report of Dr. W. Michaelsen on the fauna of South-West Australia, a
translation of which, as was recently stated in these columns, appeared in
the journal of the West Australian Natural History Society for July. Dr.
Michaelsen, who came here from Hamburg, together with Dr. Hartmeyer, of
Berlin, carried out a series of investigations in the south-west of this
State during 1905, and some of the results are embodied in this report,
the first portion of which is published in this issue.

The first portion of the report was published on Saturday; the following
is the remaining portion :—

Salt-water Animals.

The animals in salt water are uncommonly rare. In a puddle containing only
little salt artificially deepened at the edge of the dried up Hannan’s
Lake, near Kalgoorlie, we found only two species of Phyllopods of the
genera Apus (?) and Branchipus, and many larvae of a species of fly; in
the concentrated salt water of a lake on the Island of Rottnest we found
only myriads of fly-larvae, and one little beetle, that very likely got in
there by accident. Upon the whole salt is not unfavourable to the
development of animal life. We can see that in the rich fauna of the sea.
For while the steady cold of the polar seas, as well as the regular warmth
of the tropics, produce abundant animal life, we find a less abundant
fauna in the so-called temperate zones, with their change from warm to
cold. Also the intermediate condition between the almost constant salt of
the sea and the constant fresh water of rivers is very unfavourable to
animals. The estuaries of the coastal districts, as well as the salt lakes
of the interior, undergo very strong fluctuations in the proportions of
the salt, a great change between the concentration and evaporation during
the dry season, and the dilution during the rainy time. There are only a
few “euryhaline” species, which can stand this change, and so they
increase in such places, free from any struggle for existence and with
plenty of rich food, to an enormous extent as regards the number of
individuals. As a rule these animals have an uncommonly wide distribution
of genera, if not of species. Consequently as a characteristic of special
regions, such as that of S.W. Australia, they need not therefore be taken
into consideration. Furthermore, there is no barrier to the spreading of
them on the continent.

Moist-land Animals.

Under this name I include those animals which do not need clear fresh
water, but a medium constantly wetted with fresh water, moist earth,
rotten leaves, etc. The larger groups of animals belonging to this class
are the terrestrial Oligochaeta (Earthworms) and the land-planarians. Of
the smaller ones I only mention the Peripatus which we found to be a
closely allied species to that of East Australia. The land planarians
belong, as far as I can judge, to the genus Geoplana, distributed far over
the Eastern States as well as over the Malayan region, New Zealand, and
South America. This class of animals seem to agree in the methods of
spreading (but not in their geographical distribution) to the earthworms.
The latter I have thoroughly examined and will discuss them more
particularly late on.

