NO. 1431. BREEDING HABITS AND EGG OF PIPEFISH— GUDGER. 477 



an exact counterpart of this. A section through this blastoderm in 

 the phine a-h reveals the structure shown in hg. 42, Plate VII. Here 

 the two central cells stand above the basal ones, with the line of 

 demarkation on the right especially sharp. The segmentation cavity 

 (.V. e.) and the central periblast {c JJ.) are both ver}' much reduced. 



Another very common form of eight-celled blastoderm is shown in 

 tig. 11, Plate V. Here there are six cells below and two above. This 

 is evidenth' a derivative of a six-celled stage frequently met with, in 

 which two of the blastomeres of fig. 4, Plate V, divide by vertical fur- 

 rows, the other two cells undergoing no change. Later, however, a 

 division of these in a horizontal plane would give the structure shown 

 in fig. 11. Variations of this type are frequently due to the shifting 

 of this pair of upper cells. Such a divergence is shown in fig. 12, 

 Plate V, whei-e these two cells reduced in size are shifted to one end of 

 the longer axis of the ])lastoderm. Sometimes these two cells are 

 placed parallel to the main axis, l)ut over one of the central lateral cells. 

 Again they may be shifted to lie at right angles to the long axis, over 

 one of the furrows separating two lateral cells, so that one cell is at 

 the edge of the 'blastoderm. In order not to multiply figures there is 

 given only one drawing of sections from such l)lastoderms. Fig. 43, 

 Plate VIII, is a section through such a structure as fig. 12, Plate V, in 

 the plane (/-I. Here one central cell is very much higher than any of 

 the other cells. The other central cell is completely cut out of the pro- 

 toplasm and is roofed over by a protoplasmic bridge extending from the 

 high cell to the left outer cell. Following the sections to one side of 

 this, the bridge and the cell under it are found to unite. They would 

 thus seem to have been split apart from the same mass of protoplasm. 



Another eight-celled blastoderm, quite as common as either of the 

 foregoing, is represented in fig. 13, Plate VI. Here one cell has, by 

 an equatorial furrow, become cut out to lie slightly above the rest. 

 The right side of the structure is normal, save that the third cell is 

 slightly flattened at its inner edge by contact witli this central cell. 

 As in the preceding case, so here there may be variations in the 

 position of this high level cell. It may lie in the center, at the edge, 

 or at any intermediate position on the blastoderm. A section through 

 the long axis of fig. 13 would give a structure essentially like that 

 shown in fig. 43, Plate VIII, omitting the protoplasmic bridge. Klein 

 (1872, figs. 5 and 6, Plate XVI) shows essentially the same structures 

 in the same stage of the trout germ, as does Henneguy (1888) in his 

 fig. 38, Plate XVII. 



Fig. 14, Plate VI, is a seven-celled form, in which an unmistakable 

 equatorial furrow has cut off three upper from four lower cells, of 

 which three are very large. A view of this blastoderm from below is 

 shown in the next figure (fig. 15). Here the two meridional furrows 

 show quite clearly, but there is no trace of the third or equatorial 

 Proc. N. M. vol. xxix— 05 31 



