ON FOSSIL POLYZOA. 179 
that would be naturally sought after by those whose desire it is to arrange 
the various genera in a natural sequence, but in the latter suborder there 
is but little variety except in the arrangement of the cells. In the later 
work of Mr. Busk,! in the writings of Professor Smitt, and in the 
‘Brit. Marine Polyzoa’ (1880) of the Rev. Thomas Hincks, practically 
the original arrangement of the Cyclostomata is left untouched. In the 
work of Mr. Hincks there is a redistribution of genera in a very limited 
family arrangement; but the work deals manifestly with recent species, 
and with species found only in the British area. 
In his Introduction to the Marine Polyzoa Mr. Hincks refers to the 
studies of Professor Smitt in the following terms: ‘He (Smitt) has 
aimed at a genealogical classification, starting with the proposition that 
the variations of species follow the line of their development and may be 
in a great measure explained by it. The Polyzoa as compound animals 
offer great facilities for the study of the laws and causes of variation. 
The differentiation of the colony gives us a series of variations running 
from the early and simple states to the fully developed form which is the 
parallel of the series of differences amongst species. Thus the British 
species of Crisia represent the evolutionary stages of one and the same 
type, of which Smitt regards Crisia geniculata, Mil.-Hd., as the first and 
simplest. The forms of this genus he would arrange according to the 
law of their evolution in a series, the members of which, springing from a 
common origin, will hold each its evolutionary grade.’? This, on the whole, 
may be a sound working principle, though it may not be always appli- 
cable when investigating the Palsozoic Polyzoa. Ihave not the least 
doubt but that some of the Graptolites and some of the earliest types 
of Polyzoa had a common ancestral origin. I believe also that the 
uni- and multi-serial Stomatopora represent evolutionary stages of a 
more primitive type; but we are not able to show at what stage diver- 
gences or differentiation of the colony took place, for the simple reason 
that the simple and the compound colonies occupy the same horizon in 
the Lower Silurians of America. In this country we have only uniserial 
Stomatopora in the Wenlock Shales. We do not meet with multiserial 
Stomatopora until we reach the Lias. 
One of the chief difficulties the systematist has to encounter in classi- 
fying the Fossil Polyzoa is this: On what characters in the zoarium 
shall divisions be based? If every variation of the zoaria is to be 
accepted, then there can be no limits set which shall be binding alike to 
all Paleontologists, for the zoaria of species vary greatly in different 
localities and in different countries. Then, again, if—as the old workers 
have done—we accept the fenestrule, its size, shape, or character, as an 
element to guide us in the structure of genera or species, we shall still 
be at fault, for in very many of the Fenestella, both in this country and in 
America, the fenestrule varies greatly, even in the same zoariwm. There 
is, however, one element that may be safely relied on, and this I have 
chosen for my guidance—that is, the structure and the arrangement of 
the cells in the branch, or in the colony ; all other characters, structural 
or superficial, are subordinate to this. ’ 
This principle has been adopted by Mr. Hincks in his arrangement of 
Recent Polyzoa, and admirable results have followed. I shall not there- 
1 Brit. Mus. Cat. pt. iii. Cyclostomata. 
* Brit, Mar, Polyzoa (Hincks), vol. i. p. cxx. 
N 2 
