A conceptual model of the Amblyomma americanum life cycle in northeast Missouri

Authors


ABSTRACT:

The geographic distribution of Amblyomma americanum (the lone star tick) has increased as has its role as a pathogen vector. The objectives of this study were to determine seasonal activity patterns of each life stage of A. americanum in the northwestern part of the species range and the relationship of these activity patterns among life stages and degree days. Tick activity was monitored over four years since 2007 in a forest and old field habitat located in northeast Missouri. Every other week from February to December, ticks were collected using bait and drag methods. Autocorrelations demonstrated yearly seasonal patterns in each life stage and cross-correlations between life stages depicted a relationship between activity at a life stage and the previous stage's activity. Cross-correlations indicated that degree days were related to activity. These data indicated that A. americanum generally complete their life cycle in a minimum of two years in northeast Missouri, with overwintering occurring predominantly in the nymphal and adult stages. These data provide a baseline to compare the life cycle of A. americanum in northeast Missouri to populations in different parts of the species range or at different times in the region.

INTRODUCTION

Amblyomma americanum (the lone star tick) is a three-host tick that feeds on a variety of domestic and wild animals (Bishopp and Trembley 1945, Sonenshine and Stout 1971, Kollars et al. 2000, Childs and Paddock 2003) and has become an increasingly important vector of pathogens (reviewed by Childs and Paddock 2003). Its range is primarily within the southeastern United States. However, the current range of this species suggests that it has expanded north into Maine (Keirans and Lacombe 1998) and northwest into Nebraska (Diuk-Wasser et al. 2006) since the distribution was first described (Bishopp and Trembley 1945). An important determinant of the geographic range of a tick species is likely climate (Cumming 2002). As three-host ticks spend a majority of their life off-host, they are influenced by climate for development, activity, and survival (reviewed by Randolph 2004).

Developmental rates in A. americanum are dependent on temperature (Koch 1984, Barnard et al. 1985) and day length (Barnard et al. 1985, Pound and George 1988). In field studies, the development time of all life stages was negatively correlated with average ambient temperature (Koch 1984). Laboratory studies of nymphs and adult females demonstrated that temperature and day length conditions under which ticks were raised determined feeding time (Barnard et al. 1985) and day length was important in molting time of fed nymphs (Pound and George 1988). Photoperiod has been used to induce developmental diapause in nymphs (Pound and George 1988). Longer days and higher temperatures increased oviposition and the proportion of eggs eclosed in experimental laboratory studies (Barnard et al. 1985). Experimental field studies showed that pre-oviposition time and egg incubation time were temperature dependent (Patrick and Hair 1979), molting time of fed larvae was shorter at higher temperatures (Robertson et al. 1975b), and cooler temperatures retarded or prevented molting of fed nymphs (Semtner et al. 1973).

Host-seeking activity has been observed to be seasonal in studies conducted in southeastern Oklahoma (Semtner and Hair 1973b), Mississippi (Jackson et al. 1996), and southern Missouri (Kollars et al. 2000). Temperature (Semtner and Hair 1973b, Robertson et al. 1975a) and humidity (Robertson et al. 1975a, Jackson et al. 1996) influenced seasonal activity. In field monitoring studies in Mississippi (Jackson et al. 1996) and southeastern Oklahoma (Semtner and Hair 1973b), A. americanum adult and nymph activity was shown to begin in early spring and continue until mid- to late- summer, whereas larval activity commences later in the summer and ceases in late fall. Experimental field studies found that high temperatures and low humidity reduces unfed adult tick activity (Robertson et al. 1975a). Laboratory studies indicated that the lower threshold for A. amblyomma seeking hosts in a coordinated fashion ranged from 4.9 to 18.6° C and the temperature in which all activity stopped ranged from 3.3 to 15.3° C (Clark 1995). Behavioral diapause was not observed in overwintering adult A. amblyomma (Stewart et al. 1998).

