270 BULLETIN OF THE 



even in this larva with air or gas. Below it is a mass of reddish pig- 

 ment concentrated in a cluster. The size of this larva is about the 

 same as that of the last. 



On August 13, seven days after the Agalma had been placed in the 

 water, I was surprised to see, on looking for my larvse through the walls 

 of the glass vessel in which they were confined, that they had very 

 much decreased in numbers. This led to the discovery that, whereas 

 up to about this date they were found at all depths in the water, the 

 larvjE are now to be seen only upon the surface. They often cluster 

 together there, and the size of the float imparts to them a silvery color, 

 like a small bubble of air resting on the water. The reason why the 

 larvse seek the surface at this phase of their development probably is, 

 that the float has grown so large, or that the size of the primitive 

 hydrophyllium has diminished. Whatever may be the cause which led 

 the Agalmata to come to the surface, an effect which can probabl} be 

 ascribed to the two causes mentioned above combined, we find that 

 the size and general outlines of the first-formed covering-scale have 

 undergone several modifications. Fig. 5, PI. IV. shows a larval stage 

 taken August 13th, in which the size of the scale is much smaller than 

 in the larvse already described. It is found at this time in the life 

 of the larva that the border of the covering-scale has a tendency 

 to draw together, and its surface becomes grooved or furrowed. In 

 Fig. 6 we see a continuation of the same process, and in Fig. 7 still more 

 reduction in the size of this body. One or two structural features have 

 led me to regard the flat angular body on the yolk of these larvse as 

 the primitive hydrophyllium reduced in size. The tube which is found 

 in the primitive scale, especially at the marginal termination, has a yel- 

 low color with black dots. These figments were found in the tube of 

 the more reduced scale in its present condition. The small nuclei 

 spread over the surface of the primitive hydrophyllium, called in our 

 above desci-iption the nuclei of the epiblast, are easily recognized on the 

 surface of the modified scale. With the reduction in external form 

 of the plump walls of the first-formed scale, or primitive hydrophyllium, 

 there has taken place also a change of form in its internal cavity. At 

 the distal border of a scale represented in PL IV. fig. 8, the tube of the 

 scale has bifui'cated and extends in two divisions to the bell rim, where 

 both end in the neighborhood of clusters of large nematocysts or lasso- 

 cells. A yellow color was observed at these points, although the tube 

 of the scale throughout most of its course is not as markedly colored. 

 The small cell-like spots which appear on the surface of the scale and 



