Movement of Male Dermacentor andersoni (Acari: Ixodidae) Among Cattle Author(s): T. J. Lysyk Source: Journal of Medical Entomology, 50(5):977-985. Published By: Entomological Society of America URL: http://www.bioone.org/doi/full/10.1603/ME13012

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research.

SAMPLING, DISTRIBUTION, DISPERSAL

Movement of Male Dermacentor andersoni (Acari: Ixodidae) Among Cattle T. J. LYSYK1,2 Research Centre, Agriculture and Agri-Food Canada, Lethbridge, AB, Canada T1J 4B1

J. Med. Entomol. 50(5): 977Ð985 (2013); DOI: http://dx.doi.org/10.1603/ME13012

ABSTRACT Movement of male Dermacentor andersoni (Stiles) was examined among 54 pairs of artiÞcially infested donor and recipient cattle during a 3-yr period. The number of males declined at a rate independent of the initial level of infestation, while the rate of decline of females on the donor animals tended to increase with initial infestation level. Male tick movement to recipient cattle was observed in 26 of 54 (48%) of the animal pairs, but varied among years and trials. Movement tended to be greater during April compared with May and June. The daily probability of movement averaged (SD) 0.067 (0.082), and the number of males moving per day averaged (SD) 0.083 (0.228). Logistic and Poisson regression models were developed and indicated that movement was determined by interactions between the number of males on the donor animals, differences in the number of females on the donor and recipient cattle, temperature, and female age. These models can be used to incorporate movement into tick population models. KEY WORDS Dermacentor andersoni, movement, survival, anaplasmosis

Rocky Mountain wood tick, Dermacentor andersoni (Stiles), is widely distributed throughout the western United States and Canada. This tick is the principal vector of the causative agent of bovine anaplasmosis, Anaplasma marginale (Theiler), in the northwestern states including Montana, Wyoming, Idaho, Oregon, and Washington (Peterson et al. 1977). Ticks found in the adjacent Canadian provinces of British Columbia and Alberta are competent vectors of A. marginale (Scoles et al. 2006); however, natural outbreaks of the disease have never been recorded in these two provinces. Natural outbreaks associated with D. andersoni occurred in southern Saskatchewan in 1983 (ScholÞeld and Saunders 1987) and again in 2008 (Howden et al. 2010). Outbreaks in Manitoba during 2008 Ð2010 were likely transmitted by Dermacentor variabilis (Say). In the acute phase, bovine anaplasmosis causes severe anemia, reduced weight gains, and sometimes death (Zaugg 1990). Surviving animals are persistently infected and serve as reservoirs for the pathogen. Because of this, anaplasmosis remains as a barrier to the movement of cattle from the United States into Canada. A key question is whether or not the reduced incidence of anaplasmosis in western Canada is because of import regulations intended to prohibit entry of the pathogen, or to some other factor that inhibits transmission by the competent vectors present.

1 Agriculture and Agri-Food Canada, Lethbridge Research Centre, P. O. Box 3000, Lethbridge, AB T1J 4B1, Canada. 2 Corresponding author, e-mail: [email protected].

Ticks are capable of transmitting pathogens through a variety of routes including transovarially (parent to offspring), transstadially (larvae to nymph, nymph to adult), and intrastadially by the same tick switching from one host to the next. A. marginale can be transmitted transstadially and also intrastadially by males (Kocan et al. 1992a). Transstadial transmission is unlikely because immature stages feed on small mammals that are not hosts to A. marginale (Rozeboom et al. 1940) and therefore have no means to acquire the pathogen. There have been numerous demonstrations that male D. andersoni can intrastadially transmit A. marginale (Sanborn et al. 1938; Rozeboom et al. 1940; Zaugg et al. 1986; Kocan et al. 1992a,b; Scoles et al. 2005, 2006, 2008). The importance of intrastadial transmission is being increasingly recognized in this (Stiller and Coan 1995) and other tickÐpathogenÐ host systems (Bremer et al. 2005, Little et al. 2007, Ueti et al. 2008). Transmission of A. marginale by a single male tick has been recorded on several occasions (Rozeboom et al. 1940, Scoles et al. 2008). Intrastadial transmission requires that attached ticks change hosts (Little et al. 2007) under natural or seminatural conditions. Rhipicephalus (Boophilus) microplus (Canestrini) moved from two donor to a total of seven recipient cattle as larvae and adults (Mason and Norval 1981). Stiller et al. (1989) demonstrated that D. andersoni males moved from three donor to Þve recipient cattle under seminatural conditions. Rhipicephalus sanguineus (Latreille) showed considerable movement among dogs in two trials consisting of four dogs each (Little et al. 2007). Although tick move-

978

JOURNAL OF MEDICAL ENTOMOLOGY

ment was recorded in each of these studies, the numbers of hosts used and the number of independent trials were relatively small and do not allow estimating the frequency of movement or elucidating factors that might affect movement. The purpose of the current study was to measure the frequency of movement by male D. andersoni from donor to recipient cattle under variable environmental conditions. Results were used to develop logistic and Poisson regression models of male tick movement among hosts.

