UNIVERSITY
OF FLORIDA
LIBRARIES

THIS VOLUME HAS BEEN
MICROFILMED
BY THE UNIVERSITY OF
FLORIDA LIBRARIES.

I

THE LESSER ANTILLES

AMERICAN GEOGRAPHICAL SOCIETY
MAP OF HISPANIC AMERICA
PUBLICATION NO. 2

THE LESSER ANTILLES

BY

WILLIAM MORRIS DAVIS, Hon. D.Sc., Ph.D.

PUBLISHED BY
THE AMERICAN GEOGRAPHICAL SOCIETY
OF NEW YORK
BROADWAY AT 156TH STREET

1926

COPYRoGHT, 1926
BY
THE AMERICAN GEOGRAPHICAL SOCIETY
OF NEW YORK

* DOUGLAS C. MCMURTRIE, INC.
NEW YORK, N. Y.

CONTENTS
PAB
INTRODUCTORY DESCRIPTION . . . . I
PHYSIOGRAPHY OF THE ISLANDS: A SYSTEMATIC
SEQUENCE OF ISLAND DEVELOPMENT . 12
A SIMPLE SEQUENCE OF FIRST-CYCLE ISLANDS 35
SABA . . . . . . . . . 35
THE SAINTS ......... ..... 37
REDONDA ................ 39
A SIMPLE SEQUENCE OF SECOND-CYCLE ISLANDS 40
MARIE GALANTE. . . . . . . 40
SOMBRERO ........... .... 41
A MoRE COMPLICATED SEQUENCE OF FIRST-CYCLE
ISLANDS ............. . 45
MONTSERRAT .............. 45
STATIA, ST. KITTS, AND NEVIS ...... 48
DOMINICA. . . . . . . ... 63
MARTINIQUE .............. 73
ST. LUCIA ............... 79
ST. VINCENT .... ........... 106
THE VIRGIN ISLANDS ........ 109
GRENADA AND THE GRENADINES . . .. I27
SECOND-CYCLE COMPOSITE ISLANDS . . .. 138
ST. CROIX ................. 138
ANTIGUA AND BARBUDA . . . .. 144
ST. BARTHOLOMEW, ST. MARTIN, AND
ANGUILLA . . . . . . .. 166
THE GUADELOUPE GROUP ........ 170
BARBADOS ............. 172
DETAILED ANALYSIS OF ISLAND FORMS .... 76
BIOLOGICAL RELATIONS ...... . . 185
PACIFIC AND ATLANTIC REEFS ........ 194
INDEX .................... 203

LIST OF ILLUSTRATIONS
FIG PA=
I Ideal section of a Lesser Antillean bank ..... 15
2 Map of Saba Island. . . . . . 36
3 One of the Saints . . . . . . 37
4 Map of part of one of the Saints . .. 38
5 Outline of Redonda. . . . . . . 39
6 Silver Hill, Montserrat . . . . . 45
7 Map of Montserrat . . . . . ... 46
8 Center Hill, Montserrat . . . . 47
9 MapofStatia ............ 49
10 Outline of Statia. . . . . . . 50
xI Map of southeastern part of St. Kitts . . . 51
12 and 13 Map of St. Kitts . . . . . 52-53
14 Basseterre, St. Kitts, on shore slope of volcanoes 54-55
15 MapofNevis . . . . . . . 56
16 Nevis from the southwest . . . . 58-59
17 Nevis, northern side . . . . . . 6
18 Morne au Diable ............... 65
19 Map of Morne au Diable . . . . . 66
20 Southwestern coast of Dominica ..... .6869
21 Map of part of east coast of Dominica . . . 71
22 Cliffed headlands of southwestern Martinique . 75
23 Cliffed headlands of southern Martinique . 75
24 Cliff and stack of southernmost Martinique . 75
25 Map of coast of southwestern Martinique . . 76
26 Map of coast and coral reefs of eastern Martinique 77
27 Mud flow of southwestern St. Lucia . . 8-8
28 Grand Piton, St. Lucia . . . . .... 82
29 Petit Piton, St. Lucia . . . . . . 82
30 Delta plain of the Cul de Sac embayment, St. Lucia 83

LIST OF ILLUSTRATIONS
FIG. PA=T
3x Low cliffs of mid-west coast, St. Lucia ... 84-85
32 Map of east coast of St. Lucia . . . . 86
33 Highest headland cliffs on east coast of St. Lucia 87
34 Elevated cliff-base platform, near Wellington, N. Z. 87
35 Inner part of delta plain of Cul de Sac, St. Lucia 1o
36 Village of Soufriere, St. Lucia . . . o2-o3
37 Delta plain of the Roseau valley, St. Lucia . o5
38 Cliffed headlands west of Kingstown harbor, St.
Vincent ...... .... ...... o8
39 Part of St. Thomas, from the south ...... IIo- I
40 Spurs and islands on southwest coast of St. Thomas
112-113
41 Map of eastern end of St. Thomas and western end
of St. John ................ 114
42 Great Harbor, Culebra . . . . . .. 15
43 Map of Norman and Peter islands . . . 16
44 Map of Tortola and near-by islands ....... 117
45 Map of Virgin Gorda . . . . . .. 19
46 Cliffed island at entrance to St. Thomas harbor 120
47 Cockroach Island, looking east ........ 120
48 Cockroach Island, looking north . . . .. 120
49 Diagram of a partly submerged cliff ..... .128
50 Part of Union Island, one of the Grenadines . 129
51 Map of Cannouan Island . . . ... 130
52 Diamond Island, cliffed stack in the Grenadines .. 131
53 Map of southeast coast of Grenada ...... . 132
54 Ham bluff, St. Croix . . . . . .. 42
55 Lightly cliffed headland on north coast of St. Croix 143
56 Western end of St. Croix at Frederiksted . 144-145
57 Non-cliffed north coast of St. Croix at Christiansted 144
58 North coast of eastern St. Croix . . ... 45
59 Ideal section of an atoll . . . . .. 146
6o Ideal section of the uptilted and down-worn half of
the same atoll. .............. 146

LIST OF ILLUSTRATIONS
110 PA=s
61 Subdued mountains of volcanic rocks in southwest-
ern Antigua. . . . . . .. 48-149
62 Subdued spurs on south side of Willoughby Bay,
Antigua . .. . . .. . .. . 158
63 Rough plan of part of west coast of Antigua . 58
64 Rough plan of southwest coast of Antigua . .. 159
65 Map of St. Bartholomew . . . . . 166
66 Map of St. Martin . . . . . .... 67

PLATES
PLAT PAMB
I Barbados: rolling cane land near summit, looking
south ................ opp. 6
II The uplifted limestone is land of Marie Galante
opp. 38
III The younger volcanic cones of Montserrat. opp. 38
IV The southern part of Statia, looking east by south
opp. 38
V The central volcanic range of St. Kitts, looking
northeast . . . . . . . opp. 54
VI St. Kitts from Basseterre roadstead . . opp. 54
VII St. Kitts ................ opp. 54
VIII Isolated residual hills in the passage between St.
Kitts and Nevis, looking east . . opp. 55
IX Mont Carbet and the cliff spur ends of Martinique,
looking east . . . . . . opp. 78
X East coast of Martinique, looking north opp. 78
XI East coast of Martinique, looking southeast opp. 78
XII Castries harbor, St. Lucia, looking north . opp. IIo
XIII Kingstown harbor, St. Vincent looking west opp. IIo
XIV General view of harbor, St. Thomas, Virgin Islands
opp. 10o
XV Island of St. Thomas: north coast from the crest,
looking northeast ....... ... opp. nII
XVI Part of St. George's and the southwestern end of
Grenada ...............opp. 134

INTRODUCTORY DESCRIPTION

GENERAL RELATIONS OF THE ISLANDS
The insular chain of the Lesser Antilles, including
twenty-five larger islands and an uncounted number
of smaller ones, forms a curve about five hundred
miles in length between Porto Rico in latitude 180 and
Trinidad in latitude o10. It is interesting to note
that they afford the only instance on the Atlantic
side of North America of an island festoon, or arc,
such as is repeated several times off the Pacific coast
of Asia. These arcs undoubtedly express an impor-
tant deep-seated deformation of the earth's crust.
In their local structural relations the Lesser Antilles
are mostly composed of volcanic rocks, and some of
the islands are still afflicted with explosive eruptions.
These eruptions, in spite of their destructive violence,
add little to the island volume. An uncounted num-
ber of such eruptions must have taken place in the
building up of the islands from their ocean-bottom
foundation, which appears to have been subsiding
during and after their growth and which may have
long ago formed a land bridge to South America.
Most of the islands show submarine enlargement in
shallow banks, probably formed with the aid of coral
reefs as will be explained below. Several of the lower
islands consist largely or wholly of calcareous strata,

2 THE LESSER ANTILLES
which appear to represent former banks, now up-
raised or uptilted and more or less eroded and
abraded. The islands which are most uplifted and
eroded now exhibit a volcanic foundation which,
originally an island, was entirely covered with cal-
careous strata during a long period of subsidence
that preceded the later uplift. I visited the islands in
October and November, 1923.

THE CLIMATE IS TROPICAL
The climate is characterized by ardent insola-
tion and a prevailingly high temperature; yet
the sensible temperature in the shade is not excessive
when the northeast trade wind is blowing. The tops
of fair-weather trade-wind clouds are often seen ris-
ing, far away, above the horizon. They thus, like
base-down cumulus clouds over a prairie horizon,
demonstrate the rotundity of the earth on a much
larger scale than the familiar sight of ships hull-down
at sea; for their base below the horizon may be Iooo
or 200ooo feet above sea level. The Lesser Antilles lie
far down on the track of the Atlantic tropical hurri-
canes of late summer or early fall, and these scourges
rival volcanic outbursts in their destructive effects.
But the islands well to the southeast, by very reason
of their position, are now of service to their larger
neighbors on the northwest in giving early radio
warning of approaching storms. The islands also
show the value of land obstructions in provoking
rain from the trade winds; for while the lower ones
not infrequently suffer from drought, the loftier ones

INTRODUCTORY DESCRIPTION 3
have a plentiful rainfall. Unfortunately, it is pre-
cisely on the lower islands, where agriculture is most
extensively developed, that deficiency of rainfall is of
most serious import. Yet when rain does fall there it
may be heavy, beciyse condensation begins at a high
temperature, where a moderate measure of cooling
precipitates a large volume of water.

