CELLULAKINE AND OTHEK POLYZOA. 337 
o£ Jan. 26, 1911, remained in MS. at the time of liis death. Having 
examined the whole of the material from the two Stations, I can confirm all 
Levinsen's statements of fact quoted above, although I do not agree with 
him in placing the new species in Caherea. There is no doubt that Busk 
confused two species, and it is unfortunate that manonensis must be applied, 
as shown by his description, to a species from the Cape of Good Hope, not at 
present known to occur off Marion Island. 
I have adopted Levinsen's MS. name Hi'kpatricki for the new species ; and 
I give below a new diagnosis of Menipea marionensis. The two species can 
readily be distinguished under a low magnification. In Menipea mario- 
nensis the joints are placed at some distance from the axil of the bifurcation^ 
the ovicells are small and short, there are no vibracula, and none of the frontal 
avicularia are enlarged. In Ainastigia kirkpatricki the joints are at the level 
of the axil, the ovicells are large and long, conspicuous basal vibracula are 
present, and the frontal aviculaiium of the axillary zooecium (or of the 
proximal median zooecium if more than one is present) is usually much 
larger than the others. This last feature is indicated in Busk's fig. 9 
(PI. 14.), which I suppose to have been drawn from a specimen obtained 
at Stat. 144 a, and therefore to belong to Amastigia kirkpatricki. In 
describing a species obtained from more than one localitj^, it is always 
expedient to indicate the locality of the specimens figured ; and this omission,, 
by Busk and others, gives rise to many difficulties. 
With reference to Levinsen's proposal to place this species in Caherea, I 
think it desirable to exclude from this genus those species in which the- 
rootlets pass down the margins of the branches instead of down the middle 
of the basal surface. I think that C. rudis. Busk, should be placed in 
Amastigia, and that C. crassimarginata, Busk, may be referred to the saaie 
genus. It must be admitted that A. kirkpatricki has a considerable claim to- 
be included in Scrupocellaria, which it resembles in its mode of bifurcation 
(PI. 17. fig. 20). The branches are well jointed at their commencement, 
but a median series of zooecia, usually commencing near the proximal end of 
the internode, is present, in opposition to what is found in typical Scrupo- 
cellaria. The first median zooecium (E ') originates from a zooecium ("A") 
situated on the outer side of a branch with reference to the precedino- 
bifurcation. It gives to a second (E ^), which may be followed b}^ others in 
a median row. The distal member of this series (E ^ in fig. 20) behaves as. 
an axillary zooecium. 
A. kirkpatricki thns seems to unite the characters of more than one o-enus 
which may be explained by the assumption that Amastigia represents, better 
than any other genus, the ancestral character of the Scrupocellariidse, and 
that its species show approaches in various directions to those of other o-enera. 
I have suggested above that the pluriserial condition was probably antecedent 
to the biserial condition ; and on this view A. kirkpatricki is a species which, 