Materials and Methods Male movement was measured using pairs of donor and recipient cattle. Trials were conducted each April, May, and June over a 3-yr period. Cattle used were predominantly white Holstein heifer calves ranging in weight from 100 to 300 kg. Holsteins were used, as they are readily available, and the white color allows easier detection of ticks. Cattle were randomly assigned to pairs and to donor or recipient status. Each pair was housed separately throughout the trial in tick-free 12 by 12-m pens. The pens had cement ßoors and were cleaned regularly. The lot was located on the grounds of Lethbridge research Centre in an area unsuitable for wild D. andersoni. Donor animals were infested with doses of either 75 (high dose), 50 (medium dose), or 25 (low dose) male and female ticks, and recipient animals were infested with 25 females to act as a pheromone source. Two pairs of animals per dose were used in each monthly trial, for a total of six pairs of animals per trial. Overall, 54 pairs of donor and recipient cattle were used, and no cattle were used twice. Ticks used were the Þrst-generation progeny of montane ticks collected the previous summer. Montane ticks were used, as they have a tendency to feed on the upper back of the animal and would therefore be easier to detect (Wilkinson and Lawson 1965). Ticks were reared using methods described previously (Lysyk 2008). Cattle were initially infested by placing ticks within stockinette sleeves glued to the animalsÕ backs (Lysyk 2008). Ticks and sleeves were left in place until the majority of the ticks had attached. This generally required 4 Ð 6 d, but occasionally up to 11 d if the trial occurred during cold weather. Once the majority of the males were attached, the sleeves were removed, any unattached ticks removed and counted and the remainder on the back counted. Cattle were then examined at 2Ð3-d intervals, and the number of male and female ticks on each animal counted. Cattle were attached to a post using a halter and thoroughly examined over the entire body. Female ticks that appeared close to full engorgement were removed and counted to avoid contaminating the environment. Males that were detected on recipient cattle were assumed to have moved from the donor cattle at the beginning of the interval. Changes in Predictor Variables. Ambient Temperature. Mean hourly temperatures were obtained from a weather station located within 0.75 km of the study

Vol. 50, no. 5

site. Mean temperatures were calculated for each trial, as well as for each interval within a trial. Female Aging. Mean hourly temperatures were also used to calculate a standardized age scale for engorgement, as time to engorgement can be affected by hourly temperature (Lysyk 2008). Cumulative normalized time at the beginning of each sample interval was calculated for each trial using hourly temperatures and the equation Age ⫽ ⌺hr⫻⌬t where h ⫽ hour, ⌬t ⫽ 1/24, and r is the daily rate of development. The latter was calculated using the rate equation r ⫽ 1/(exp(3.0732 ⫺ 0.0535␪ ⫹ 0.0008␪2) where ␪ ⫽ hourly ambient air temperature (Lysyk 2008). The expected cumulative engorgement at the beginning of each interval and trial was calculated using a weibull distribution such that F(Ai) ⫽ 1 ⫺ exp(⫺((Ai ⫺ 0.625)/0.34)1.81) where Ai ⫽ cumulative normalized age at the beginning of the ith interval (Lysyk 2008). The expected proportion engorgement (Pi) was calculated for each trial and interval as F(Ai ⫹ 1) ⫺ F(Ai). Calculated values of r were adjusted as r’ ⫽ r/x to account for solar insolation, as the original equation was derived using cattle held indoors at constant temperatures. The value of the adjustment factor x was determined empirically by measuring how well the model Þt the observed data when values of x were varied from 0.75 to 1.00 by 0.01. For each value of x, the expected number of engorged ticks, Ei, was calculated for each interval and trial by multiplying the total number of engorged ticks removed during each trial by Pi. The sum of (Oi ⫺ Ei)2 was calculated where Oi ⫽ the observed number of ticks that engorged during an interval. This was repeated for all values of x, and the value that produced the lowest sum was selected as the adjustment factor. Tick Numbers. The total number of males (males on donor ⫹ males on recipient), number of females on donors, and the number of females on the recipients at the beginning of each interval were transformed to ln(y ⫹1). A fourth variable, the female differential, was calculated as ⌬F ⫽ ln(RF⫹1) ⫺ ln(DF⫹1) where RF ⫽ the number of females on the recipient cattle and DF ⫽ the number of females on the donor cattle. The variable ⌬F was used as an index of the difference between female numbers on recipient animals compared with donor animals. A value of ⌬F ⬎ 1 indicated that the number of females on the recipient animals exceeded the number on the donor animals while ⌬F ⬍ 1 indicated the number of females on the donor animals was greater than on the recipients. Linear regression (Proc Mixed, SAS Institute 2010) was used to estimate the relationship between numbers alive and time and also to determine whether changes in each variable over time were consistent among doses. The day the bag was removed was designated as day 0, and time of each sample calculated as days since bag removal. The model used was ln(y ⫹ 1) ⫽ ␤0 ⫹ ␤0L ⫻ (XL) ⫹ ␤0M ⫻ (XM) ⫹ ␤1 ⫻ T ⫹ ␤1L ⫻ (XL) ⫹ ␤1M ⫻ (XM) where T ⫽ days since bag removal, XL is a dummy variable that has the value 1 for low-dose animals, 0 otherwise, XM has the value 1 for mediumdose animals, 0 otherwise. An autoregressive error

LYSYK: MOVEMENT OF MALE D. andersoni

September 2013 Table 1.