THE EVENTFUL HISTORY OF THE ISLANDS
Historically and politically the Lesser Antilles
are of importance because of their early discovery
and settlement. Columbus came upon several
of them before the end of the fifteenth century;
and some were occupied by European settlers in
the third or fourth decades of the sixteenth cen-
tury. They have been acquired individually by
several European powers; and as their value in-
creased with the increase of population and products
their ownership has often been bitterly disputed,
especially among the English, Spanish, and French
through the last half of the eighteenth century and
the early years of the nineteenth. During these years
some of the islands changed ownership several times,
but they were eventually-at least until our war with
Spain in 1898 and our purchase from Denmark in
1915 brought some of them into American owner-
ship-divided among five European nations; hence
nowhere else in the Western Hemisphere are so
many European languages spoken in so close con-
tiguity. The Guadeloupe group and Martinique,
separated by Dominica, as well as the smaller island

4 THE LESSER ANTILLES
of St. Bartholomew and the southern half of St.
Martin, are French; Statia, Anguilla, the northern
half of St. Martin, and the little cone of Saba are
Dutch; Vieques and Culebra, small dependencies of
Porto Rico, were long under Spnish rule; and the
western members of the Virgin group, including St.
Croix, were Danish. The other islands, a good major-
ity of the whole chain, are British possessions. It is
curious to note that on several of the British islands
formerly possessed by France nearly all the local
names are still French, many of them memorializing
early French governors or other officials; and that the
Creoles, or whites of pure French blood, still preserve
their own language and hold themselves rather aloof
from the English, as if they were of a somewhat
different social order. The negroes on those islands
still speak a French patois; but that will probably be
largely replaced in a generation or two by English
through the influence of the schools as well as because
of the greater value of English in official relations and
business transactions.
From a literary and romantic point of view, the
islands are of peculiar interest; if the question of
piracy were taken up it might be shown that their
insularity and their partition among several Euro-
pean powers promoted freebooting; but the story is
too long to be entered upon here.
The small-scale manner of acquisition, in which the
fate of each Lesser Antillean island seems to have
been determined separately, chiefly by naval battles,
affords a striking contrast to the large-scale manner

INTRODUCTORY DESCRIPTION 5
in which the ownership of great continental areas has
been settled; as when the whole of the St. Lawrence
region and a great part of the wilderness to the north
of it were transferred from French to English owner-
ship with the fall of Quebec and Montreal after the
middle of the eighteenth century, or as when the vast
Louisiana territory was acquired for the United
States by purchase from France early in the nine-
teenth century. As to the names "Leeward" and
"Windward" for the two administrative divisions of
the British group, no justification can be found in the
relation of the islands to persistent trade winds. The
Leeward Islands include Tortola and other eastern
members of the Virgin group, Sombrero, St. Kitts,
Nevis, Antigua, Barbuda, Montserrat, and Domin-
ica; the Windward include St. Lucia, St. Vincent,
and Grenada, with the little Grenadines. Northern
and Southern islands would have been more appro-
priate names.

THE AMERICAN, EUROPEAN, AND AFRICAN POPULATION
The islands were originally the home of the
Caribs, who in early times were almost exterminated
on various islands by cruel wars, such as have often
marked the encounters of civilized invaders with
savage natives the world over. A small number of
so-called Caribs still remain on St. Vincent; but
their chief survival is in shell heaps along the shore
and in such words as "Caribbean" and "cannibal."
Since the aborigines have practically disappeared,
the much more numerous modern inhabitants

6 THE LESSER ANTILLES
of the islands are chiefly negroes or colored popu-
lation with more or less negro blood. Whites are
in largest number on Barbados, where they con-
stitute about a fifth of the total; on Grenada they
are only about a fifteenth; and on Tortola their
proportion is lower still. While a small part of the
colored population have acquired a considerable
competence, and some of these an education, the
great majority are poor and ignorant. Those on
Vieques, next southeast of Porto Rico, where Spanish
influence still remains, are pitiably debased in their
manner of living; the contrast there to be seen be-
tween the fine residences of plantation owners or
managers and the wretched hovels of the laborers is
most deplorable.
The survival of superstition among the negroes is
illustrated by a story told me by a recent white
resident on the mountainous island of Tortola, the
chief British member of the Virgin group. A young
negress there was accused of murdering a child.
Before going to court, she killed a lizard and placed
its tongue between the sole of her foot and her shoe;
when a question was about to be asked her by the
magistrate, she prevented him from speaking by
bearing down on the lizard's tongue; and so she was
discharged. Whether or not the story is true, it was
generally believed on the island by those who heard
it. Such superstition, it is to be hoped, will gradually
yield to education. Certainly there was an air of prog-
ress in the groups of neatly dressed and intelligent-
looking children on their way to school, as I saw them

~\

PI.. I--Barbados: rolling cane land near summit, looking south,

*'

~i

-

INTRODUCTORY DESCRIPTION 7
on a morning drive on Grenada. On Antigua, where
an appearance of neatness prevails in many villages,
the negro families too often live in small, one-room
cottages, some of which are built of wattle. Barbados
is the best cultivated and most thriving of the
British islands; but from there and elsewhere there is
a continuous small stream of negro emigition,
mostly to New York.
Problems of island government are difficult of solu-
tion. They involve the development and mainte-
nance of a self-supporting and contented population.
Experienced officials are in charge of affairs, but their
experience has perhaps been gained on other islands
in other oceans; and it is to be questioned whether
such experience always includes a sufficient ac-
quaintance with West Indian psychology. For
example, on one of the islands the department of
Agriculture undertook a few years ago to aid the
small farmers by lending them moderate sums of
money to be expended in farm betterment and to be
returned in small annual payments; but some of the
money was spent for other objects, and in three years
the returns had been so inconsiderable that legal
proceedings were proposed as a means of collecting
the payments then overdue. Surely the trouble here
was by no means wholly financial, but largely
psychological; such trouble might be better met by a
psychologist trained in business administration than
by an administrator untrained in psychology.
The distribution of population on certain islands
presents several points of interest. It is only on the

8 THE LESSER ANTILLES
relatively low limestone islands like Antigua, Grande
Terre, and Barbados that the whole surface is well
enough occupied for the partition of estates and the
location of villages to be developed on an areal plan.
On the mountainous islands, like St. Thomas, the
chief American member of the Virgin group, Guade-
loupe and Martinique among the French possessions,
and Dominica, St. Lucia, St. Vincent and Grenada
among the British possessions, villages are almost
wholly limited to the shore and plantations to the
shore-facing valleys. Here, therefore, the treatment
should be circumferential and radial, rather than
real. The high interior of these mountainous islands
remains a forested wilderness, only here and there
cultivated in small patches. On St. Lucia, St. Vin-
cent, and Grenada, which have more indented shore
lines than Dominica and Guadeloupe, shore settle-
ments are wholly determined by the bays in front of
them and by the available valley-floor areas behind
them. The larger towns on all the islands are on the
leeward, or western, coast, where navigation was
safer for small sailing vessels on which the islands long
depended for oversea transportation; consequently a
visitor on a modern steamer, which touches only at
the chief ports, seldom sees the windward, or eastern,
side of an island.
The arrangement of highways is in accord with the
distribution of population. The low islands are
crossed by a network of roads, many of which are of
excellent surface; the high islands are served chiefly
by coastwise roads, along or near the shore, with

INTRODUCTORY DESCRIPTION 9
valley roads leading inland and occasionally crossing
over the axial mountain range. On several of the
more embayed high islands, however, the coastwise
roads are by no means level, for they lie a moderate
distance inland so as to pass behind the bay heads
and therefore have to make many ascents and de-
scents over the inter-bay spurs. On Dominica and
St. Lucia the coastwise roads fail to make the entire
circuit of the island by reason of the boldness of parts
of the coast.
Public water service is well developed on Barbados
and Antigua, so that nearly every village has a supply
piped to it from one or more central reservoirs; but
on the mountainous islands each chief town or village
has only a local water supply from its own valley.
Cisterns supplied from house-roof rainfall are also
largely in use on certain islands. At St. Thomas small
rectangular areas on hillsides are cemented to serve
as catchment surfaces; two such areas are indicated
in Figure 4o.

DIFFICULT ECONOMIC PROBLEMS
Economically the islands suffer from their small
size; from the small variety of their products,
which are almost wholly agricultural; and from
the consequent limitation of their industries. The
low limestone islands, the little-dissected lower
slopes of the younger volcanic islands, and the delta
plains in former embayments of the older volcanic
islands are quite generally given over to the cultiva-
tion of sugar cane. On certain islands limes are

ro THE LESSER ANTILLES
largely grown. A curious treatment of the limes is
seen when, after being gathered from the trees and
heaped on the roadside, they are very gently rolled
on a grater, one at a time, by negro women, to extract
a valuable essence from the outermost rind. Lime-
juice is an export of some importance, especially from
Montserrat and Dominica. Cacao has been a profit-
able crop, especially on Grenada, St. Vincent, and St.
Lucia; but recently the lower price of West African
cacao in the London market has placed the West
Indian cacao planters in a difficult position. Domin-
ica is reputed to be an important potential producer
of many kinds of fruit; but its rugged surface makes
road construction difficult and transportation of fruit
from tree to port without injurious bruising almost
impossible. There are no mineral products worth
naming, for Trinidad with its asphalt lake is not here
considered. Although a number of wells have been
drilled for oil on the central crest of Barbados and a
pipe line twelve miles long is said to have been laid
to large cisterns at Bridgetown, the chief port,
petroleum is not yet an important export.
Modern methods of ocean transportation and com-
munication have been disadvantageous to the islands
in some respects. The port formally called Charlotte
Amalia but generally known as St. Thomas, as it is
the only town on the formerly Danish island of that
name, was in the days of sailing vessels a center for
distribution of European goods to other islands and
for collection of their products for transport to
Europe; but transshipment largely ceased with the

INTRODUCTORY DESCRIPTION 1r
advent of steamships. However, St. Thomas con-
tinued for a time to be a port of call for steamers to
receive cable instructions from their owners; but
orders are now sent by radio to passing vessels. The
town remained for a time longer of some importance
as a coaling station, and many workers of both sexes
found intermittent occupation in carrying coal in bas-
kets on their heads, African fashion. But this grimy
work is now falling off with the arrival of fewer ves-
sels and with the increasing replacement of coal by oil,
which is pumped from supply vessels into tanks and
taken aboard by gravity. The raising of sugar cane
used to be profitable on St. Thomas before the aboli-
tion of slavery; but the abandoned plantations are
now overgrown with bushes, and the old windmills,
often set on hills the better to catch the wind, are in
ruins. The best fields of St. Thomas are now planted
with Guinea grass, an excellent pasturage for cattle.
At Castries, the chief port of St. Lucia, the work of
coaling vessels is still so eagerly sought for by the la-
boring population that a curious cry, begun by those
on the water front and taken up by others in the
streets, rises from the town when an incoming steamer
is signaled from the hill at the harbor entrance.

PHYSIOGRAPHY OF THE ISLANDS

A SYSTEMATIC SEQUENCE OF ISLAND
DEVELOPMENT
My chief interest in the Lesser Antilles lay in their
physiographic features. These I had opportunity of
observing either briefly or deliberately during my
voyage, when stops varying from half a day to nearly
a fortnight were made on ten of the islands and views
of about as many more were had from the steamer's
deck. The observations thus made have been supple-
mented since my return by an examination of many
charts and by a study of pertinent descriptions. Thus
the islands have gradually come to take their place
in my mind as members of a developing sequence ;
so that, the sequence once being known, the general
features of any island may be concisely learned from
the stage it occupies in the sequence; and, the general
features being thus learned, the special features can
be easily apprehended as individual modifications of
the general features.
The sequence in its simplest form begins with a
volcanic island, built up from a slowly subsiding sea
floor by a succession of active eruptions; after erup-
tive growth ceases, the island gradually diminishes by
erosion as well as by subsidence, while a lagoon-
i See, for a brief statement of their origin, W. M. Davis: The Forma-
tion of the Lesser Antilles, Proc. Natl. Acad. of Sci. Vol. Io, 1924. pp.
205-211.

