Flea, rodent, and plague ecology at Chuchupate Campground, Ventura County, California

June, 2002 Journal of Vector Ecology 107 Flea, rodent, and plague ecology at Chuchupate Campground, Ventura County, California Richard M. Davis 1, Randall T. Smith 2, Minoo B. Madon 3, and Erika Sitko-Cleugh 4 1 Vector-Borne Disease Section, California Department of Health Services, 4840 Market Street, Suite D, Ventura, CA 93003 2 County of Ventura, Environmental Health Department, 800 S. Victoria Road, Ventura, CA 93009 3 Greater Los Angeles County Vector Control District, 12545 Florence Ave., Santa Fe Springs, CA 90670 4 Biological Sciences Department, California Polytechnic State University, San Luis Obispo, CA 93407 (present address: California Department of Fish and Game, Monterey, CA) Received 17 August 2001; Accepted 14 November 2001 ABSTRACT: Chuchupate Campground in Ventura County, California, was closed to the public for 18 years (1982 to 2000) because of uncontrolled vector fleas and persistent plague antibody titers in rodents. The primary purpose of this study was to clarify the plague ecology of Chuchupate Campground by identifying involved rodents and their vector fleas and by determining many of their ecological parameters: abundance, flea and host preferences and diversities, and flea seasonality. Rodents and fleas were identified to species, some fleas were tested for Yersinia pestis, and rodent bloods were analyzed for the presence of antibodies to Y. pestis. During this study, 20 flea species were identified from 10 rodent and one lagomorph species collected. Five species of rodents were seropositive for plague during 13 of the 17 years in which plague testing was conducted. A likely reservoir species was not determined, but evidence of plague resistance was discovered in Merriam s chipmunks (Tamias merriami) and dusky-footed woodrats (Neotoma fuscipes). The susceptible rodent and flea complexes at Chuchupate are the California ground squirrel (Spermophilus beecheyi) and its fleas, Oropsylla montana and Hoplopsyllus anomalus, Merriam s chipmunk and its flea, Eumolpianus fornacis, and the dusky-footed woodrat and its flea, Orchopeas sexdentatus. Host preference, diversity, and seasonality of fleas are discussed, as well as the pivotal role of woodrat houses and nests as foci for hosts, fleas, and plague. Journal of Vector Ecology 27(1): 107-127. 2002. Keyword Index: Plague ecology, rodents, fleas, host preference, diversity, seasonality. INTRODUCTION Plague, caused by the bacterium Yersinia pestis, is maintained in natural foci of the disease in wild rodents through transmission between rodents by their flea ectoparasites. The wild rodent reservoirs are species that are susceptible to the infection but resistant to the disease. In California, as many as 18 species of rodents from various geographic regions have been implicated in the epidemiological cycle of plague. Many species of rodents and other small mammals are susceptible to infection, but are not necessarily important reservoirs of infection. Animal hosts of plague are classified as enzootic (maintenance) hosts, and epizootic (amplification) hosts (Poland and Barnes 1979). The maintenance group includes rodents that are relatively resistant to plague. In this group, mortality from plague infection is low, although a field population may demonstrate high positivity rates. Die-offs, commonly seen among more susceptible rodent species, are rare in this group. Humans can acquire the bacteria from fleas that have previously fed on infected rodents, primarily in an area of a die-off. Plague is occasionally introduced into populations or areas of more susceptible species, especially in areas with an overlap of individuals or populations of two or more species. When this happens in a species that is highly susceptible to plague, an epizootic with high mortality may occur. The susceptibility to plague of a given species may vary geographically or temporally, especially with variations in density of host populations, flea populations, or both. In addition, increasing evidence of resistance to plague in certain rodent species has been suggested or documented (Eskey and Haas 1940, Holdenried and Quan 1956, Hudson et al. 1964, Lang and Wills 1991, Nelson 1980, Politzer 1954).

108 Journal of Vector Ecology June, 2002 In 1900, a ship from Hong Kong docked at San Francisco and released Norway rats carrying plague infected fleas which caused a small epidemic in San Francisco that was eventually controlled (Barnes 1982). However, the bacterium persisted in the Norway rat population and eventually spread to the native rodent populations in the San Francisco Bay area. It was reported that California ground squirrels died in large numbers as early as 1903, and by 1908 the first infected California ground squirrel was found in nearby Contra Costa County (McCoy 1908, Wherry 1908). Today, Y. pestis is found throughout much of California. Plague epizootics usually involve one or two rodent and flea complexes that are essential for amplification (Barnes 1982, Dennis et al. 1999, Eskey and Haas 1940). In California, susceptible rodents generally include California ground squirrels (Spermophilus beecheyi), chipmunks (Tamias spp.), and woodrats (Neotoma spp.). California ground squirrel fleas typically include Oropsylla montana and Hoplopsyllus anomalus (Clover et al. 1989, Davis 1999, Eskey and Haas 1940, Lang and Wills 1991). This latter host and flea complex has been associated with most human plague cases reported in California (Barnes 1982, Lang 1993, 1996, Nelson 1980). The second most important complex of plague ecology in California involves chipmunks (Tamias spp.) and their fleas Eumolpianus spp. (Barnes 1982). Woodrats (Neotoma spp.) and their fleas, Orchopeas sexdentatus and Anomiopsyllus spp., are the third most important complex of plague ecology in California (Barnes 1982, Clover et al. 1989, Lang 1993, 1996). Deer mice (Peromyscus maniculatus) are considered to be a primary reservoir for plague (Goldenburg et al. 1964, Larson et al. 1996, Nelson 1980, Nelson and Smith 1976). Characteristic fleas of deer mice include: Aetheca wagneri, Atyphloceras m. multidentatus, Catallagia spp., Malaraeus telchinus, Peromyscopsylla spp., and Rhadinopsylla sectilis sectilis (Clover et al. 1989, Johnson and Traub 1954, Lang 1993, Larson et al. 1996, Lewis et al. 1988). Voles, especially the California vole, Microtus californicus, are considered to be another reservoir of plague (Goldenberg et al. 1964, Miles et al. 1957). The frequently associated flea of the California vole is Malaraeus telchinus (Miles et al. 1957). Gage et al. (1995) pointed out that plague epizootics occur most often in areas where there are a number of host species living in diverse, patchy habitats. Chuchupate Campground certainly has those characteristics with varying habitats and ecotones, 12 species of rodents (10 species