The effect of climate on development and activity likely has a direct impact on the length of the life cycle of A. americanum. Under standardized laboratory conditions (24° C, ≥90% humidity and 16:8 light: dark cycle), A. americanum could be forced through their life cycle in < 22 weeks if ticks were able to attach and feed almost immediately after ecdysis but most ticks required 28–30 weeks (Troughton and Levin 2007). In field studies, first-year adults were generally not active (Semtner et al. 1973a). Under field conditions, the life cycle of A. americanum extends over multiple years (Hair and Howell 1970). As seasonal patterns of activity for all life stages vary geographically (Smith and King 1950, Hair and Howell 1970, Kollars et al. 2000), life cycle may also vary.

This study used a multi-year data set to statistically determine if there was a temporal pattern of activity within each life stage of A. americanum in two different habitats in northeast Missouri. In addition, we examined if these patterns within a life stage were associated with activity patterns of other life stages and degree days, a measure that depicts the relationship between development and temperature. These data were used to develop a conceptual model of the A. americanum life cycle in northeast Missouri for comparison with populations in other parts of the range of this tick.

MATERIALS AND METHODS

Two permanent 120×70 m sampling grids were established in Kirksville, Adair County, MO (40° 10′ 31″ N, 92° 36′ 10″ W). One grid was in a secondary oak-hickory forest and the other was in an old field, dominated by non-native grasses with a few widely spaced trees and shrubs. In each grid, ticks were collected on eight regularly spaced 30 m north-south drag samples and eight regularly spaced dry ice-baited traps. For drag sampling, 1 × 1 m flannel cloths, cut into ten equal width strips to ease movement through vegetation, were pinned to a wooden dowel. For dry ice bait sampling, approximately 200 g of dry ice was placed in the middle of a 1×1 m flannel cloth laid on the ground and let to sublime for one h. Cloths were then sealed in individual plastic bags and transported back to the lab. Ticks were recovered from each cloth, preserved in 95% ethanol, and identified to species and life stage. Both drag and bait sampling were used as each method targets separate classes of active ticks (Petry et al. 2010). A mark-recapture study conducted in 2011 (York and Foré, unpublished), following marking procedures of Kramer et al. (1993), did not find any significant decreasing trend in nor recapture of nymphs or adults, suggesting that removal sampling may have little effect on total population size in our study site.

Each grid was sampled simultaneously approximately every other week from 2007–2010 except when snow or ice covered the ground; usually beginning late February or early March and ending the beginning of December. Sampling occurred between 10:00 and 15:00 and was not conducted in the rain, but adjusted + or – a couple of days. The first sample in this study was collected on February 22, 2007 and the last sample was collected on December 2, 2010.

Cumulative average degree days for a thirty-day period prior to each sampling date were calculated using equation 1, where Tmaxi and Tmini are the maximum and minimum daily temperatures for a specified day (i) prior to the sampling date and Tbase is the base temperature (0° C).

(1)

image

Maximum and minimum temperature data for Kirksville, MO, were accessed from the National Oceanic and Atmospheric Administration National Climatic Data Center (cdo.ncdc.noaa.gov). In this study, a base ambient temperature of 0° C was used. This temperature was used in models for Ixodes scapularis (Ogden et al. 2005) and is below levels of activity for A. amblyomma in laboratory studies (Clark 1995). If average daily temperature was below 0° C, then the value for that day was set at zero (Ogden et al. 2005).

Time series analyses (Shumway 1988) were used to examine patterns of tick activity in each the forest and field sampling sites. Our questions were: 1) are seasonal patterns correlated over years?; 2) are patterns of activity of one life stage correlated with another life stage?; and 3) are life stage patterns correlated with degree days? Data for each sampling date were summed from both drag and bait sampling within each sampling site. Values of zero were inserted for 26 missing sampling dates in December, January, and February of the 105 sampling dates used. Tick activity was assumed to be zero for these dates as threshold activity temperatures for nymphs and adults were 9.6° C and 7.3° C, respectively (Clark 1995), and observed average temperatures were below those thresholds. Additionally, we observed no ticks in samples collected in February and December (Figure 1). Dates for missing sampling periods in December, January, and February were selected using an average of two-week intervals from the last known sampling date, which was the normal interval between sampling dates throughout the study.

Figure 1.

Time series of the total number of observed A. americanum ticks at A) adult, B) nymphal, and C) larval stages collected in each sampling session using bait and drag methods in forest (♦) and field (○) study sites in Kirksville, MO, over a four-year period (February, 2007–December, 2010).

