10 ARVID RER: MOLANDER, ALCYONACEA. 
ratively insignificant, and the crowding together of the polyps not being so prominent; 
this endodermal net work of canals and cell-strings becomes strongly developed. In the 
case of Xenia wandeli, the endodermal cell-strings are tightly packed, whereby they 
have got the appearance of flat bodies lying between the coelenterons. A considerable 
proportion of the Xeniidae are provided with siphonozooids. In simular cases we also 
find in this family a direct connection, on the one hand, between the siphonozooids, and 
on the other, between the siphonozooids and the autozooids. These direct, horizontal 
canals are of secondary nature and of the same nature as the direct canals within Cla- 
vularia. This direct connection occurs above the endodermal cell-strings. 
Fam. Ålcyontidae has been divided by KÖKENTHAL (1906) into two sub-groups, 
Nidaliinae and ÅAlcyomiinae. With respect to the former, he notes, among other charac- 
teristies, direct connection between the coelenterons together with a system of canals 
and strings, which form an indirect system. This latter sub-group is characterized, a- 
mong other things, by a lack of direct connection between the polyps, but there exists 
a well-developed endodermal, indirect connection. 
A closer investigation of the canal system within Alcyoniidae proves, however, 
that, in this case, KUKENTHALI”s conception is not correct. 
If we now investigate the canal system within Nidaliinae, we find there, first of 
all, a fairly well-developed system of endodermal canals and strings — the indirect system. 
Further, there is seen, here and there, a direct connection between the coelenterons. 
KÖUKENTHAL looks upon these direct canals as constant, and, as such, of systematic value. 
But, from the investigations I have made on direct canals within other groups, I have 
reason to suppose — although I have not had the opportunity of investigating Nidalia 
or Nidaliopsis — that these direct canals have come into existence as a cönsrgenee of 
the crowding together of the coelenterons. 
Now we come to the Alcyoniinae. KÖUKENTHAL says that, in this case, the direct 
connection is wanting. In the majority of cases I found direct connection between the 
polyps here, too, and it is impossible to distinguish this in any way from the direct con- 
nection which KUKENTHAL found in the case of Nidaliinae. A specimen, Alcyonmium Val- 
diviae, described by KUKENTHAL (1906), has distinct direct connection, as appears from 
his description and figures. KÖKENTHAL says: »Schon oben an der Zweigspitze lassen 
sich kurze, entodermale Kanäle wahrnehmen, die einen Polyp mit der anderen verbinden, 
aber auch an Querschnitten vom unteren Teile eines Astes lassen sich solche Verbindungs- 
kanäle nachweisen». A rich collection of polyps on short lobes has produced these se- 
condary direct canals in this specimen. 
In the case of Alcyonium digitatum, too, I have found direct tubes, although less 
frequently, and chiefly in the case of stunted, ill-developed specimens, which have not yet 
put out branches or lobes, but have a Nidalia-like appearance (PI. II fig. 24). The mem- 
branous Alcyonium-forms, too, sometimes have direct canals, as I have mentioned above 
(PI. III fig. 31.) Of other groups connected with Alcyoniinae I have examined only 
Anthomastus, viz. Å. purpureus. This genus is characterized, among other things, by 
the presence of siphonozooids. The few large autozooids are situated at a considerable 
distance from each other. As in the case of Xemiidae with siphonozooids, we find here, 