ISLAND DEVELOPMENT 13
enclosing barrier reef is built up around it; and finally,
when the island is completely submerged, it is suc-
ceeded by an atoll, essentially as imagined by Darwin
in his theory of coral reefs. But this simple sequence
is affected by complications of three kinds. In the
first place, an island instead of being completed in
small size by a relatively short series of eruptions
from a single vent, may, after subaerial erosion is
more or less advanced on the earliest-formed cone,
resume its eruptive growth from the same or from a
new vent and become a larger composite island. In
the second place, a simple or a composite island and
its reef system may be uplifted or uptilted at any
stage of its development by deformational forces,
which thereupon interrupt the first-cycle sequence
and introduce a second cycle; and the island thus up-
raised will then be again characterized by erosion,
subsidence, and reef growth during the new cycle of
development upon which it thus enters.
In the third place, on recognizing that the initial
eruptive formation of different islands took place at
various times from early Tertiary almost to Recent,
it will be understood that, whatever the stage of
either cycle reached by an island when the Glacial
period intervenes, the waves of the lowered and
chilled ocean, exemplifying certain postulates of the
Glacial-control theory of coral reefs as stated by
Daly, first truncate the encircling reef, because its
corals are weakened or killed by the chilled waters;
then continue their abrasive work by planing down
the lagoon floor wherever it is laid bare, thus convert-

14 THE LESSER ANTILLES
ing it into a rimless calcareous platform; and even-
tually, though for a short time compared to that
needed for the mature dissection of an island, attack
its headlands and cut them back in cliffs. Next, as
the ocean rises to its normal level in Postglacial time,
the abraded calcareous platform will be more or less
aggraded, chiefly by organic detritus, and converted
into a shoaling bank; but, inasmuch as the Lesser
Antilles are believed to stand in the marginal belt of
the Atlantic coral seas where, as will be explained in
a later section, the ocean temperature has not been
very favorable to reef growth in Postglacial time, no
new bank-border reefs but only discontinuous, island-
bordering reefs are developed. Coral growth in the
Lesser Antillean chain is therefore now limited for the
most part to discontinuous reefs along or near the
island shores, well back from the bank border. The
bank border, therefore, usually remains as rimless as
when it was abraded. These relatively new and dis-
continuous bank reefs must be regarded as feeble
novices compared to the stalwart veterans of the
Pacific coral seas, which are now still vigorously con-
tinuing the growth that was begun in Preglacial time.
They were not truncated in the Glacial period,
although they were probably more or less dissected
by solvent erosion when the ocean was lowered. It
is only in the marginal belt of the Pacific coral seas
that feeble novice reefs are found, like those of the
Lesser Antilles.
Thus interpreted, the circum-insular banks of the
Lesser Antilles are complex structures, as shown in

ISLAND DEVELOPMENT 15
Figure I. Their foundation must be in nearly all
cases a volcanic mass, AF, which should include a
part of the originally submarine slope, AB, below sea
level, S1, at the time when the volcano became ex-

Fig. i-Ideal section of a Lesser Antillean bank.
tinct, and a part of the originally subaerial slope,
EM, which was more or less eroded before it was
submerged, and possibly also an early-cut cliff and
platform, ECB. The greater part of the terrace-like
mass of the banks will consist of a barrier reef, RR,
with its external talus deposits, TT, and its enclosed
lagoon deposits, LL, the latter being extended in-
wards by the delta deposits, DD, in embayed valleys,
YY. But a surface fraction of the reef and its lagoon
floor will have been cut away by low-level abrasion
in the Glacial epochs, with respect to the lowered
ocean, S3, and thus a gently inclined calcareous plat-
form, PM, will have been produced, widened inward
by a cliff-base belt of abraded volcanic rocks, MV,
and outward by a detrital embankment, KK. In
Postglacial time the abraded platform will have been
aggraded by novice reefs, N, and detrital deposits,
GG, into the present bank. over which the spurs of

T6 THE LESSER ANTILLES
the island will be cut off in the plunging cliffs, FV.
The bank alone is a matter of observation by sound-
ing; the abraded platform at a small depth below the
bank surface and the sloping volcanic foundation at
a greater depth are matters of inference. The three
terms, foundation, platform, and bank will be used
as thus defined throughout this essay. It is easily
conceivable that platforms of low-level abrasion may
have been produced at successive levels in successive
Glacial epochs around a progressively subsiding
island and that then they may have been aggraded in
successive non-Glacial epochs; but this complication
of the problem is not further considered.
The easy possibility of thus imagining the inferred
complex structure of the submarine banks around the
Lesser Antilles has of course nothing to do with the
truth of the theory from which the structure is
deduced; but the truth of the theory can be better
tested if these and all its other consequences are
clearly imagined.
In order to give fuller warrant for certain state-
ments made in the preceding paragraphs, several
sections of a general nature will be here introduced.
The first sets forth in brief form a climatic classifica-
tion of volcanic islands, so that the larger relations of
the islands belonging in the above-mentioned "mar-
ginal belt" of the coral seas may be better under-
stood; another describes the position and breadth of
the marginal belts of the Pacific and Atlantic oceans
as defined by their islands; a third shows that oceanic
volcanic islands usually exhibit the effects of sub-

ISLAND DEVELOPMENT T7
sidenre; and a fourth qualifies a broad statement
made in the first regarding the absence of cliffed
islands from the coral seas.

A CLIMATIC CLASSIFICATION OF ISLANDS
If volcanic islands are formed in the frigid seas they
will bear glaciers; if in the warmer parts of the torrid
seas they will be encircled by coral reefs; if in the
intermediate cool seas they will neither bear glaciers
nor be encircled by coral reefs. The limits of these
seas are subject to change in consequence of geologi-
cal changes of climate. Thus if islands are formed in
the cold seas adjoining the frigid seas, they will, like
the Faroes, show the effects of glacial erosion during
the Glacial period, although they bear no glaciers to-
day. Similarly, islands that stand near the margin of
the warm seas and are today more or less completely
reef-encircled will show the effects of having lost their
reefs during the Glacial epochs of the Glacial period
when the ocean was somewhat chilled and its surface
was somewhat lowered. Just as the islands of the
cold seas should exhibit forms due to normal erosion
in Preglacial or Interglacial epochs, modified by
glacial erosion in the Glacial epochs and little affected
by Postglacial changes, so the islands in the marginal
belt of the warm or coral seas should exhibit forms
due to Preglacial or Interglacial erosion, similar to
those of reef-encircled, coral-sea islands, but modified
by low-level abrasion and by erosion with respect to
the lowered ocean in the Glacial epochs when reef
protection was withdrawn; and the latest abrasional

i8 THE LESSER ANTILLES
modifications thus introduced should be little modi-
fied by Postglacial changes.
Islands in the frigid and the cold seas need not be
further considered here; but the features of those in
the three other seas demand attention. St. Helena
in the South Atlantic, recently examined by Daly,2
is a typical example of a maturely sculptured island
in the cool seas. It has been elaborately dissected by
its short but steep streams and severely abraded on all
sides by the unobstructed ocean waves. The work
done by the streams in eroding their radial valleys
appears to be comparable with that of the waves in
cutting back the shore cliffs, as if both destructive
processes had been at work continuously since the
island was formed. Various reef-encircled islands in
the torrid Pacific are equally typical of the true coral
seas, for the absence of cliffs on their spur ends shows
that they have long been reef-protected.
On the other hand, certain islands standing farther
from the equator show features that in part resemble
those of the always-reef-encircled islands of the true
coral seas and in part resemble those of the never-
reef-encircled islands of the cool seas. The most
significant feature of the first kind is a circum-
insular submarine bank, so like the lagoon floor of
reef-encircled islands that it may be regarded as
having been formed as such a floor and only sub-
ordinately modified by low-level abrasion in the
Glacial epochs and by aggradation in Postglacial
SR. A. Daly: The Geology of Ascension and St. Helena Islands,
Geol. Mag.. Vol. 59. 1922. pp. 146-156.

ISLAND DEVELOPMENT 19
time. The most significant features of the second
kind are the punging cliffs around the island shore,
the base line of which appears to lie at or a little
below the inner margin of the submarine bank.
These cliffs may be of greater or less dimensions
according as their islands were exposed for a longer
or shorter period to low-level abrasion; that is,
according as the islands stand on the polar or the
equatorial side of their marginal belt. Postglacial
features include only small changes due to wave work
along the present or normal shore line, to bank aggra-
dation, and to reef upgrowth. It is noteworthy, how-
ever, that apart from shore fringing reefs, the reefs
associated with the banks of cliffed, marginal-belt
islands today are more or less discontinuous and
that, instead of being built up from the outer margin
of the bank, they rise toward or to sea level from the
middle or inner part of the bank; they will therefore
be called bank reefs, as above noted, to distinguish
them from true barrier reefs of the coral seas. The
fact that the discontinuous reefs of marginal-belt
islands thus stand, like timid novices, well back from
the outer edge of the bank, instead of rising, like the
stalwart veterans of the true coral seas, from the
edge of the bank suggests that the ocean temperature
of the marginal belts has been less favorable to reef
growth in Postglacial time than it was in Preglacial
and Interglacial time. It is as if the temperature of
the water did not become high enough for reef growth
until after the ocean surface had risen so much as to
leave the outer border of the abraded platform at a

20 THE LESSER ANTILLES
depth where reef-building corals could not establish
themselves. A contributing cause to the delay in reef
growth is found in certain islands, to be mentioned
below, in the eastern part of the marginal belts where
the dominant ocean currents come from cooler lati-
tudes and are therefore unprovided with floating coral
larvae. This cause of delay would not be present in
the western part of the marginal belts, where the
dominant currents come from the coral seas and
should therefore be well charged with coral larvae:
there, the delay in reef establishment and the result-
ant location of reefs well back from the margin of
their banks, as well as their frequent failure to reach
present ocean level, appear to be due wholly to the
failure of the Postglacial ocean to rise to as high a
temperature around these islands as it had in Pre-
glacial and Interglacial times.

THE MARGINAL BELTS OF THE CORAL SEAS
Although islands with cliffed shores and submarine
banks are not numerous enough to define the mar-
ginal belts of the Pacific precisely, they indicate that
the belts have a width of about 50 of latitude between
230 and 280 north and 280 and 320 south latitudes
in the western and central parts of the Pacific and
that they become confluent by looping across the
equator in the neighborhood of the Marquesas Islands
in the eastern half of the ocean. The most remarkable
series of such islands and banks is found in the north-
western extension of the Hawaiian chain. The banks
are from 10 to 30 miles in diameter and are seldom

ISLAND DEVELOPMENT 21
over 30 or 4o fathoms in depth. The reefs, so far as
they occur, stand back from the edge of the banks;
they may be called bank atolls if no central island
occurs, or bank barriers if a central island survives.
This series includes six bank atolls, three banks with-
out conspicuous reefs but with cliffed central islands
or pinnacles, five banks without reefs or central is-
lands, and'two intermediate forms. The facts that
some of these banks have no reefs and that all the
central islands are reduced to so small a size that they
have a continuously cliffed shore line without embay-
ments place these examples on the polar side of their
marginal belt. In the western North Pacific, several
islands south of Japan have well embayed shore lines
and cliffed inter-bay spurs with fringing reefs and
rather extensive reefless submarine banks.' These is-
lands presumably stand near the equatorial side of
their belt.
In the South Pacific, Norfolk Island, northwest of
New Zealand in latitude 290, is about 13 square miles
in area and 1o5o feet in height; it is strongly cliffed
and is partly bordered by a small fringing reef; the
smaller Phillip Island, also strongly cliffed, is three
miles away: both islands rise near the center of a vast
bank, 6o by 20 miles in extent, with depths of 20
fathoms near the island and of 40 or 50 fathoms near
the bank edge. Lord Howe Island, between Norfolk
Island and Australia in latitude 320 S., is smaller,
with cliffs 8oo feet high; it rises from a bank 12 by 8
W. M. Davis: Drowned Coral Reefs South of Japan, Proc. Nail.
Acad. of Sci., Vol. 9, 9a23, pp. 58-62.

