Ciliate nuclear architecture, in particular the sequestration of a transcriptionally silent germline genome, allows for the accumulation of mutations that are "hidden" from selection during many rounds of asexual reproduction. After sexual conjugation, these mutations are expressed, potentially resulting in highly variable phenotypes. Morphological traits are widely used in ciliate taxonomy, however, the extent to which the values of these traits are robust to change in the face of mutation is largely unknown. In this study, we examine the effects of mutations accumulated in the germline genome to test the mutational robustness of four traits commonly used in ciliate morphological taxonomy (number of somatic kineties, number of post-oral kineties, macronuclear size, and cell size). We find that the number of post-oral kineties is robust to mutation, confirming that it should be preferentially used in taxonomy. By contrast, we find that, as in other unicellular and multicellular species, cell/macronucleus size changes in response to mutation. Thus, we argue that cell/macronucleus sizes, which are widely used in taxonomy, should be treated cautiously for species identification. Finally, we find evidence of correlations between cell and macronucleus sizes and fitness, suggesting possible mutational pleiotropy. This study demonstrates the importance of, and methods for, determining mutational robustness to guide morphological taxonomy in ciliates.