. Species composition influences on the feeding success of mosquitoes. Ann. and feeding suc- cess of mosquitoes Entomol. Soc. Am. 64513-516. attracted to caged dogs in Indiana. J. Am. Mosq. Control Assoc. l:181-185. Edman, J. D., L. A. Webber and A. A. Schmid. 1974. Roberts, R. H. 1965. A steer baited trap for sampling Effect of host defenses on the feeding pattern of Culex insects affecting cattle. Mosq. News 25:281--285. nigripalpus when offered a choice of blood sources. J. Villavaso, E. J. and C. D. Steelman. I97Oa. A collapsible Parasitol. 60:874-883. dog-baited trap. Mosq. News 30:39-42. Fletcher, M. G., E. C. Thrner, J. W. Hanson and B. D. Villavaso, E. J. and C. D. Steelman. 1970b. Laboratorv Perry. 1988. Horse-baited insect trap and mobile insect and field studies for the southern house mosquito, Culex sorting table used in a disease vector identification pipiens quinquefasciatus Say, infected with the dog study. J. Am. Mosq. Control Assoc. 4:431-435. heartworm, Dirofilaria immitis (Leidy), in Louisiana. J. Gillies, M. T. 1980. The role of carbon dioxide in host- Med. Entomol. 4:47 l-476.