112 
PACIFIC SCIENCE, Vol. XIII, April, 1959 
and that "some species of Talitridae with 
truly terrestrial habits bear the impression of 
their terrestrial life.’’ Figure 1, based on avail- 
able records of distribution in the Pacific area, 
does not claim to be complete, but does give 
some idea of the extent of this fauna. 
The distribution of species in the Indo- 
Pacific generally, as at present known, reveals 
little pattern, due mainly to the fact that most 
records are of one or two species from each 
of many island localities. 
It is likewise practically impossible to m.ake 
anything of distributional patterns in Poly- 
nesia and Micronesia, but it is interesting to 
speculate how much of this apparently chaotic 
distribution, making due allowance for mis- 
identifications, is due to the voyages of the 
Polynesians. The custom of carrying ample 
stocks of fern root, taro, yam, gourds, and 
kumara on long journeys in the very humid 
and moist atmosphere of their canoes, makes 
them likely agents of distribution. When 
more is known of the amphipod distribution 
pattern, it may be interesting to attempt a 
reconciliation v/ith ethnological theories of 
migration and botanical theories of plant 
distribution by natives in the Pacific. 
One might predict an endemic fauna in each 
island or country where the terrestrial species 
occur, and where information is available this 
does seem to be the case. Chilton (1909) re- 
corded five new species of Orchestia not found 
elsewhere from Auckland, Campbell, Snares, 
and Stewart Island. Stephenson (1938) de- 
scribed Orchestia patersoni (as Talorchestia pa- 
tersoni) from Stewart Island. This has since 
been taken on Snares Island, Bench Island in 
Foveaux Strait, and Solander Island, all quite 
close to the original type locality, and there 
appear to be slight racial differences between 
some of these localities (Hurley, 1957). 
The New Zealand fauna includes six de- 
scribed species endemic to the mainland 
(Hurley, 1957} and at least two further 
species as yet undescribed, as well as the six 
species already m.entioned from the outlying 
islands. Of these, only Talitrus sylvaticus is 
known to occur elsewhere. If evidence can be 
produced to show that the Australian form of 
r. sylvaticus has not been introduced there in 
the last 200 years from New Zealand, then 
perhaps the solution to its distribution is to 
be looked for in the geological history of the 
two countries. It has been suggested that a 
New Zealand land m.ass was connected to 
Australia and farther north as late as Mid- 
Cretaceous time but probably not later, and 
had good connections much earlier which 
probably extended to take in various northern 
islands including New Guinea, New Britain, 
the Solomon Islands, and New Caledonia 
(Fleming, 1957). 
One further possibility is that of transpor- 
tation by birds, a theory of amphipod distri- 
bution for which Segerstrale, et al. (1954) has 
produced considerable geographical and ex- 
perimental evidence. While there is evidence 
of considerable migration of bird populations 
between New Zealand and Australia, it must * 
be said that the distances and the difficulties 
involved make this theory much less appealing 
in this instance than, for example, in explain- 
ing anomalous discontinuous distribution 
within New Zealand. 
Certain affinities have been suggested in the 
past between the New Zealand and South 
African faunas. At present, the terrestrial am- 
phipods of the two countries are considered 
specifically distinct, and the evidence, in my 
opinion, suggests parallel evolution from lit- 
toral ancestors in the two countries. 
There is one surprising fact in this distri- 
butional pattern. It is the complete absence 
of any records from Central and South Amer- 
ica, which is especially striking since the 
forest conditions there are just those which 
would be expected to support a large fauna of 
cryptozoic amphipods. Nor do any of the 
North American records suggest an indige- 
nous fauna. This absence does not seem due 
to lack of collecting. Allee (1926^?) in a com- 
prehensive study of leafmould fauna of Barro 
Colorado Island, Panama Canal Zone, lists no 
amphipods whatsoever. Other papers I have 