Beginning and end dates, trial duration, time to bag removal, and average temperature during each trial

Year

Trial

Start

Finish

Duration (d)

Days in bag

n Observations

2007

1 2 3 1 2 3 1 2 3

April 20 May 18 June 15 April 18 May 17 June 13 April 17 May 16 June 12

May 5 June 2 June 30 May 10 May 31 June 28 May 5 May 30 June 27

15 15 15 22 14 15 18 14 15

4 4 4 11 4 4 6 4 4

4 4 4 5 5 5 5 5 5

2008 2009

979

structure was assumed for each animal pair. The parameters ␤0 and ␤1 are estimates of the intercept and slope for high-dose animals. The intercept and slope for low-dose animals were estimated as (␤0 ⫹ ␤0L) and (␤1 ⫹ ␤1L) and as (␤0 ⫹ ␤0M) and (␤1 ⫹ ␤1M) for medium-dose animals. The estimates of ␤0L and ␤0M are position parameters and indicate the reduction in the intercept or starting point for the low and medium doses respectively relative to the high dose. These were included in the decay models for donor males, donor females, and ⌬F, as dose varied among the donor animals, but were not included for recipient females, as the dose was constant for recipient animals. Models were tested for common intercepts and slopes. A reduced model was Þt if a common slope was justiÞed. The daily loss rate for males, females on donor animals, and recipient females was calculated as 1 ⫺ exp(␤1 ⫹ ␤1LXL ⫹ ␤1MXM) where variables and parameters were as deÞned earlier. Frequency of Movement. Initial analysis examined movement among the animal pairs and the number of days in which movement was detected. ␹2 tests were used to determine whether the number of animal pairs with movement and the number of sampling days when movement was detected varied among doses, trials, and years. Factors Affecting Movement. Both logistic regression and Poisson regression (Atkins and Gallop 2007) were used to quantify movement in relation to numbers of male ticks on the donor animals (DM) at the beginning of each sample interval, the female differential (⌬F), temperature during the sample interval, female age at the beginning of the sample interval, and all possible two-way interactions. Poisson regression was used to model the number of males that moved in each animal pair, Ni, during an interval. Logistic regression was used to model the probability of movement for an animal pair during an interval using the variable Yi that was assigned the value 0 if Ni ⫽ 0 and 1 if Ni ⱖ 1. The Logistic and Poisson models (Proc Genmod, SAS Institute 2010) were developed in a similar fashion using an autoregressive error structure for each animal pair to account for temporal autocorrelation and ln(sample interval length) as an offset variable to adjust for the sample interval length (Woodward 1999). Predictor variables were selected in a sequential fashion. All one-variable models were estimated and the variable with the greatest Wald

⬚C Mean ⫾ SD

Min.

Max.

9.6 ⫾ 5.8 11.8 ⫾ 5.3 16.7 ⫾ 5.7 4.6 ⫾ 8.3 14.0 ⫾ 6.3 16.6 ⫾ 5.3 6.0 ⫾ 7.3 12.8 ⫾ 6.4 17.0 ⫾ 5.2

0.0 1.5 6.8 ⫺10.7 3.0 5.7 ⫺8.3 0.2 8.1

24.4 25.8 29.6 21.6 31.5 28.2 24.8 25.0 27.6

statistic (parameter/SE) selected for inclusion. The remaining variables were then individually Þt to a model containing the Þrst selected variable, and the second variable with the greatest Wald statistic selected and retained. This was repeated until no additional variable had Wald statistic with P ⱕ 0.05. Results were interpreted graphically (Atkins and Gallop 2007). Results Changes in Predictor Variables. Ambient Temperature. Variation in ambient temperatures among the trials is summarized in Table 1. Average hourly temperatures ranged from 4.6 to 9.6⬚C, 11.8 Ð14.0⬚C, and 16.6 Ð17.0⬚C during the April, May, and June trials of each year, respectively. Tick attachment was delayed during the April trials of 2008 and 2009 because of the low temperatures; therefore, the bags were left in place for 11 and 6 d, respectively. Temperatures were sufÞciently warm during the remaining trials to ensure sufÞcient attachment, and bags were removed after 4 d. Overall, the proportion of unattached males when bags were removed ranged from 0 to 5.3%, and the proportion of unattached females ranged from 3.3 to 11.3% across trials. Female Aging. Initial examination of the cumulative proportion of engorged females indicated a systematic bias, with females from the April trials engorging several days later than females from the May and June trials (Fig. 1A). Differences among the trials were reduced when normalized time was used as the time scale, as this accounted for variation in temperature among trials (Fig. 1B); however, engorgement appeared to occur earlier than predicted, likely because of the original equations being based on ambient temperature alone, and not accounting for the warming effects of solar insolation. Satisfactory agreement between observed and predicted cumulative engorgement was obtained when normalized age was adjusted by dividing by 0.82 (Fig. 1C). Tick Numbers. Tick numbers on the animals were dynamic during the trials and were well described using the regression models (Fig. 2; Table 2). The parameter estimates for all response variables are shown in Table 2. The number of males per pair declined following bag removal (Fig. 2A) as indicated by the signiÞcant negative slope (␤1; Table 2). The rate