The earthworms are the farmers among the animals. They stick to one place.
Only slowly, step by step, their independent spreading take place. Dry
land stretches, salty grounds, and seas are impassable barriers for them.
That is how the configuration of the continents and the seas, and the
climatic circumstances of passed geological periods, are ascertained by
their recent geographical distribution. Looking back from the latter we
can now construct those former barriers hindering their spreading as well
as the ways of spreading. The geographical distribution of the South-West
Australian earthworms can give us therefore, the best instruction of the
historical-geographical character of this district. Our collection of
earthworms contains 34 endemic species, besides a number of imported ones,
mostly European forms. These endemic species belong to five different
genera. As no earthworm was hitherto known in this district it is no
wonder that all these species are new ; for the distribution of endemic
species is as a rule very limited. However, all the genera also occur in
other States of Australia. The phyletically oldest genus Eodrilus,
represented only by one species, has a world-wide distribution in
scattered areas ; in Queensland, N.W. and Central Australia (in oases) are
to be found living single species of that genus. The other four genera are
Plutellus, Megascolides Notoscolex, and Megascolex, the phyletically lower
genera of the sub-family Megascolecinae derived from Eodrilus. The chief
dominion of these genera is Australia, including Tasmania. Here these
genera, together with their progenitor Eodrilus, predominate without
rivalry. Furthermore, they prevail in Ceylon, while a few of them as
outposts are found in India, in New Zealand and the Chatham Islands, as
well in the west of North America. S.W. Australia, therefore, belongs
certainly to the Australian region of terrestrial Oligochaeta, and has no
special character of its own nor nearer connections with the nearest extra-
Australian districts, than the Indo-Malayan and the Ceylon districts.
However, in slight details the South-Western Australian terrestrial,
Oligochaeta show some peculiarities, just as those of the other States of
Australia somewhat differ from one another. Firstly, the absence of the
genus Diporochaeta in S.W. Australia is remarkable ; that genus you find
so often in Victoria and Tasmania, and to a smaller extent in dispersed
colonies in North Queensland, India, and in New Zealand. South-west
Australia is in its fauna of terrestrial Oligochaeta somewhat like certain
districts of New South Wales and perhaps also of South Australia. The
different sub-districts of south-west Australia show some certain
peculiarities amongst them. The three principal genera Plutellus,
Notoscolex and Megascolex (the small fourth genus Megascolides may be
joined to Plutellus) have quite distinct limited dominions. Plutellus,
combined with Megascolides, occupies the whole district from Northampton
to Albany as far as its hydrographical conditions permit the existence of
earthworms ; Notoscolex is only to be found in the middle part from Perth-
York to Bridegtown, Megascolex only in the south part from the Harvey to
Albany. These different dominions are by no means caused by the
physiographical characteristics, as the degree of moisture of the soil
within their limits is not parallel to the lines of equal annual rainfall.
For instance, the dominion of Megascolex contains the locality of
Broomehill, the farthest outpost in the dryer inland, besides the very
rainy stretches of the southern coast. These dominions are only decided by
the historical geographical data, except on their common inland or N.E.
border, which is fixed by the general dryness. I will discuss these
historical geographical data in the next chapter.

Erdgeschichtliche “Conclusions.”

South-west Australia has only been connected, as regards the important
distribution of endemic terrestrial Oligochaeta (earthworms), with the
Eastern States of the Australian Continent, from which they have spread.
We can perceive three different phases of settling. The phyletically
oldest genus of Megascolecinae, Plutellus (Megascolides) occupied the
whole district. The phyletically younger genus, Notoscolex, and the
youngest Australian genus, Megascolex, could only take possession of
smaller districts, which differed from one another. At the time of the
immigration of these younger forms the passage to these districts seems to
have been a rather difficult one, for apparently the spreading obtained
from only few species which resolved themselves into a larger number of
younger species while spreading over a certain district. As the starting
points for the spreading of these two younger genera are very different,
the question is whether these temporary but distinctly separated settlings
have come on different roads into our district. Probably the configuration
of the surface has been, in earlier geological periods, much more
complicated, be it that the continent has run out towards the south-west
into separated peninsulas, or that temporary isolation of certain sub-
districts by separation of bigger isles has occurred.

The new knowledge about the further distribution of the genera of
earthworms, which concerns the whole of Australia, confirms my former
opinion that there are between Australia and the southern points of the
other continents (Africa and South America) no relations which can be
declared to depend only upon a direct connection between them. There is no
need for the supposition of an ancient great Antarctic Continent which
connected Australia, Africa, and South America as some scientific men
still suppose. In fact certain relations between the southern points of
these continents cannot be doubted ; but they do not depend upon a direct
connection of the land. They consist on the one part of euryhaline forms
for which the salt sea is no barrier, which can be transplanted by the
west wind drift over the stations on the different isles lying there
between one continent and another (littoral Oligochaeta : Lumbricillus,
Marionina, Enchytraeus, Michaelsena, and Microscolex) ; and on the other
part of ancient forms, which were pushed away by the phyletically younger
forms, developing and spreading in the larger continents of the northern
hemisphere and of the tropics. Such a form is the Eodrilus of south-west
Australia, whose generic companions have survived in the southern points
of Africa and South America, and on the islands of New Zealand, New
Caledonia, and Madagascar, which were separated in early geologic times
from the continents. Even in same [sic] out of the way districts north of
the equator some species of this genus occur, as for example in the
Cordilleras of Central America and Mexico as well as in the Kameroon
Mountains. The distribution of these forms in Australia does not entail
the immigration of them from the south. They are altogether wanting in the
well-searched districts of Tasmania, Victoria, and New South Wales (South
Australia cannot be included here because of the slight explorations made
there), that is to say just in the corner of the continent, which must
have formed the connection with this problematic antarctic continent. The
Australian localities, as well as the others of this genus, are typically
ancient. Species of this genus are found in the oases of Central
Australia, in the background behind the vast dry district in S.W., N.W.,
and N.E. Australia, as in other continents pressed against the wall by the
enormous development of phyletically younger genera. These Eodrilus
outliers might be put in connection with the other species of this world-
wide distributed genus by an earlier bridge between Australia and S.E
Asia. For the explanation of their origin there is no need for the
building up in vast oceans of enormous continents.