22 THE LESSER ANTlULteS
miles in extent and as dee&p as the Notrflk Island
bank. Balls Pyramid, tot far away, is an extraor-
dinary stack or pinnacle, its height of IrS6 feet being
greater than its shorter sea-level diameter; it stands
on a bank three by seven miles in extent. Middleton
and Elizabeth Reefs are small bank atolls, north of
Lord Howe Island in latitude 29. The reduction of
these high islands to continuously diffed shores sug-
gests that, as in the case of the Hawaiian examples,
low-level abrasion has here been repeatedly at work
and that the islands therefore stand near the polar
side of their marginal belt. Raoul Island, in the Ker-
madec group, is embayed with ciffed spur ends and is
surrounded by a submarine bank: it probably be-
longs in the marginal belt.
The Marquesas Islands are cliffed and embayed,
and the few soundings that are charted near them
suggest the presence of circuminsular banks of
moderate depth; hence the members of this group are
provisionally regarded as standing in the cross-
equator loop of the Pacific marginal belts, between
the cool seas on the east where currents come from
the temperate zones of both hemisphere and the
true coral seas farther west where the same currents
have become warm enough for reef formation by
their long and slow passage through the torrid zone.
The delay in the establishment of reefs around these
islands may be due largely, as above noted, to the
absence of coral larvae in the great ocean current that
approaches the islands from the cool seas.
The marginal belt of the Atlantic will be more fully

ISLAND DEVELOPMENT 23
described in a later section. Suffice it to state here
that the temperate and torrid Atlantic as a whole
appears to be a cool reefless ocean, corresponding to
the reefless eastern Pacific; that the coral seas of the
Atlantic are limited to the Caribbean Sea and part of
the Gulf of Mexico; and that the marginal belt of
the Atlantic seems to pass from the eastern coast of
equatorial Brazil to the Lesser Antilles, with an
abnormal northeastward loop that takes in Bermuda.
Like the Marquesas Islands, the Lesser Antilles are
approached by an ocean current that comes from the
cool seas, where no reefs are present to furnish
floating coral larvae.

STABLE AND UNSTABLE VOLCANIC ISLANDS
It has been briefly stated in a preceding section that
the submarine banks which occur around the cliffed
islands of the marginal belts resemble the lagoon
floors of reef-encircled islands, modified by low-level
abrasion in the Glacial epochs and by subsequent
aggradation and reef growth. It may now be added
that the formation of such lagoon floors is believed to
be due to the slow accumulation of deposits, largely of
organic origin, upon slowly sinking foundations-
usually volcanic islands-from which the enclosing
reefs have maintained themselves at sea level by up-
growth, essentially according to Darwin's theory of
coral reefs. The main reasons for this belief are found
by comparing the expectable features of stable and of
unstable islands with the features of actual islands in

24 THE LESSER ANTILLES
the coral seas, the cooler seas, and the marginal belts,
in brief as follows.
Stable islands in the cool seas will, in their matu-
rity, be commensurably eroded and abraded, like
St. Helena; but in their later stages, when the
streams are shortened by the recession of the cliffs,
abrasion will gain the upper hand; the uplands of
subaerial erosion will become smaller; when they
disappear the island will be a cliff-rimmed stack or
pinnacle and eventually will be cut down to a shallow
submarine platform by normal abrasion, or to a
somewhat deeper platform by low-level abrasion or
to a shallow normal platform rimmed around by a
low-level platform. A fair number of partly dis-
sected and abraded volcanic islands, with shallow
banks of slightly different depths around them, and
hence such as might be produced around slowly sub-
siding islands better than around stable islands, are
known in the cool seas, especially in association with
the volcanic groups of the eastern North Atlantic;
but practically no examples of shallow banks of uni-
form depth, representing the platforms of completely
truncated stable islands are charted in the cool seas.
Inasmuch as volcanic islands, however ancient,
would if stable remain perpetually as shallow banks,
the rarity of such forms in the cool seas is noteworthy.
The various banks of the North Atlantic have differ-
ent depths, usually greater than 50 fathoms.
Among unstable islands of the cool seas, two classes
must be distinguished-rising islands and sinking
islands. Rising islands would present a succession of

ISLAND DEVELOPMENT 25
wave-cut benches. If such islands occur, they are
rare. Sinking islands would be benched and reduced
in size as they sank, and their diminished summits
would eventually disappear. Just before their dis-
appearance, they might show as slender stacks
rising over submerged platforms of less or greater
depth. Two remarkable examples of this kind are
believed to occur south of Japan: One is Sumisu-
shima, or Smith Island, 31%2 N., 1400 E., consisting
of two pinnacles, the larger one having a smaller sea-
level diameter than its height of 445 feet; soundings
of 147 and 306 fathoms are charted within three
miles. The other is Sofu-gan, or Lot's Wife, 30 N.,
i40oy2 E., which is again of less sea-level diameter
than its height of 326 feet; soundings of 1o5 and IIo
fathoms are charted two miles away. This rock was
first reported in 1788; hence it cannot be a spine of
the temporary kind that rose above Mt. Pel6; it must
be the result of advanced abrasion.
After subsiding islands have disappeared in the
cool seas, they might long escape discovery, except
where many soundings are taken. A few rather
deeply submerged cones are known in the Pacific;
and, as above noted, a considerable number of small
banks at depths of 50 or more fathoms are charted
in the North Atlantic; but all these submarine
eminences may be growing or extinct volcanoes not
built up to sea level, instead of former islands sub-
merged. Hence their bearing on the problem here
considered is uncertain.

THE LESSER ANTILLES

THE FORMATION OF ATOLLS
The possible occurrence of stable islands in the
coral seas is of especial interest in connection with
the formation of atolls. Such islands could serve as-
atoll foundations only if, as has been stated by Daly,'
they are of so ancient origin as to have been worn
down to low relief with deeply weathered soils when
the Glacial period supervened and only if they were
reefless in the Glacial period,' so that after being re-
duced to submarine platforms by low-level abrasion,
reefs might grow up around the platform borders and
enclose lagoon floors of very accordant depths. But
if all the numerous Pacific atolls have been thus
produced, it would be natural to expect that some
islands of less ancient origin would have been less
completely peneplained in Preglacial time and there-
fore less completely truncated by low-level abrasion
Sin the Glacial epochs; and these residual islands
would today survive in the center of barrier-reef and
almost-atoll lagoons; they would be cliff-rimmed, and
they could have no embayments of greater width or
of greater rock-bottom depth than could have been
excavated by low-level erosion in the Glacial epochs.
As the actual reef-encircled islands of the coral seas
do not present such features, and as the lagoon
floors they enclose show significant variations in
R. A. Daly: The Glacial-control Theory of Coral Reefs, Proc. Amer.
Acad. of Arts and Sci.. Vol. 51, 1915, pp. 157-251.
According to Daly, the absence of reefs in the Glacial epochs would
be due. in part at least, to the lowered temperature of the ocean; but,
under the conditions here assumed, that cause of reeflessness would be
inoperative in the true coral seas.

ISLAND DEVELOPMENT 27
depth,6 long enduring stability of reef foundations is
to be doubted.
On the other hand, if the volcanic islands of the
coral seas are unstable, some might rise and others
sink. The rising islands would bear reef benches of
less or greater width at various altitudes. A fair
number of such benched islands are known, especially
among the non-volcanic islands of the Australasian
archipelago. It is highly significant of the origin of
the reef benches that their limestones rest, in all cases
that have been closely studied, unconformably upon
the underlying foundation rocks; thus proving an
earlier episode of subsidence and reef formation be-
fore the later episode of elevation and reef emergence.
Other unstable islands in the coral seas might
subside either at a nearly uniform rate or inter-
mittently. If the subsidence were not too rapid, an
up-growing reef would be formed around an aggrad-
ing lagoon floor; the subsiding island would diminish
in size; its increasingly mature valleys would become
well embayed; the rock-bottom depth of the embay-
ment would come to be much greater than the
measure of ocean lowering in the Glacial epochs; and
the inter-embayment spur ends would not be
cliffed. Raiatea and Borabora, Matuku and Totoya,
which may be cited as typical islands of the coral seas,
as well as various others that might be named, exem-
plify these features in a striking manner. With further
subsidence an island would be reduced to one or
W. M. Davis: Coral Reefs and Submarine Banks, Journ. of Geo..
Vol. 26, 1918, pp. 198-223, 289-309, and 385-41.x

28 THE LESSER ANTILLES
more non-cliffed mountain-top islets, encircled by
an almost-atoll reef. The islets of Truk in the
western part of the Carolines and of Mangareva, or
Gambier, south of the Paumotus are of this nature;7
the absence of cliffs on their shores, in contrast to the
strongly cliffed little islands of the marginal belts,
confirms the view that the reduction of good-sized
volcanic islands to the present small size of almost-
atoll islets is due to subsidence of unstable volcanic
cones and not to the abrasion of stable cones.
After an island disappears by subsidence in the
coral seas, its former presence is still indicated by an
atoll. Herein lies one of the chief differences between
the coral seas and the cool seas; for in the cool seas
a sunken island leaves no memorial. The rarity of
volcanic islands in the cool seas of the Pacific in con-
trast to the abundance of atolls in the Pacific coral
seas may perhaps be thus explained; but as the con-
trast may also be explained by an original difference
in the number of volcanic islands formed in those
seas, no demonstration of subsidence is thus provided;
nevertheless, the contrast is still worthy of considera-
tion. If subsidence be accelerated after a barrier reef
has been formed, it will be drowned and be succeeded
by a fringing reef on the new shore line, as was long
ago perceived by Darwin although he then knew of
no actual fringing reefs thus associated with sub-
sidence. They are, however, abundant in the Philip-

Idem: The Small Islands of Almost-Atolls, Nature, Vol. o05, 1920,
pp. 292-293.

ISLAND DEVELOPMENT 29
pine Islands.8 One of the best examples is along the
west side of the non-volcanic island of Palawan,
where the drowned barrier reef and the unusually
deep lagoon behind it are both well defined by numer-
ous soundings.9 If rapid subsidence take place after
the atoll stage of reef upgrowth has been reached,
the atoll reef will be submerged. The reef may again
grow up the surface if its submergence is small; it
will be permanently drowned if the submergence is
greater. A remarkable group of submerged atolls lies
north of Fiji; some of their reefs may again reach the
surface.
All these features appear to me to contradict the
view that volcanic islands are as a rule stable and to
confirm the view that they are prevailingly unstable
and exhibit a decided tendency to subsidence. The
only fact which gives support to the view that such
islands are prevailingly stable is the moderate varia-
tion in the depth of reef-enclosed lagoons; this being
explained under the Glacial-control theory of coral
reefs by the assumption that, even in the region here
called the coral seas, Preglacial reefs were cut away
and their lagoon floors were reduced to a standard
depth by low-level abrasion in the Glacial epochs.
That explanation is unacceptable because the
absence of cliffed spur ends on the reef-encircled
islands of the coral seas makes it unreasonable; and
it is also unnecessary, because the variation of lagoon
a Idem: The Fringing Reefs of the Philippine Islands, Proc. Natl.
Acad. of Sci.. Vol. 4, I9i8, pp. 197-204.
Idem: Subsidence of Reef-encircled Islands, Bull. Geol. Soc. of
Amer., Vol. 29, i918, pp. 489-574.

