OUP user menu

Population Dynamics of Sand Flies (Diptera: Psychodidae) at Two Foci of Leishmaniasis in Texas

(CC)
Chad P. McHugh, Brian F. Ostrander, Russell W. Raymond, Sara F. Kerr
DOI: http://dx.doi.org/10.1603/0022-2585-38.2.268 268-277 First published online: 1 March 2001

Abstract

Sand flies were collected at a focus of leishmaniasis in Medina County, TX, from April through October 1997 and at a focus in Bexar County, TX, from April 1998 through December 1999. Lutzomyia diabolica (Hall) were collected from April through November with peak abundance in July. The male:female ratio of Lu. diabolica was 1:6.2. Almost all female Lu. diabolica in the collections were unfed. One gravid Lu. diabolica contained 49 ova. Female Lu. anthophora (Addis) were active from February through December with three peaks in abundance suggestive of successive generations. Unfed and gravid Lu. anthophora were collected in about equal numbers. Gravid females contained from 1to 64 ova per female. The male:female ratio was 1:1.8, with male Lu. anthophora collected in all months. One female Lu. anthophora was found infected with Leishmania in July 1999. Lutzomyia texana (Dampf) were collected from April through October with peak abundance in April during 1997. The male:female ratio was 1:1.4, with most females unfed. Two gravid Lu. texana contained 32 and 102 ova. An undescribed species of Lutzomyia was found only at the Medina County site from May through September 1997. Trapping sites four times per month versus two times per month in 1999 did not appear to adversely affect the abundance of Lu. diabolica or Lu. anthophora. There were marked differences in the species composition and relative abundance at the different sites, indicating that the spatial distribution of sand flies is patchy in nature.

  • Lutzomyia
  • temporal abundance
  • population ecology
  • Texas
  • leishmaniasis

Leishmania mexicana Biagi, a parasitic protozoan causing cutaneous lesions in humans, is enzootic in Texas and Arizona. Woodrats, Neotoma micropus Baird in Texas and N. albigula Hartley in Arizona, are the reservoir hosts (McHugh et al. 1990; Kerr et al. 1995, 1999) and the sand fly Lutzomyia anthophora (Addis) appears to be the enzootic vector of parasites among these rodents (McHugh et al. 1993).

Twenty-nine human cases of leishmaniasis have been reported in Texas (McHugh et al. 1996), and an additional human case has been diagnosed (C.P.M., unpublished data), but the identity of the zoonotic vector to humans remains controversial. Lutzomyia anthophora is not believed to be anthropophagic and Lu. diabolica (Hall), an avid human feeder, was suggested as a possible bridge between the enzootic cycle and humans (Lawyer and Young 1987). Both species fulfill a number of criteria (Killick-Kendrick 1990) used to incriminate phlebotomines as vectors of Leishmania spp. Laboratory colonies of both species have been used to determine host preferences and to demonstrate that both are competent vectors of L. mexicana (Endris et al. 1984, 1987; Lawyer 1984; Lawyer and Young 1987). The known geographic ranges of both species (Young and Perkins 1984, McHugh 1991) parallel that of human cases and both have been collected at a case residence (Lawyer 1984). To date, the only isolates of L. mexicana from the sand fly fauna of the United States have been from Lu. anthophora (McHugh et al. 1993).

Lindquist (1936), Addis (1945b), Endris (1982), Endris et al. (1984), and Lawyer (1984) used laboratory colonies of Lu. anthophora and Lu. diabolica to document a number of biological characteristics, including fecundity, temperature-dependent development rates, sex ratios, survivorship, host preferences, blood and sugar feeding, and autogeny.

Data on the population ecology of these insects, essential to incriminate vector species and understand the dynamics of the Leishmania transmission cycle, is fragmentary or anecdotal. Addis (1945a) stated that adult Lu. anthophora were active from late March through late November in Uvalde, TX, but gave no quantitative indication of abundance. Parman (1919) reported that Lutzomyia spp. were active from early September through late November in Uvalde, with peak activity from 25 September through 10 October. He apparently relied on biting collections and gave a range of one specimen in several nights to 25-30 each night, but few other details. Lindquist (1936) reported collections of Lu. diabolica from early May through mid-November at Uvalde and in June, August, September, October, and November in Del Rio, TX, ≈125 km to the west, but provided no quantitative details. Lawyer (1984) documented the abundance of Lu. anthophora, Lu. diabolica, Lu. Texana, and an undescribed Lutzomyia sp. at eight sites in southcentral Texas, but his sampling was limited temporally. Lawyer (1984) also reported the fecundity of field-collected Lu. diabolica used to establish a laboratory colony. We describe herein the temporal abundance of sand flies, as well as other aspects of their biology, at two foci of leishmaniasis in southern Texas.