In the same manner the phyletically endemic earthworms of Australia point
out a direct land connection between Australia and South-East Asia. The
almost identical generic relationships of the earthworm fauna of Ceylon
(not in the same period in India) and Australia affords further proof. One
could be tempted to join Ceylon to the Australian earthworm region. Only
the existence of some rare Indian and Indo-Malayan forms in Ceylon
authorises a separation of the Cingalese earthworm region This

Relation of Ceylon to Australia

is the more striking, as we miss in the intermediate Malayan Archipelago,
even in New Guinea every trace of the phyletically older Megascolecine
genera, characteristic of Ceylon and Australia. On the other hand this
wide separation between the two districts of the elder Megascolecines,
Australia and Ceylon, is easily explained, if we examine carefully the
earthworm fauna of the intermediate districts. Here in the Indo-Malayan
region the younger offspring of the Megascolecine line prevails, that is
the genus Pheretima. Like other sub-families this genus Pheretima is very
prolific in spreading. (This power of colonising is demonstrated by the
exported forms. In many places of the old world Pheretima species have
been imported with plants by man. These imported forms have settled
themselves and pushed away the endemic earthworms of these places, at
least out of all cultivated localities. For instance, I find in a rich new
collection from Madagascar (certainly not belonging to the Pheretima
region) more species of Pheretima than of endemic Malagasy genera. Certain
circumstances which cannot be discussed here show with certainty that
these Pheretima species are not indigenous in Madagascar but imported.) It
has pushed away or even exterminated in the Indo-Malayan district the
elder Megascolecines, its ancestors.

The separation of Australia and Ceylon from the Indo-Malayan region was
only just in time to save them from being overwhelmed by the genus
Pheretima, and thus kept alive one of the most interesting Oligochaeta
fauna. It is still to be mentioned that some of the phyletically oldest
Megascolecine genera have spread far beyond the Indo-Malayan and Cingalese
regions, the genera Plutellus and Megascolides up to the Cordillera of
North America (and even into Pennsylvania?), the genus Diporochaeta to
India and on the other hand over New Zealand and into the Chatham Isles.