surveyed), and 20 species of fleas, several known to readily transmit plague bacilli. From 1982 (when plague was first discovered at Chuchupate) through 1996, 49 surveys of rodents and fleas were conducted at Chuchupate. An additional 18 intensive surveys from 1997 through 2000 were also conducted. Data reported herein are a result of 67 surveys conducted during 17 years of the 19 year period. Eskey and Haas (1940) postulated that by knowing which species of fleas were efficient vectors of plague bacilli may make it possible to eventually forecast the likelihood of the infection becoming established in a particular locality from studies of the rodent population and their flea infestation. In addition, Murray (1971) stated that contributions to the known seasonality of vector fleas should enhance the ecological and epidemiological knowledge of plague, and may help to more accurately target flea control before flea population density increases, thus preventing epizootics and possible human exposure. The control of plague transmission has been traditionally accomplished by controlling the vector flea populations with insecticides. However, several attempts to control the flea populations at Chuchupate proved unsuccessful, therefore leaving the campground closed. Since Chuchupate Campground was unique in its uncontrolled plague status, this study attempted to identify the rodent and flea ecology that made it unique and which contributed to frequent plague epizootics at the campground. Host preference, diversity, and seasonality of the fleas were examined, as well as the pivotal role of woodrat houses and nests as foci for hosts, fleas, and plague. MATERIALS AND METHODS Study site Chuchupate Campground is located between the Transverse Mountain Range and the Tehachapi Range in northeast Ventura County, California (34 47.29 N, 119 00.13 W). It is approximately 105 km NW of Los Angeles and 6 km SW of Frazier Park at an elevation of 1890 m on Frazier Mountain. The approximate 5.5 ha campground is in the Los Padres National Forest managed by the U. S. Forest Service. It is situated on a NW facing slope covered by a unique mixture of mature desert montane/mixed coniferous forest dominated by singleleaf pinyon pine (Pinus monophylla), Jeffrey pine (Pinus jeffreyi), and desert scrub oak (Quercus johntuckeri). In 1946, a fire destroyed the forest surrounding Chuchupate Campground, but the campground was saved, creating a unique island ecosystem of mature pines and oaks unlike the surrounding areas. The paved roads in the campground were used as dividing lines to separate the area into four sections labeled according to slope and habitat as upper,

June, 2002 Journal of Vector Ecology 109 middle, lower, and meadow. The upper section of the campground is dominated by a Jeffrey pine woodland with an understory of semidesert woodland species including desert scrub oak, canyon live oak (Quercus chrysolepus), interior live oak (Quercus wislizenii), hoary-leaved coffee berry (Rhamnus californicus tomentella), and desert ceanothus (Ceanothus greggii). The middle section has an overstory of Jeffrey pine and singleleaf pinyon pine, and an understory of desert scrub community dominated by rabbitbrush (Chrysothamnus nauseosus) and Great Basin sagebrush (Artemesia tridentata). The lower section is a semidesert woodland dominated by desert scrub oak and desert ceanothus, with a few Jeffrey and pinyon pines. The meadow is comprised of unidentified reeds and sedges. Plague and treatment history The first plague surveillance of Chuchupate Campground was conducted in June 1982, during which 93% of the California ground squirrels were found to be seropositive for Y. pestis, as well as positive fleas collected from those ground squirrels. Due to the high risks for plague transmission, the campground was closed to public use. Since 1982, continued detection of plague antibodies in the rodents and unsuccessful flea control efforts kept the campground closed. Rodent sampling Rodents were live-trapped with Sherman and Tomahawk live traps (H. Sherman Traps, Tallahassee, FL, and Tomahawk Live Trap, Tomahawk, WI). Traps were generally set in the afternoon and collected by noon of the following day to ensure both diurnal and nocturnal rodents had ample opportunity to be trapped. All captured animals were anesthetized with ethyl ether and brushed vigorously to collect fleas for assessing flea density, species diversity, and possible infectivity with plague. Rodent blood samples were drawn by cardiac puncture for plague antibody testing. Animals were sexed, reproductive status noted, and marked with a numbered ear-tag or microchip (National Band and Tag, Newport, KY, and AVID, Norco, CA). All animals were released at their capture sites after recovering from anesthesia. Flea collection Fleas from each capture were counted, preserved, identified, and some were tested. Flea identifications and numbers were summarized to determine a flea index (average number of fleas per host), percent infestation, host preferences, and seasonality. Flea populations declined following treatment with a chitin inhibitor starting in June 1997, therefore possibly affecting later collections and some flea indices. Yersinia pestis determinations Rodent blood samples were tested for Y. pestis antibodies by a passive hemagglutination test (Chu 2000) at the California Department of Health Services (CDHS). Seropositive results were expressed in reciprocal titers, denoting the concentration as determined by titration. Reciprocal titers below 1:32 (1:16 in early surveys) were considered nonspecific reactions, and not positive. Selected fleas were pooled by species and tested for Y. pestis by laboratory mouse inoculation (Chu 2000) at CDHS. Infected animals were then tested by fluorescent antibody tests and culture at CDHS. Data Analysis All summary data presented in tables and figures were calculated from raw data collected in the field. Rodent population estimates were made using the Jolly- Seber model. RESULTS Rodents Intensive rodent trapping and flea collection was conducted monthly from July 1988 through June 1991 and June 1997 through November 1997 and from March 1998 through August 1998. During those periods, a total of 883 individuals from 10 species of rodents was captured. Additional surveys during the study resulted in an additional 680 captures. During the entire study period 1,563 rodents were captured during 6,057 trap nights. Average trap success during the entire period was 26% (26 captures per 100 traps), and ranged from a high of 62% to a low of 7%. Trapping was most successful in areas dominated by large bushes, such as hoary-leaved coffeeberry and desert scrub oak and least successful in areas dominated by rabbitbrush and Great Basin sagebrush and in the meadow. Ten species of rodents were captured during the surveys. In addition, three brush rabbits (Sylvilagus bachmani) and two ornate shrews (Sorex ornatus) were captured. Western gray squirrels (Sciurus griseus) and signs of Botta s pocket gopher (Thomomys bottae) were seen, but none captured. The dusky-footed woodrat was the most frequent rodent species captured (46% of the total captures) (Table 1), and was trapped every month. The deer mouse also was trapped every month and comprised 21% of all rodents captured. Merriam s chipmunks and California ground squirrels were seldom captured during the coldest months (October/November through March/