980

JOURNAL OF MEDICAL ENTOMOLOGY

Fig. 1. Cumulative proportion of total engorged ticks removed in relation to (A) days since the beginning of each trial, (B) unadjusted normalized age, and (C) adjusted normalized age (age/0.82). Circles ⫽ 2007, triangles ⫽ 2008, squares ⫽ 2009. White symbols ⫽ Trial one (April), gray symbols ⫽ Trial two (May), and black symbols ⫽ Trial three (June). Solid line is predicted value.

of decline was similar among pairs treated with different doses of ticks as indicated by the nonsigniÞcant dose ⫻ time interaction (F ⫽ 0.5; df ⫽ 2, 249; P ⫽ 0.60);

Vol. 50, no. 5

therefore, parameters ␤1L and ␤1M were not estimated (Table 2). The slope with respect to time was used to calculate an estimated loss rate of 0.24 per day (1 ⫺ exp(⫺␤1)). The number of females on the donor animals also declined after bag removal; however, there was a signiÞcant dose ⫻ time interaction (Table 2), indicting the slope with time was not consistent among doses. Females on the animals treated with low and medium doses had slopes 0.11 and 0.05 greater compared with high-dose animals (Fig. 2B). The slopes from Table 2 were used to calculate loss rates for each dose as 0.30, 0.27, and 0.22 for females on high-, medium-, and low-dose animals, respectively. The number of females on recipient animals declined at a lower rate (0.12) than females on the donor animals (Fig. 2C). The differences in loss rates of females on the donor and recipient and donor animals caused ⌬F to increase with time with signiÞcant dose ⫻ time interactions (Table 2). The variable ⌬F exceeded 1 between days 2Ð3 for the low-dose groups, and between days 8 Ð9 and days 6 Ð7 for the medium- and high-dose groups, respectively (Fig. 2D). The linear model Þt the data adequately (r2 ⫽ 0.34), although some curvilinearity with respect to time was apparent (Fig. 2D), especially for the medium and high doses. However, changing to a cubic model increased r2 slightly to 0.40, and did not appreciably affect the descriptive conclusions. Frequency of Movement. Movement occurred on 26 of 54 (48%) of the animal pairs. For most pairs, movement was detected on a single day (n ⫽ 19); however, Þve animal pairs had movement detected on two occasions, one pair had movement detected on three occasions, and one pair had movement detected on four occasions. Fifteen animal pairs had a total of one tick each that moved; seven pairs had two ticks that moved; and one pair each had three, four, Þve, and six ticks that moved. The percentage of pairs with movement was greater during 2009 (12 of 18 ⫽ 67%) compared with the combined percentage for 2007 (8 of 18 ⫽ 44%) and 2008 (6 of 18 ⫽ 33%) (␹2 ⫽ 3.7; df ⫽ 1; P ⫽ 0.05). The percentage of pairs with movement during the April trials was 67% (12 of 18) and was greater than the combined 44% (8 of 18) and 33% (6 of 18) pairs with movement during the May and June trials (␹2 ⫽ 3.7; df ⫽ 1; P ⫽ 0.05). The percentage of pairs with movement was similar among doses (␹2 ⫽ 1.9; df ⫽ 2; P ⫽ 0.38) and was 61% (11 of 18), 39% (7 of 18), and 50% (8 of 18) (44%) for pairs challenged with the high, medium, and low doses, respectively. Observations were made on 12, 15, and 15 sample days during each of 2007, 2008, and 2009, respectively, for a total of 42 different sampling days. Six animal pairs were examined each sample day. No movement was detected on any animal pair on 45% (19 of 42) of the sample days. Movement was detected on one, two, three, and four pairs of animals on 33% (14 of 42), 14% (6 of 42), 5% (2 of 42), and 2% (1 of 42) of sample days, respectively. Days with movement were more frequent for animals infested with the high dose (15 of

LYSYK: MOVEMENT OF MALE D. andersoni

September 2013

981

Fig. 2. Changes in (A) total number of males, (B) number of females per donor animal (DF), (C) number of females per recipient animal (RF), and (D) ⌬F ⫽ ln(RF⫹1) ⫺ ln(DF⫹1) in relation to days since the beginning of each trial. Circles ⫽ low-dose animals, triangles ⫽ medium-dose animals, and squares ⫽ high-dose animals. Short dash, dashÐdot, and solid lines are the predicted values for the low, medium, and high doses, respectively.

Factors Influencing Movement. Daily Probability of Movement. The proportion of animal pairs with tick movement averaged (SD) 0.143 (0.167) and ranged from 0.0 to 0.67 across sample days. After adjustment for the length of the sample interval, the daily probability of movement averaged (SD) 0.067 (0.082) and ranged from 0.0 to 0.33. Logistic regression indicated that the daily probability of movement was determined by interactions between DM and ⌬F, DM and

42) compared with the combined medium (7 of 42) and low (8 of 42) doses (␹2 ⫽ 4.8; df ⫽ 1; P ⬍ 0.05). The percentage of days with movement varied among years (␹2 ⫽ 6.6; df ⫽ 2; P ⬍ 0.04) and was 58% (7 of 12), 33% (5 of 15), and 73% (11 of 15) during 2007, 2008, and 2009, respectively. Days with movement were most frequent during the April trials (11 of 14 ⫽ 79%) compared with the May (8 of 14 ⫽ 57%) and June (4 of 14 ⫽ 29%) trials (␹2 ⫽ 7.0; df ⫽ 2; P ⬍ 0.05). Table 2.