30 THE LESSER ANTILLES
depths, although moderate in comparison with ocean
depths, is actually considerable; and in certain large
lagoons the present depth is, in spite of Postglacial
aggradation, even greater than the liberal measure of
ocean lowering in the Glacial epochs adopted in the
Glacial-control theory. The great lagoon enclosed by
the magnificent barrier reef around the island of
Tagula, east of New Guinea, is a striking instance of
this kind."
If the marginal-belt islands are considered alone,
most of their features may be explained by assuming
the islands to be either stable or unstable; but in view
of the evidence for their subsidence given by the
estimated rock-bottom depth of some of their embay-
ments and in view also of the more general evidence
given for prevalent subsidence by the islands of the
coral seas proper, the stability of the marginal-belt
islands seems improbable. It is made still more so
when the persistent occurrence of extensive shallow
banks around them is contrasted with the extreme
rarity of such banks in the cool seas; for if the
marginal-bank islands were stable and their banks
were based chiefly on platforms produced by the
abrasion of formerly larger islands, then the cool seas,
being of vastly greater area than the marginal belts,
ought to contain at least a fair number of similar
banks based on platforms produced by the abrasional
truncation of large stable islands there situated;
but the cool seas do not contain such banks. Hence
t* Idem: The Barrier Reef of Tagula, New Guinea, Annals Assoc. of
Amer. Geogrs., Vol. 12, 1922. pp. 97-151.

ISLAND DEVELOPMENT 31
the banks around the marginal belts must be other-
wise explained; and they are best explained as chiefly
formed by upgrowing reefs and accumulated lagoon
deposits around islands originally about as large as
the banks but much reduced in size by subsidence;
the reefs, the lagoon floors, and the island shores
being brought to their present form by low-level
abrasion in the last Glacial epoch, with small Post-
glacial modifications. It is certainly a remarkable
feature of the oceans that shallow banks of fairly
accordant depths, after being practically absent from
the vast areas of the cooler seas, are suddenly found
to be prevalent in association with the relatively small
number of islands that stand on the border of the
coral seas; and it is also highly significant that the
islands which rise within these banks are cliffed, while
the islands of the coral seas are not. When all these
considerations are viewed together, the evidence for
the prevalent, long-continued, and slow subsidence of
marginal islands and for the former upgrowth of lagoon-
enclosing coral reefs around them gains much strength.

CLIFFED ISLANDS IN THE CORAL SEAS
It has been briefly stated in an earlier paragraph
that the islands of the true coral seas are not cliffed
and that cliffed islands are found only in the marginal
belts of those seas and in the cooler seas. This state-
ment must now be modified; for it appears that vol-
canic islands in the coral seas also are usually cliffed
in their youth before protecting reefs are established
round them. The reason appears to be that the active

32 THE LESSER ANTILLES
outwash of detritus from the deepening valleys of
young non-embayed islands furnishes an abundant
supply of cobbles, gravel, and sand by which beaches
are formed all around the island shore. Any incipient
fringing reefs are thus smothered, the attack of the
waves on the shore is unobstructed, and cliffs fronted
by gently sloping rock platforms a little below sea
level are produced. The rock platforms are extended
a little seaward by detrital embankments, and the
width of a platform is usually five or six times as
great as the height of the cliff behind it.
Abrasion will continue as long as the island stands
still; but as soon as it subsides enough to embay the
valley mouths, the stream detritus will be pocketed
there; and then the firm face of the partly submerged,
or "plunging," cliffs will afford a fitting support for
corals to grow upon and in time to form a fringing
reef. If subsidence then ceases, the little bays will
soon be filled with deltas, and thereafter the detritus
will again tend to form circuminsular beaches and to
smother the reefs; whereupon the abrasion of a plat-
form backed by cliffs will again be instituted but at a
somewhat higher level with respect to the island mass
than before. On the other hand, if subsidence con-
tinues at a rate rapid enough to prevent the filling of
the valley-mouth bays and yet not rapid enough to
drown the corals, the fringing reefs will grow up as
the island diminishes in size and form an offshore
barrier reef, and the enclosed lagoon will be more or
less aggraded. After the island sinks below sea level
the reef will be an atoll.

ISLAND DEVELOPMENT 33
Reunion, in the southern Indian Ocean, is a good
example of a cliffed and beach-begirt island, out from
whose deep valleys an immense volume of detritus is
washed and offshore from which only a few small reef
patches are charted." Tutuila, the chief American
member of the Samoan group in the Pacific, appears
to have had its slopes deeply dissected by streams and
to have had its shore cut back in high cliffs fronted by
a wide platform while it stood in its reefless youth
several hundred feet higher than now. The absence of
portecting reefs at that time cannot be explained by
a lowering of ocean temperature, because in that case
many other islands in the coral seas of the Pacific
ought also to be cliffed, and they are not. The
absence of reefs around Tutuila then, as around
Reunion now, is best explained by the formation of
beaches composed of outwashed detritus. After ex-
tensive cliffing a slight submergence appears to have
embayed the valleys a little and thus permitted the
upgrowth of an offshore reef, now recognized by shal-
lower soundings on the outer part of the broad
circuminsular bank; but soon an acceleration of the
submergence drowned that reef, embayed the valleys
more deeply than before, and determined the occur-
rence of a fringing reef of later generation at the
higher level of the present shore line. The submer-
gence thus indicated cannot be advisedly explained
by Postglacial ocean rise, because the present bank,
which is undoubtedly aggraded to a less depth than
I Idem: Clift Islands in the Coral Seas. Proc. Nall. Acad. of Sci..
Vol. 2, x9I6, pp. 283-288.

34 THE LESSER ANTILLES
the supposed underlying rock platform, has soundings
of 6o fathoms or more back of the drowned reef."
Tahiti was sharply dissected and even more strongly
cliffed in its youth, especially on its windward coast,
than Tutuila; but there the offshore reef, which ap-
pears to have been established, like the now drowned
reef of Tutuila, on the cliff-base platform when sub-
sidence set in, has continued its upgrowth with con-
tinued subsidence, so that now, when the measure of
subsidence as indicated by the inferred rock-bottom
depth of the larger embayed valleys has reached 6oo
or 80o feet, it is a typical barrier reef. Here again
the absence of reefs during the former higher stand
of the island cannot be explained by a lowering of
ocean temperature, because in that case the other
islands not far away in the Society group should also
be cliffed, and they are not. Recently a pause in the
subsidence of Tahiti appears to have taken place; for
the formerly embayed valley mouths are at present
nearly all delta-filled, and the delta fronts have ad-
vanced into the lagoon and thus formed a confluent
belt of alluvial lowlands fronting the plunging cliffs
nearly all around the island.13
It thus appears reasonable to conclude that young
volcanic islands even in the warm coral seas experi-
ence cliff cutting during a part of their youth, before
subsidence sets in. A corollary of this conclusion is
that the maturely dissected, reef-encircled islands of
2 Idem: The Coral Reefs of Tutuila, Sciene, Vol. 53, 1921, pp. 559-
565.
a Idem: Les falaises et les rcifs coralliens de Tahiti, Ann. de Giogr.,
Vol. 27. I918, pp. 241-284.

ISLAND DEVELOPMENT 35
the coral seas, which today have non-cliffed spur ends
between their well developed embayments, were ex-
posed to abrasion when they were younger and stood
higher and that the cliffs then cut are now invisible
because they are completely submerged. It may
strain a reader's credulity to accept, even provision-
ally, so surprising a corollary; but that measure of
reason-guided credulity is more consistent with what
is now known of volcanic islands and the reefs that
encircle them than an unreasoned incredulity.
If the facts and arguments and conclusions pre-
sented in these four intercalated sections are now
borne in mind, the relations of the Lesser Antilles to
the coral-reef problem as a whole may be much
better apprehended than if those islands, to the
description of which we now return, are studied alone.
A SIMPLE SEQUENCE OF FIRST-CYCLE ISLANDS
SABA, A YOUNG VOLCANIC ISLAND
Three small members of the Lesser Antilles in
different stages of a simple sequence of first-cycle
development may be first presented. I saw them
only from passing steamers. Saba (Fig. 2), the sim-
plest island in the whole chain, represents an early
stage: it is a young volcanic cone, from two to two
and a half nautical miles in diameter and 2820 feet
high. As yet it is little dissected by streams but rather
sharply cliffed by waves around the shore and bor-
dered by a narrow submarine shelf, hardly wide
enough to be called a bank. It naturally shows no
signs of subsidence in the way of embayments even if

THE LESSER ANTILLES

Fig. 2-Saba Island (from U. S. Hydrogr. Office Chart No. ion).

its subsidence has begun, for its little valleys all have
hanging mouths in the cliff face, well above the sea
level. Naturally it has no coral reefs, because, as has
just been explained, they cannot be established on a
detritus-covered, cliff-base shelf. The inhabitants,
of Dutch descent, have learned to speak English
rather than their forefathers' tongue; they cultivate
the uplands and also build small boats which are
lowered down the cliffs to the harborless shore and
sold for inter-island traffic.

ISLAND DEVELOPMENT 37
THE SAINTS, EMBAYED AND CLIFFED RESIDUALS
The Saints, south of the lofty island of Guade-
loupe, represent a well advanced stage in the first-
cycle sequence of small-island development. They
are the separated residual summits, one of which
is shown in Figure 3, of an elaborately dissected
and partly submerged volcanic mass about five
miles across over all, which rise with cliffed head-

Fig. 3--One of the Saints, south of Guadeloupe, with moderately
cliffed headlands between well-developed drowned-valley embayments;
looking west.

lands from a rimless bank eight or ten miles in
diameter and from 30 to 40 fathoms in depth at its
outer border. Part of one of the islands is mapped in
Figure 4. Here the mature subaerial sculpturing of
the residuals is manifestly the work of a much longer
period than that required for the abrasion of the im-
mature headland cliffs. It is therefore believed that
the Saints show in their well carved summits and
slopes the work of a long continued interval of sub-
aerial erosion on a slowly subsiding island, which was
protected from wave attack during nearly all of that
interval by an upgrowing barrier reef; and that the
rimless bank which now extends around them repre-
sents the reef and its enclosed lagoon floor, as
modified during the Glacial epochs by low-level
abrasion, which also accounts for the immature head-

THE LESSER ANTILLES

a
N
a

U

-' q

0 0

I -
a\

Fig. 4-Part of one of the Saints (from U. S. Hydrogr. Office Chart
No. 362).

-. 4

I I IllV Lllilitt'l Iimt-stone Wand of Made (alantv.

V.
* ~ *; *.: I I

Pi.. III-The younger volcanic cones of Montserrat.

L r.

' ),A
y^:-. .

