18 
takes place in the epistomial, the peristomial, and the bivalvular 
ooecia. In all sucli species in which tbere exists an atrium the 
closing of the atrial opening by the zooecial operculum, will at 
the same time, when the operciilar apparatus of the ooecium is 
drawn back by means of its muscles, form a perfectly safe passage 
from the ooecium to the zooecium. But even in such cases where 
there is no atrium and the distal margin of the operculum cannot 
reach the free margin of the ooecium, the distance between these 
two margins is so short that the egg by a protrusion of the ten- 
tacular sheat can have no difficulty in getting safely into the ooe¬ 
cium. In some Ctenostomata ( Valkeria, Bowerhankid) the egg after 
the death of the elder polypide is inclosed in the tentacular sheat 
of a newly formed one having a rudimentary intestinal canal, and 
therefore it is probable that also in the Cheilostomata an admission 
of the egg into the tentacular sheath precedes its passage into the 
ooecium. 
Systernatical remarks. The main causes of the great difficulties 
connected with the classification of the Cheilostomata are partly the 
great variation found in the different characters and partly the faet 
that a number of characters (f. i. the median pore and the different 
forms of orifice) either have risen independently or have developed 
in a similar manner (parallel characters) in widely different families 
and genera. Families and genera have hitherto mainly been based 
upon a single easily perceptible character, and as most of the lea- 
ding characters hitherto chosen are partly such parallel characters, 
partly characters taken from structural features subject to consider- 
able variation, the result has been, that most families and genera 
contain a number of heterogeneous species. One of the principal 
characters used in the classification of Hineks is the form of the 
orifice, but from more reasons this character is not fit to be a lea- 
ding one, for in the first place, I have found the same forms of 
orifice within a number of different families, secondly it is impos- 
sible to put up sharply separated types of orifices. So for instauce 
the forms of orifice regarded as characteristic for the genera Le- 