␤i ␤0 ␤0L ␤0M ␤1 ␤1L ␤1M r2

Relationship between number of ticks on animals and time since bags were removed TM

DF

⌬F

RF

Estimate ⫾ SE

t

Estimate ⫾ SE

t

Estimate ⫾ SE

t

Estimate ⫾ SE

t

4.06 ⫾ 0.16 ⫺0.93 ⫾ 0.20 ⫺0.40 ⫾ 0.20 ⫺0.27 ⫾ 0.01 NS NS 0.64

25.4* ⫺4.7* ⫺2.0 ns ⫺20.0*

4.36 ⫾ 0.20 ⫺1.43 ⫾ 0.28 ⫺0.60 ⫾ 0.28 ⫺0.36 ⫾ 0.03 0.11 ⫾ 0.04 0.05 ⫾ 0.04 0.67

22.0* ⫺5.1* ⫺2.2* ⫺14.5* ⫺3.1* 1.3ns

2.60 ⫾ 0.13 NA NA ⫺0.13 ⫾ 0.01 NA NA 0.20

20.1*

⫺1.64 ⫾ 0.27 1.33 ⫾ 0.39 0.31 ⫾ 0.39 0.25 ⫾ 0.03 ⫺0.13 ⫾ 0.05 ⫺0.10 ⫾ 0.05 0.34

⫺6.0* 3.4* 0.8 ns 7.8* ⫺2.7* ⫺2.1*

⫺9.5*

Effect

df

F

df

F

df

F

df

F

D T D⫻T

2, 51 1, 251 NS

11.1* 401.3*

2, 51 1, 249 2, 249

13.2* 458.4* 5.0*

NA 1, 251 NA

90.8*

2, 51 1, 249 2, 249

6.5* 92.2* 4.1*

TM, total number of males (males on donor ⫹ males on recipient); RF, the number of females on the recipient cattle; DF, the number of females on the donor cattle; and ⌬F, ln(RF⫹1) ⫺ ln(DF⫹1). Model used was ln(y ⫹ 1) ⫽ ␤0 ⫹ ␤0L ⫻ (XL) ⫹ ␤0M ⫻ (XM) ⫹ ␤1 ⫻ T ⫹ ␤1L ⫻ (XL) ⫹ ␤1M ⫻ (XM) where T ⫽ days since bag removal, XL is a dummy variable that has the value 1 for low-dose animals, 0 otherwise, XM has the value 1 for medium-dose animals, 0 otherwise. NS, nonsigniÞcant. *SigniÞcantly different from 0 at P ⱕ 0.05.

982

JOURNAL OF MEDICAL ENTOMOLOGY

Vol. 50, no. 5

Table 3. Results of logistic and Poisson regression relating tick movement to environmental factors Logistic Estimate ⫾ SE

Poisson Z

Estimate ⫾ SE

Z

Intercept ⫺2.5831 ⫾ 0.5146 ⫺5.02 ⫺3.1848 ⫾ 0.4545 ⫺7.01 ⌬F NS 0.3827 ⫾ 0.1247 3.07 DM ⫻ ⌬F 0.1734 ⫾ 0.0578 3.00 NS DM ⫻ Age 0.5387 ⫾ 0.1965 2.74 0.7629 ⫾ 0.1555 4.91 Age ⫻ Temp. ⫺0.0774 ⫾ 0.0260 ⫺2.98 ⫺0.0823 ⫾ 0.0237 ⫺3.48 DM, ln(males on donor ⫹ 1); ⌬F, ln(RF⫹1) ⫺ ln(DF⫹1); RF, the number of females on the recipient cattle; DF, the number of females on the donor cattle. NS, nonsigniÞcant. All parameters signiÞcant at P ⱕ 0.05.

Age, and Age and temperature (Table 3). Interactions between two predictor variables indicate that the response to one variable is inßuenced by the level of the second variable. The probability of movement increased with DM, but the rate of increase with respect to DM was inßuenced by age and ⌬F. The increase with respect to DM was lower when the population of females was young and greater when the population of females was older (Fig. 3A). Similarly, the rate of increase with respect to DM was lower when ⌬F favored the donor animals (low ⌬F) compared with the recipient animals (high ⌬F) (Fig. 3B). The probability of movement tended to decrease as temperature increased, but this was also modulated by age of the female populations. Movement was more likely to occur if the female population was physiologically older than when younger (Fig. 3C). Numbers of Males Moving Per Day. Movement of a single male was detected on 28 occasions, by two males on six occasions, and by three and four males each on one occasion. In total, 47 of 2700 (1.7%) of the males placed on the donor animal were recovered from the recipient animals; however, this reßects the initial number of males placed on the animals, rather than the populations at the time movement was detected. These were quite dynamic as indicated earlier. After adjusting for the length of the sampling interval, the average (SD) number of males moving per day was 0.083 (0.228) with a range of 0.0 Ð1.50. The number of males that moved per day was related to ⌬F as a main effect and the relationship was straightforward, as ⌬F acted independently of other variables (Table 3). Each unit increase in ⌬F resulted in a 47% increase in the predicted rate of movement (Fig. 4A). The daily rate of movement increased with DM, but the relationship was inßuenced by female age (Fig. 4B), with the rate of movement greater in the presence of older females compared with younger females. The relationship between daily rate of movement and temperature depended on female age (Fig. 4C). In the presence of older females, the rate of movement was greater at low temperatures and decreased as temperature increased; however, the rate of movement was lower and relatively constant with respect to temperature in the presence of younger females.

