116 
PACIFIC SCIENCE, Vol. XIII, April, 1959 
by the damp layer of humus which cuts off 
all light and movements of air, further reduc- 
ing the comparatively small fluctuations of 
temperature and humidity that prevail 
throughout the forest” (Lawrence, 1953). 
The microclimate may change a little but 
presumably the animals living therein can 
meet these changes by retreating a little far- 
ther and deeper into the leafmould and even 
into the soil beneath. The annual rhythm of 
adaptation to climate is then vertical and not 
horizontal as in the case of, say, the larger 
vertebrates. For smaller animals, this is an 
inestimable advantage. In unfavourable con- 
ditions their route of escape is not barred by 
large intractable areas of a biotype in which 
they cannot survive. In short, they can adjust 
themselves to seasonal changes without leav- 
ing their very specialised niche. When condi- 
tions do change abnormally, they may still 
survive in islands in the forest floor formed, 
for example, by deep pockets of leafmould, 
and by fallen and rotting logs which protect 
the ground beneath. This is well illustrated 
where forests have been felled or burned off. 
Under the remnant logs, there is usually a 
damp area with a relict cryptozoic fauna, ab- 
sent in the immediately surrounding pasture 
or scrub. Examination will usually reveal iso- 
pods, millipedes, pseudoscorpions, opiliones, 
centipedes, endemic worms (found very rarely 
in New Zealand pastures, the pasture species 
being exotic ones), and, if the area is large and 
moist enough, amphipods. That land isopods 
tend to collect in such protected places 
has been recorded by Allee (1926) who has 
further shown that isopods will often conserve 
moisture still longer by aggregating together. 
This is of positive survival value since aggre- 
gated isopods lose moisture more slowly than 
isolated ones. 
In areas of remnant native bush, often the 
only places where endemic terrestrial isopods 
are present in any numbers are the ecological 
islands provided by decaying trees and logs 
lying on the ground, an effect noted by Dendy 
(1895) for the cryptozoic fauna generally. "It 
is far easier to find cryptozoic animals in 
partially- cleared localities, where they are col- 
lected together under the remaining fallen 
logs, than in virgin forest, where there is so 
much cover that the animals are widely scat- 
tered, and the search becomes very laborious.” 
On the east coast of the North Island, New 
Zealand, where there are many such remnants, 
it is common to find the conglobating and 
relatively thick-skinned isopod, Cubans danae, 
in the space between the log and the compact 
soil below, where there is a little but not a 
great deal more moisture than in the sur- 
rounding unprotected ground. The compara- 
tively thin-skeletoned and nonconglobating 
Styloniscus otakensis is only found in the much 
moister, rotting wood of the log, associated 
with ants, peripatus, termites, millipedes, and 
other invertebrates with similar environmental 
preferences. 
In Australian conditions, Birch and Clark 
(1953) have noted that amphipods occur in 
wet years "in the leaf litter of stands of timber 
in drier habitats such as Casuarina, Eucalyptus, 
and Angophora forests. But in most years the 
leaf litter in these forests is too dry for Talltrus 
to survive.” However, the fact that repopula- 
tion occurs in wet years suggests a survival 
population in the "drier habitats” or in close 
proximity. 
The fact that no great change in feeding is 
necessary would facilitate entry to the ter- 
restrial environment. The decaying seaweed 
diet of supralittoral amphipods is not mark- 
edly different from the decaying leaf diet of 
the terrestrial ones. Birch and Clark (1953) in- 
clude the amphipod Talltrus sylmtlcus 3.mongst 
the litter feeders or "decomposers,” com- 
menting that it feeds on leaves only after they 
have been on the ground for some months. 
Terrestrial amphipods, like supralittoral 
ones, will turn cannibalistic and carnivorous 
in laboratory conditions when other food is 
not available, but in normal conditions they 
seem to be ectophagous rather than preda- 
tory. Lawrence (1953) remarks that "the in- 
testine of Talltrus {Talltroldes) eastwoodae 