RESULTS

In northeast Missouri, A. americanum activity followed a yearly pattern with seasonal fluctuations in total tick collections (Figure 1). The forest site experienced more fluctuation in the number of ticks collected than the field site, but the seasonal pattern of activity for both sites was similar over the years. In both study sites, adults and nymphs were observed to begin host-seeking activities in mid-March that persisted until the end of July and mid-August, respectively (Figure 1). While larvae appeared as early as May at the field site in 2007, their activity was typically observed to begin in late July and continue until November at both sites (Figure 1). Auto-correlations of tick activity within a life stage among years in forest and field sites were similar, but only forest data is presented (Figure 2, field site data available upon request). Yearly activity patterns of adults and nymphs were similar with the total number of active ticks at a given sampling time positively associated with the number of active ticks at lags of one to four (2–8 weeks) and 23–26 (46–52 weeks) sampling periods in the future (Figures 2a, 2b). Adults had an additional positive association at a lag of 22 (44 weeks) in the future. Larval activity was positively correlated at a lag of one, two, and 26 sampling periods (two to four, and 52 weeks) in the future (Figure 2c).

Figure 2.

Auto-correlograms of observed A. americanum activity at A) adult, B) nymphal, and C) larval life stages in the forest site in Kirksville, MO. Bars represent the auto-correlation function (ACF) coefficient at a given time lag (two-week time step), and those that lie outside the solid lines are statistically significant (p < 0.05). Auto-correlograms of observed activity in the field site show similar results.

Cross-correlations of activity patterns between life stages in the forest site showed that adult activity had a significant positive correlation to larval activity at lags of five to ten sampling periods (10–20 weeks) in the future (Figure 3a). This implies that the present adult activity is positively correlated with the larval activity 10–20 weeks in the future. Larval activity had a significant positive correlation to nymphal activity at lags of 17–23 sampling periods (34–46 weeks) in the future (Figure 3b). Finally, nymphal activity had a significant positive correlation to adult activity at lags of zero to two (0–4 weeks) and 22–26 sampling periods (44–52 weeks) in the future (Figure 3c), implying that the present nymphal activity is positively correlated with the adult activity in zero to four and 22–26 weeks in the future.

Figure 3.

Cross-correlograms between the observed activity of a given life stage of A. americanum and its subsequent life stage at the forest study site in Kirksville, MO. Cross-correlation function (CCF) coefficients represented by bars at positive time lags (two-week time step) demonstrate the relationships of A) adult activity to larval activity, B) larval activity to nymphal activity, and C) nymphal activity to adult activity. Bars that extend beyond the black lines are statistically significant (p < 0.05). Cross-correlograms of activity between life stages at the field study site showed similar results.

Degree days had significant positive correlations to adult activity in the forest site at lags of 16–24 sampling periods (32–48 weeks) in the future (Figure 4a), implying that the current degree days is positively correlated with the adult activity in 32–48 weeks in the future. There were significant positive correlations between degree days and nymphal activity at the time of sampling at lags of zero (0 weeks) and 17–25 sampling periods (34–50 weeks) in the future (Figure 4b), implying that the current degree days is positively correlated with nymphal activity currently and 34–50 weeks in the future. Degree days had significant positive correlations to larval activity at lags of zero to six sampling periods (0–12 weeks) in the future (Figure 4c), implying that the current degree days is positively correlated with the larval activity in 0–12 weeks in the future.

Figure 4.

Cross-correlograms comparing degree days accumulated 30 days prior to sampling with observed activity of a given life stage of A. americanum at the forest study site in Kirksville, MO. Cross-correlation function (CCF) coefficients represented by bars at positive time lags (two-week time step) demonstrate the relationships of A) degree days to adult activity, B) degree days to nymphal activity, and C) degree days to larval activity. Bars that extend beyond the black lines are statistically significant (p < 0.05). Cross-correlograms comparing degree days to observed activity at the field study site showed similar results.