110 Journal of Vector Ecology June, 2002 Table 1. Frequency distribution of rodent captures at Chuchupate Campground. Number captured includes recaptures. Rodents Number Captured % of Total Dusky-footed woodrat (Neotoma fuscipes) 725 46 Deer mouse (Peromyscus maniculatus) 330 21 Pacific kangaroo rat (Dipodomys agilis) 174 11 Calif. ground squirrel (Spermophilus beecheyi) 136 9 Merriam s chipmunk (Tamias merriami) 89 6 Brush mouse (Peromyscus boylii) 63 4 California vole (Microtus californicus) 16 1 Calif. pocket mouse (Chaetodipus californicus) 16 1 Pinyon mouse (Peromyscus truei) 10 <1 Western harvest mouse (Reithrodontomys megalotis) 4 <1 Total 1,563 April), as they generally undergo winter torpor. Locations of capture were recorded to determine rodent habitat preferences. Deer mice were captured in all habitats sampled. Dusky-footed woodrats were mainly captured in areas with large woody shrubs under which they build their stick houses; whereas California ground squirrels, Pacific kangaroo rats and Merriam s chipmunks were seen and captured in all habitats except the meadow. Brush mice and pinyon mice were only captured in the upper loop of the campground, whereas California voles were captured only in or adjacent to the meadow. The pocket mice and harvest mice were captured only in the lower loop adjacent to the meadow. The Jolly-Seber mark/recapture method was used to estimate population sizes for the entire campground for those species or combination of species that had sufficient numbers of captures and recaptures during the 1997-1998 surveys. Recapture data were sufficient for population estimates of dusky-footed woodrats and all rodents combined. The mean of the population estimates for woodrats was 28.6 with a range of 17 to 51, while the mean of the estimates for all the rodents was 83.4 with a range of 41 to 160. In conjunction with the high population of woodrats, 72 woodrat houses were located in the campground (~12/ ha). Fleas During the entire study from 1982-2000, a total of 2,442 fleas was collected. These fleas are represented by five families, 17 genera, and 20 species. Some species represent extensive range extensions and new host relationships. Of the total fleas collected, only a small number of fleas (156) was not identified or not identified to species. The total number of each species of flea collected and their associated host is listed in Table 2. Seasons during which fleas were collected were divided into spring (March - May), summer (June - August), fall (September - November), and winter (December - February). Snow occasionally prevented sampling during the winter months, especially during February. Most of the seasonality data was from the 1988-1991 surveys and the more recent 1997-1998 surveys, with additional survey data helping to fill in some of the missing months. Seasonality is summarized in Table 3. Flea hosts were categorized as either the normal or primary host, as an occasional host, or as an accidental host (Table 4). Hosts were considered to be primary whenever 50% or greater of that particular species was collected from a single host species. A host was determined as occasional if it was represented by between 6 and 49% share of the total found. A host was determined as accidental if the percent share was 5% or less. In a few cases, there were insufficient data to determine any relationship. See Tables 2, 5, and 6 for numbers of fleas collected on each host and for percent share distribution by flea species or by host species. The following list of flea species is in alphabetic order for ease of reference. Aetheca wagneri This species comprised 9% of the total flea fauna, with the majority collected during the summer months. Most (76%) were found on deer mice. Occasional hosts included dusky-footed woodrats (12%) and brush mice (9%). Aetheca wagneri occurred accidentally on chipmunks, harvest mice, brush mice, pinyon mice, and voles. It was recovered from deer mice all year. Anomiopsyllus falsicalifornicus congruens Only 23 specimens were collected, of which 91% were from

June, 2002 Journal of Vector Ecology 111 Table 2. Total numbers of fleas collected by species and per host at Chuchupate Campground from 1982-2000. HOSTS FLEAS Merriam s chipmunk Calif. ground squirrel Calif. pocket mouse Pacific kangaroo rat western harvest mouse brush mouse deer mouse pinyon mouse dusky-footed woodrat California vole brush rabbit Totals A. wagneri 1 1 19 155 3 24 1 205 A. f. congruens 2 21 23 A. n. nudatus 3 1 1 5 247 257 A. e. longipalpus 4 22 26 A. m. multidentatus 7 1 17 1 26 C. luski 1 1 2 10 5 19 C. i. Interrupta 7 7 E. e. eumolpi 1 1 E. fornacis 113 3 1 117 H. anomalus 49 1 50 H. o. linsdalei 2 7 4 1 14 M. telchinus 1 2 3 12 13 62 92 M. sp. aff. divisus 1 17 249 267 M. cummingi 27 1 28 O. nesiotus 2 4 6 O. sexdentatus 1 292 293 O. montana 3 766 4 19 1 793 P. h. adelpha 14 26 1 4 45 R. s. sectilis 1 1 7 1 6 16 Thrassis sp. 1 1 Totals 123 815 0 31 3 45 262 6 927 66 8 2286 woodrats, and the remainder (9%) from deer mice. They were recovered only from March through June. Anomiopsyllus n. nudatus This subspecies was the fourth most abundant flea collected, and totaled 11% of the total fleas collected. It was collected every month except February (which was seldom sampled because of snow), and was primarily a spring flea with a peak in numbers in March and April. This species was found primarily on dusky-footed woodrats (96%), and with accidental occurrence on chipmunks, kangaroo rats, brush mice, and deer mice. Atyphloceras echis longipalpus Only 26 specimens of this subspecies were collected primarily during the winter, of which 85% were from woodrats. Deer mice were an occasional host with 15% share. Atyphloceras m. multidentatus Only 26 specimens of this subspecies were collected, primarily in the fall from woodrats (65%). An occasional host was the deer mouse (27%). Accidental hosts included pinyon mice and voles. Catallagia luski Only 19 specimens of this species were collected. Although it was collected in small numbers, the primary host appeared to be the deer mouse (53%), and the few specimens were collected more commonly in the spring. Brush mice and woodrats were occasional hosts, and chipmunks and kangaroo rats were accidental hosts. Cediopsylla inaequalis interrupta This subspecies was collected only once (7 specimens), in June from a brush rabbit, one of its normal hosts.