Fig. 3. Predicted probability of movement based on the logistic regression model. (A) Probability of movement in relation to the donor male ⫻ age interaction solved at age ⫽ 0.42 (solid line), 0.72 (long dash line), 1.03 (medium dash line), and 1.34 (short dash line). (B) Movement in relation to the donor male and female differential interaction solved at ⌬F ⫽ ⫺2.38 (solid line), ⫺0.99 (long dash line), 0.39 (medium dash line), and 1.78 (short dash line). (C) Probability of movement in relation to the temperature ⫻ age interaction solved at age ⫽ 0.42 (solid line), 0.72 (long dash line), 1.03 (medium dash line), and 1.34 (short dash line).

Discussion Loss of males and females included losses because of mortality and undetected emigration while losses of females also included drop of engorged females. The daily loss rate for females on the recipient animals may have been lower than the loss rate for the low dose donor animals, as the recipient animals did not have males added and would not have had losses because of engorgement and drop. The loss rate for females

September 2013

LYSYK: MOVEMENT OF MALE D. andersoni

Fig. 4. Predicted number of males moving per day based on the Poisson regression model. (A) Number of males moving per day in relation to the female differential. (B) Number of males moving per day in relation to the donor male ⫻ age interaction solved at age ⫽ 0.42 (solid line), 0.72 (long dash line), 1.03 (medium dash line), and 1.34 (short dash line). (C) Number of males moving per day in relation to the temperature ⫻ age interaction solved at age ⫽ 0.42 (solid line), 0.72 (long dash line), 1.03 (medium dash line), and 1.34 (short dash line).

placed on the donor animals likely increased with dose because the greater tick numbers increased host irritation resulting in great self-licking and grooming by the cattle. Typically, loss of D. andersoni on cattle prevented from grooming is usually ⬍0.02 per day and only occasionally reaches 0.07 per day (Lysyk 2008, Lysyk et al. 2009a,b) considerably lower than observed here. However, grooming mortality can be substantial. Hereford calves separated from their mothers lost 19 Ð 69% of D. andersoni over a 5Ð 6-d interval,

983

compared with loss of 92Ð96% by calves not separated from their mothers (Rich 1973). The differences were attributed to grooming by the mother. Loss of Rhipicephalus (Boophilus) microplus (Canestrini) was greatly reduced in cattle prevented from self-grooming (Snowball 1956). Grooming was also believed to be the major cause of resistance of cattle to Amblyomma hebraeum Koch (Norval et al. 1988). Despite Despite these relatively large loss rates, movement occurred in nearly 50% of the animal pairs. This alone suggests that there is considerable potential for transmission of anaplasmosis in western Canada, as a single male tick is capable of transmitting the pathogen (Rozeboom et al. 1940, Scoles et al. 2008). There are relatively few studies to compare results with, as methodology and measure of movement vary among studies. However, studies indicate that inter-host transfer occurs. In total, 71 R. (B.) microplus were recovered from naõ¨ve recipient cattle that had been housed with donor cattle previously infested with 30,000 larvae (Mason and Norval 1981). Although the numbers moving represented a small (⬍0.24%) percentage of the initial population, ticks were recovered from three of seven (42%) naõ¨ve recipient cattle. Stiller et al. (1989) infested each of three donor calves with 450 male D. andersoni, and found that males moved to each of Þve recipients. Approximately 4.9% of the ticks were recovered, and 53% of the males recovered had transferred hosts. This suggests that up to 2.6% of the released ticks transferred hosts, only slightly more than the 1.7% in the current study. Rhipicephalus sanguineus moves even more readily among canine hosts. In two trials, four fourths and three fourths dogs acquired ticks from a different host under group housing (Little et al. 2007). Recovery rates after 5Ð7 d were high, ranging from 19 to 34% of the ticks initially placed on the hosts. From 19 Ð35% of recovered ticks had transferred hosts. Both the percentage of animals receiving immigrant ticks and the percentage recovery were greater than in the current study. This may reßect differences in tick and host biology as well as in methodology. The current study also measured movement in terms of several additional outcomes that were intended to provide Þner levels of detail. The Þrst measure was the daily probability of movement of male ticks from a donor to recipient animal, and is useful for assessing and modeling transmission of anaplasmosis. The number of males that moved per day was derived to provide more detailed information if modeling efforts required. Both measures of movement were inßuenced by the interaction between female age and temperature. Movement was greatest when age was highest and temperature lowest. This likely occurred because as females age, they release a pheromone that causes males to detach and mate (Sonenshine et al. 1974), resulting in males becoming available to move and possibly change hosts. The negative relationship with temperature may at Þrst seem counterintuitive as arthropod activity typically increases with temperature; however, cattle may huddle together during periods of cold weather (Hahn 1985). This may have

