154 ARTHUR VVILLEy. 



M. Julin's work overlaps mine in regard to Ciona intes- 

 tinal is, and it follows inevitably from his statements, although 

 he does not say it in so many words, that protostigma A is 

 not homodynamous with protostigma B. I think these two 

 protostigmata are homodynamous structures, and that is how 

 the matter rests. 



The comparison of the protostigmata in M. manhattensis 

 and C. intestinalis may be tabulated as follows, the stig- 

 mata being denoted by Roman numerals in their serial, and 

 by capital letters in their individual capacities : 



Protostigmata. Cioua intestinalis. Molgula manliattensis. 



I . . . A . . . .A 



II . . . C . . . . C 



III . . . D . . . . B 



IV . . . B . . . . D 

 V . . . E . . . . E 



VI . . . F . . . . F 



Gtarstang's Observations on Botkyli.us.^ 



If it is desired to form any idea as to the probability of one 

 method of formation of the branchial stigmata of Ascidians 

 having a more primitive character than another, it is necessary 

 to take note of Garstang's observations on the formation of 

 the stigmata in the oozoid and blastozoid of Botryllus. 

 Reference has been made above (p. 142) to the known fact 

 that in Clavelina and other Ascidians the definitive stigmata 

 arise by separate perforations, without the intervention of 

 protostigmata. Here, therefore, we have two different methods 

 to consider, and to decide between them which is the inore 

 primitive. The (piestion can only be approached from a basis 

 of probability, and in this sense it has been answered un- 

 equivocally by Garstang to the effect that the formation of 

 protostigmata is a primitive character. Garstang found that 



> W. Garstaug, "On the Development of the Stigmata in Ascidians," 

 ' Proc. Roy. Soc.,' vol. lii, 1892, p. 505. In this paper the term protostig- 

 mata was first introduced. 