PL. IV-Thie southlerni part of Statia, looking east by south. ()rangetown lies at tile foot of the volcano

ISLAND DEVELOPMENT 39
land cliffs, and during Postglacial time by aggrada-
tion. The fact that the cliffs plunge somewhat below
present sea level, as shown by the charted depths
near them, gives support to this view. The submer-
gence of the carved island and the conversion of its
summits into separate islets are ascribed to sub-
sidence, not to Postglacial ocean rise, because the
amount of carving and probably the depth of carving
also-for the slopes of the islets have the appearance
of pitching down to a considerable depth below sea
level-seem to be much too great to have been ac-
complished by low-level erosion during the Glacial
epochs.
REDONDA, A SOLITARY REMNANT
A solitary and uninhabited islet known as Redonda
(Fig. 5) between Nevis and Montserrat, less than

Fig. 5-A rough outline of the residual island of
Redonda, north of Montserrat; looking north.

a mile in diameter and about Iooo feet in height,
rises from an imperfectly charted bank, several
miles across, with depths of 30, 40, or 50 fathoms.
This little remnant appears to represent a penulti-
mate stage in the first-cycle sequence of insular
forms; that is it is the central islet of an almost-
atoll, the atoll reef having been cut away and the

40 THE LESSER ANTILLES
central islet having been modified by the waves of
the lowered and chilled Glacial ocean. The scar of a
landslide, about 700 feet high at the top, and a heap
of fallen detritus are seen on the southeastern slope.
Manganese ore is said to have been quarried here.
SUBMARINE BANKS OR ABRADED ATOLLS
Finally, the ultimate stage of the first cycle
is exemplified by several small, rimless and island-
free banks, which appear to be true atolls modified,
like the forms of earlier, stages, by low-level abrasion.
One, three miles across, lies some miles to the east
of Redonda, with depths of from 43 to 86 fathoms;
another, to the north of Redonda, is five miles long,
with depths of 25 or 30 fathoms. A few others
of similar size and depth are charted elsewhere
in the chain. Evidently, if this schematic interpreta-
tion is correct, each one of these rimless banks began
as a simple young volcanic island, like Saba, and then
passed through the mature stage of elaborate sub-
aerial sculpture and partial submergence, represented
by the Saints, and the penultimate stage of approach-
ing extinction, represented by Redonda, before
reaching the ultimate stage of atoll development,
with its episodes of low-level abrasion followed by
imperfect reef growth and aggradation.
A SIMPLE SEQUENCE OF SECOND-CYCLE ISLANDS
MARIE GALANTE, AN UPLIFTED ATOLL
Two simple examples of second-cycle islands
may be adduced. The first is Marie Galante, a

ISLAND DEVELOPMENT 41
limestone island in the Guadeloupe group, nine
miles in diameter, with a remarkably even profile
at a height of 670 feet. As represented on H. O.
Chart 363, the relief seems stronger than it would be
inferred to be from the view of the island (P1. II)
reproduced from a recent photograph by Mr. G. S.
Miller, Jr., of Washington.
This island was seen only in the distance: no
detailed studies have been made of it. It appears to
be an upraised atoll, somewhat benched during and
moderately dissected since its recent emergence, not
yet provided with a bank of second generation.

SOMBRERO, AN UPLIFTED AND CLIFFED ATOLL
The second example I did not see at all; it is
Sombrero, southeast of the Virgin group, a narrow
limestone island a mile long and from 25 to 40
feet high. Its shore cliffs plunge to depths of 8 to 14
fathoms, and it is surrounded by a bank three by four
miles across and 30 or 40 fathoms deep at the outer
edge. Julien spent several months on the island
nearly half a century ago,1 and gave an excellent
account of it as an uplifted atoll, originally about as
large as the bank of second generation that now sur-
rounds it, but reduced to its present size by wave
attack. He concluded that its "marine deposits ap-
pear to have been formed upon the area, oscillating
vertically, of the bottom of a lagoon, more or less
inclosed," as if it had the basin form peculiar to
14 A. A. Julien: On the Geology of the Key of Sombrero, Annals
Lyceum of NaL. Hist. of New York, Vol. 8, 1867, pp. 251-278

42 THE LESSER ANTILLES
atolls. The rarity of atolls in the West Indies is
recognized, but it is held that Sombrero must have
been one, because of the repeated occurrence of
lagoon-like limestones upon it; and it is added that
"many of the isolated keys and banks with which
its [the West Indian] archipelagoes abound, may re-
veal to future examination the possession (in their
former history if not at present) of a true atoll con-
struction." The parenthetical remark here included
is, in my opinion, fully justified by the features of
Antigua, to be described below. Six beds of lime-
stone were recognized on Sombrero, bearing many
fossils like the forms now living on the bank around
the abraded island, and indicating as many subsi-
dences with deposition followed by emergences with
erosion.
The following extracts from Julien's account
illustrate the detailed analysis of the island's history
that he felt warranted in making. At an early stage
in its growth "'the whole subsiding area was covered
with a close and uninterrupted madrepore reef.
From the few species and individuals of the shells in
the coral-bed, from their fragility and that of the
corals, and from the absence of fragments, it may be
inferred that the reef grew in comparatively quiet
water. . The quietness of the locality could not
have been due simply to great depth. We are forced
to believe that some barrier encircled the reef
. varying at different periods in height relative
to the bottom of the lagoon. Indeed such a barrier
must necessarily have been formed on the outer edge

ISLAND DEVELOPMENT 43
of the oscillating area, on the first occasion that the
superincumbent sea was sufficiently shallow to sup-
port coral life, creating an atoll when it reached the
sea level. . Coral life ceased over the central
area, with the exception of many scattered dumps
and the frequent superposition of a more delicate
species. These remnants were next overlaid with
coarse sand from the shore of the lagoon and the
barrier islets, and thereby killed. The greater quiet-
ness and depth of the water favored the abundant
growth of a few species of fragile shells, whose un-
broken condition proves that they grew where they
now lie." Similar inferences are presented regarding
the several overlying limestone beds. The formation
of the sixth bed was followed by two oscillations of
level without submergence; the guano deposits which
then accumulated in crevices in the limestone have
been quarried and shipped away. "Thus then this
little rocky islet stands out in the open ocean, a
solitary pillar, like those of the Temple of Serapis,
marking old convulsive throbs and prolonged oscilla-
tions of the deep-sea bottom."'
Sombrero must have had, according to the scheme
of development proposed in the present essay, a
somewhat more extended experience than'the small
rimless banks which represent the ultimate stage of
one-cycle islands; for after passing through the first-
cycle stages represented by Saba, the Saints, Re-
donda, and an atoll, it appears to have been intro-
duced into a second cycle by a moderate uplift and
then to have passed through the young second-cycle

44 THE LESSER ANTILLES
stage of Marie Galante, preparatory to reaching its
own mature stage of advanced abrasion. Its emer-
gence would therefore seem to have been early enough
for it to have suffered low-level abrasion in at least
one Glacial epoch, while the emergence of Marie
Galante appears to have been so recent that it hardly
bears the marks of even the latest Glacial epoch of
low-level abrasion.
This interpretation is necessarily in large measure
hypothetical, for most of the conditions and processes
it involves are lost in the past. Nevertheless, in so far
as the interpretation is correct, the past conditions
and processes involved in the interpretation were
facts in their own time just as truly as are the
observable conditions and processes of today. That
the hypothetical interpretation may be accepted as
essentially correct appears from the simple manner in
which it assembles and correlates a variety of insular
features which at first sight seem to have little in
common.

A MORE COMPLICATED SEQUENCE OF
FIRST-CYCLE ISLANDS
EARLY STAGE: MONTSERRAT, A COMPOSITE ISLAND
Several examples belonging in a more complicated
sequence may now be presented, beginning with a
composite island of easy analysis. Montserrat (Fig.
7), nine by five miles across, was seen only from a

Fig. 6-Silver Hill, a maturely dissected volcanic mass at the northern
angle of Montserrat. The cliff of the most exposed headland is about 450
feet high and is one of the highest cliffs seen on any island.

steamer that passed near its western and northern
coast; but five or six cones in different stages of
erosion and hence of different dates of eruption
were recognized. It has therefore had a more
complicated history than the small, one-cone island
of Saba. Its oldest cone, Silver Hill (Fig. 6), occupies
the northern corner of the island, with a height of
1285 feet; it is a cone no longer but is so maturely
eroded as to have lost all its initial form; its exposed
shore is cliffed, and it is fronted by a two-mile rimless
bank. Center Hill (Fig. 8), a few miles to the south-
west, about five miles in diameter and 2450 feet in

THE LESSER ANTILLES

---------^--sf -Ma^ --- 40!J
FI5.' M di, waoi 62O10a~e s
Fig. 7-Montserrat (from U. S. Hydrogr. Office Chart No. ion).

FIRST-CYCLE ISLANDS 47
height, is much less eroded; for, while its loftier
slopes are sharply incised by deep, close-set, steep-
sided valleys between which the original upper surface
of the cone is reduced to acutely serrated ridges, its

Fig. 8-Center Hill. on the northwest coast of Montserrat. showing the
deeply incised inner valleys between sharply serrated ridges, the moder-
ately dissected lower slopes, and the low cliffs of the shore. The higher
cliffs in the left foreground are part of Silver Hill. Fig. 6.

lower and gentler slopes still preserve their initial
form little changed between the same valleys, which
are there wider spaced and much less deep; its shore
line is moderately cliffed. The Soufrieres, a compact
group of the youngest and loftiest cones, 3002 feet
at the highest point, still smoking and little trenched,
occupy the southern part of the island with low,
beach-based bluffs along the shore and with no sub-
marine bank offshore; their southeastern slope pitches
down at an angle of 35 or 40o, as if part of the cone
there had slipped into the sea. This young cone
does not therefore possess the long basal slope of very
gentle declivity seen on Nevis, to be described below,
and it thus gives support to the idea shortly to be
presented that the gentle basal slope of Nevis is a

48 THE LESSER ANTILLES
consequence of the growth of its young cone on a
preexistent bank of moderate depth, while' the
steeper slopes of the Soufrieres on Montserrat result
from their descent into deep water. On the other
hand, Center Hill descends by gentler slopes north-
westward and eastward to its bluffed shore, as if its
volcanic beds were there spread out on a bank now
buried on the south of Silver Hill, like the bank that
is not buried to the north of it. Only Silver Hill, the
oldest, maturely dissected cone at the northern angle
of Montserrat, shows signs of submergence in its
slightly embayed valleys. No reefs occur around this
island; those that are believed to have once rimmed
the bank in front of Silver Hill have been cut away;
and none can yet have been formed on the beached
shores of the younger cones. Plymouth, the chief
town of Montserrat, lies where, in consequence of the
projection of a low lava point a little farther north,
the shore is moderately concave on the bluffed slope
of a young cone of small size on the west coast. The
smooth lower slopes of Center Hill and other similar
cones are cultivated. The saddles between adjacent
cones guide cross-island roads.

A MORE ADVANCED STAGE: STATIA, ST. KITTS,
AND NEVIS ON ONE BANK
The next example includes three composite
islands that rise from a single bank, 47 miles in
length. I landed on the middle member, St. Kitts,
and drove around its younger cones; the southern
member, Nevis, was well seen from passing steamers:

FIRST-CYCLE ISLANDS 49

Fig. 9-Statia (from British Admiralty Chart No. 487).

50 THE LESSER ANTILLES
the northwestern member, Statia, properly St.
Eustatius, was seen only in the distance. The
last-named island (Fig. 9) is a volcanic doublet
consisting of two cones of unlike age. A sketch made
from St. Kitts through very hazy air is given too
definite an expression in Figure io. The older cone

Fig. xo-A rough outline of Statia. as seen through hazy air from the
northwest end of St. Kitts: the huge inclined slabs of limestone, known
as the White Wall," on the shore of the cone, appear to have been lifted
up from a preexistent submarine bank when the volcano was formed.

of Statia is now a well dissected mass, a mile or more
in diameter and 960 feet high, with a somewhat ir-
regular and moderately cliffed shore line. It is over-
lapped on the southern side by the younger cone, two
or three miles in diameter and 1950 feet high; the
long concave slopes of this cone resemble those of
Center Hill on Montserrat but are less dissected and
less cut back along their shore. According to Molen-
graaff,15 the younger cone bears on its southern side
some huge monoclinal slabs of limestone, known as
the "White Wall," containing shallow-water fossils
and rising with strong inclination to a height of 9oo
feet. From this as well as from the moderate inclina-
tion of its basal slopes, one may infer that the younger
volcano was built up on a bank of calcareous deposits
is G. A. F. Molengraaff: De geologic van het eiland St. Eustatius,
Leiden. 1886.

FIRST-CYCLE ISLANDS 51
that had already been formed in association with the
older volcano.
Similarly, St. Kitts (Figs. 12 and 13), formally
named St. Christopher and colloquially reduced to
' Singkits,"' consists in its older southeastern part of
six or eight small volcanic residuals, whose area is too

Fig. ix-The southeastern part of St. Kitts (from British Admiralty
Chart No. 487).