112 Journal of Vector Ecology June, 2002 Eumolpianus eumolpi eumolpi This subspecies was represented by a single specimen collected from a chipmunk during July. Eumolpianus fornacis This species comprised 5% of the total fleas collected. It was collected primarily on chipmunks (97%), and whenever chipmunks were collected (April through November), with a population peak in June. Accidental hosts were woodrats and a single specimen from a vole. Hoplopsyllus anomalus Fifty specimens of this species were collected, of which 49 (98%) were from its primary California ground squirrel host. One specimen was collected from a woodrat, an accidental host. Most were collected during the warmer summer months. During the initial survey in June 1982, H. anomalus accounted for 25 (50%) of the total 50 collected, all from ground squirrels. Hystrichopsylla occidentalis linsdalei Only 14 specimens of this subspecies were collected, all from March through July. Although collected in small numbers, deer mice appear to be the primary host (50%), while the other hosts are considered occasional - woodrats (29%), brush mice (14%), and voles (7%). Malaraeus telchinus The 92 specimens of this species were collected from 6 different host species, but primarily from voles (67%). Occasional hosts were deer mice and woodrats, and accidental hosts were chipmunks, harvest mice, and brush mice. Although seasonality would appear to be summer and fall, it could not accurately be determined since voles hosted the greatest numbers and they were captured only from July through November. Of the M. telchinus collected from voles, 41% were collected in October. Megarthroglossus sp. aff. divisus This possible new species comprised 12% of the total fleas collected at Chuchupate, was the third most abundant flea collected, and was primarily recovered from woodrats (93%). The deer mouse would be considered an occasional host, and brush mice as accidental hosts. Collections were made from September through April, but reached peak population densities in late fall/early winter. One woodrat in September was infested with 25 specimens and another in October with 30 specimens. Meringis cummingi Only 28 specimens were collected, 27 (96%) from kangaroo rats. A single specimen was collected from a deer mouse in April. Those from kangaroo rats were collected in April, July, and primarily September. Opisodasys nesiotus Only 6 specimens were collected, 2 from brush mice and 4 from deer mice. Although few specimens were collected, it is suggestive that Peromyscus ssp. are the preferred hosts. Specimens were collected in March, June, November, and December, too few to determine seasonality. Orchopeas sexdentatus (following Lewis 2000) This species was the second most abundant flea collected, comprising 13% of the total. These fleas were collected almost entirely from dusky-footed woodrats, (one specimen was collected from a deer mouse). They were collected in nearly every month of the year except for December and February, suggesting they may be a warmer season flea, or that those omissions were an artifact of infrequent surveys during the winter. Oropsylla montana This species was the most abundant flea collected, comprising 35% of the total. Of all the O. montana collected, 97% were recovered from California ground squirrels. This flea, known primarily from ground squirrels, was also found in small numbers on Merriam s chipmunks, deer mice, duskyfooted woodrats, and a brush rabbit, all accidental hosts. These fleas were collected every time ground squirrels were captured (April through November), but primarily during the summer months. During the initial survey in June 1982, O. montana accounted for 250 (32%) of the total 793 specimens collected. During that survey, the 14 ground squirrels captured averaged 19.6 O. montana per animal. Peromyscopsylla hesperomys adelpha Only 45 specimens of this subspecies were collected. Although few in number, the deer mouse would be considered the primary host (58%), while the brush mouse (31%) and woodrat (9%) would be considered occasional hosts. One specimen was collected from a pinyon mouse. Specimens were collected during every month except July, suggesting that it is likely present all year. Rhadinopsylla sectilis sectilis Only 16 specimens of this subspecies were collected. Although few in number, the deer mouse and woodrat would be considered primary hosts (44 and 38% prevalence respectively). One specimen each was collected from a harvest mouse, a brush mouse, and a pinyon mouse. Although few in number, specimens were collected during the cooler months from November through May, suggesting that this is likely a winter and spring flea. Thrassis sp. Only one specimen was collected from a Pacific kangaroo rat in October 1997. Unfortunately this specimen was damaged, but it was determined to be either Thrassis bacchi or Thrassis aridis. Although this represents a significant range extension, it was determined to likely be T. aridis (subspecies undetermined) because their normal hosts are species of Dipodomys (Hubbard 1947, Lewis et al. 1988). Rodent infestation The flea index (average number of fleas per rodent) was calculated for the entire study period, 1982-2000,

June, 2002 Journal of Vector Ecology 113 Table 3. Seasons and months each flea species was recovered at Chuchupate Campground. (x = month recovered). FLEAS Spring Summer Fall Winter M A M J J A S O N D J F A. wagneri x x x x x x x x x x x x A. f. congruens x x x x x A. n. nudatus x x x x x x x x x x x A. e. longipalpus x x x x x x x A. m. multidentatus x x x x x x x C. luski x x x x x C. i. Interrupta x E. e. eumolpi x E. fornacis x x x x x x x x H. anomalus x x x x x H. o. linsdalei x x x x x x M. telchinus x x x x x x x x x M. sp. aff. divisus x x x x x x x M. cummingi x x x O. nesiotus x x x x O. sexdentatus x x x x x x x x x x O. montana x x x x x x x x x P. h. adelpha x x x x x x x x x x x R. s. sectilis x x x x x x x Thrassis sp. x and for 1997-1998 (Table 7). Percent infestation or prevalence (percentage of rodents captured with fleas) for 1997-1998 was also determined (Table 7). The diversity of flea species for each host is shown in Table 4. The dusky-footed woodrat and the