THE ASCIDIANS 235 



is the fiict that in many places they are contracted into a more or less 

 regular spiral, the coils of which are so close that, seen as they always 

 are through the material in which they are imbedded, they appear to 

 be made up of discrete bodies which may easily be mistaken for cells. 

 By staining them in various ways and clearing them in glycerine and 

 clove oil, and, especially, by following the individual fibers far enough 

 (for often they may be traced all the way into the body of the animal) 

 their real character can be determined beyond question. One point 

 concerning them, however, I have not been able to make out with cer- 

 tainty. In the terminal bulbs are frequently seen what appear to be 

 definite loops formed by the central fibers, and in places between the 

 bulb and the point of branching the fiber sometimes appears to be dou- 

 ble. Both these facts suggest the possibility and support the belief 

 that each fiber is double ; that is, extends down the rootlet to the end 

 and there turns back upon itself. These fibers are, without much 

 doubt, fibers from the mantle that have grown into the peduncle first of 

 all, and then out into the rootlets ; and it would seem probable that in 

 this process they have looped in rather than grown in at a free end. 

 One may imagine that the thing was accomplished by a process similar 

 to the looping- out of the mantle fibers on the atrial languet in some 

 species of Amaroucium or Distaplia for example, or over the incuba- 

 tory pouch of Colella. The evidence for this interpretation, however, 

 is not conclusive. In many places in the rootlets the fibers are cer- 

 tainly not double when fully mature ; whether the two strands may 

 have become fused into one here, or one of them become wholly degen- 

 erated, I am unable to say. 



The most puzzling thing in the structure of the rootlets is the fact 

 that the muscle fibers are actually ejubedded in the test, and are not sit- 

 uated in a cavity with an epithelial wall derived from the ectoderm 

 of the mantle. From all analogies one would expect the rootlets to be 

 in reality ectodermal vessels, each with a thin layer of test. I have, 

 however, searched in vain for epithelial cells of any sort either in the 

 rootlets or the larger portions of the peduncle. Even after staining 

 with htematoxylin, which can always be relied on to bring out the 

 nuclei in epithelial tissues of the Ascidians, not even a suggestion of 

 nuclei has appeared in these structures. All the outside transparent 

 material is apparently test and nothing else. Moreover, by splitting 

 the peduncle at its larger, proximal end next to the body in the plane 

 of the fiber bundles, one obtains conclusive evidence that the muscle 

 fibers are fully surrounded by the test material. In such a place as 

 that shown in p1. xxviii, fig. 16, for example, one or two fibers are 