THE LESSER ANTILLES

Figs. 12 and 13-St.Kitts(from U.S.

FIRST-CYCLE ISLANDS

Hydrogr. Office Chart No. ion).

54 THE LESSER ANTILLES
small to show embayed valley heads but whose slopes
pitch into the sea in a manner betokening advanced
erosion while they stood higher than now and abra-
sion when the ocean was lower than now: they are
attached to one another by sea-level beaches and
occupy a stretch six miles in length over all: some of
them are shown in Figure 11. Adjoining them on the
northwest is a group of lofty young volcanoes occupy-

BR lMSlO / ~ HELL

Fig. 14-Basseterre, the chief town of St. Kitts, on the shore

ing an area of 12 by 5 miles, the highest rising to an
altitude of 4314 feet. Here the steep upper slopes,
verdant with well-watered vegetation, are sharply
trenched by close-set valley heads, while their gentler
lower slopes, which, like the similar slopes of Statia,
suggest the preexistence of a bank on which the lavas
and agglomerates of the lofty young cones were
spread out, are moderately incised by the lower
courses of the same valleys, there shallower and
wider-spaced. The shore is cut back in low, ragged
cliffs but is without valley embayments. According
to Cleve,1" large slabs of limestone bearing shallow-
water marine fossils cloak the flanks of Mt. Brim-
P. T. Cleve: On the Geology of the Northeastern West India Islands,
Handle. Kongl. Svensk. Velensk. Akad., Vol. 9 (N. S.), 1870, No. 12, pp.
1-48.

PI.. V-Th'le central volcanic range of St. Kitts, looking northeast. Mt. Brimstone on the left.

I
lVI

':
.~ ~P $,

-'

'~~ , ~ ;:i,,
i '1

-..go

) ~~it

PL. VI-St. Kitts from Basseterre roadstead. Continuous with PI. VII.

PL.. VII-St. Kitts.

PL. VIII-Isolated residual hills in the passage between St. Kitts and Nevis, looking east.

FIRST-CYCLE ISLANDS 5
stone, a small parasitic cone on the west side of one
of the higher cones; hence here again, as in the case of
Statia, it may be inferred that a bank had been
formed in association with the older volcanic residuals
before the younger cones were built. As the bank is
continuous over the island-free stretch between the
younger parts of Statia and St. Kitts, it is probable
that it is there based on several completely sub-

C- __ ___ "'" .. ..*.

dope a group of young volcanoes; looking northwest and north.
merged volcanic masses. Basseterre (Fig. 14), the
chief town of St. Kitts, lies on the southern slope of
the young cones, its open roadstead being protected
from the trade-wind swell by the beach-tied volcanic
residuals on the east.
Finally, Nevis (Fig. 15), the southeasternmost of
the three composite islands here considered, consists
of three mature residuals of earlier eruption and ero-
sion, between and above which a younger cone of
much later origin has been built up. It is of wonder-
fully graceful form, the finest island of its kind in the
Lesser Antilles. The steep but slightly furrowed
slopes of its verdant crater rim, which rises to a
height of 3596 feet, descend toward the shore with
ever decreasing declivity, little cut by radial con-

THE LESSER ANTILLES

LZI4 6Z34

w4fufentr 9str#;mnm inr %,pute of the fre ItLent cloudI-cap
%btiw#r(,r ut.I I-hie %direan b 'Jaldo Tetore line is o4
.itfloiI4 (waLt c(,ntiIr, five Iy w-vwn milk in diinwtivr,
very slightlyy (ita back herec and tlhre in bIluffs afew

FIRST-CYCLE ISLANDS
feet in height. The maturely carved residual, Come
Peak, on the east of the crater summit (Fig, 16) is
higher than the others; it is completely unwrapped in
the flanks of the younger cone; the second. Sadle
Hill, on the south, of medium height, is inwrapped on
three sides but reaches the sea on the fourth, where it
is cut back in cliffs a hundred feet or more in height;
the third (Fig. 17), the lowest of the three. forms a
salient on the north coast where it advances toward
the beach-tied residuals of St. Kitts. the nearest of
which is only a few miles distant. Although Nevis
has no uplifted bank limestone. the very gentle
declivity of its lower slopes suggests. as already inti-
mated, that it was built up from a shallow sea bot-
tom; its contrast in this respect with the steeper
slopes of Saba and of the Montserrat Soufrinres
which have no bank around them and which therefore
must have been built up from deep water. is certainh-
significant.
The bank from which these three composite islands
rise extends two miles northwest of Statia and ten
miles south of Nevis, with a total length of 47 mikes
a breadth of eight or ten miles, and a marginal depth
of 3o or 40 fathoms. Its trend is sympathetic with
that of the Lesser Antilles as a whole and with that
of the axes of the elongated volcanic islands, like
Guadeloupe, Lomninica, and several others. Saba
stands in line with it, not far to the northwest.
Hence, as already suggested, the tlk is thought to
have been built up as a rt'v-ea'nckwl l IAvY ko t in
asmulation with a suIbsidin4 series if earlier vocaukic

58 THE LESSER ANTILLES
islands, some of which appear to be wholly submerged
in the island-free stretch between Statia and St.
Kitts and in the southern end of the bank beyond
Nevis; and, as in the case of the Saints, the bank here
is believed to have recently lost its enclosing reef by
low-level abrasion. The younger volcanoes appear
to have been piled up on the bank in Glacial or Post-
glacial time. Discontinuous reefs are charted on the
bank near some of the islands; one fringes the north-
SADDLEMn(

Fig. x6-Nevis, as seen from the southwest. The young cone is built
are seen in Saddle and Cone mountains.

east coast of St. Kitts for two miles; another re-
sembles a submerged barrier reef, as it extends with
a depth of six or eight fathoms for 15 miles along the
northeast border of the bank from St. Kitts to Nevis
and encloses depths of Io or 12 fathoms.

ADJUSTMENT OF BANK DEPTHS TO PRESENT
OCEAN LEVEL
An important problem is opened up by the frequent
recurrence of submarine banks, associated with
marginal belt islands, which slope gradually seaward
to depths of about 40 fathoms around their outer
border; for that depth indicates that the detritus to
be moved and the marine agencies for moving it are
in adjustment with respect to present sea level. Daly

FIRST-CYCLE ISLANDS 59
has found from a study of many charts that the
depth at which a marked increase of slope occurs on
the outer part of continental shelves is not Ioo
fathoms, as it is often said to be, but is, as in the case
of marginal-island banks; about 40 fathoms. My own
studies lead to the same conclusion with respect to
the depth at which a change is made from a gentle
slope to a steep pitch in the exterior submarine pro-

N C LM

on the ruins of an older volcanic mass, two surviving summits of which

file of Pacific coral reefs. This appears to mean that
the activity of waves and currents is in the long run
such that the fine detritus which prevails on the
outer part of a bank or shelf will not be allowed to
accumulate there at less depths than 40 fathoms but
will be gradually shifted to the bank border where it
will settle down upon the steeper slope and build it
outward up to the standard 4o-fathom depth. The
longer a land mass remains stable, the broader may
the continental shelf or bank in front of it become;
but its border depth will remain of about the same
measure.
The same relation should have obtained between
the level of the lowered ocean in the Glacial epochs
and the depth at the outer border of the detrital
embankments that were then built of the material

60 THE LESSER ANTILLES
abraded in the production of low-level platforms.
Hence, if the ocean was lowered 30 fathoms, the
border of the embankment then formed should lie at
a depth of about 70 fathoms below present sea level.
But, as a matter of fact, bank-border depths of 70
fathoms are exceptional. The depth at which a
change of declivity takes place between the gentle
slope and the steeper pitch on continental shelves, on
HuivftCA'I4. HIU-

Fig. 17-The northern side of Nevis. looking east.

marginal-belt banks, and on the exterior profile of
the veteran coral reefs in the tropical Pacific is gener-
ally 40 or 50 fathoms, not 70 fathoms. Consequently,
if we accept the fact of low-level abrasion during the
Glacial epochs, whatever embankments were then
formed appear now to have been so well aggraded in
Postglacial time that they are no longer recognizable:
the built-up banks are now as a rule in good adjust-
ment with normal ocean level. It is perhaps sur-
prising that detritus should have been provided
during the relatively short Postglacial epoch in suffi-
cient quantity to obliterate the effects of the low-
level abrasion that was accomplished during a longer
time; but no other satisfactory interpretation of the
facts has been reached. Evidently the smaller the
measure of ocean lowering and the less the amount
of abrasion then accomplished, the more readily will

FIRST-CYCLE ISLANDS 6r
banks be brought to normal depths in Potglacial
time. Two corollaries follow from these considera-
tions. First, the depth of existing banks around
marginal belt islands gives no safe indication of the
depth of the abraded platforms that are supposed to
be beneath them, because the thickness of Post-
glacial aggrading deposits is unknown. Second, the
depth of the larger reef-enclosed lagoon floors in the
coral seas of the Pacific should not be regarded as
giving a rough measure of an abraded platform be-
neath them, not only because abrasion appears not to
have taken place there but also because. even if it di.d
the uncertainty as to the thickness of Postglacial
deposits prevents a safe inference as to platform
depth from lagoon-floor depth.

THE CONTIN NTAL SHELF OF BRAZIL
The continental shelf along the coast of Brazil is a
specific illustration of the adjustment of offshore
profiles to present marine agencies. The region is well
shown in H. 0. charts 153, 1504. 1532. 167o. 1671,
and 1672, According to Branner,1' the latest move-
ment of the coast was an earl Pliocene depression
the so-called stone reefs which rise to sea level near
the shore, as well as the continental shelf outside of
the reefst appear to have gained their present forms
since that time, Branner prepared a chart, on which
submarine contour lines show that the shelf. sloping
SJ, C, ItanWer The StoWe Rwis oi tBazi. Thti Ci*GlogicaQ aad
Ci(j(F>MC) ReU00" NNIt A W Ca (wA4 >a th, Coal Rsq& 4AMU.
AwnwA Qf CIM^ S90C at RnnWW-A ('wtkitd. V4 Xg4.

62 THE LESSER ANTILLES
gently seaward and ten or more miles in width, ex-
tends along the coast for more than 13 degrees of lati-
tude, or about 8oo miles; its outer border has a depth
of 30 or 40 fathoms, beyond which a steep pitch de-
scends to iooo fathoms or more in a further distance
of five or ten miles. South of the area charted by
Branner the shelf increases in width to 6o miles in
latitude 160 S. and to 80 miles in latitude 19 and
20o S., but its border depth remains about the same.
This broader part of the shelf is occupied by several
discontinuous bank reefs, to which reference will be
made below.
It would therefore seem that, when the shore line
was withdrawn toward the border of the Brazilian
shelf in the Glacial epochs, a low-level bench should
have been cut there, unless the shelf was then de-
fended by growing corals. If it was thus defended at
that time, a bank barrier reef should now have been
built up on the shelf surface, five miles or more out
from the present stone reef as the ocean slowly rose in
Post-glacial time. As no such reef is found, it may
be inferred that corals were not growing on the shelf
border when the ocean was lowered. In the absence
of protecting corals, the shelf should have been
benched by the lowered ocean; yet no trace of a low-
level bench is now to be seen. It would therefore
seem that whatever low-level bench was cut along the
coast of Brazil in the Glacial epochs has been oblit-
erated during the Postglacial restoration of a normal
shelf profile.
A similar obliteration of any low-level platform

FIRST-CYCLE ISLANDS 63
that was abraded during the Glacial epochs around
the border of the bank supporting Statia, St. Kitts,
and Nevis appears to have taken place; and the same
statement can be made for various other banks yet
to be described. Hence the measure of ocean lowering
in the Glacial epochs by about 30 fathoms, as cal-
culated by Daly, cannot be independently tested by
the submarine profiles of the marginal-belt banks
here under examination.