DISCUSSION

Seasonal patterns of activity among the life stages of this northern population of A. americanum were similar to southerly populations in central Mississippi (Jackson et al. 1996), southeastern Oklahoma (Hair and Howell 1970), and southeastern Missouri (Kollars et al. 2000). Adult activity was reported to begin in February in Mississippi (Jackson et al. 1996); March in our study, in southeastern Missouri (Kollars et al. 2000) and in Oklahoma (Semtner and Hair 1973b), and April in Oklahoma (Hair and Howell 1970) and end of August in all studies. Nymphal activity was observed from April to September in our study, in Oklahoma (Hair and Howell 1970, Semtner and Hair 1973b) and in southeast Missouri (Kollars et al. 2000), but was longer in Mississippi encompassing May to October (Jackson et al. 1996). Larval activity in our study began in July as did populations in Mississippi (Jackson et al. 1996) and southeastern Missouri (Kollars et al. 2000), but began earlier in Oklahoma (Hair and Howell 1970). The cessation of larval activity began as early as October in southeast Missouri (Kollars et al. 2000) and Mississippi (Jackson et al. 1996) and as late as November in our study. Hair and Howell (1970) did not report sampling to cessation of larvae.

Our four-year data set suggests that the life cycle of A. americanum in northeast Missouri generally takes a minimum of two years to complete. The cross-correlation between the number of active adults and the number of active larvae 10–22 weeks in the future (Figure 5a) suggests that adults emerge in the spring, mate, and lay eggs that hatch in the same year (year 0). The lag between adult and larval activity derived from field-collected data is slightly longer than the eight weeks required for A. americanum larvae to hatch in a laboratory under standardized conditions (22–24° C, 90% RH, 16:8 L:D) (Troughton and Levin 2007). Experimental field studies in Oklahoma (Koch 1984) found that larvae could overwinter in some years, but survival was partly dependent on the date of female feeding. Preliminary experimental studies in our study site suggest that larvae may have very low winter survival (Hauser and Foré, unpublished). The emergence of unfed nymphs in the spring concurrent with adults and the cross-correlation with number of active larvae with number of active nymphs 36–46 weeks in the future suggests that year 0 larvae feed before winter and emerge in year 1 as unfed nymphs. Cross-correlations with unfed nymphs with unfed adults 44 to 52 weeks in the future suggest that unfed adults emerge in year 2 to mate. These findings are consistent with the two to three year life cycle suggested for an Oklahoma population (Hair and Howell 1970). Koch (1984) found that unfed adults and nymphs had 91% and 59% survival, respectively, over the first winter, but unfed nymphs were much less likely than adults to survive a second winter in Oklahoma field experiments.

Figure 5.

The conceptual model of the A. americanum life cycle depicts the relationships between observed activity and A) significant positive cross-correlations of subsequent activities, as well as B) significant positive cross-correlations of degree days to life stage activity at the forest study site. The average observed duration of activity for each life stage (white bars) indicates the time of year each life stage is observed. The significant positive correlations of activities between life stages (A, black bars) suggest a temporal relationship (in weeks) between a life stage and the subsequent life stage (hashed line). The significant positive correlation of degree days with activity (B, black bars) suggest a temporal relationship (in weeks) between activity and degree days (solid line).

The observed lags between life stages is partly dependent on developmental time. The relationships between accumulated temperature prior to sampling and the number of active ticks suggest that temperature may be important for development (Figure 5b). Our field data suggest that degree days in the previous year is positively associated with the number of ticks seen at the time of sampling for both unfed nymphs and adults and is likely a result of temperature effects on development rates from the previous life stage. Molting time of engorged larvae (Robertson et al. 1975b) and nymphs (Semtner et al. 1973, Robertson et al. 1975b) is influenced by the month of release into the field, suggesting increasing temperature decreases development rate. If nymphs molted after June, they do not become active adults in the same season (Robertson et al. 1975b) and, of those that molt earlier in the season, few are unlikely to become part of the active adult population (Semtner et al. 1973).

The focus of this model was seasonality and the potential influence of temperature accumulated over time and, therefore, does not capture the variation in number of individuals among years. This conceptual model of the A. americanum life cycle in northeast Missouri, based on four years of monitoring, serves as a foundation for comparison with other populations. One working hypothesis is that changes in climate are likely to affect the length of the life cycle in populations. Shorter winters may allow for earlier emergence and faster development of larva such that more of the larval population acquires blood meals prior to winter. Shortening of the life cycle is likely to influence disease transmission rates by changing the demographic structure of the tick population.

Ancillary