478 



PROCEEDINGS OF THE NATIONAL MUSEUM. 



VOL. XXIX. 



furrow. The segmentation cavity (.v. e.) is so small as to be almost 

 negligible. Unfortunately, no section of this figure can be given, but 

 a comparison between it and tig. 4'2, Plate VII, will make clear its 

 internal make-up. 



These nine figures of the eight-celled stage have been introduced to 

 show (1), the great irregularities which enter into the segiuentation of 

 the pipefish egg at this stage: {'A), that these all result from the posi- 

 tion of the third furrow, which, ordinarily meridional and parallel to 

 the first and perpendicular to the second plane of division, here 

 becomes equatorial, and (3), that the irregularities thus resulting may 

 l)e reduced to four types, which may be traced to the very close of 

 segmentation. In order to estal)lish definitely these points, a table is 

 given showing the relative numbers of the different kinds of eight- 

 celled blastomeres which have been counted. 



From these eight-celled blastomeres are derived four types of seg- 

 mentation which persist to the close of segmentation. From figs. 7, 

 8, and 9 come two t3'pes of flat structures; from figs. lU, 11, 13 (with 

 the eighth cell in center) there comes a high-arched type of blasto- 

 derm, and from figs. 12 and 13 (with the eighth cell at one end) a 

 type of blastoderm thick at one end and tapering toward the other. 

 These structures will l)e more clearly shown in the next section. 



Table sliowiiKj relattre numbers^ of bhistodeniis for each fi/j/e of the eiglit-ce/led xtai/e of the 



Plpefifih egg. 



[Types referable to figures on Plates V and VI.] 



Killed in— 



Perenyi . . . . 

 Formalin . . . 

 Perenyi . . . 

 Formalin . . . 



Do 



Sub-acetic . . 

 Perenyi 



Total. 



« Six-celled. 



SIXTEEN-CELLED STAGE. 



Intermediate between the eight and sixteen-celled stages are found 

 many blastoderms with twelve, fourteen, and fifteen cells. These are 

 in fact more abundant than blastoderms with exactly sixteen cells. 



Figs. 16 and 17, Plate VI, .show the two most regular sixteen-celled 

 stages that have been found, yet they do not have the regular struc- 

 ture of the corresponding stages shown for Serranus by Wilson (1891) 

 and for Cristiceps by Fusari (1890). These blastoderms have been 



