196 R. T. JACKSON — STUDIES OF PAL.EECHINOIDEA. 



continuation of column 7, without any means of distinguishing the fact 

 that the resulting column was really made up by the combination of tAVO 

 separate columns. 



The whole matter of columns requires critical consideration in this new 

 light. Columns are the result of the superimposing of plates in successive 

 rows. While columns are probably the clearest and easiest way of com- 

 paring the introduction and position of plates in an interambulacrum, 

 the real feature of variation and progressive development is the row, the 

 columns being the resultant of successive rows exposed in a limited area 

 to lateral thrusts and pressure. Considering it in this light, at a certain 

 period in the life of the genus, species and individual a row of one plate 

 is built; this is the first row. In the next stage a row of two plates is 

 built, then a row of 3, succeeded l)y 4, etcetera. When the higher num- 

 bers are attained, several rows of a given number of plates are built 

 before the introduction of a row of a higher number of plates. When a 

 row is built containing one plate more than the preceding row, some one 

 plate by lateral thrusts is forced into a pentagonal form, which we nat- 

 urally consider, therefore, the initial plate of a new series or column, as 

 expressed in these pages. The plate which thus adopts the pentagonal 

 form, as shown by tabulation of cases of odd-numbered columns in Melo- 

 nites(pii^e 163), is the median plate of the row in over 95 per cent out of 

 138 examples cited. The reason that this plate is dynamically selected 

 is apparently because the middle plate of a row is that one which received 

 equal pressure from both sides, and also being in the middle of an area 

 built on a curved horizontal plane a newly added plate can there most 

 easily effect an entrance to permanent position. When a row is built of 

 one more pldte than the })receding row possessed, we cannot probably 

 say with positiveness which plate of the new row is the superadded one, 

 although plates which are to form new columns are probably added in 

 the middle of tlie area, as this conclusion is supported by such extremely 

 definite arrangement in the method of addition of columns. 



In the dorsal portion of the interambulacrum of this specimen of 

 Oligoporus danse (plate 6, figure 31) the form of the plates is gradually 

 changed from the hexagonal outline of older plates to the characteristic 

 rhombic form of young newly introduced plates, as described in Melonites 

 multiporus (plate 3, figure 13). This change is efi'ected as in Melonites by 

 the gradual shorter and shorter length of the upper and lower sides of 

 the hexagons. An entirely rhombic plate is not shown in this specimen ; 

 but such would doubtless have existed if the specimen had been pre- 

 served a little nearer to the dorsal termination of the area. 



The spines of Oligoporus danx have not been previously described. In 

 a specimen of this species in Yale University Museum a number of spines 