984

JOURNAL OF MEDICAL ENTOMOLOGY

increased contact between donor and recipient cattle, resulting in greater opportunity for ticks to change hosts. The effects of ⌬F, the female differential, are also probably related to pheromone production and mate availability. Once the majority of females on donor animals were replete, males may have responded and moved to the greater number of females on the recipient animals. This agrees with observations on R. sanguineus where dogs that were infested mainly with females received a greater number of immigrant ticks compared with dogs infested with only males (Little et al. 2007). Abundance of males on the donor animals (DM) inßuenced the daily probability of movement and the number of migrating males per day, but the effect was mainly through interactions with age in both cases, and also with ⌬F in the latter case. Movement tended to increase with DM, but the effect was accelerated in the presence of older females on the donor animals and also in the presence of a greater female differential, likely for reasons already discussed. Overall, these results suggest there are limited barriers to the movement of D. andersoni among cattle under western Canadian conditions. Anaplasmosis, if introduced into areas of Canada where D. andersoni is present, would likely experience natural transmission as recently occurred in areas of Canada where D. variabilis is abundant. Results further suggest that movement is affected by a complex suite of interacting variables. Preventing movement may require more than one approach such as controlling numbers on potential donor animals. It may also require efforts to make recipient animals less attractive by manipulating the tick population on these. Results presented in this study will allow incorporation of movement into tick population models to explore these possibilities. Acknowledgments I am grateful to C. HimslÐRayner, R. C. Lancaster, M. K. Toohey, D. Schmaltz, K. Smith, and D. Pittman, Lethbridge Research Centre (LRC), for expert technical assistance and animal care. This work was supported by Beef Cattle Research Council Grant “Seasonal activity of wood ticks, Dermacentor andersoni, a vector of bovine anaplasmosis.” Cattle experiments were conducted accordance with Canadian Council of Animal Care Guidelines and protocols 0720, 0803, and 0905 “Movement of male D. andersoni among cattle” approved by the LRC Animal Care Committee. Ticks were reared under protocols 0708, 0804, and 0908 “Rearing ticks on small mammal hosts for colony maintenance and experimental purposes.”

References Cited Atkins, D. C., and R. J. Gallop. 2007. Rethinking how family researchers model infrequent outcomes: a tutorial on count regression and zero-inßated models. J. Fam. Psychol. 21: 726Ð735. Bremer, W. G., J. J. Schaefer, E. R. Wagner, S. A. Ewing, Y. Rikihisa, G. R. Needham, S. Jittapalapong, D. L. Moore, and R. W. Stich. 2005. Transstadial and intrastadial experimental transmission of Ehrlichia canis by male Rhipicephalus sanguineus. Vet. Parasitol. 131: 95Ð105.

Vol. 50, no. 5

Hahn, G.L.R. 1985. Weather and climate impacts on beef cattle. pp. 85Ð 89. In Beef Research Report 2, USDA-ARS Series 42, USDA-ARS, Ames, IA. Howden, K. J., D. W. Geale, J. Pare´, E. J. Golsteyn–Thomas, and A. A. Gajadhar. 2010. An update on bovine anaplasmosis (Anaplasma marginale) in Canada. Can. Vet. J. 51: 837Ð 840. Kocan, K. M., W. L. Goff, D. Stiller, P. L. Claypool, W. Edwards, S. A. Ewing, J. A. Hair, and S. J. Barron. 1992a. Persistence of Anaplasma marginale (Rickettsiales: Anaplasmataceae) in male Dermacentor andersoni (Acari: Ixodidae) transferred successively from infected to susceptible calves. J. Med. Entomol. 29: 657Ð 668. Kocan, K. M., D. Stiller, W. L. Goff, P. L. Claypool, W. Edwards, S. A. Ewing, T. C. McGuire, J. A. Hair, and S. J. Barron. 1992b. Development of Anaplasma marginale in male Dermacentor andersoni transferred from parasitemic to susceptible cattle. Am. J. Vet. Res. 53: 499 Ð507. Little, S. E., J. Hostetler, and K. M. Kocan. 2007. Movement of Rhipicephalus sanguineus adults between co-housed dogs during active feeding. Vet. Parasitol. 150: 139 Ð145. Lysyk, T. J. 2008. Effects of ambient temperature and cattle skin temperature on engorgement of Dermacentor andersoni. J. Med. Entomol. 45: 1000 Ð1006. Lysyk, T. J., D. M. Veira, J. P. Kastelic, and W. Majak. 2009a. Inducing active and passive immunity in sheep to paralysis caused by Dermacentor andersoni. J. Med. Entomol. 46: 1436 Ð1441. Lysyk, T. J., D. M. Veira, and W. Majak. 2009b. Cattle can develop immunity to paralysis caused by Dermacentor andersoni. J. Med. Entomol. 46: 358 Ð366. Mason, C. A., and R.A.I. Norval. 1981. The transfer of Boophilus microplus (Acarina: Ixodidae) from infested to uninfested cattle under Þeld conditions. Vet. Parasitol. 8: 185Ð188. Norval, R.A.I., R. B. Floyd, and J. D. Kerr. 1988. Ability of adults of Amblyomma hebraeum (Acaria: Ixodidae) to feed repeatedly on sheep and cattle. Vet. Parasitol. 29: 351Ð355. Peterson, K. J., R. J. Raleigh, R. K. Stroud, and R. L. Goulding. 1977. Bovine anaplasmosis transmission studies conducted under controlled natural exposure in a Dermacentor andersoni ⫽ (venustus) indigenous area of eastern Oregon. Am. J. Vet. Res. 38: 351Ð354. Rich, G. B. 1973. Grooming and yarding of spring-born calves prevent paralysis caused by the Rocky Mountain wood tick. Can. J. Anim. Sci. 53: 377Ð378. Rozeboom, L. E., G. W. Stiles, and L. H. Moe. 1940. Anaplasmosis transmission by Dermacentor andersoni Stiles. J. Parasitol. 26: 95Ð100. Sanborn, C. E., G. W. Stiles, and L. H. Moe. 1938. Anaplasmosis transmission by naturally infected Dermacentor andersoni male and female ticks. North Am. Vet. 91: 31Ð33. SAS Institute. 2010. SAS OnlineDoc 9.3. SAS Institute, Cary, NC. Scholfield, L. N., and J. R. Saunders. 1987. Attempted transmission to cattle of Anaplasma marginale from overwintered Dermacentor andersoni ticks. Can. J. Vet. Res. 51: 379 Ð382. Scoles, A. G., A. B. Broce, T. J. Lysyk, and G. H. Palmer. 2005. Relative efÞciency of biological transmission of Anaplasma marginale (Rickettsiales: Anaplasmatacae) by Dermacentor andersoni (Acari: Ixodidae) compared with mechanical transmission by Stomoxys calcitrans (Diptera; Muscidae). J. Med. Entomol. 42: 668 Ð 675. Scoles, G. A., T. F. McElwan, F. R. Rurangirwa, D. P. Knowles, and T. J. Lysyk. 2006. A Canadian bison isolate of Anaplasma marginale (Rickettsiales: Anaplasmatace-

