76 THÉEI, NORTHERN AND ARCTIC INVERTEBRATES. I. SIPUNCULIDS. 
tail the tentacular erowns of the first-mentioned group, but, according to KOREN and 
DANTELSSEN, the two forms which constitute it, viz. Ph. abyssorum and Lilljeborgii, 
are provided, the former with from 20 to 24 tentacles, the latter with from 8 to 10. 
However, my own inspection of Ph. abyssorum has given rise to the opinion in my 
mind, that its tentaecles are in a low developmental state, presenting themselves as 
small, irregular knobs or elevations (Figs. 72, 73) round the oral disk. As to Ph. 
Lilljeborgii, the tentacles of which I have not been able to study, I believe the same 
to be the case, though in a higher degree still. Thus, at is seems to me, the shape 
of the tentacular crown does not afford any true character distinguishing the two 
last-mentioned species from each other, nor from the other doubtful species. It 
evidently presents a series of tramsitional stages, a thing of great scientific interest 
in itself, but not exactly convenient for the systemadtist. 
Nevertheless, with regard to the outer appearance, there exist some other dif- 
ferences which must form criteria of species. Thus, the Ph. abyssorum is in posses- 
sion of from 10 to 12 (or, according to my counting, 14), somewhat irregular rings 
of powerful hooks (Figs. 74, 75), while Ph. Lilljeborgii is totally devoid of hooks. 
Also Ph. glacialis appears to be well differentiated from the others by its very narrow 
proboscis, resembling thread in fineness and whipcord in texture, and which is pro- 
vided behind the oral disk with a broad girdle of hooks. Compare the Synopsis given 
above on Pages 59 and 60. 
I we now pass on to the retractor-museles and their points of attachment, it 
becomes evident that here too no true species characters can be given. In all the 
forms which belong to the "”abyssorum-section"', there are, as a rule, two ventral re- 
tractor-muscles, which have more or less grown together, in some cases so much so, that 
they seemingly present themselves as a single retractor. Referring to Pages 16—18, 
where I have treated more in detail the question of the retractors and their number, 
I merely wish to accentuate here that in the section of Gephyrea in question normal 
animals with only one retractor never occur, and that the statements of KOREN and 
DANIELSSEN with regard to Ph. glaciale and Sarsi are erroneous. In all the cases 
the two roots, which are remnants of the two separate, primary, retractors, embrace 
the nerve-cord. 
In order to adduce proof of my statement that the retractors are of but question- 
able value for the systematic arrangement of the species in question, I refer to the 
annexed table, which represents an excerpt of my annotations upon more than 200 
individuals of Ph. Sabellarie and improvisum, which I have had the opportunity of 
investigating. Seeing that the proboscis, as a rule, is more contracted than the trunk 
itself, I have thought it suitable to compare the post-anal portion of the trunk with 
that of the retractors. In the same table the degree of coalescence is indicated by 
approximate measurements of the length of the split between the retractor roots. 
