Vol. XI] BERRY— FOSSIL CHITOXS 



491 



variation. On the other hand the features separating the 

 palmulatus complex from crassicostatus seem dependable 

 enough. These two facts are brought out rather forcefully 

 in the following summary of observations made on Dr. Clark's 

 numerous Santa Monica specimens. 



Sixty-six head valves of the Long Wharf Canyon mirabilis 

 have 9 ribs and 9 slits to correspond, though the two outermost 

 ribs are in each case apparently duplex, and counting each of 

 these as 2 would raise the number of ribs to 1 1. Fifteen others 

 have 9 ribs each, but the number of slits could not be accurately 

 determined. Two have 9 ribs, 9 slits to correspond, and a 

 supernumerary slit between two of the ribs. Four have the 

 usual 9 ribs, but 10 slits, due to a double slit opposite one of the 

 ribs, the position of the rib so favored being variable. A single 

 valve has 10 ribs and 10 corresponding slits. 



Ten median valves show variations in the number of longi- 

 tudinal ribs on each side as follows : 18-18, 17-16, 16-16, 16-15, 

 15-15. 15-15, 14-14. 14-14, 14-14, 10-10. The number ap- 

 parently depends in considerable degree upon the size of the 

 specimen. 



One tail valve shows 12 longitudinal riblets to a side on the 

 central area, 2 show 11,2 show 10, 8 show 9, 8 show 8, 5 show 

 7. and 2 show 6. none showing less than 6. Here again there 

 seems to be a certain degree of correlation, not entirely ab- 

 solute, between the number of these riblets and the size of the 

 specimen. Twenty-five tail valves show 5 radial ribs, all at 

 first simple, but soon divided by a strong median furrow, sec- 

 ondary furrows being sometimes developed later on each side 

 of the primary one. In the outermost pair of ribs on one or 

 two of these specimens the central furrows commence so early 

 that the total number of ribs could easily be stated as 7. Two 

 valves clearly have 7 ribs. Fifteen specimens have 6 ribs, the 

 2 central ribs in such cases being usually very evidently homo- 

 logous with the single median rib of the 5-ribbed specimens. 

 This is shown by the pair in question being very close together, 

 while neither develops its median furrow as early as the re- 

 maining ribs do. Turning to the insertion plates, 7 valves 

 have 25 slits each, 9 have 24, 9 have 25, 4 have 22. 1 has 21, 

 4 have 20, 4 have 19, 2 have 18. and 1 has 16. The mode 