September 2013

LYSYK: MOVEMENT OF MALE D. andersoni

cae) is not transmissible by Dermacentor andersoni (Acari: Ixodidae), whereas ticks from two Canadian D. andersoni populations are competent vectors of a U.S. strain. J. Med. Entomol. 43: 971Ð975. Scoles, G. A., J. A. Miller, and L. D. Foil. 2008. Comparison of the efÞciency of biological transmission of Anaplasma marginale (Rickettsiales: Anaplasmataceae) by Dermacentor andersoni Stiles (Acari: Ixodidae) with mechanical transmission by the horse ßy, Tabanus fuscicostatus Hine (Diptera: Muscidae). J. Med. Entomol. 45: 109 Ð114. Snowball, G. J. 1956. The effect of self-licking by cattle on infestations of cattle tick, Boophilus microplus (Canestrini). Crop Pasture Sci. 7: 227Ð232. Sonenshine, D. E., R. M. Silverstein, E. C. Layton, and P. J. Homsher. 1974. Evidence for the existence of a sex pheromone in 2 species of Ixodid ticks (Metastigmata: Ixodidae). J. Med. Entomol. 11: 307Ð315. Stiller, D., and M. E. Coan. 1995. Recent developments in elucidating tick vector relationships for anaplasmosis and equine piroplasmosis. Vet. Parasitol. 57: 97Ð108. Stiller, D., M. E. Coan, W. L. Goff, L. W. Johnson, and T. C. McGuire. 1989. The importance and putative role of Dermacentor spp. males in anaplasmosis epidemiology: transmission of Anaplasma marginale to cattle by ad libitum interhost transfer of D. andersoni makes under

985

semi-natural conditions. In Proceedings of the 8th National Veterinary Hemoparasite Disease Conference, St. Louis, MO. Ueti, M. W., G. H. Palmer, G. A. Scoles, L. S. Kappmeyer, and D. P. Knowles. 2008. Persistently infected horses are reservoirs for intrastadial tick-borne transmission of the apicomplexan parasite Babesia equi. Infect. Immunol. 76: 3525Ð3529. Wilkinson, P. R., and J. E. Lawson. 1965. Difference of sites of attachment of Dermacentor andersoni Stiles to cattle in southeastern Alberta and in south central British Columbia, in relation to possible existence of genetically different strains of ticks. Can. J. Zool. 43: 408 Ð 411. Woodward, M. 1999. Epidemiology. Study design and data analysis. Chapman & Hall/CRC, Boca Raton, FL. Zaugg, J. L. 1990. Seasonality of natural transmission of bovine anaplasmosis under desert mountain range conditions. J. Am. Vet. Med. Assoc. 196: 1106 Ð1109. Zaugg, J. L., D. Stiller, M. E. Coan, and S. D. Lincoln. 1986. Transmission of Anaplasma marginale Theiler by males of Dermacentor andersoni Stiles fed on an Idaho Þeld-collected, chronic carrier cow. Am. J. Vet. Res. 47: 2269 Ð 2271. Received 11 January 2013; accepted 14 June 2013.