If we now regard the earthworm Fauna of New Zealand

we see that this territory must have been detached in a geological period
very remote from the separation of the continental masses of Australia and
S.E. Asia. Only some of the oldest Megascolecine genera Diporochaeta and
Megascolides, have been able to spread over this district ; the latter
found the South Island of New Zealand already detached and could only take
possession of the North Island. In the detached New Zealand region there
then developed out of the most archaic genus Eodrilus two special genera
Maoridrilus and Rhodedrilus, which do not occur outside this district.
Except that there is another Eodrilus-offspring, the sub-family
Octochaetinae, which only exists in India (and Ceylon?) is quite extinct
in the intermediate districts. As there is no trace to be found of these
Octochaetines in Australia, where they would have been protected from the
destructive competition of the younger Pheretima, we must suppose that the
connection, New Zealand and the Continent of Asia, very early broken, did
not touch Australia. New Zealand may very likely have had some connection
with the Indo-Malayan district, perhaps through New Guinea. The appearance
of some old Australian forms (Megascolides and Diporochaeta) is not
contrary to this view. These genera in earlier times much more widely
distributed, may have arrived by a longer route over New Guinea to New
Zealand, if not directly as a one-sided transmission of the fauna. Such a
one-sided transmission of the fauna (“einseitige Faunen zuschiebung”)
without mutual exchange of it may have gone on in the following manner :—A
part of the continent “A,” perhaps a peninsula of it, is separated, the
connecting land bridge sinking down beneath the level of the sea. This
part of the continent “A” becomes an island, on which only live animals of
the continent “A.” Farther on this island coalesces with another land “B.”
The fauna of the island derived from the continent “A” enters the land “B”
and mixes itself with the fauna [of] this land; but not a single animal of
the land “B” attains to the continent “A.” Such one-sided transitions of
animals without mutual exchange are not infrequent, and have often taken
place in the Indo-Malayan-New Zealand districts, with their strongly
varied and changed configurations in the course of the geologic periods.

The Marine Fauna of the South-West Coasts.—Physiography : In this part,
too, I shall begin with a description of the physiographical peculiarities
important to the animals.

The coast of S.W. Australia is for the greatest part a level coast
(“Ausgleichskuste”) with monotonous, long stretches of sands. Only in some
places the older rocks come to the surface on the coast and cause steep
cliffs. In the youngest geological districts, for instance at Cottesloe,
some miles to the north of Fremantle, these rocks and cliffs are formed of
Pleistocene limestone. But where the Australian tableland comes right down
to the sea, as on the south coast at Albany, gneissic and other primitive
rocks form steep stretches along the coast.

There are many considerable bays. These are especially marked and very
intricate on the south coast. Easterly from Cape Leeuwin, where the much-
worn edge of the Australian tableland is washed by the sea, you often find
almost circular bays, nearly enclosed from the sea by rocks, stretching
far out into the ocean, as for instance Princess Royal and Oyster Harbours
at Albany. Ordinary bays are to be found where the rounded limestone hills
form simple promontories as in the case of Geographe Bay behind Cape
Naturaliste. Where these lines of hills stretch far into the sea, deeper
but less rugged bays are formed as in Cockburn Sound, which opens to the
north behind the crest of Garden Island, which in recent times has been
separated at the south point from the continent. The estuaries of the
larger rivers also increase the number of bays. A peculiar formation seems
to me to occur in the greatest of all bays of W. Australia, Shark Bay.
Behind a steep cliff, rising from the deep sea (which not far from the
coast is 40 fathoms deep) to the height of more than 300ft. above the
level of the sea, the land slopes down inland into a shallow flat valley
with several slight longitudinal ridges. It has the appearance as if at
the breaking-down of the coastline on the outside the new front of the
remaining land had risen, and at the same time had exerted pressure upon
the hinterland, causing a sinking and a slight folding of the latter. On
the sea breaking through the weaker and lower parts of the outside edge
only the deeper parts of the valley were filled, while the higher portions
of the longitudinal foldings formed narrow peninsulas running in north and
south.

The features of the coast are still more complicated by sevaral [sic]
isles and groups of isles, generally the remnants of hills of the
Pleistocene limestone you find on the mainland (such as Rottnest and
Garden Islands). In places these islands are based on the primary rocks
like the isles near King George’s Sound on the south coast.

