ZooL.— Vol. III.] BANCROFT— COMPOUND ASCIDIANS. 1 59 



(figs. 16-23) furnishes variations corresponding to the 

 following species: — 



B otry litis gemmeus Savigny, 1816, p. 203. 



Figure 21 agrees quite well with the description, except 

 in that the dorsal lines are not complete. At the time of 

 drawing, however, most of the zooids of the same colony 

 had the dorsal bands complete. 



Botryllus moi'io Giard, 1872, p. 629, PI. XXX, figs. 8, 9. 

 Figure 23 (white banded zooids) agrees well with Giard's 

 description; and the mother colony (fig. 16) agreed still 

 better a few generations before the drawing was made. At 

 the time when it was drawn, the colony was rapidly degen- 

 erating, and the two dorsal bands had become fused into 

 one. 



Botrylhis morio var. S Della Valle, 1877, p. 25. 



Figure 23 (zooids without white bands) agrees almost per- 

 fectly with the description; and fig. 10 of Family I also 

 agrees well, but here the lines causing the reticulation are 

 not produced entirely by an absence of pigment, but have 

 a pigment of their own. Since both these figures agree so 

 well with Della Valle's description, we are justified in uniting 

 the families I and II, and saying that all of the described 

 species just enumerated should be united into one species, 

 as the color variations represented by them are found among 

 the members of two families, the variations of which overlap.^ 



Summarizing the results of this comparison, then, it has 

 been -proved that B. auroli7ieatus var. i, B. marionis, 

 B. calendula, and B. rubens must undoubtedly be united 

 into one species, and B . gemmeus and B. morio into another ; 

 and furthermore, that there is ample evidence for uniting all 

 the six species 7nentioned into one. 



iFor other cases in which the variations in the two families overlap, compare fig. 10 

 with fig. 22; and figs. 11 and 12, both of the same colony, with figs. 18 and 19, also drawn 

 from one colony. As the zooids of each of these last mentioned pairs of figures belonged 

 to the same colony, all gradations between the two conditions were observed, and, had 

 that been the object, it would have been easy to select zooids from different families 

 resembling each other still more closely. 