Materials and Methods

Study Sites and Sampling Schedule.

During 1997, collections were made at a privately-owned ranch (29° 21′ N, 99° 21′ W) a few kilometers northwest of D’Hanis in Medina County, TX, and ≈90 km west of San Antonio. A 67-yr-old male (case #24 of McHugh et al. 1996) who lived part-time on the ranch was diagnosed with leishmaniasis in 1989, and infected woodrats were found on the ranch in 1994 and 1995 (S.F.K., unpublished data). The ranch was located on the northern edge of the south Texas Plains at an elevation of 300 m above mean sea level. The vegetation was a mix of native brush and improved pasture, dominated by live oak (Quercus virginianus), mesquite (Prosopis sp.), pricklypear cactus (Opuntia sp.), grasses, and forbes. Large mammals on the ranch included beef cattle, white-tailed deer (Odocoelius virginianus), and wild hogs. Rodents present included Neotoma micropus, Peromyscus spp., and Sigmodon hispidus. A number of other mammals including opossums, skunks, armadillos, and squirrels also occur in this area (Davis and Schmidly 1994), but were not observed. The soils on the ranch were predominantly Mercedes clay that has low permeability and high water holding capacity (USDA 1977). A few acres were Valco clay loam, with a moderate permeability. Sampling began on 8 April using a mix of solid-state army miniature light traps and Hock new standard miniature light traps (John W. Hock, Gainesville, FL) supplemented with dry ice. Twelve sites, all near Neotoma nests, were sampled in April. Additional sites were added monthly until a total of 24 per month was being sampled by July. Trapping was discontinued at the end of October at the request of the landowner.

In 1998 and 1999, trapping was conducted on Lackland annex (29° 22′ N, 98° 41′ W), a 3,034-ha military training site in western San Antonio, Bexar County, TX, at an elevation of ≈205 m. There is some development on the eastern and central portions of the annex, but the remainder of the area is largely native brush, with persimmon (Diospyros sp.), mesquite, live oak, white-brush (Aloysia sp.), pricklypear cactus, and grasses. Some areas were disturbed as the result of road grading and training exercises. The mammalian fauna on the annex was similar to that at the Medina County ranch except for the absence of cattle and a paucity of rodent species other than woodrats (Neotoma micropus). Woodrat nests were common in scattered clumps of cactus, but also were built in brush piles and at the bases of persimmon trees. The soils in this area are Houston Black gravelly clays that have a high water holding capability and slow permeability (USDA 1962).

Four areas of the annex—the northwest, southwest, and southeast corners and the west-central perimeter—were sampled in 1998 using Hock new standard miniature light traps supplemented with dry ice. In total, 34 sites were trapped in April, and 40 sites, 10 in each area, were sampled once monthly for the remainder of the year.

In 1999, sampling was concentrated in the northwest and west-central areas of the annex where infected woodrats were found (Rosolowski 1999) and where sand flies were collected consistently during 1998. A total of 10 sites in these areas was sampled twice monthly (20 trap nights per month). An additional five sites a few hundred meters to the east of the west-central area were sampled four times monthly (20 trap nights per month). The purpose of the sampling four times per month was to determine the following: how much improved resolution more frequent sampling would provide; and whether more frequent trapping would negatively impact sand fly abundance by removal sampling. All traps in both years at the annex were placed at active woodrat nests.

Eight funnel traps, designed to collect sand flies exiting rodent burrows, also were used four times a month (32 trap nights per month) in 1999. Nests sampled by funnel traps were ones at which a Leishmania-infected woodrat previously had been collected, but at which a light trap was not being used. The funnel trapping was intended to increase the probability of collecting Leishmania-infected sand flies, but it also provided an opportunity to compare the effectiveness of different sampling methods. Funnel and CO2-baited light traps traps were set shortly before sunset and collected shortly after dawn.

