MONOGRAPH OF THE EXISTING CRINOIDS. 211 



The most primitive type of columnal has about its crater a raised band mark- 

 ing the position of the original annulus from which the rest of the columnal has 

 been built up. This band, however, is only preserved in comparativclv rare 

 instances, and usually only hi the columns of small and delicate forms, such as 

 Khizocrinus lofotensis (fig. 135, p. 205). 



The primitive form of the terminal stem plate is a circular disk (fig. 532, pi. 3), 

 and this is the form first taken in all young pentacrinoids. In some pcntacrinoids, 

 and in a few of the stalked species, this form is maintained with little or no varia- 

 tion, but in many pentacrinoids the originally circular disk grows not by a uniform 

 extension of its entire border, but by more or less definitely localized additions of 

 calcareous matter, so that it becomes lobate or, in extreme cases, sharply digiti- 

 form (figs. 533-540, pi. 3). 



The terminal stem plate in Promachocrinus is strong!}' lobate or more or less 

 digitiform, suggesting that of the species of Ilathromctra. This type of stem plate 

 always accompanies greatly elongated columnals in pentacrinoid larvae. If the 

 coluninals are very short the terminal stem plate approaches a circular form, length- 

 ening columnals being correlated with an increasingly lobate outline, which finally 

 becomes digitiform. 



Pentacrinoid larva? with short columnals and a more or less circular terminal 

 stem plate, in other words, with a column of comparatively slow growth, never 

 show any trace of radicular cirri; but pentacrinoid larva; with very long columnals 

 and a strongly digitiform terminal stem plate, that is, with a very rapid stem 

 growth, often form additional attachments further up the column (figs. 540, 

 541, pi. 3). 



Radicular cirri are entirely distinct from the other type of cirri (fig. 127, 

 p. 197); they are most perfect at the base of the column and rapidly become 

 smaller and less perfect toward the crown. The true cirri arc always absent 

 from the base of the column, first appearing, usually in a deficient series of 

 more or less imperfect individuals, just beyond (reckoning from the terminal 

 stem plate) the first stem syzygy, the most perfect and the best developed being 

 just under the crown. 



The radicular cirri are merely special processes developed from the overgrowth 

 and expansion of the terminal stem plate, and are always confined to the region 

 below the first stem syzygy; the true cirri represent five dorsal processes, or groups 

 of processes, one from each of the five metameric divisions of the body. 



Radicular cirri are probably to be interpreted as originally a terminal stem 

 plate which is reduplicated through a number of columnals on account of the very 

 rapid growth of the latter; that is, a number of the earliest columnals possess a 

 tendency, progressively decreasing, to expand laterally at the ends; but on account 

 of the fulcral ridge such expansion can only take place at two points, so that it 

 forms two long processes, one on either side. 



The radicular cirri themselves are best considered as representing a step in 

 development beyond the digitiform typo of terminal stem plate; this form of stem 

 plate is developed from the circular through the lobate as a result of a great increase 

 hi the rate of growth; further increase in the rate of growth results in immensely 

 increasing the length of the digitiform processes, which become jointed and branched. 



