The standard method for assessing blood cell characteristics using an ocular micrometer is time-consuming and limited. We used the Nikon NIS Elements imaging software and May-Grünwald-Giemsa staining to determine whether automated image analysis is suitable for rapid and accurate quantitative morphometry of erythrocytes. Blood was collected during four seasons from 126 geometric tortoises and the blood smears were evaluated for cell (C) and nuclear (N) characteristics of the erythrocytes. We measured area, length (L), width (W), perimeter, elongation and pixelation intensity, and calculated L/W and N/C areas. Erythrocyte size differed among cohorts; females, the larger sex, had smaller erythrocytes than either males or juveniles. Males had more elongated erythrocytes than females and erythrocytes of adults were more elongated than those of juveniles. Erythrocyte size and shape influence the efficiency of gas exchange owing to surface area to volume ratios, which are greater for small, elongated cells than for large, round cells. The high N/C ratio and low pixelation intensities of males and juveniles indicate that they may have had more immature erythrocytes in their circulation than females. The use of pixelation intensity to indicate the presence of immature erythrocytes was validated by seasonal differences that corresponded to the biology of the tortoises. Pixelation intensity was lowest in winter. We found that automated image analysis is a rapid and reliable method for determining cell size and shape, and it offers the potential for distinguishing among developmental stages that differ in staining intensity. The method should be useful for rapid health assessments, particularly of threatened species, and for comparative studies among different vertebrates.