378 University of California Publication* in Zoology. t v L - 6 



a manner as not to injure the spines. In addition to the effect of 

 reagents, disintegration results in the early loosening and dis- 

 appearance of the spines. In no case where the living worm has 

 been examined have the spines been absent. 



The distribution of the spines affords a useful basis of dis- 

 tinction. In G. rugosa (Spencer, 1889), "spines are scattered 

 over the whole surface of the body"; in G. nigrosetosa (Haswell, 

 1902) "they are mainly confined to the dorsal surface, except at 

 the anterior end. Over the rest of the ventral surface are 

 scattered a very few, much smaller than those on the dorsal sur- 

 face." In G. urna (Lonnberg, 1891), the distribution is confined 

 to the extremities and the lateral margins of the body, leaving 

 the middle of the body, from just in front of the canal opening 

 to the level of the acetabulum, free from spines. In G. fimbriata 

 the condition is as described by Lonnberg for G. urna, except 

 that spines are absent from the greater part of the lateral margin 

 of the body, and do not occur on the lateral frills. Possible 

 variation with age and the ease with which spines may be lost 

 by slight disintegration or action of reagents, makes any dis- 

 tinction resting on them somewhat impracticable. Here again 

 we have differences coordinated with, but scarcely useful as, 

 specific distinctions. 



3. Size of tail rosette. This character appears to depend less 

 on the state of contraction than do the width and amplitude of 

 the lateral folds, and if reduced to a definite statement in terms 

 of some other dimension of the animal, would probably be found 

 to vary about a mean characteristic for each species. The 

 astonishing illustrations given by Lonnberg (pi. 33, figs. 1-4) 

 would, if corroborated, not necessarily invalidate the worth of 

 this characteristic, provided comparisons were made between 

 specimens in a similar state of contraction, G. rugosa (pi. 33, 

 figs. 5, 6) in a contracted state has a rosette about one-fourth its 

 greatest body-width in diameter. In G. nigrosetosa, the rosette 

 is about five-sixths the body- width; in G. urna about one-half; 

 while in G. fimbriata it varies from two-thirds to three-fourths 

 or even more. 



4. Length of uterus. This is a character which it is exceed- 

 ingly difficult to apply. The length of the much-contorted 



