384 HAROLD H. PLOUGH. 



the time of the operation which later goes to form the skeleton. 

 Many of these larvae lived for eight days or even more, but al- 

 ways remained enlarged ball-like larvae (Fig. 2). In perhaps 

 half of these larvae a normal tripartite gut was formed, while in 

 the others either a small portion or none at all was differentiated. 

 Some of these larvae were probably derived from blastomeres like 

 the upper one in Fig. I C. 



The larvae listed in the fourth column of Table I. are of special 

 interest. They were briefly described in my preliminary report. 

 All showed the development of a skeleton but this was incomplete. 

 The commonest type was a four day pluteus with one long 

 postoral arm and the other absent. The oral portion of the 

 skeleton was usually missing entirely. Fig. 3 shows several such 

 larvae compared with a complete one half size larva of the same 

 stage. Such larvae of Paracentrotus have been figured by Runn- 

 strom (1914) in earlier stages. It seems impossible to escape the 

 impression that such larvae are derived from one of the blasto- 

 meres of an egg divided like B in Fig. i. 



The larvae listed in the last column of Table I. were normal but 

 of approximately one half size (Fig. 3). They were complete in 

 every observable detail, and were usually as viable as normal 

 control plutei. Such larvae are probably derived from cells 

 divided along the egg axis ( Fig. i A}. It was observed that 

 isolated blastomeres developed at a slower rate than normal con- 

 trol eggs. The first cleavage appeared fifteen to twenty minutes 

 after the second in the normal egg, and the gastrulae were some- 

 times two hours or more behind. 



Finally it should be stated that I have never found larvae with 

 defective skeletons among the normal controls unless these were 

 subjected to abnormal treatment. Shaken eggs sometimes give 

 as many as 15 per cent, of abnormal skeletons, but never eggs 

 handled as indicated. Almost every group of control eggs shows 

 a few larvae which do not gastrulate, or which do not form skele- 

 tons, but these seldom live more than two days and usually less. 



PAIRS OF BLASTOMERES FROM THE SAME EGG. 



If the results summarized above seem inconclusive because of 

 the large number of blastomeres which fail to develop to the 



