CXXVlll INTRODUCTION. 



trusted. There remain the characters of the cell itself 

 and the habit of erowth. It can hardly be deemed 

 doubtful which of them should have precedence in a 

 natural system; we may go very much further, indeed, 

 and say that in such a system the latter must hold a very 

 secondary and subordinate place. The essential structure 

 of the cell, as one of the primary zooidal forms, must 

 certainly be accounted the most important point, both in 

 itself and as a clue to relationship. The mere habit is, so 

 to speak, a superinduced condition^ which may be different 

 in the most nearly related and similar in the most diver- 

 gent forms; and groups based on it, instead of fitting in 

 with natural afiinities, are found to traverse them at all 

 points. Thus the venerable family of the Escharidce 

 (auctt.) is a mere jumble of incongruous elements, and 

 no more represents the natural relationships of the forms 

 which compose it than would a group of plants founded 

 on the colour of their flowers. In the same way the genus 

 Lepralia (auctt.) is a miscellaneous collection, made up, as 

 it were, of numerous distinct nationalities, included with- 

 in a purely artificial boundary -line. The older systems, 

 indeed, for the most part, inverted the true order, and 

 gave the first place to secondary, but easily recognizable, 

 characters, thus sacrificing nature to convenience. 



It is an important fact that within the compass of a 

 tingle species the most diverse modes of growth may 

 occur. Thus in Membranipora arctica (of which I have 

 examined numerous specimens obtained by the Dutch 

 Arctic expedition) we meet with a crustaceous state, in 

 which it creeps as a network over stones, a Hemescharine, 

 in which it rises into unilamellate expansions, and an 

 Escharine, in wliich it forms well-compacted bilamellate 