Historically, the weather at both the Medina County ranch and the Bexar County site is generally subtropical with hot, humid summers and mild winters. Precipitation is bimodal with peaks in May and September, with annual totals of 72.2 cm in Medina County (USDA 1977) and 71.3 cm in Bexar County (National Climatic Data Center, Asheville, NC). Weather data for 1997 were collected at the Hondo municipal airport, ≈18 km from the ranch study site (National Climatic Data Center, Asheville, NC). Weather data for 1998 and 1999 were collected at Kelly Air Force Base (AFB), ≈9.5 km to the east of the Lackland annex study area (National Climatic Data Center, Asheville, NC).

Processing of Collections.

Collections were returned to the laboratory where the insects were anesthetized with either CO2 or triethylamine. Specific determinations were made using the key of Young and Perkins (1984). Sand flies were shaken in a vial of 70% isopropyl alcohol for about 30 s to remove scales and surface sterilize. Female sand flies were dissected in a drop of Schneider’s Drosophila medium supplemented with 20% fetal bovine serum as described by McHugh et al. (1993) to determine their gonotrophic state and to check for infection with parasites. Females were classified as unfed if neither blood in their gut nor developing ova were present. Females were classed as bloodfed if there was any visible blood in the gut or as gravid if late-stage eggs were present. In 1998 and 1999, the number of ova present was determined for a subset of gravid females. If parasites were observed, the gut was teased apart and transferred to a tube containing 1 ml of the supplemented Schneider’s medium (McHugh et al. 1993). Polymerase chain reaction (PCR) using primers 13A and 13B and methods described by Rogers et al. (1990) and Rosolowski (1999) was used to provide generic identification of parasites.

Data Analysis.

In 1997 and 1998, a monthly trap index of flies per trap night was calculated separately for each sex of each species. In 1999, two indices were calculated; one for each of the four times per month collections and one for each of the twice a month collections. To ensure a consistent index of abundance, retrapping was conducted when it was felt that weather or trap malfunction made trap counts unreliable.

To test the null hypothesis that sampling four times a month in 1999 gave the same pattern of collections as did sampling twice a month, we used chi-square (α = 0.05) test to compare all months in which both sampling schemes collected enough female flies to satisfy the assumptions of the chi-square test. For Lu. anthophora that was the period February through October and for Lu. diabolica that was May through August.

Results

Weather.

The first half of 1997 was slightly cooler and wetter than average, with below normal temperatures in the months of April through July, and more precipitation than normal in the months of February through June. Rainfall of almost 30 cm fell in June and contributed to an excess of 9.1 cm for the year. Rainfall was below average July through September.

Total rainfall of 101 cm, >30 cm above the historic average, at Kelly AFB in 1998 indicated that the year was exceptionally wet. In fact, after a January and February with slightly higher than normal rainfall, precipitation during the following 5 mo was well below average with higher than average temperatures, resulting in drought-like conditions. October was exceptionally wet with almost 44 cm, over half the yearly average of rain, most of which fell in a very short time. As a result, there was widespread flooding in southern Texas. If the excess of 36 cm over the historic October average were discounted, the annual precipitation in 1998 would actually be below average.

In much of Texas, 1999 was exceptionally hot and dry with drought conditions in much of the state. Total rainfall at Kelly AFB was 30 cm below average, with every month except June and July at, or well below, historic means. Observed mean maximum temperatures were at, or above, historic levels every month except July.

Temporal abundance.

A total of 172 trap nights using CO2-baited light traps was conducted from April through October 1997, resulting in the capture of 1,558 sand flies comprising four species (Table 1). One hundred forty-one trap nights documented the temporal abundance of sand flies at the Medina County ranch during that time. Lutzomyia diabolica was the most frequently collected species during 1997. Females were first captured on 29 April, peaked in July at 25.5 females per trap night, and then were present in low numbers until the end of October when sampling ceased (Fig. 1). Male Lu. diabolica were present from mid-May through mid-September with a peak of 1.2 males per trap night in July. Male and female Lu. texana were most common (3.9 males/trap night, 5.5 females per trap night) (Fig. 1) when sampling started in April and were collected through October. Lutzomyia anthophora females were captured in low numbers during the months of April, May, July, September and October. The abundance of female Lu. anthophora never exceeded 0.5 females per trap night. Male Lu. anthophora were present only in May, July and September. Females of Lutzomyia species "A" were captured in low numbers in May, June, July, and September.

