838 



That what has hitherto been termed enamel, is in fishes a com- 

 pound structure, separable into ganoin and kosmine (fcoo-yuetv, to 

 adoTTi) ; the former being transparent and laminated, but otherwise 

 structureless, whilst the latter consists of minute branching tubes re- 

 s^^'mbling the dentine of true teeth. 



That the kosmine covering the osseous scales of so many ganoid 

 fish, as in Lepidotus semiserratus, Megalichthys Hibberti, &c., is 

 homologous and identical with the substance composing the dermal 

 teeth of the true placoids, such as the Dog-fish, Thornback, &c., 

 only that, whilst in the former the areolae of kosmine are aggregated 

 upon bony scales, in the latter they are implanted in the soft integu- 

 ment, without the intervention of any bony matter. It follows from 

 this, that the distinction of " ganoid " and " placoid " is scarcely a 

 physiological one, inasmuch as the scales of many so-called ganoid 

 fish, such as Dapidius orbis, Acipenser, &c., exhibit little or no trace 

 of either ganoin or kosmine ; that in many of the Placoids these sub- 

 stances are very largely developed ; and that a series of well-defined 

 links exist, passing through the common Thornback, the common 

 Spotted Dog-fish, Hybodus reticidatus, Macropoma Mantelli, Da- 

 pidius gramdosus, Holoptychius, Diplopterus and Megalichthys, by 

 which the ganoid and placoid forms merge in one another. 



That ganoid scales consist of variously modified osseous lamellae, 

 the result of successive additions made chiefly to the lower surface 

 of each ; but also, under particular circumstances, either to a part, 

 or to the whole of the upper surface. 



That these lamellae have not been the result of any process of 

 excretion, or depositions from a secreting surface, as supposed by 

 M. Agassiz, but that they have been formed by the calcification of 

 the lower laminae of an investing vascular periosteum ; and that 

 consequently the phaenomena attending the structure and growth of 

 these ganoid scales contribute in a material degree to establish the 

 correctness of the views recently promulgated by Professor Sharpey 

 respecting the growth and development of human bone ; the gra- 

 dual formation of Haversian canals being traced with great ease 

 from the simple laminae seen in the scales of Lepidosteus, Lepidotus, 

 &c., through Aspidorhynchus, Acipenser, HoloptychiuSy &c. to their 

 high degree of development in Megalichthys. 



That the study of the microscopic structure of the dermal ap- 

 pendages of fish may, when carried on with due caution, be made a 

 valuable auxiliary, both in distinguishing between allied species, and 

 in establishing the existence of important affinities, even when ap- 

 plied to otherwise insignificant fragments ; but that it is capable of 

 being overstrained, and of leading to erroneous conclusions, if any 

 classifications are founded upon it irrespective of the other portions 

 of the fish to which the scales belong, because of the unequal ratio 

 in which the various parts of an organism may have been developed. 

 Thus, whilst Lepidosteus osseus presents one of the simplest forms 

 of ganoid scales, it has the concavo-convex vertebral articulations of 

 the Ophidians ; on the other hand, in many species, as in Megalich- 

 thys and Holoptychius, whilst the structure of each part of the exo- 



