26 THE DEVELOPMENT OF 



the mesoderm. 1 Neither at this nor at any other stage have 

 I seen any " secondary mesoderm " like that described by 

 Reichenbach ('76 and '86, and Ishikawa '85). 



With growth the mesoderm spreads in all directions ex- 

 cept that the bands mentioned above do not meet in the 

 middle line of the thoracic region until a comparatively late 

 stage (see fig. 40), though they do farther forward (fig. 

 35). .At stage C these bands have reached the level of the 

 stomodeum and partially surround its inner extremity, 

 thus giving origin to the muscles of the gastric mill and 

 O3sophagus, to be developed later (fig. 35m). Still other 

 portions extend further forward reaching (figs. 37 and 41) 

 to a point in front of the eye. Behind the thoracico-abdom- 

 inal fold the mesoderm is more developed, while at the 

 anterior margin of the fold and near the tip of the abdomen 

 there appear certain large cells (figs. 36, 37, 41, 51 gc) 

 belonging to both meso- and ectoderm (and which, in fact, 

 appeared still earlier, fig. 29) . These cells must be regard- 

 ed as budding cells and at once suggest comparisons with 

 the large mesoderm cells described by so many authors 

 among the annelids and notably by Dr. Whitman ('78 and 

 J 87). It is in just these regions that growth occurs. The 

 young shrimp when it hatches from the egg has its abdo- 

 men with the normal number (7.) of segments. The cepha- 

 lothorax, too, is complete in its appendages in front (see 

 fig. 27) ; but the series stops with the second maxilliped, 

 leaving six pairs to be produced in subsequent growth. 

 The anterior region of these growth cells nearly, and I be- 

 lieve (though I cannot say with certainty) exactly corre- 

 spond with this region which is afterward to produce the 

 missing portions of the thorax. So, too, in the tip of the ab- 

 domen we find another region for the intercalation of new 



1 By an error, the planes of the sections in figure 29 are wrongly 

 numbered. The figures 30 and 31 should be transposed, as will be 

 seen by a comparison of the corresponding figures. 



