In Mercer and Crawford counties where our study was conducted, 224,760 baits were distributed (Fig
In Mercer and Crawford counties where our study was conducted, 224,760 baits were distributed (Fig.?1). being adults from 25-75% as unknown. Since the impact of age effect might be influenced based on our selection of a cutoff value for age classification, we looked at the sensitivity of the cutoff on the age effect in the model selection. None of the age classifications we examined were competitive models for the difference in RVNA seroprevalence, unless we used the original age classifications recorded at the time of capture. Table 1 Beta estimates, standard errors (Std error) and test statistics for the covariates that relate to the probability (Pr) of a raccoon ( em Procyon lotor BY27 /em ) being an adult based on 1,584 raccoons sampled in Virginia from 2014C2016. thead th rowspan=”1″ colspan=”1″ Co-variates /th th rowspan=”1″ colspan=”1″ Estimate /th th rowspan=”1″ colspan=”1″ Std error /th th rowspan=”1″ colspan=”1″ Z value /th th rowspan=”1″ colspan=”1″ Pr( |Z|) /th /thead Intercept?1.65930.3808?4.360.0000Weight1.25810.078815.970.0000Sex1.25870.13929.040.0000Day of year?0.01930.0015?13.140.0000 Open in a separate window Open in a separate window Fig.?2 Histograms of raccoons ( em Procyon lotor /em ) sampled in Pennsylvania and assigned a probability of being an adult based on sex, weight, and time of year captured. Raccoons classified in the field as adults (blue bars; n = 191) are shown, along with raccoons classified in the field as juvenile or unknown (yellow bars; n = 73). The light green is where the two histograms overlap. In the analysis, raccoons with a less than 25% probability of being an adult were classified as juveniles, raccoons with a greater than 75% probability of being an adult were classified as adults, and raccoons with a 25C75% probability of being adult were categorized as unknown. Although we had few raccoon recaptures between trapping GATA3 periods, we wanted to estimate raccoon densities to calculate the number of baits available for consumption. We used closed population capture-recapture models implemented in Program MARK which assumed a population closed to births, deaths, emigration, and immigration to estimate raccoon densities by zone [29, 30]. A behavioral effect on capture rates (e.g., trap happiness or trap shyness) was included by estimating initial capture rate and recapture rate separately and allowing for population size to vary by BY27 study zone. Abundance was converted to density by dividing by the study area (7.8 km2), and variance was converted using the Delta method [31]. A simple linear model was used to examine the difference in RVNA seroprevalence and tetracycline deposition in the teeth between pre-bait and post-bait trapping. We evaluated the impacts of average difference in age, sex, and weight on antibody response. Additionally, we examined several hypotheses about the relationship of bait density and change in RVNA occurrence including a positive BY27 linear relationship. One hypothesis was that there would be a difference between 75 baits/km2 and 150?baits/km2, but little difference between 150 baits/km2 and 300 baits/km2 (Den75). Another hypothesis was that there would be little difference between 75?and 150 baits/km2, but there would be a difference between 150 and 300 baits/km2 (Den300). We also considered that there would be a difference between all levels of bait densities but the difference between 75 and 150 baits/km2 would not be the same as that from 150 to 300 baits/km2 (Denfac). Model selection was performed using Akaike’s Information Criterion (AICc) and correcting for small sample sizes [32]. There was little support for any of the additive models examined (all model weights for additive models were 0.05); therefore, we present the results for the top ten models only (all with a model.