View this table:
Table 1
Fig. 1

Temporal abundance and trophic status of female Lu. diabolica (A) and Lu. texana (B) collected in Medina County, TX, during 1997.

Only 573 individuals were collected during 354 traps nights in 1998 at the Bexar County site (Table 2); data from 334 trap nights were used to describe temporal trends in sand fly abundance. Lutzomyia anthophora was the predominant species; females were present from April through November, with small peaks in abundance in May and July and highest numbers (3.9 per trap night) in September (Fig. 2). The November collections included some gravid females, indicating Lu. anthophora was reproductively active in early winter. Male Lu. anthophora had a similar pattern, but at slightly lower numbers (Fig. 2). One female Lu. diabolica was collected in April, one in August, and five in September. No male Lu. diabolica were collected during 1998. The few Lu. texana females that were present were collected in April, September, and October. Male Lu. texana were collected in April and August through October.

View this table:
Table 2
Fig. 2

Temporal abundance of female (A) and male (B) Lu. anthophora collected in Bexar County, TX, during 1998.

During 1999, 480 trap nights captured 4,895 individuals (Table 3). Lutzomyia anthophora females were active from early February through early December (Fig. 3). The last gravid females were collected in October. There appeared to be three peaks in the abundance of female Lu. anthophora, most clearly seen in data from the sites sampled four times per month. The first peak was in May-June, followed by a midsummer peak in July-August, and an early fall peak in September-October. Male Lu. anthophora were collected in all months. Trends in male abundance (Fig. 4) were similar to females, but included an additional peak of activity in February-March and late activity in November-December. Male and female Lu. diabolica were active April through November (Fig. 5). The four per month collections showed peaks in abundance of females in May, June, and July.

View this table:
Table 3
Fig. 3

Temporal abundance and trophic status of female Lu. anthophora sampled twice (A) and four times (B) monthly.

Fig. 4

Temporal abundance of male Lu. anthophora sampled twice (A) and four times (B) monthly.

Fig. 5

Temporal abundance and trophic status of female Lu. diabolica sampled twice (A) and four times (B) monthly, and temporal abundance of male Lu. diabolica sampled four times monthly (C).

The chi-square test comparing the trends in female abundance in the two and four times a month collections was highly significant for both Lu. anthophora (χ2 = 63.2, df = 8, P < 0.0005) and Lu. diabolica (χ2 = 29.4, df = 3, P < 0.0005); however, this was not because the four per month sampling was depleting the populations of sand flies. In fact, the numbers of both Lu. anthophora and Lu. diabolica at the four per month area were lower than expected early in the year, but showed a marked increase relative to the two per month collections from July onward.

Funnel Traps.

A total of 384 funnel traps nights in 1999 yielded only 133 sand flies; 77 male and 53 female Lu. anthophora, two gravid Lu. texana, and one unfed Lu. diabolica. Lutzomyia anthophora were collected from late March through late November.

Spatial Abundance.

Species composition and relative abundance varied between the two foci and among areas at the Bexar County focus. At the Medina County ranch site in 1997, Lu. diabolica was the predominant species, whereas Lu. anthophora barely was represented (Table 1). The reverse was true in Bexar County in 1998; Lu. diabolica was present in low numbers, and Lu. anthophora was collected most commonly (Table 2). In 1999, sampling concentrated on two areas of the Lackland annex separated by only a few hundred meters. The collections at those sites (Table 3) were remarkably different from each other. The 10 sites sampled twice monthly in 1999 yielded moderate numbers of Lu. anthophora (total 951) and a few Lu. diabolica (total 48), a pattern seen in 1998 at other areas of the annex. At the five sites sampled four times a month in 1999, Lu. anthophora were >1.5 times as abundant (total 1,524), and Lu. diabolica were collected at a rate almost 50 times (total 2,345) that at the other area just a few hundred meters distant. Lutzomyia texana, which was present in modest numbers at the ranch in 1997, were uncommon at the Bexar County site in 1998 and 1999. Lutzomyia species "A" collected in Medina County in 1997 was never collected in Bexar County.

Sex Ratio and Trophic Status.

Combining light trap data all sites and years, the male:female ratios for Lu. anthophora, Lu. diabolica, and Lu. texana were 1:1.8, 1:6.2, and 1:1.4, respectively. For Lu. anthophora collected in funnel traps during 1999, the male:female ratio was 1.5:1.