The bottom of the sea sloping from the coast shows in its character the
same undulations as the formation of the coast. There are vast submarine
stretches of loose fine or coarser sands, without any vegetation, and
almost without any animals, corresponding to the uninhabited sandy strand
of the coast. In other places on the firmer sands a rich flora of kelp
seaweed has grown, mostly Florideae and smaller Fucaceae, smaller than the
European Fucus. While the botanist may get a rich collection here, the
zoologist does not cover his expenses, for these seaweed “meadows” are
very poor in animal life, though not quite as poor as the plain sands. A
trifle more favourable to the life of animals are those sands on which the
stronger seaweed has settled. These plants with a hard stalk offer a good
support to many sedentary animals like sponges, bryozoa, and others, and,
also at the same time, protection and food for many wandering animals. But
a far richer fauna is found on the rocky and stony bottoms which form the
submarine continuations of the rocky coasts and islands, and which you
find with certainty in all the so-called “passages,” the breaking of the
sea through rocky barriers, for instance in the South Passage between Dirk
Hartog and the continent, as well as in the South Channel between Garden
Island and the continent. Here there is a paradise for the zoologists,
especially in the warmer northern districts, where the means of support,
of protection, and of food, are further increased by the strong coral
growths. A very favourable facies for abundant animal life also occurs in
the slimy mud which is to be found, especially in those parts of the bays
not swept by strong currents, as in the well protected and almost closed
bays of the south coast, for instance in. Princess Royal and Oyster
harbours at Albany, and also in the wider bays of the west coast, such as
Koombana Bay, at Bunbury, in Cockburn Sound behind Garden Island, and in
Sharks Bay. Especially striking is the number of the sponges and of the
ascidians in the slimy bottoms of the S.W. Australian seas.

The difference in the level of the sea at ebb and flow is only a very
small one. For Fremantle I read in the sailing hand book :—“Tide not
exceeding 33in.” That gives a relatively very small difference in the
height of the water ; in fact, I did not see a single ebb-tide during a
whole winter season which approached to the above amount. But the ebb-tide
is a little lower during the summer months, when the prevailing land winds
aid the reflux of the water. It is remarkable that they do not have a half
day’s cycle of tide but a whole day’s in south-west Australia. Through
this practical absence of tides the zoologist misses good opportunities
for working—for on other coasts collecting during the lowest ebb-tides is
the most productive—but also finds that what in other districts would
produce a rich and thoroughly characteristic fauna is here very barren.

The Fauna.—As the prevailing trend of the Western Australian coast is from
north to south, we must look for the principal geographical relations of
its littoral fauna in the same directions. These are all that can be
talked about before the exact scientific examination of the collections.
For in only these two directions the fauna, corresponding to the
variations in temperature of the water, shows striking differences in its
general habits. The discussion of these geographical relations falls
naturally into the questions concerning the antarctic, the sub-antarctic,
and the northern sub-tropical fauna, but these questions cannot possibly
be discussed without further preparation. While that part of the fauna
declining towards the tropical region contains, as a rule, very
characteristic and striking forms, the part towards the colder regions of
the south is characterised by the absence of these tropical forms. All I
can at present state is that there are certain forms, which point to the
antarctic or sub-antarctic relations of the southern districts, and which
characterise the southern fauna in a distinctive manner, such as the
Tunicates (Colella.) But the real discussion of this question must be
postponed until after a more exact study of the collections has been made.

The question of the direction of the tropical and. the sub-tropical fauna
of the coasts of Western Australia remains. Even the most superficial
study of the collections shows that the northern part is richer in
typically tropical animals than the southern, that the characteristic warm-
water-fauna gets much poorer in the south, and by-and-bye some of its
characteristic groups disappear altogether. Let us consider, firstly,

