CLARK: THE CIDARIDAE. 189 



range arc really unknown. There appears to be a variety of baculosa common in 

 the East Indies, in which the primaries are cross-banded as in this species, and 

 this form has been confused with annulifera. Now if de Loriol ('73) and Mor- 

 tensen ( : 03) were correct, it would be clear that Lamarck's annulifera is this 

 variety of baculosa, and in that case the present species should be called lutkeni, 

 as de Loriol clearly figures and describes it under that name. Mortcnsen says 

 he has examined Lamarck's type and it is baculosa, but A. Agassiz examined all 

 of Lamarck's types some forty years ago and satisfied himself that the present 

 species is Lamarck's annulifera. In a disagreement such as this it is obvious 

 that the earlier investigation is the one least liable to error, for there had been 

 considerably less time for a chance confusion of labels or specimens. Both 

 de Loriol and Mortensen apparently overlook the fact that A. Agassiz examined 

 Lamarck's types in Paris and that there has never been the slightest reason for 

 supposing that he made any mistake in associating Lamarck's name with this 

 species. Until it can be shown that such a mistake was made, the name it has 

 borne so long should be retained for this species. So far as we now know, it is 

 an Australian and East Indian form, and does not occur in the Red Sea or along 

 the African coast. The Museum of Comparative Zoology has two fine specimens 

 from the Gulf of Siam, received from the Copenhagen Museum. They were 

 collected by Mortensen, and labelled by him " Stephanocidaris bispinosa." The 

 species is apparently nearly as variable as baculosa, both in coloration and in 

 the form of the primary spines ; in some cases the secondaries are green and 

 the primaries cross-banded with purple and green, but in other specimens the 

 secondaries are pale brown and the primaries are dull with less distinct markings. 

 The secondaries usually (perhaps always?) have a median longitudinal stripe, 

 darker than the ground color. The primaries are frequently flattened and wid- 

 ened at the base, tapering to the tip and quite thorny, much as in Stephenocida- 

 ris, but they are often nearly cylindrical with few thorns. I am not satisfied 

 that the varieties recognized by Dbderlein are sufficiently constant to warrant 

 their recognition by name. 



Phyllacanthus baculosa. 



Cidarites baculosa Lamarck, 1816, Anim. s. Vert., 3, p. 55. 

 Phyllacanthus baculosa A. Agassiz, 1872, Rev. Ech., pt. 1., p. 150. 



Plate If, figs. 4, 5, Rev. Ech., A. Agassiz, 1873. Plate 59, figs. 1-5, Semon'g 

 gesain. Ech., Doderlein, 1903. 



Common, variable, and widely distributed as is this much-discussed and per- 

 plexing species, its true characters and the limits of their variability are still little 

 understood. It seems useless in the present state of our knowledge to attempt 

 to recognize varieties, and we can only say that with all the diversity of coloration 

 and of primary spines, the deep red or purple spots on the collar of the primaries 

 is an obvious character almost always present. It is true de Loiiol ('83) and de 

 Meijere (:04) have described specimens with a narrow unspotted collar, but it is 