Unfed (46.7% of total females collected) and gravid (52.0%) females were about equally represented in light trap collections of Lu. anthophora. Bloodfed Lu. anthophora rarely were collected. Collections of female Lu. diabolica were composed almost exclusively of unfed individuals; only 1.8% of female Lu. diabolica were bloodfed or gravid. Of female Lu. texana collected, 91.6% were unfed and 8.4% were gravid. No bloodfed Lu. texana were collected. The trophic status of female Lu. anthophora collected in funnel traps during 1999 was similar to that seen in light trap collections, with 43.4% unfed and 52.8% gravid.

Fecundity.

Reliable counts of ova in a total of 123 gravid Lu. anthophora yielded estimates of 36.1 (SD 12.2; range, 1-64) and 37.7 (SD 16.6; range, 1-61) in 1998 and 1999, respectively. If those females with <10 ova were excluded as having recently oviposited and retained the eggs present, the means increased to 37.3 (SD 10.7) and 42.6 (SD 9.6), respectively. Of two gravid Lu. diabolica examined in 1999, one had two ova that probably were retained after oviposition and the other contained 49 eggs. Two Lu. texana collected in March 1999 contained 32 and 102 ova.

Leishmania infections.

In total, 347 Lu. anthophora, 494 Lu. diabolica, 83 Lu. Texana, and 10 Lutzomyia sp. "A" were examined for Leishmania infection over the 3 yr. A single, bloodfed Lu. anthophora collected at Lackland annex in July 1999 was infected with promastigotes morphologically similar to Leishmania. The gut was transferred to Schneider’s Drosophila medium, but the isolate was overgrown quickly with fungus. The filtrate of the tube was screened using PCR (Rogers et al. 1990, Rosolowski 1999) and determined to be Leishmania sp. No Leishmania-infected Lu. diabolica, Lu. texana, or Lutzomyia sp. "A" were found.

Discussion

Temporal Abundance.

Although total numbers of individuals varied between sites and among years, the trends in temporal abundance were fairly consistent for each species. Female Lu. anthophora were active from February through November, months during which the mean monthly low temperature was ≥10°C. February is a slightly earlier date for the onset of activity than that mentioned by Addis (1945a). He gave late March through November as the period of activity for Lu. anthophora at Uvalde, TX, at about the same latitude as our study sites. Male Lu. anthophora also were active in January and December, albeit in low numbers. Peaks in abundance of females in 1999 were at ≈2-mo intervals. Based on the development rate data of Endris et al. (1984), which indicated 2 mo is a reasonable generation time, and the fact that egg hatch in Lu. anthophora is synchronous, with all eggs laid by a female hatching within 24 h (Endris et al. 1984), each peak probably represented a separate generation. Male Lu. anthophora had two additional peaks, indicating that this species may have from three to five generations per year. Efforts are underway to develop a predictive model for the temporal abundance of Lu. anthophora using our 1999 data, the temperature-dependent development rate data of Endris et al. (1984), daily weather data, and the stochastic techniques of Curry and Feldman (1987).

Lutzomyia diabolica activity appeared to be restricted to the warm months of the year. This species was not active until April when the mean monthly low temperature was ≥15°C. Although a few individuals were collected as late as November, most adult activity was finished by September. This confirms the observations of Lindquist (1936) who reported that Lu. diabolica was active from mid-May through November in Uvalde, TX. There were three peaks of abundance at monthly intervals in 1999. This is probably too short a time for the peaks to represent successive generations. Endris (1982) reported that first-instar Lu. diabolica did not eclose synchronously, with as much as 70% of eggs laid by a single female failing to hatch within 30 d of oviposition. It is possible that the observed peaks resulted from episodic waves of egg hatching or an overlay of episodic hatching and production of subsequent generations.

Female Lu. texana were collected from April though October. In 1997, the only year when enough females were collected to see temporal trends, peak abundance (5.5 females per trap night) was seen in April, the first month sampling was conducted. Of the females collected in October of that year, almost half were gravid, indicating reproductive activity during that month. Collections at the Medina County ranch in 1995 (C.P.M., unpublished data) indicated that Lu. texana were even more numerous in January when they averaged 10 females per trap night. This species was reported as being active year-round near Brownsville, TX, to the south of our sites, but not active during the winter months at Del Rio, TX, at about the same latitude as our study sites (Eads et al. 1965).

