20 ARVID R. MOLANDER, ALCYONACEA. 
This brings us to the question of the possibility of being able really to distinguish 
between these membranous forms and the Cornulariidae. 
According to KUKENTHAL, the family of Cornulariidae is characterized in the follow- 
ing manner: »Polyps separate from each other, connected at the base by free stolons 
or by a membranous disc containing a small number of canals. » This definition is not 
satisfactory. 
As long as the stolons are single and the polyps are scattered, confusion between 
the membranous forms and the Cornulariidae ought scarcely to occur. So far, how- 
ever, the membranous transformation seems seldom to advance. But when the 
stolons unite with the membranes and the polyps collect in larger groups, the Cornula- 
riidae show great similarity to the membranous forms of other families. The point to 
be remembered is that, within the Cornulariidae, the thin membrane, which at the base 
only connects the polyps, contains unimportant coenenchym, which makes a retraction 
of the polyps impossible. Further, we must notice the spicules, which have their definite 
character. If we investigate the solenia in the membranous form and the Cornularii- 
dae, we meet with a decided difference: within the membranous forms we find a much 
more vigorous development of the horizontal canals, which, for the most part at least, 
fill the membrane with a well-developed anastomosis. The vertical canal and cell-strings 
also occur. In the case of the real Cornulariidae, the horizontal canals are limited in 
number, and the greater part consists of mesogloea, except in the case of Cornularia, 
which has simple stolons. Where there is also a closer gathering of the polyps, this ar- 
rangement is retained with an entirely basal connection between the polyps, without ver- 
tical cell-strings, while, with the membranous form, an endodermal, vertical tissue of ca- 
nals is developed. If we keep strictly to the facts now mentioned, we shall find it possible 
to distinguish the real Cornulariidae from the membranous transformations of other fam- 
ilies. I regret that it has not been possible for me to investigate all the species of fam. 
Cornulariidae, but by means of my investigations I have already found support for my 
supposition that many species classed in this family do not belong to it. WRIGHT and 
STUDER propose the classification of Clavularia into Stolonifera and Membranipoda, a 
classification which is rejected by SCHENK, as, in the case of a number of Clavularia spe- 
cies, he simultaneously found specimens with both stolon and membranous development. 
A thorough investigation of all Cornulariidae, according to WRIGHT's and STUDER'S 
principles, would certainly be full of interest as regards systematization. A minor mem- 
branous extension of the stolon is to be found with most of the species of Clavularia, for 
instance, at the bases of the polyps, but my opinion is that the vigorous membranous 
development within the real Cornulariidae is wanting, and that the numerous species 
which have been classed here as having vigorous membranous extension are foreign to 
to the group. The membranous transformations are a consequence of the special form 
of the substratum and the statical change thereby caused, but in the case of Clavularia 
there are to be found membranous and stoloniferous forms mingled with each other, 
independently of the shape of the substratum. I have also found only minor membran- 
ous extensions in the case of the real Clacularia species I have investigated. The cause 
