452
D.M. COHEN:
historical and ecological. Both have common ground in the questions posed above. The
raw data for both approaches is the same, but the objectives are very different.
Historical zoogeography traditionally has described distribution patterns of phyletic
groups (i.e. species, genus, family or other taxonomic category) with the objective of
applying these data to understanding phylogeny. Examples may be found in the taxonomic
revisions cited later in this paper. In the historical-phyletic approach the taxa analyzed
may live in different habitats, have different ecological requirements and varying modes
of dispersal. The unifying feature is phylogeny, which is rigidly formalized through the
hierarchy of taxonomic classification. Ecological information, when considered at all, is
not accorded similar treatment, in some instances because it is simply unavailable, in
others because there is no adequate method of coping with it within the analytical scheme.
Considerable emphasis is placed on looking to the past to explain present distributions.
Although some attempt is made to use data accruing from the methods of historical
zoogeography for predictive purposes, this is a subsidiary objective.
Ecological zoogeography is concerned with environmental parameters and distributions, both of individual species and of ecological assemblages of species. One of the
earliest uses of the ecological method in fish zoogeography was that of MURRAY and
HJORT (1912), who stressed the importance of changes in relative abundance of species
taken by bottom trawl in different biogeographic regions of the North Atlantic shelf
and slope. The geographical distribution of broadly defined complexes that include
species with commercial value has been discussed by RAss (1959, 1965). More recently,
formal approaches to the problem of characterizing faunas have been given by FAGER
and LONGHURST (1968) for demersal fishes in the Gulf of Guinea, DAY and PEARCY
(1968) for demersal fishes off the coast of Oregon, and EBELING et aI., (1970) for deep
pelagic animals off Southern California. In this ecological approach, numerous taxa may
be considered, but all are generally members of the same community. The unifying
feature is habitat, which is rigidly selected by the sampling program. Taxonomic information is necessary but of secondary importance. Distributions are described chiefly in
terms of existing environmental parameters, their origins are rarely questioned. High
predictive values has been predicated (LONGHURST, 1969) and is perhaps the chief contribution that zoogeography can make to the rational exploitation of marine resources.
Although reasons for studying plant and animal distribution are diverse, the approaches have common ground in the data necessary to adequately answer their questions,
and it seems obvious that early emergence of what MACARTHUR and WILSON (1967) call
a "truly comprehensive theory of biogeography" will be of advantage to all.
The approach taken in the following discussion of Indian Ocean fish has been
dictated by the limited amount of accurate taxonomic data in revisional studies, which
can serve as sources for describing distribution patterns. Ecological information varies
from virtually non-existent for some kinds of shore fishes to fairly adequate for some
oceanic species. I have attempted to compare species or genera that are found in roughly
similar habitats, although the groupings are very broad indeed. Various authors (for
example PARIN and BEKLEMISHEV, 1966) have proposed models, for purposes of fish
zoogeography, that describe spatial and temporal variation in the location of a species.
These kinds of data are important in explaining distribution, but except for some commercially important species, there is little such information on the Indian Ocean fauna.
Temperature is generally acknowledged as the single most important determinant
of fish distribution. The 20° surface isotherm in winter generally coincides with the limits
D.M. COHEN:
historical and ecological. Both have common ground in the questions posed above. The
raw data for both approaches is the same, but the objectives are very different.
Historical zoogeography traditionally has described distribution patterns of phyletic
groups (i.e. species, genus, family or other taxonomic category) with the objective of
applying these data to understanding phylogeny. Examples may be found in the taxonomic
revisions cited later in this paper. In the historical-phyletic approach the taxa analyzed
may live in different habitats, have different ecological requirements and varying modes
of dispersal. The unifying feature is phylogeny, which is rigidly formalized through the
hierarchy of taxonomic classification. Ecological information, when considered at all, is
not accorded similar treatment, in some instances because it is simply unavailable, in
others because there is no adequate method of coping with it within the analytical scheme.
Considerable emphasis is placed on looking to the past to explain present distributions.
Although some attempt is made to use data accruing from the methods of historical
zoogeography for predictive purposes, this is a subsidiary objective.
Ecological zoogeography is concerned with environmental parameters and distributions, both of individual species and of ecological assemblages of species. One of the
earliest uses of the ecological method in fish zoogeography was that of MURRAY and
HJORT (1912), who stressed the importance of changes in relative abundance of species
taken by bottom trawl in different biogeographic regions of the North Atlantic shelf
and slope. The geographical distribution of broadly defined complexes that include
species with commercial value has been discussed by RAss (1959, 1965). More recently,
formal approaches to the problem of characterizing faunas have been given by FAGER
and LONGHURST (1968) for demersal fishes in the Gulf of Guinea, DAY and PEARCY
(1968) for demersal fishes off the coast of Oregon, and EBELING et aI., (1970) for deep
pelagic animals off Southern California. In this ecological approach, numerous taxa may
be considered, but all are generally members of the same community. The unifying
feature is habitat, which is rigidly selected by the sampling program. Taxonomic information is necessary but of secondary importance. Distributions are described chiefly in
terms of existing environmental parameters, their origins are rarely questioned. High
predictive values has been predicated (LONGHURST, 1969) and is perhaps the chief contribution that zoogeography can make to the rational exploitation of marine resources.
Although reasons for studying plant and animal distribution are diverse, the approaches have common ground in the data necessary to adequately answer their questions,
and it seems obvious that early emergence of what MACARTHUR and WILSON (1967) call
a "truly comprehensive theory of biogeography" will be of advantage to all.
The approach taken in the following discussion of Indian Ocean fish has been
dictated by the limited amount of accurate taxonomic data in revisional studies, which
can serve as sources for describing distribution patterns. Ecological information varies
from virtually non-existent for some kinds of shore fishes to fairly adequate for some
oceanic species. I have attempted to compare species or genera that are found in roughly
similar habitats, although the groupings are very broad indeed. Various authors (for
example PARIN and BEKLEMISHEV, 1966) have proposed models, for purposes of fish
zoogeography, that describe spatial and temporal variation in the location of a species.
These kinds of data are important in explaining distribution, but except for some commercially important species, there is little such information on the Indian Ocean fauna.
Temperature is generally acknowledged as the single most important determinant
of fish distribution. The 20° surface isotherm in winter generally coincides with the limits