Lutzomyia species "A" was present only in the 1997 collections in the months of May through September, but was never numerous enough to see clear temporal trends in abundance.

Sand flies were present in fairly low numbers during 1997 and 1998, leading to concerns that our trapping may deplete the population and therefore influence temporal trends in abundance. Sampling four times per month in 1999 did not appear to have an adverse impact on the abundance of sand flies in the area; but because the dispersal capability of sand flies and the range over which the light traps attract sand flies is unknown, it is impossible to estimate over how large an area populations were being sampled.

Funnel Traps.

Funnel traps were a relatively ineffective method of collecting sand flies, collecting an average of only 0.35 individuals per trap night over the entire year. This contrasted with an average of 10.2 individuals collected per CO2-baited light trap night during 1999. The reason for the poor success of funnel traps in our study remains unclear. Lutzomyia anthophora use animal burrows as resting sites; Young and Perkins (1984) reported collecting ≈600 individuals from a single nest and estimated that an equal number escaped capture. Chaniotis and Anderson (1968) found that funnel traps were "very effective" in collecting sand flies exiting ground squirrel burrows in California. Comer and Corn (1991) determined that funnel traps were as effective (6.2 flies per trap night) as light traps in collecting Lu. shannoni Dyar in Georgia.

Spatial Abundance.

Differences in the species composition and relative abundance at the two foci indicated that sand fly populations were distributed patchily in nature. This may have resulted from differences in the animal fauna at the sites; cattle and a number of rodent species present at the Medina County ranch site were absent at the Bexar County site. Inasmuch as host abundance was not quantified at the sites and the host preferences of the sand flies remain largely unknown, this explanation remains to be explored. Additionally, comparisons of species composition at Medina County in 1997 and Bexar County in 1998 may be criticized because year-to-year temporal changes in populations were a possible confounding factor. The best indication of patchiness was the 1999 data that showed remarkable differences in abundance of Lu. anthophora and Lu. diabolica at two closely situated areas on the Lackland annex. Lawyer (1984) also commented on the disjunct distribution of sand flies at Garner State Park in Uvalde County, TX. He found numerous Lu. diabolica resting in latrines and biting at night in upland areas of the park, while they were absent in lower parts of the park bordering the Frio River. The factors that influence the distribution of sand flies are unknown, but could include availability of hosts, oviposition sites, carbohydrate sources, soil type, and resting sites.

The differences between abundance at two closely situated sites in Bexar County during 1999 also indicated that Lu. anthophora and Lu. diabolica did not disperse broadly or the light traps used to collect the insects were attractive over only a short distance. As noted above, nothing is currently known of the dispersal capability of those species or the effective range of the light traps for sand flies.

Sex Ratio and Trophic Status.

Endris et al. (1984) reported a male:female ratio of 1:1 for colonized Lu. anthophora. Lawyer (1984) reported slightly more males than females (1.4:1) in nine generations of colonized Lu. diabolica. Lawyer (1984) used several collecting methods to sample a field population of Lu. diabolica and reported male:female ratios ranging from 4:10 to 6:10 in resting collections. There are several possible reasons why sex ratios in our light trap collections of Lu. anthophora (1:1.8, male:female) and Lu. diabolica (1:6.2, male:female) and those of Lawyer (1984) departed from a 1:1 ratio. First, survivorship of males may be less than that of females, therefore there is a shorter time (=less probability) that any individual male is in the environment and available to be sampled. Addis (1945b) determined that female Lu. anthophora lived longer than males, with an average life span in the laboratory of 9.2 d compared with 5.3 for males. The opposite was true for virgin males and females of Lu. diabolica, which had mean longevity of 14.9 and 11.7 d, respectively (Lawyer 1984). Second, there may be an impact of diet on sex ratio, presumably acting by influencing differential survivorship of immatures. Lawyer (1984) found that varying diets resulted in male:female ratios of from 1:1.7 to 2.4:1 in colonized Lu. diabolica. Third, male sand flies may not be as attracted to CO2-baited light traps as are females, or males may not be as active nocturnally as females. There are no published data to support or refute these last hypotheses. A male:female ratio of 1:1.4 for our collections of Lu. texana is similar to the ratio of 1:1.2 reported by Eads et al. (1965) for collections made in Brownsville, Cameron County, TX.

