Studies on Mesozoic faunas suggest that the diversity in the Middle Triassic has well exceeded the diversity in the Early Triassic. Recent and current reinvestigations show that this is true with respect to diversity at the genus level in basal actinopterygians. However, the diversity of chondrichthyans and coelacanths and their position in the recovery process are much more difficult to assess. For several reasons, only a few fish assemblages meet the criteria necessary for testing: 1. The majority of known localities has not been well enough sampled or well enough studied for their inclusion in conclusive numerical analyses. 2. The results from numerical analyses of diversity at the species level are not meaningful if only the best-explored sites are considered. 3. Below the family level, diversity of fish faunas generally increases during the late Early Triassic. Yet relatively rare and poorly preserved Early Triassic palaeoniscoids had been overlooked and remained unidentified in the collections. 4. Counts of specimens per species across localities, relative abundance per locality at the family level, and sampling bias are identified as underestimated parameters when assessing diversity. 5. The morphological revision of prominent fish groups and their paleobiogeography suggests that patterns of diversity of ichthyofaunas during the late Early Triassic and early Middle Triassic are more complex than previously suggested.

The current doctrine suggests a relatively slow faunal recovery; many Triassic fish genera appear before the Smithian and become "cosmopolitan". This model would be indicative of an early and sudden dispersal event. Yet such an event contrasts with the record and evolution of certain actinopterygian groups, with the poor dating of most Early Triassic ichthyofaunas, and with the scarce but remarkable record of palaeoniscoids in the Triassic.