16 ARKIV FÖR ZOOLOGI. BAND 8. NIO 10. 



provided with such canals, in which blood-vessels run. In 

 this portion of the plate they form quite a network. On the 

 surface of the plate there are several minute pores into which 

 the canals open. The fibrillated portion is traversed by only 

 a few canals. To consider the canaliferous part of the homo- 

 geneous portion of the plate as a distinct layer is quite 

 wrong. Several authors have described a layer of cosmine 

 in fish-scales. I hope to discuss the question of the struc- 

 ture and homology of fish-scales in a separate paper, here I 

 will merely state as my opinion, that the canaliferous layer 

 has no specific character. It is developed in scales of plate- 

 shape which lie close to each other. When we take into 

 consideration, that there are vessels passing from the interiör 

 to the exteriör stratum of the corium, it is quite evident 

 that when the scales develop into large juxtaposed plates 

 the vessels must be enclosed in them, forming canals. 



When a plate of an Ostraciontid is examined, dark lines 

 alternating with light ones will at once be observed (Fig. 9). 

 They run parallel to the sides of the hexagon in each »tri- 

 angle». They are broken off by the »lines». Williamson, 

 who noticed them first, described them as angular concentric 

 bundles of fibres (p. 660), the dark lines being membranous, 

 the light calcareous. This view is quite wrong. An exami- 

 nation of sections through the plate shows at once that the 

 difference between the dark and light lines is caused 'by the 

 direction of the bundles of fibres (Fig. D — F and 10). A 

 plate viewed from above presents dark lines where the ver- 

 tical bundles run, light ones along the horizontal bundles 

 which are parallel to the sides of the hexagon. Where the 

 vertical bundles cross those running from the centre of the 

 plate towards the side of the hexagon, the latter diverge form- 

 ing meshes for the vertical bundles. A vertical section through 

 a »triangle» shows dark lines where the bundles parallel to 

 the side of the hexagon are cut off, light ones where the 

 vertical bundles run. The latter are not cut off in such a 

 section. The dark lines seen from above do not correspond 

 to those presented by a vertical section from the apex of 

 the triangle perpendicular to the side of the hexagon. Both 

 these and the light lines are due to an optical phenomenon. 

 They have nothing to do with membranous or calcareous 

 parts. I may draw attention to another point. These alter- 