deer mouse demonstrated great flea diversity, each with 15 species. The next greatest diversities were for the brush mouse with 9 species of fleas and Merriam s chipmunk with 7 species. Merriam s chipmunk (Tamias merriami) Merriam s chipmunks had a flea index of 1.8 and a high percent infestation (75%) during 1997-1998, and an overall flea index of 1.4. Chipmunks had a diversity of 7 flea species. The most abundant flea found on the chipmunks was E. fornacis, which comprised 92% of the fleas collected. The remaining 6 species of fleas recovered were very few in number. California ground squirrel (Spermophilus beecheyi) California ground squirrels had the highest overall flea index (6.6) and a high percent infestation (79%) during 1997-1998. During the 1982-1988 surveys, the flea index was 9.4 and ranged from 2.0 to 19.6 (in 1982). Of the rodent species with fleas, the California ground squirrel had a low flea species diversity, consisting of only 2 species, O. montana and H. anomalus. Oropsylla montana comprised 94% of the fleas collected from ground squirrels. California pocket mouse (Chaetodipus californicus) None of the 16 California pocket mice collected had any fleas. Pacific kangaroo rat (Dipodomys agilis) Many of the kangaroo rats captured in earlier surveys were not examined for fleas as kangaroo rats are generally considered highly resistant to plague (Holdenried and Quan 1956, Nelson 1980). As a result, only 59 of the 174 specimens collected were examined for fleas and none for Y. pestis. Pacific kangaroo rats had an overall flea index of 0.7, and a low percent infestation (24%) during 1997-1998. The species diversity was relatively low with five species recorded. Meringis cummingi represented 87% of all the fleas recovered. The other 4 species were recorded in very small numbers. Western harvest mouse (Reithrodontomys megalotis) Three of the four specimens collected had 1 flea each, represented by two A. wagneri and 1 M. telchinus. Brush mouse (Peromyscus boylii) Brush mice had a low overall flea index of 0.7, a low percent infestation of 37% during 1997-1998, yet a relatively high diversity of nine flea species. The two common flea taxa were A. wagneri (42%) and P. h. adelpha (31%). The remaining 7 flea species were found in very low numbers. Deer mouse (Peromyscus maniculatus) Deer mice had an overall flea index of 0.9, a relatively low percent infestation of 42% during 1997-1998, but very high diversity of 15 flea species. Aetheca wagneri was the

114 Journal of Vector Ecology June, 2002 Table 4. Host preference of each flea species at Chuchupate Campground, its recorded importance in the literature in plague epidemiology (*), and flea species diversity (Flea Species/Host). (++ normal or primary host; + occasional host; - accidental host;? insufficient data). HOSTS FLEAS Merriam s chipmunk Calif. ground squirrel Calif. pocket mouse Pacific kangaroo rat western harvest mouse brush mouse deer mouse pinyon mouse dusky-footed woodrat California vole brush rabbit * A. wagneri - - + ++ - + - A. f. congruens + ++ * A. n. nudatus - - - - ++ A. e. longipalpus + ++ * A. m. multidentatus + - ++ - C. luski - - + ++ + C. i. Interrupta ++ * E. e. eumolpi ++ * E. fornacis ++ - - * H. anomalus ++ - * H. o. linsdalei + ++ + + * M. telchinus - - - + + ++ * M. sp. aff. divisus - + ++ M. cummingi ++ - * O. nesiotus + ++ * O. sexdentatus - ++ * O. montana - ++ - - - * P. h. adelpha + ++ - + * R. s. sectilis?? ++? ++ Thrassis sp. ++ Flea Species/Host 7 2 0 5 2 9 15 4 15 5 2 most abundant species, representing 59% of the fleas recovered. In relationship to the numbers of A. wagneri recovered, the remaining 14 flea species were found in lower numbers. However, the deer mouse was considered the primary host for the following 6 species: A. wagneri, C. luski, H. o. linsdalei, O. nesiotus, P. h. adelpha, and R. s. sectilis. Pinyon mouse (Peromyscus truei) Only 6 fleas were collected from 10 pinyon mice, with an overall flea index of 0.6. Three of the six specimens were A. wagneri, and the other three species were recovered as only single specimens. Dusky-footed woodrat (Neotoma fuscipes) Dusky-footed woodrats had an overall flea index of 1.3, and a percent infestation of 54% during 1997-1998. They were infested with a high diversity of 15 flea species. The most numerous species found were O. sexdentatus (31%), Megarthroglossus sp. (27%), and A. n. nudatus (27%). The remaining 12 flea species were represented by proportionately lower numbers. However, the duskyfooted woodrat was considered the primary host for the following 7 species: A. f. congruens, A. n. nudatus, A. e. longipalpus, A. m. multidentatus, Megarthroglossus sp., O. sexdentatus, and R. s. sectilis. California vole (Microtus californicus) California voles had the second highest overall flea index (2.7), and the highest percent infestation (100%) during 1997-1998. The species diversity of fleas was relatively low, represented by 5 species. Malaraeus telchinus comprised 94% of the fleas from California voles. The

June, 2002 Journal of Vector Ecology 115 Table 5. Percent share distribution of each host species per each species of flea at Chuchupate Campground (Percentages rounded to whole numbers; totals may not equal 100%). HOSTS FLEAS Merriam s chipmunk Calif. ground squirrel Calif. pocket mouse Pacific kangaroo rat western harvest mouse brush mouse deer mouse pinyon mouse dusky-footed woodrat California vole brush rabbit A. wagneri <1 <1 9 76 1 12 <1 A. f. congruens 9 91 A. n. nudatus 1 <1 <1 2 96 A. e. longipalpus 15 85 A. m. multidentatus 27 4 65 4 C. luski 5 5 11 53 26 C. i. Interrupta 100 E. e. eumolpi 100 E. fornacis 97 3 <1 H. anomalus 98 2 H. o. linsdalei 14 50 29 7 M. telchinus 1 2 3 13 14 67 M. sp. aff. divisus <1 6 93 M. cummingi 96 4 O. nesiotus 33 67 O. sexdentatus <1 100 O. montana <1 97 <1 2 <1 P. h. adelpha 31 58 2 9 R. s. sectilis 6 6 44 6 38 Thrassis sp. 100 remaining 4 flea species were recovered as only single specimens. Plague testing During the course of the study, 61 surveys were conducted in which rodent blood samples were collected and tested for antibodies to Y. pestis. No surveys were conducted during 1987 and 1995. Nineteen of the 61 surveys (31%) revealed