Abstract
Examination of recent brachiopods, Terebratulina, Terebratella and Eucalathis, illustrates earlier observations that the young lophophores are centrifugal in Terebratulacea and centripetal in Terebratellacea, and suggests that both median septa and certain median structures free of the valve-floor are initiated by centripetal lophophores. This is turn suggests that the absence of the centronella-stage in Cancellothyridae, etc., is due not to tachygenesis but to different early lophophores from those of the Centronellidae, etc.; that the abnormal loops of Hamptonina may have been due to abnormal lophophore-orientation as well as retarded development; and that lophophore mutation could occasion changes of up to superfamily rank without leaving intermediate forms. It is concluded from consideration of the loop-lophophore relationship discussed above, against the background of present knowledge of Palaeozoic and certain Triassic loop-bearing brachiopods, that such earlier forms are neither Terebratulacea (in which they are usually placed) nor Terebratellacea, but that after re-study it should be possible to group them as superfamilies in which the brachial structures have less value for classification than in the Mesozoic superfamilies quoted.