NOTES:ON RETICULARIAN RHIZOPODA, 39 
forate, I have convinced myself by observations extending 
over many years, and that a large proportion of the Arenacea 
are likewise imperforate I fully believe ; but the exceptions 
in the latter case are so numerous and varied, that the arena- 
ceous group cannot be included as a whole in a sub-order of 
which the distinctive character is the imperforate test, and 
if omitted the term becomes at once misleading; therefore 
it is manifestly better to abandon a primary subdivision 
based solely on the condition of the investment in respect 
to perforation. 
The adoption of three sub-orders, instead of two, depending 
on shell-texture rather than on mere perforation, as latterly 
proposed by Von Reuss and adopted by Rupert Jones, meets 
the difficulty in part, but is still open to objection. On the 
one hand there is a considerable group of true Miliolz, our 
knowledge of which is much extended by the “ Challenger ” 
collections, that have rough arenaceous tests; and on the 
other, the large and important family Textularide is practi- 
cally unprovided for, inasmuch as it is sometimes truly 
arenaceous, sometimes hyaline and perforate, and sometimes 
externally sandy but with an inner perforate shell. Von 
Reuss meets this difficulty by dividing certain genera 
and placing the two halves in different sub-orders; thus 
Textularia appears as Textularia amongst the hyaline forms 
and as Plecanium amongst the arenaceous, and Bulimina in 
the same way as Bulimina and Ataxophragmium ; but this 
is cutting the knot rather than untying it, and even were so 
summary a method practically convenient, the proposal to 
divide a natural group like that comprising the Textularian 
and Bulimine types in order to meet the exigencies of an 
artificial distinction, is not one to be lightly adopted. 
Passing from sub-orders to families, even greater anoma- 
lies are apparent in Von Reuss’s scheme, especially amongst 
the hyaline forms. For example, Sperddina is found at almost 
the opposite end of the scale to the Rotaline genera; Nodo- 
saria, Cristellaria, and Polymorphina are placed in separate 
families, whereas, in point of fact, they form an absolutely 
continuous series; and Zeztularia, Bulimina, and Cassidu- 
lina are similarly separated. These appear to me fatal 
objections to the details of the classification. 
In the collateral English arrangement there is no infringe- 
ment of natural relationship apparent in the constitution of 
‘the families, except, perhaps, the association of Textularia 
and its immediate allies with Glodigerina and the Rotaline 
genera. Apart from this, its chief drawback is that the 
divisions are too large to be zoologically convenient. 