positive titers ranging from 1:32 to 1:8192. Only three rodents were found dead during the entire study (a dusky-footed woodrat in both 1983 and 1990, and a Merriam s chipmunk in 1986). Two of the 3 specimens tested were florescent antibody positive (the one woodrat from 1983 and the chipmunk), thus representing the only evidence of any possible die-offs from plague. Five species of rodents demonstrated positive Y. pestis antibody titers at Chuchupate Campground during 13 (76%) of the 17 years in which plague testing was conducted (Table 8). Only 4 (24%) of the 17 years surveyed revealed no positive rodents. California ground squirrels were positive 11 (85%)of the 13 years, Merriam s chipmunks were positive 8 (62%) of the 13 years, dusky-footed woodrats were positive 7 (54%) of the 13 years, deer mice were positive 4 (31%) of the 13 years, and brush mice were positive 3 (23%) of the 13 years. In May 1998, a Merriam s chipmunk had a titer of >1:256 (insufficient blood to titer further). This chipmunk had been caught during a previous survey in July 1996, at which time it had a titer of 1:8192. Another chipmunk had a titer of 1:32 in July 1997, and 1:32 again in June 1998. A dusky-footed woodrat with a titer of 1:4096 in July 1999, again tested positive with a titer of

116 Journal of Vector Ecology June, 2002 Table 6. Percent share distribution of each species of flea per each host species at Chuchupate Campground. (Percentages rounded to whole numbers; totals may not equal 100%). HOSTS A. wagneri A. f. congruens A. n. nudatus A. e. longipalpus A. m. multidentatus C. luski C. i. interrupta E. e. eumolpi E. fornacis H. anomalus H. o. linsdalei M. telchinus M. sp. aff. dividus M. cummingi O. nesiotus O. sexdentatus O. montana P. h. adelpha R. s. sectilis Thrassis sp. FLEAS Merriam s chipmunk 1 2 1 1 92 1 2 Calif. ground squirrel 6 94 Calif. pocket mouse Pacific kangaroo rat 3 3 87 3 3 Western harvest mouse 33 67 Brush mouse 42 2 4 4 7 2 4 31 2 Beer mouse 59 <1 2 2 3 4 3 5 6 <1 2 <1 2 10 3 Pinyon mouse 50 17 17 17 Dusky-footed woodrat 3 2 27 2 2 <1 <1 <1 <1 1 27 31 2 <1 <1 California vole 2 2 2 2 94 Brush rabbit 88 13

June, 2002 Journal of Vector Ecology 117 Table 7. Flea index (average number of fleas per individual host) at Chuchupate Campground for the entire study period, 1982-2000, and for 1997-1998, and percent infestation or prevalence of fleas for each rodent species during 1997-1998. Rodents Flea Index Flea Index % Infestation 1982-2000 1997-1998 1997-1998 California ground squirrels 6.6 6.2 79 California voles 2.7 4.8 100 Merriam s chipmunks 1.4 1.8 75 Dusky-footed woodrats 1.3 1.6 54 Deer mice 0.9 0.7 42 Western harvest mice 0.8 0.5 50 Pacific kangaroo rats 0.7 1.0 24 Brush mice 0.7 0.6 37 Pinyon mice 0.6 0 0 California pocket mice 0 0 0 1:256 in June 2000. Rodents captured prior to 1996 were not marked, so persistence of antibody titers could not be determined. However, the 2 chipmunks and 1 woodrat that retested positive during 1997-2000 surveys suggest possible plague resistance in those species. Prior to the 1997-1998 surveys, numerous flea pools from California ground squirrels tested positive. During the 1997-1998 surveys, flea pools from other rodent species were tested, but only 1 flea pool was found to be positive for plague. This positive flea pool consisted of 3 fleas, M. telchinus, from a single California vole captured in July 1997. That particular vole was not seropositive to Y. pestis at that time, nor upon recapture in August and October 1997. DISCUSSION Determining the interactions between the 10 possible rodent host species and the 19 species of fleas recovered from rodents at Chuchupate Campground is an integral part of understanding the persistence of plague there. The variety and densities of rodent hosts, the array of rodent species involved in plague, the mixture and possibly changing status of maintenance and amplification species, and the diverse flea fauna, especially many known to have been involved in plague epidemiology, makes an interesting and complex plague scenario at Chuchupate. The findings of the flea and host ecology and relationships begin to provide insight into Chuchupate s persistent plague problem and past difficulties in controlling the flea populations using traditional methods. Eskey and Haas (1940) and Murray (1971) suggested that studies of rodent populations, their flea infestations, and flea seasonality, along with knowledge of their vector competency, might eventually allow one to determine the likelihood of plague becoming established in a particular locality. This knowledge can also allow a more accurate targeting of flea control before their populations increase, thus preventing epizootics and possible human exposure. In the following discussion, each species of flea, their host preferences, seasonality, and their capacity to transmit Y. pestis (Tables 3 and 5) are discussed and compared with data from the literature. The key rodent species, their flea composition, and involvement with plague are also discussed and compared with data from the literature. Flea ecology Aetheca wagneri This flea was present throughout the year, but was mostly collected during the warmer months. Numerous authors (Hubbard 1947, Larson et al. 1996, Linsdale and Davis 1956, Lewis et al. 1988, Lang 1993, Morlan 1955) stated it was present all year around, but the timing of peak populations varied considerably geographically and likely altitudinally. The peak in A. wagneri numbers during the warmer months at Chuchupate is likely a result of its high altitude. All authors stated that A. wagneri was most commonly collected from P. maniculatus, which was also true for Chuchupate with 76% occurrence on that host. It was found to infest a total of 7 rodent species, primarily deer mice, but occasionally brush mice and woodrats. Aetheca wagneri can serve as a vector of Y. pestis (Eskey and Haas 1940), and has been found naturally infected with Y. pestis (Clover et al. 1989, Larson et al. 1996, Morlan 1955). Hubbard (1947), Lang (1993), and Lewis et al. (1988) reported little evidence that this flea is important