The Collections from Shark Bay,

the most northern point of our district, situated about the middle of the
western coast of Australia and cut by the parallel between 25 degrees and
26 degrees of southern latitude. Shark Bay has a fauna nearly tropical in
character. There are in first line several species of reef corals, which
give this character to the bay, several species of Turbinaria, partly
growing to large, imposing structures, and medium sized Pocillopora,
several Astraeidae, usually small, one Fungia, etc. A dead, worn piece of
a Madrepora, too, was found on the beach. But these corals do not appear
here, as far as I know, as reef-builders. The richest patch of corals
noticed is that on the south coast of Surf Point at the south end of Dirk
Hartog, where there were scattered colonies of Pocillopora and Astraeidae,
whose locations were separated from each other by several yards. Whether
these colonies approach more closely at the northern side of Surf Point,
which is washed by the breakers, as I presume to be probable, must be left
undecided, because the violence of the breakers made it impossible to
study that part of the coast. According to the literature at my disposal
the most southern of the real coral reefs of the Western Australian coast
are to be found near Anderson Point, which is somewhat to the north of the
tropic of Capricorn, and more than two degrees north of the northern end
of Shark Bay. Other tropical forms of Shark Bay are some composite thrifts
(Zoantharia), horn-corals (Gorgonidae), and a number of Alcyonaridae of
the genera Sarcophytum, Dendronephthya and Nephthya. Some specimens of the latter, which grow in a cauliflower-shaped colony of more than two feet in diameter, are the most gigantic Alcyonidae I ever saw. Of the other tropical animals of Shark Bay I may mention the small pearl-shell (Meleagrina imbricata), which maintains a small colony of pearl fishers at Denham on Shark Bay. The large true pearl-shell (Meleagrina margaritifera) is not endemic in the district of Shark Bay, but the Government attempts at introducing it have apparently been very successful.

If we follow the coast southward, we soon see a rapid diminution of these
tropical forms. Here again we examine the reef corals, which are so highly
sensitive to the temperatures of the surrounding water. Some of the Shark
Bay corals, as, for instance, the genus Fungia, were not to be found
further south. The genus Pocillopora has still some small scattered
colonies on the isle of Rottnest, in the Fremantle district. Some small
Astraeidae still go a little way further south, the farthest, as far as we
know, being in the Warnbro Sound, in 32 degrees south latitude. The most
southerly was a certain species of Turbinaria. We not only found this
species at Shark Bay and in the Fremantle district, near Garden Island,
but as far south as Koombana Bay, off Bunbury, in about 33 degrees south
latitude. A lucky haul brought up several living fragments of an
apparently very magnificent growth. Through this discovery the known
frontier of the distribution of the reef corals (not the real coral reefs)
on the Western Australian coast is carried some good six degrees of
latitude to the south. At this point the real southern termination of the
reef corals, probably, is nearly reached. The change in the fauna is still
increased by the disappearance of all the forms of animals whose existence
is due to the existence of the corals. The number of these is very
considerable, for the corals form large and strong shelters, which offer a
special facies at the bottom of the sea for a very peculiar fauna. I will
only speak of one example of the commensuals of the corals, a Polychaet
worm, a species of Eurythoe. The specimens from Turtle Island (near Port
Hedland, on the North-West Australian coast), are very magnificent, at the
most 18 cm. long. Several specimens from Shark Bay are all smaller than
the least of those of N.W. Australia, being at the most 7cm. long, while
the largest specimens taken at Fremantle were only 2 cm. long. No Eurythoe
has been found further southward.

The warm water forms, independent of the corals, also show a speedy
diminution to the S. We found only one single horn-coral (Gorgonia) so
plentiful in Shark Bay, a certain Melithaeid (Mopsella) in occasional
specimens further southward off the cliffs of Rottnest, in the Fremantle
district, and in Oyster Harbour, at Albany, on the south cost. Just as
striking is the disappearance of the Alcyonidae, which, in Shark Bay, grow
to such splendid proportions. Further southward we only found one single,
miserable specimen, a Nepthyid some few cm. long.

While the Western Australian coast shows us a very simple and rather
regular change from the tropical to the temperate zone, the faunistic
picture becomes much more complex if we regard the group of isles lying to
the west of the coast, viz., the Houtman’s Abrolhos. These isles are
situated at a distance of a distance of about forty sea miles westward of
Champion Bay, between 28.15 degrees and 29 degrees of south latitude. A
exact description of these isles and of their physiographic faunal
relations has been given by R. Helms. After studying the zoological
materials of the Abrolhos in the Western Australian Museum, in Perth, I
can confirm Helms’s assertion about the marine fauna of the Abrolhos. The
isles are abundantly surrounded by well-developed coral reefs, and show
otherwise in their fauna a thoroughly tropical character, and that just
opposite and only forty miles distant from a coast of which the fauna can
scarcely be called sub-tropical. In order to find a similar tropical fauna
alongside the coast we must travel about 6deg. of latitude to the north,
far beyond Shark Bay, up to the tropic of Capricorn. How is this contrast
between the faunal character of these islands and the coastal district at
the same degree of latitude to be explained? What is the foundation of this

Tropical Character of the Abrolhos?

