On the structdbe and classification op the Tremataspidae. 5 



scopic structure of the shield is carefully worked out, his results agreeing substantially with 

 Schmidt's. He distinguishes four layers: I) The thin glassy, prismatic or enamel layer; 2) the 

 bony layer, containing Haversian canals; concentric Haversian lamellae, and bone cells. 

 3) The layer of large medullary chambers with fewer bone cells in the matrix, and 4) the 

 inner layer, consisting of parallel lamellae of crossed fibres and elongated lacunae. He re- 

 gards the frontal pit of Thyestes and Tremataspis as containing an organ comparable with 

 the vesicular outgrowth of the roof of the secondary forebrain (paraphysis as described by 

 Selenka); and considers it probable that a similar frontal organ occurs in Gephalaspis, Au- 

 chenaspis and Pteraspis. The anterior lateral openings are regarded as nasal pits, comparable 

 with the lateral openings (the so called orbits or nasal pits of Huxley) oi Pteraspis. The or- 

 bital opening indicates the presence of a primitive, unpaired eyeball comparable with the 

 originally unpaired lateral eyes of true Vertebrates. The Parietal organ (Mitteloffnung, 

 F. Schmidt) is the opening for the parietal eye and is represented in G Agassizii by the 

 post orbital valley. 



The two small openings behind the parietal organ are the openings of the Ducti endo- 

 lymphatic!. 



The posterior pair of lateral openings is compared with the pseudobranchial openings 

 of certain Selachians and possibly with the marginal cells of EuJceraspis. The six marginal 

 incisions on the anterior margin of the ventral shield are regarded as indications of an equal 

 number of gill openings. 



In 1893, Rohon adds new data on the structure of the oral region and sensory open- 

 ings, and describes the structure and arrangement of the trunk scales. The parietal opening 

 and the two pairs of lateral openings, are shown to have scalloped margins. The parietal 

 and nasal openings have a spongy floor formed by a continuation of the inner layer of the 

 shell. He is now doubtful about the significance of the openings. 



A part of the posterior dorsal line of surface furrows (lateral line organs) is erro- 

 neously described as the last trace of a transverse segmentation of the dorsal shield. 



He distinguishes ten polygonal oral plates, four on each side and two median ones. 

 The mouth is erroneously supposed to be a large semicircular slit between the anterior row 

 of plates and the anterior margin of the dorsal shield. Three rows of trunk scales are de- 

 scribed; a dorsal, median, and a ventral, or marginal, row; also a large number of small 

 scales in the caudal region. 



In 1899, Rohon maintains that the ant-orbital opening, or «frontal-organ» and the 

 postorbital opening or «Foramen-parietale» of Tremataspis mark the location respectively 

 of a primitive «Paraphyse» and parietal organ, comparable with those openings, or organs, 

 in other fossil or recent fishes, and in the Asterolepidae, Stegocephalidae and recent Saurians. 

 He also maintains that the orbits of Thyestes and Tremataspis point to the origin of the paired 

 eyes of recent Vertrebrates from a single unpaired organ. 