118 Journal of Vector Ecology June, 2002 Table 8. Summary of plague serosurveys conducted at Chuchupate Campground, 1982-2000. Year No. Positive Surveys / No. Surveys Species Positive 2000 2 / 2 S. beecheyi, T. merriami, N. fuscipes, P. boylii 1999 3 / 3 T. merriami, N. fuscipes, P. maniculatus 1998 3 / 6 T. merriam, P. maniculatus 1997 1 / 6 S. beecheyi, T. merriami, P. maniculatus, P. boylii 1996 1 / 1 S. beecheyi, T. merriami 1995 0 / 0 1994 1 / 1 S. beecheyi, N. fuscipes 1993 1 / 2 S. beecheyi 1992 0 / 1 1991 0 / 6 1990 0 / 11 1989 0 / 11 1988 2 / 6 S. beecheyi, T. merriami 1987 0 / 0 1986 1 / 1 S. beecheyi, T. merriami 1985 1 / 1 S. beecheyi, T. merriami, N. fuscipes 1984 1 / 1 S. beecheyi, N. fuscipes, P. maniculatus, P. boylii 1983 1 / 1 S. beecheyi, N. fuscipes 1982 1 / 1 S. beecheyi, N. fuscipes Table 9. Host preference of each flea species on rodents likely associated with woodrat stick houses at Chuchupate Campground, and its recorded importance in the literature in plague epidemiology(*). (++ normal or primary host; + occasional host; - accidental host;? insufficient data). brush mouse deer mouse dusky-footed woodrat * A. wagneri + ++ + A. f. congruens + ++ * A. n. nudatus - - ++ A. e. longipalpus + ++ * A. m. multidentatus + ++ C. luski + ++ + * H. o. linsdalei + ++ + * M. telchinus - + + * M. sp. aff. divisus - + ++ * O. nesiotus + ++ * O. sexdentatus - ++ * P. h. adelpha + ++ + * R. s. sectilis? ++ ++

June, 2002 Journal of Vector Ecology 119 in transmission, and its role in maintaining Y. pestis is likely to be of only minor importance. Specimens of A. wagneri were not tested for Y. pestis during this study. Anomiopsyllus falsicalifornicus congruens Although only 23 specimens were collected, 91% were on dusky-footed woodrats, and the remainder on deer mice. All were collected in the spring and summer months (March-June). Augustson (1943), Lewis et al. (1988), Linsdale and Davis (1956), and Stark and Kinney (1969) listed this flea as primarily an ectoparasite of woodrats, especially the dusky-footed woodrat, and their associated species of Peromyscus. Most authors also stated that this flea was more commonly encountered in woodrat nests. Linsdale and Davis (1956) stated that it was present in woodrat nests in large numbers throughout the year, reaching peaks in January/February and in May, and in January and June on animals. They also recorded an average of 46.8 fleas per woodrat nest while only 2.2 per animal. There is no evidence this flea is involved in plague epidemiology. Anomiopsyllus n. nudatus This flea was present throughout the year, with a peak in March/April, and was found primarily (96%) on dusky-footed woodrats. Records in the literature stated that A. nudatus was most frequently encountered on woodrats, and from winter and spring months (Lang 1993, Morlan 1955). It was also reported as being more commonly recovered from woodrat nests (Augustson 1943, Morlan 1955), although the high numbers recovered from woodrats at Chuchupate would indicate that they may not be as restricted to the nest as is A. f. congruens. During this study, no A. nudatus were tested for Y. pestis. Although Hubbard (1947) reported that this flea was never used in Y. pestis experiments, Eskey and Haas (1940) did infect it in the laboratory. Barnes (1982) and Pollitzer and Meyer (1961) indicated that an unidentified species of Anomiopsyllus was found naturally infected with plague bacilli from woodrats in New Mexico. Since Morlan (1955) recovered large numbers of this flea from woodrats and their nests in New Mexico, it is suspected that the reference to Anomiopsyllus spp. may indeed be A. n. nudatus. Atyphloceras echis longipalpus Only 26 specimens were collected, but 85% were from woodrats, primarily those trapped during the winter months. The most commonly recorded preferred hosts are woodrats, but with occasional records from Peromyscus spp. and Microtus sp. (Augustson 1943, Hubbard 1947, Lewis and Lewis 1994, Linsdale and Davis 1956, Morlan 1955). Morlan (1955) reported finding A. echis in January, March, April and November and Linsdale and Davis (1956) reported a possible peak in December. Linsdale and Davis (1956) reported this flea as being primarily in nests of dusky-footed woodrats. There is nothing reported in the literature indicating that this flea is a vector of Y. pestis, and no specimens were tested during this study. Atyphloceras m. multidentatus Only 26 specimens were collected, 65% from woodrats and 27% from deer mice. None were ever recovered during summer months, suggesting this may be a cool weather flea at Chuchupate. Numerous authors reported A. m. multidentatus more commonly during the cooler months, primarily from Peromyscus spp., Neotoma spp., and occasionally California voles. Clover et al. (1989) reported this flea as a nest flea of dusky-footed woodrats and of deer mice found in association with woodrats. Lewis et al. (1988) indicated that its role in pathogen transmission is uncertain. However, Eskey and Haas (1940) infected this flea in the laboratory, and Pollitzer and Meyer (1961) reported it to be naturally infected with plague bacilli from Peromyscus in the western US. There were no A. m. multidentatus tested for Y. pestis during this study. Catallagia luski Only 19 specimens were collected, of which the majority were from deer mice (53%) and during the spring. Schwan and Nelson (1983) and Lewis and Haas (2001) reported this flea from several species of Peromyscus and also from cooler months. Little is known of the biology and ecology of Catallagia sp., in part because it is a nest flea and rarely collected from hosts (Hubbard 1940, Lewis and Haas 2001, Schwan and Nelson 1983). In addition, many species of Catallagia spp. appear to be cold season fleas and are restricted to higher elevations. Nothing was reported in the literature indicating that this flea is a vector of Y. pestis, and no specimens were tested during this study. Cediopsylla inaequalis interrupta Seven C. i. interrupta were collected, all from one brush rabbit in June 1997; therefore seasonality is indeterminate. Its preferred hosts are rabbits and hares (Lewis et al. 1988, Linsdale and Davis 1956), but it has been found in small numbers (3.3%) on California ground squirrels (Lang 1993). This flea apparently does not transmit Y. pestis (Lang 1993, Lewis et al. 1988), and Hubbard (1947) reported that this species is extremely difficult, if not impossible, to infect with Y. pestis under laboratory conditions. No specimens were tested during this study. Eumolpianus eumolpi eumolpi This flea, represented by a single specimen, was recovered from a Merriam s chipmunk collected in July. This flea has also been recorded from chipmunks at nearby Mt. Pinos area. Lewis et al. (1988) described it as a common parasite of sciurids, with chipmunks being the preferred host. Eskey and Haas (1940) reported that this flea represented 82% of the fleas from Tamias spp. They also determined that

