540 LEWIS. 



Kelvin, in his Robert Boyle and Baltimore Lectures, called an "orthic 

 tetrakaidecahedron." Such a solid has been drawn in Figure 1, c-i. 

 It has already been stated that there is reason to consider the top and 

 bottom of the cell as hexagonal, the sides and angular points of one of 

 these hexagons being perpendicularly above or below those of the 

 other. If this is done, Figure 1, c, represents either the top or basal 

 view of the solid, seen in projection and not in perspective. The 

 central hexagon is surrounded by alternating squares and hexagons — 

 three of each. A lateral view, with a side and not an angle of the top 

 and basal hexagons toward the observer, is shown in Figure 1, /. 

 To pass from the top to the bottom, two surfaces must be traversed, — 

 a square with a hexagon below it, or a hexagon with a square below it. 

 A lateral view, with an angle of the top and basal hexagons toward 

 the observer, is seen in Figure 1, g. If, instead of a hexagon, one of the 

 squares should be regarded as the top, then the solid, seen from above, 

 would appear as in Figure 1, h. A The lateral view with an angle of the 

 square toward the observer has precisely the same appearance; but 

 if a side of the square is toward us, the lateral aspect is as in Figure 1, i, 

 which is the same as Figure 1, g, turned about. These, then, are the 

 appearances to be sought in actual cells if they possess the tetrakaide- 

 cahedral form. The way in which such solids combine to fill space is 

 shown in Figure 3 (Plate 1). 



A further and very instructive review of the literature has been 

 provided by D'A. W. Thompson, but it leaves him unable to decide 

 whether the cells of vegetable parenchyma are dodecahedra or tetra- 

 kaidecahedra. He finds that these cannot be distinguished in ordi- 

 nary sections, but suggests that it might be done through maceration. 

 "Very probably," he concludes, "it is after all the rhombic dodeca- 

 hedral configuration which, even under perfectly symmetrical condi- 

 tions, is generally assumed." 



Maceration appears to be impracticable, 5 but there is no special 

 difficulty in reconstructing, by Born's familiar wax-plate method, 



4 In this view the solid is seen to be a cube strongly truncated by an octa- 

 hedron. Any mineral which crystallizes in the isometric system is capable of 

 exhibiting such a combination, and it is not uncommon in gold, galena, pyrite, 

 salt, fluorite, cuprite and diamond. But the truncating of the cube by the 

 octahedron, or vice versa, in crystals is often slight, and the faces are usually 

 unequal, so that the production of an orthic tetrakaidecahedron would be very 

 rare indeed. For this information the writer is much indebted to Professor 

 B. L. Robinson and Professor Palache. Crystals of alum, however, may be 

 nearly orthic as seen, for example, in Figure 548, Plate 70, of Adams' Micro- 

 graphia (4th ed., London, 1771). 



5 M. H. Dut rochet (Recherches . . . sur la structure intime des animaux et 



