5 { 
Lhe Use of Graphs in Paleontology. 359 
of plates in a transverse row (Fig. 1, IV-VII). Is this seriation 
merely due to the ingenuity of the human mind, cr does it 
represent a series of changes which the plates may actually undergo 
in evolution? The following strands of evidence taken together 
create a presumption that the latter is the true explanation, 
(1) While plates like I occur as early as the Ordovician, others 
do not become manifest until a later date. Presumably, therefore, 
this is the most primitive condition. 
Fic. 1.—The ambulacral plates of various Palmozoic echinoids: I, Bothrio- 
cidaris; Il, Hyattechinus: IlI, Lepidocentrus : IV, Lepidocidaris ; 
V, Lovenechinus ; VI, Oligoporus; VII, Lepidesthes (after Jackson). 
(2) Plates like those shown in III, for example, are not confined to 
forms which are closely allied, but also occur in forms which are 
distantly related. They are, therefore, not peculiar to any one 
particular line of descent. 
(3) Among closely related forms some may show shape II, some 
shape III, and others such conditions as IV-VI, etc. Evidently a 
number of steps in the series may occur within the narrow limits 
of a few closely allied lines of descent. 
(4) Some of the modern echinoids exhibit in early development 
plates like I, which at later stages become like ITI. 
(5) Two or more shapes of plate, and any number from two to 
twenty, may be present in different parts of an ambulacrum of the 
same individual. 
These last two facts indicate that some or all of these shapes may 
occur in the same line of descent. In a similar way the direction of 
change in structural units of any group of organisms may be traced. 
Such structural units, and the directions of change they follow, 
1 See Osborn, Amer. Naturalist, 1917 , p- 449 et seq. For full discussion of 
separability of units of organic structure. 