Perhaps the currents of the sea have a certain influence in
differentiating the fauna. The southern port of the Western Australian
coast is much influenced by the cold S.W. current which, branching off
from the sub-Antartical west-wind drift, flows along this coast. The
northern part of the Western Australian coast is, on the other hand,
touched by a warm current of the sea, which coming down from the north
coast of Australia, rounds off the north-western corner of the continent,
and then goes on to the south. According to the map this warm current does
not even reach as far as Shark Bay. However, in fact, we must suppose that
it goes further southward and prevails over the cold south current unto
that point where the coast recedes a little to the east (Steep Point, a
little south of Shark Bay). Then if one follows the real southern
direction of the current, one finds that deflected at an acute angle from
the coast it strikes the Abrolhos Islands. The supposition of a northern
current going so far south makes the tropical character of the Abrolhos
comprehensible, and at the same time the colder temperature of the waters
of the coast lying opposite would be explained by a branch of the cold
south current going along the coast. Thus two contrary currents pass by
one another, a cold southern near the coast and a warm northern out to
seaward.

But the circumstances become still more complicated if one regards another
circulation of the sea water, that is the “cold swell” (kalter Auftrieb)
rising from the depth of the sea, which is generally found on the sub-
tropical west coasts of continents. On the Western Australian coast, too,
there are east winds prevailing, as a rule, for a considerable part of
each day. These push the warmer surface water seaward, and so cause the
rising of the colder bottom water to replace the driven water at the,
surface. This rising of the colder bottom water, the “cold swell” takes
place quite close to the coast, and of course influences mostly the line
of the coast and its fauna. On the surface the originally colder water
soon gets a higher temperature, so that by the time it has been driven by
the east winds over the space between the coast and the Abrolhos it has
grown warm. In this manner, too, the difference between the marine fauna
of the Abrolhos and that of the coast may be explained.

The effect of the rising “cold swell” is also very likely kept back from
these isles by the

Peculiar Configuration of the Sea Bottom.

in the district of the Abrolhos. The sea between the coast and the isles
is very shallow, upon the average about 20 fathoms, at the most 30
fathoms. Not until further seaward than the Abrolhos, just as on the
Continental coast further to the north and to the south, does the bottom
of the sea slope quickly to a much greater depth. The isles and reefs of
the Abrolhos lie on the top of a slightly submerged promontory. The rising
of the “cold swell” in consequence of the surface water being driven
seaward, can naturally only take place where the bottom of the sea rises
directly to the continental coast, not just in front of this grooup [sic]
of isles. For here the driven surface water gets filled up by the surface
water of the shallow sea lying behind. This water driven away from the
shallow sea of course must be filled up by the other water, and this must
come from the coastal stretch southward and northward of the shallow sea
above the submerged promontory. Here, it is true, you find the water of
the “cold swell,” but the way it must take as surface water before
reaching the Abrolhos is much longer than the straight line from the coast
to the isles. Furthermore, the rapidity of the drift of this surface water
becomes lessened by the barrier formed by the isles and their many reefs
only just below the surface ; this barrier has only comparatively narrow
and for the greater part very shallow passages for this drift. The time
during which the originally cold water stays as surface water above the
submerged Abrolhos promontory is comparatively long. During the intense
heat in the summer time when the seaward winds causing the “cold swell”
are generally prevalent, this time must be sufficient to bring the water
to the temperature necessary the development of the coral reefs.

This discussion of the problem concerning the striking tropical character
of the Abrolhos represents, of course, only an attempt to its solution. A
definite statement of the causes of these circumstances cannot be made
before a careful examination of the sea currents has been made and the
different temperatures of the waters in this district are fully known....