EEVISION OF PALEOZOIC STELLEEOIDEA. 31 



Not only this, but we may confidently add that the varied organi- 

 zation of those of the Middle Ordovicic proves that there is a long 

 previous liistoiy of which paleontology knows nothing. The origin 

 of the xisteroidea goes back in all probabiUty to the Proterozoic, as 

 may be inferred from the complex metamorphosis of the starfish 

 larva. In all of the Cambric, however, not a single starfish has been 

 found, and this is the more surprising when one considers the abun- 

 dance of life unearthed from these sediments in so many places 

 throughout the world. This may mean that a preservable starfish 

 skeleton was not evolved until Lower Ordovicic time, vvhere they 

 will surely be found. 



The most primitive fossil starfish. — Of the Ordovicic starfishes, the 

 genus Hudsonaster is the most primitive, but with the full quota of 

 parts in the primary skeleton that an unmodified or unspecialized 

 starfish must have. Among Hving starfishes this primitive skeleton 

 is rarely seen, but occurs in Heterasterias volsellata, "and in a few 

 other species, especially when young " (Verrill 1914:27). There are 

 other genera with a simpler skeleton {Stenaster and Tetraster) , hut these 

 are clearly cryptozonian forms that have originated in aphanerozonian 

 stock not unlike Hudsonaster. The oldest Hudsonasters (see plates 1 to 

 6) are small animals about 10 mm. in diameter, with thick, highly con- 

 vex, closely adjoining plates, aU. of which are devoid of spines ex- 

 cepting the adambulacrals, which have simple, minute, articulating 

 spines. On the ventral side there are in each radius double columns 

 of rectangular ambulacralia, with the pieces of each ambulacrum 

 arranged opposite to one another, or nearly so. Outside of these are 

 single columns of adambulacralia of about the same number as the 

 ambulacrals; they are, however, somewhat larger and more trans- 

 verse. These are laterally bounded by much larger, highly convex, 

 granulated, very prominent marginals, wHch, as they border the 

 animals and are simple ossicles (there are not here two superposed 

 marginaha as is so common in living forms), are the inframarginal 

 columns. In the axils of the rays lies a single large marginal plate, 

 the axillaiy, and these occupy the entke interbrachial areas. Ai-ound 

 the inner sides of these axillaries the adambulacraha continue, the 

 two basal pieces being the largest and the essential elements of the 

 oral armature. All of the ventral ossicles are, therefore, seen to bo 

 of the primary skeleton. 



On the dorsal side of Hudsonaster, the rays have medially promi- 

 nent columns of radials, while on each side of these are other columns 

 of thick ossicles, alternating with the radiaUa, and these are the 

 supramarginals. The former ossicles continue closely adjoining to 

 near the center of the disk, where hes a single large plate usually called 

 in this memoir the central disk plate, or, more rarely, the centro-dorsal. 

 In the same way, the supramarginals abut in the axils upon a single 



50601°— Bull. 88—15 3 



