Conulariids provide good examples of how biometric and other characters that have been used to diagnose fossil species can be affected by taphonomic processes, possibly leading to the erection of taphonomic taxa, or taphonomic artifacts. Based on analysis of the taphonomy of Conularia quichua Ulrich from the Devonian Ponta Grossa of southern Brazil, we argue that caution must be exercised when using biometric and other characters to diagnose conulariid species. For example, measurements of the spacing of the transverse ribs must be corrected for compaction of the theca parallel to its long axis. C. quichua oriented at high angles to bedding almost always exhibit this kind of deformation, which if not corrected for results in substantial additional measurement error. Similarly, the value of the apical angle of C. quichua differs between compressed and uncompressed specimens, making it difficult to measure this character with a high degree of consistency and reproducibility. Other characters (geometry of the transverse ribs, presence or absence of interspace ridges and nodes) used to diagnose conulariids are susceptible to modification and information loss through weathering. In reviewing published descriptions of other conulariids, we have found that certain species may be taphotaxa. Future descriptions of new species should be based on collections encompassing the known spectrum of preservational patterns. Also, the erection of new conulariid taxa should be based as much as possible on complete or nearly complete specimens, and morphometric comparisons should be made using specimens showing similar patterns of preservation.