120 Journal of Vector Ecology June, 2002 this flea was an efficient vector of plague bacilli, and Barnes (1982) and Pollitzer and Meyer (1961) stated that it has been found naturally infected with Y. pestis and involved in epizootic plague amplification. Eumolpianus fornacis This species of flea was collected primarily from chipmunks, and whenever chipmunks were collected (April through November), with peak numbers in June. Eumolpianus fornacis has been recorded primarily on Merriam s chipmunks in southern California from Monterey County south (Hubbard 1947, Linsdale and Davis 1956). Linsdale and Davis (1956) also recorded this flea from brush mice and dusky-footed woodrats, but 78% were from Merriam s chipmunk. They also recorded them in small numbers (1.9/animal compared to 1.4/chipmunk at Chuchupate) year-round in Monterey County, a site much lower in elevation than Chuchupate. Hubbard (1947) stated that this species is closely related to Eumolpianus eumolpi, a flea considered to be a vector of Y. pestis and known to readily bite humans. At Chuchupate, this flea represented 92% of all the fleas collected from Merriam s chipmunks, and with the chipmunk s continual involvement in plague ecology, it is highly likely that this flea is a vector at Chuchupate. Barnes (1982) reported E. fornacis as a principal vector of Y. pestis in chipmunks in the Pacific states. It should be noted, however, that 4 of the 5 flea species found on accidental chipmunk hosts at Chuchupate are also known to transmit Y. pestis. During this study, E. fornacis was seldom collected in sufficient numbers to frequently test for plague, and the plague bacillus was not isolated in them. Hoplopsyllus anomalus Fifty specimens were collected, all but 1 on ground squirrels. Hoplopsyllus anomalus normally occurs in large numbers on California ground squirrels in the dry, warmer months (Barnes 1982, Davis 1999, Holdenried et al. 1951, Lang 1993, Rutledge et al. 1979, Rykman et al. 1954, Rykman 1971), which was also true at Chuchupate with their occurrence from May through October. At Chuchupate, the low numbers of H. anomalus are likely the result of the high altitude and long, cold winters. Burroughs (1947), Eskey and Haas (1940), and Wheeler and Douglas (1941) reported H. anomalus as a poor vector of Y. pestis in laboratory studies. It was suggested that it is an inefficient vector, but is compensated for by its frequent excessive numbers on ground squirrels (Burroughs 1947, Eskey and Haas 1940). Hubbard (1947), however, reported it as an efficient vector and transmitter of Y. pestis. No specimens of H. anomalus were tested for Y. pestis during this study because too few were collected. Hystrichopsylla occidentalis linsdalei This flea was represented by only 14 specimens at Chuchupate, all from March through July, and the majority (50%) from deer mice. Although numbers were low, woodrats, brush mice, and voles were considered occasional hosts. The only recorded seasonality was by Linsdale and Davis (1956), who noted that it was present all year but in low numbers. The majority of specimens collected at Chuchupate were from the spring and early summer months. The possible difference between Chuchupate and their reported all-year seasonality is likely the difference in elevation (457-838 m versus 1890 m). Lewis and Lewis (1994) reported this flea as restricted to coastal California, and infesting species of Microtus, Peromyscus, and Neotoma. Linsdale and Davis (1956) found it to be the most common flea on California voles and the second most abundant flea they recorded. Their more coastal and lower elevation site may reflect the differences found compared to those at Chuchupate. It is interesting to note that this flea is nearly always collected as a single specimen from its hosts, possibly because it may be a nest flea as reported by Clover et al. (1989). This flea is a known vector of Y. pestis (Clover et al. 1989, Lewis and Lewis 1994), but no specimens were tested for Y. pestis during this study. Malaraeus telchinus The 92 specimens of this flea were collected from 6 different host species, but primarily from California voles (67%). Occasional hosts were deer mice and dusky-footed woodrats (13 and 14% respectively). Seasonality at Chuchupate was undetermined, since their primary host (voles) were mostly captured during the summer and fall months. Specimens were collected every season, which would agree with numerous authors who all stated that M. telchinus were found throughout the year and seldom with distinct peaks. The many authors also listed numerous host species and as many as 11 species of rodent hosts from one site (Linsdale and Davis 1956), but Peromyscus spp. appeared to be the most prevalent recorded hosts. Linsdale and Davis (1956) found this flea to be the most abundant on California voles, with a flea index of 3.0 per vole. The overall flea index on voles at Chuchupate was 2.7, and M. telchinus represented 94% of the fleas from voles. Clover et al. (1989) indicated that this flea was found in the nests of dusky-footed woodrats. Eskey and Haas (1940) reported M. telchinus as a poor vector of Y. pestis under experimental conditions. Burroughs (1947) and Kartman and Prince (1956) also reported that this flea appears to be an inefficient plague vector. Marchette et al. (1962) suggested that it might effectively transmit the infection via mass infestation. Pollitzer and Meyer (1961) reported this flea was naturally infected with Y. pestis from Peromyscus spp. and Microtus spp. During the study,