The different proportions of unfed, bloodfed, and gravid females in the collections probably reflected differences in the behavior of the different trophic classes in each species. Unfed females commonly were collected for all species, indicating they were nocturnally active, perhaps in search of hosts or sugar meals. Lindquist (1936) reported that Lu. diabolica were nocturnal feeders and were active from about 2000 hours to midnight at Uvalde, TX. Likewise, Lawyer (1984) reported that the species was nocturnal, but he also reported one instance of a Lu. diabolica bloodfeeding at 1400 hours in full sunlight. In contrast, Addis (1945a) reported that bloodfeeding Lu. anthophora were most active in the morning, with activity declining by the afternoon and nonexistent at night.

Gravid Lu. anthophora also were well represented in our light trap collections, indicating that females were nocturnally active. The purpose of this activity, perhaps searching for oviposition sites or sugar meals, is unknown. Gravid Lu. texana were uncommon and very few gravid Lu. diabolica were collected. This indicated that gravid females of these species either were not active, at least not nocturnally, or remained in habitats such as rodent nests and were not exposed to the light traps.

Bloodfed females of all species were uncommon in our collections. One possible explanation is that these species produce the first batch of eggs autogenously and few survive to take a bloodmeal for subsequent ovipositions. Lawyer (1984) examined Lu. diabolica for evidence of autogenous egg development and found none. Endris et al. (1984) did not report autogeny in a colony of Lu. anthophora, although they did not specifically rule it out. No information on the autogeny status of Lu. texana or Lutzomyia species "A" is available. Lawyer (1984) reported that bloodfed Lu. diabolica comprised ≈31-46% of females resting in latrines, indicating that, at least in that species, failure to find them in light traps was probably due to a trap bias reflecting insect behavior.

Fecundity.

There is close agreement between fecundity data from field populations and those derived from laboratory colonies of Lu. anthophora and Lu. diabolica. A mean fecundity of 37-42 eggs per female Lu. anthophora found in our study is similar to that reported by Endris et al. (1984), who observed a mean of 35.0 (SD 17.3) for 15 generations of a laboratory colony. Endris et al. (1987) observed means of 35.9, 36.3, and 39.9 eggs per female for groups of colonized Lu. diabolica. Lawyer (1984) reported an average of 36.6 eggs from 164 field-collected Lu. diabolica. Our single, female Lu. diabolica with 49 eggs is in reasonable agreement with those data. Our two female Lu. texana averaged 67 eggs; there are no other published field or laboratory data on fecundity of that species.

Leishmania Infections.

The detection by PCR of Leishmania in a female Lu. anthophora is the only instance of an infected sand fly collected in the United States, other than two Lu. anthophora females infected with L. mexicana reported by McHugh et al. (1993). Although no specific identification of the present isolate was made using isoenzyme analysis, the previous detection of L. mexicana in woodrats at the site strongly indicates that the isolate was that species.

Lawyer and Young (1987) demonstrated transmission of L. mexicana by Lu. diabolica and argued that that species was the most likely vector in Texas. Our failure, and that of Lawyer and Young and (1987), to find infected Lu. diabolica does not eliminate the possibility that this anthropophagic species is a zoonotic vector to humans. Given the dearth of human cases, it is possible that infections in Lu. diabolica and transmission to humans only occurs under certain circumstances, such as unusually large populations of Lu. diabolica. Because the host preferences and vector competence of Lu. texana and Lutzomyia sp. "A" are unknown, it is impossible to interpret the significance of the negative findings with these species.

Although our study documented several important characteristics of sand fly populations in southern Texas, many additional aspects of their biology remain to be investigated. As mentioned above, nothing is known of the dispersal capability of the insects or the range over which they are attracted to CO2-baited light traps. Nor do we know the source of their sugar meals, their longevity, or their host utilization. Ideally, when these data are available, they will be correlated with data on infections in woodrat populations (Rosolowski 1999) and used to explain the spatial and temporal distribution of Leishmania in nature.

Acknowledgements

We gratefully acknowledge the help of Michelle Corbett in collecting and processing collections in 1997. Access to the sampling sites in Medina and Bexar counties also is acknowledged. This research was supported by National Institutes of Health grants (AI)39725, GM55337, and GM50080.

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/), which permits non-commercial reuse, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions{at}oup.com

References Cited

View Abstract