RESUMO
Hematology is a common tool for wildlife health assessments. Manual leukocyte counts are required in reptiles, however, disagreement between quantification methods has been observed in some chelonians. This study determined agreement between two methods of leukocyte quantification, eosinophilic leukopet, and blood film white blood cell (WBC) estimates, in free-living eastern box turtles (Terrapene carolina carolina). Simultaneously, we assessed the impact of handling duration on both leukocyte quantity and corticosterone levels. We collected blood at capture (<2 min from disturbance) and again before release 30-150 min later from 92 box turtles at six sites in Illinois and Tennessee. Constant and proportional error was present in the leukopet results for WBC, lymphocytes, and basophils compared to the estimate method. Both methods were in agreement for heterophils, monocytes, and eosinophils. Agreement between the methods was significantly more likely at WBC counts below 23,241/µl. All hematologic parameters were significantly higher at the final blood draw compared to the initial blood draw using both WBC determination methods, except relative eosinophil and basophil counts. Corticosterone levels varied with time, with maximum concentrations reached at 54 min postcapture, followed by a rapid return to baseline levels. Corticosterone level was not significantly associated with any hematologic parameter or sex. This study provides a framework for understanding the effects of animal handling methodology and diagnostic modality when evaluating hematologic health in box turtles.