DOMINICA AND ITS SMALL BANK
Dominica is a superb example of an elaborately
dissected, composite volcanic island. It is 27 miles in
length, north and south, and 12 miles in width-a
grand mountain range of impressively bold forms
rising from the sea to an altitude of 4747 feet. Receiv-
ing an abundant rainfall it is covered with luxuriant
vegetation. The western coast was viewed from the
passing steamer on my outward voyage; and on the
return, opportunity was taken, while the steamer lay
at anchor off Roseau, the chief port on the southwest
coast, to make a short automobile trip into the interior.
The dissection of the island as a whole is so far ad-
vanced that it was not possible to determine how
many volcanoes compose its axial range. Certain
rather wide-spaced valleys on the west coast, con-
suming more and more of the initial volcanic slopes
as their depth increases inland, have reduced the
intervalley sectors, which must have originally
headed far up toward the volcanic centers, to short
sectors of buttress-like form. These culminate in

64 THE LESSER ANTILLES
summits about halfway from the shore to the island
axis, where peaks of decidedly greater height still
rise. It is probable that the buttress summits are con-
nected with the axial range by sagging intervalley
ridges, the degraded representatives of the upper and
inner half of the original full-length sectors.
The western half of the island thus seems to in-
dude a marginal range with short-sector summits of
a considerable height, each of which, rising at the
apex of a sloping isosceles triangle, has its base along
the shore. It may be noted that the deeply dissected
volcanic island of Huaheine, in the Society group of
the Pacific, possesses a similar marginal range of tri-
angular buttresses, of larger or smaller size according
to the spacing of the valleys, all around its oval cir-
cuit; but there each buttressed summit looks down
over a lower central area, which probably represents
the hilly remains of a caldera floor or of a central
mass of less resistance than the overlying lavas of the
buttress range. In contrast to this centrally ex-
cavated island, the central peaks of Dominica are
still so high that they dominate the summits of the
marginal range, at any rate on the west side of the
island.
The northern and southern ends of the island pre-
sent instructive features. Morne au Diable, a recent
addition at the northern end, is a well individualized
volcanic cone, roughly outlined in Figure 18 and
mapped in Figure 19, about three miles in diameter
and 2917 feet high; it is maturely dissected by
radial, consequent ravines and well cut back around

FIRST-CYCLE ISLANDS 65
its exposed side in model-like sea cliffs, up to rooo
feet in height, at the end of each radial spur. Unlike
the low, spur-end cliffs on St. Lucia and certain other
islands, yet to be described, which are small affairs
compared to the embayed valleys between them,
the high, spur-end cliffs of Morne au Diable have
such dimensions as to suggest that they, hie the
more mature cliffs of St. Helena, represent a measure

Fig. 18-Morne au Diable, a maturely dissected volcano at the
northern end of Dominica.

of abrasion closely comparable with the measure of
erosion represented by the valleys. The valleys
ought to have hanging mouths, but, as they seemed
to mouth at sea level, it is suspected that a slight
submergence-perhaps due to Postglacial ocean rise
-has taken place since the cliffs and valleys gained
essentially their present forms. Although this cone
is of more advanced dissection than Saba and has
perhaps been slightly submerged, as just suggested,
no reefs rise around its shore. Their absence may be
plausibly explained by the abundance of its out-
washed detritus, for in spite of its inferred slight sub-
mergence its valley mouths did not seem to be em-
bayed. Furthermore, as it stands in the marginal

THE LESSER ANTILLES

Fig. g--Morne au Diable (from U. S. Hydrogr. Office Chart No.
1318).

belt of the Atlantic coral seas, a more potent cause of
the absence of reefs is probably to be found in the in-
sufficient rise of ocean temperature in Postglacial
time.
THE SOUTHWESTERN CLIFFS OF DOMINICA
The southern end of the west coast of Dominica, a
simplified view of which is given in Figure 20, sketched
from the steamer at anchor off Roseau a little farther

FIRST-CYCLE ISLANDS 67
north, exhibits two great radial spurs-shown in the
right half of the sketch-cut off in slanting terminal
facets, thus repeating the features of jMorne au
Diable but on a larger scale as well as in a more ad-
vanced stage of sculpture. The more advanced sculp-
ture is shown by the undulation of the spur crests and
by the excavation of steep-pitching ravines, widening
upward into spatulate forms, on the spur sides and
on their faceted ends. A rather strong submergence
is clearly indicated here by the separation of the
spurs at the shore by valleys filled with sloping mud
flows to a width of a quarter or half mile. The sub-
mergence thus inferred may well have been as much
as 500 or 8oo feet, and it must therefore be explained
by island subsidence. If the island were restored to
the higher stand it held during the production of the
inter-spur valleys and spur-end cliffs, the measure of
abrasion seen in the cliffs would, as in the case of
Morne au Diable, appear to be comparable with the
measure of erosion seen in the valleys; but in view of
the mature stage of sculpture here attained before
submergence, as compared with the immature dis-
section of Morne au Diable, it is probable that the
sculpture of the southern end of the island was well
advanced even before the recent cone at the northern
end was formed.
At that early time in the history of the island, the
temperature of the surrounding sea must have been
favorable to reef growth, in spite of the absence of
reefs as indicated by the great cliffs of abrasion; for
an early stage of this still lofty island must have been

68 THE LESSER ANTILLES
contemporaneous with a more advanced and reef-
defended stage of a worn-down island like St. Lucia,
to be described below. Hence the early absence of
reefs around Dominica-or at least around its south-
ern end--may be best explained by the abundance of
detritus outwashed from its valleys. It is profitable

Fig. 2o-The southwest-

to recall here that Tahiti, the chief island of the
Society group, also exhibits spur-end cliffs of a size
commensurate with its inter-spur valleys and that,
in spite of its situation in the coral seas of the Pacific,
its cliffs were cut at a time when neighboring older
islands, like Borabora, were undoubtedly reef-
encircled; hence the absence of reefs around Tahiti
during the abrasion of its great cliffs is to be explained
by the outwash of detritus to its shores. Thus con-
firmation is found for the same explanation of the
cliffs of southwestern Dominica. Furthermore, like
Dominica, Tahiti has subsided since its cliffs were
cut; but unlike Dominica, Tahiti is now surrounded
by an up-built barrier reef. The reason for this con-
trast may be found partly in the mud flows by which

FIRST-CYCLE ISLANDS 69
the Dominica valleys have been deluged; for such
flows are destructive of any preexistent reefs, as will
be seen on St. Lucia; but partly also in the situation
of Dominica in the Atlantic marginal belt, where any
Preglacial reefs that may have been formed would
have been cut away by low-level abrasion in the

ern coast of Dominica.

Glacial epochs and where, even in Postglacial time,
the temperature for reef growth has been reached
tardily if at all.
WEST COAST OF DOMINICA
Several valleys of the west coast have benched
lateral slopes, apparently the result of alternating
epochs of mud-flow eruption and torrent erosion
after the coast had been deeply dissected in the
production of the buttressed marginal range above
described. The benches would appear to represent
successive mud flows, each of which occupies a
valley cut in an earlier flow and which is in turn
cut by a valley occupied by a later flow; except
that the first and largest valley would seem to have

70 THE LESSER ANTILLES
been cut in the original slopes of the west coast. It
would, perhaps, be safer not to imply by the word
"successive" that all the flows subsequent to the
deep dissection of the coast are now represented by
valley-side benches. Many minor flows may have
been buried under later and greater flows; and
similarly, many little valleys in minor flows may have
been entirely consumed in the excavation of greater
valleys in greater flows. Hence the above description
may be conservatively modified to read: The
benches would appear to represent a series of de-
creasing maxima in an irregular succession of greater
and smaller mud flows; and a similar phrasing would
apply to the valleys. The most pronounced examples
of these forms were seen in the large valley next north
of the great faceted spurs; they are roughly outlined
in the left half of Figure 20. Other examples of fairly
good definition were noted farther north; Roseau
occupies the higher part of a torrent delta at the
mouth of the benched valley next north of the one
sketched. Fine sections of the mud flows were seen in
the valley-side cliffs inland from Roseau.
Along the west coast the shore of the inter-valley
slopes, as well as that of the benches that slant
forward from certain valleys, is generally cliffed to
heights of 50 or Ioo feet. This small measure of
abrasion is by no means comparable with the work
of near-by valley erosion. It may therefore be sug-
gested that abrasion of the inter-valley shore here-
abouts has been weakened or delayed by the issue of
the bench-making mud flows from inland centers of

FIRST-CYCLE ISLANDS 71
eruption, as well as by the torrential outwash of detri-
tus from the valleys; also that such cliffs as were ear-
lier cut from time to time are now submerged by the
subsidence of the island, evidence for which is found

Fig. 3--An embayed and cliffed part of the east coast of Dominica
(from U. S. Hydrogr. Office Chart No. 1318).

72 THE LESSER ANTILLES
in the width of the mud flows on the southwest shore
between the great faceted spurs, as told above. It
may be added that the torrential outwash of detritus
is still manifestly effective today in delaying abrasion,
as is shown by the slightly convex advance of the
Roseau delta farther seaward than the shore to the
north and south.1s According to the chart, the east
coast (Fig. 21) is more embayed and more cliffed
than the west coast.

THE DOMINICA BANK
The bank adjoining Dominica is chiefly limited
to the east coast, where it has a width of three or
four miles. As that coast did not come under my
observation, it is impossible for me to correlate the
bank with the adjoining coastal features; but, from
analogy with other islands, it may be inferred
that offshore reefs have there had better oppor-
tunity of formation in the past than appears to have
been the case on the west coast. The absence of a
bank along the west coast seems, as above implied,
to be due to the fact that eruptions of mud flows
have there been intermittently continued to com-
paratively recent times, thus preventing reef growth.
Dominica is therefore not far advanced in the
scheme of development here outlined.
The above-described features of Morne au Diable
are of some interest as illustrating the results of ero-
uA newspaper report dated October 6, 1924, states that torrential
rains on Dominica caused the Roseau torrent to flood the town on its
delta, with loss of life there and damage to the estates in the valley.

FIRST-CYCLE ISLANDS 73
sion and abrasion in a simple and manifest manner.
The explanations given for the features of the west
coast, including the marginal range of buttressed
summits as well as the mud-flow benches in the larger
valleys, should be taken as only tentative until con-
firmed by more detailed observation. The inter-
pretation of the faceted spur ends of the southwest
coast is also provisional; for, as already stated, it is
based only on inspection from offshore. But if it is
supported by further investigation, especially around
the end of the island to the southeast coast, it may
well be regarded as the most significant lesson of the
island. Dominica would then be classed as the only
example thus far discovered among the larger islands
of the Lesser Antilles still showing high cliffs that
were cut during the earlier, reefless stages of its
development, before subsidence had set in effectively
and while its shore was subjected to severe abrasion
because the outwash of detritus from its deepening
valleys prevented the establishment of reefs around it.

THE COMPOSITE ISLAND OF MARTINIQUE
My view of this mountainous island was limited to
a part of its southwestern coast, as the steamer was
running from Fort de France, the chief town, to the
southernmost point. The island measures 35 by 13
miles and rises to a height of 4428 feet. It is adjoined
by a bank half a mile wide on the west and several
miles wide on the east, where the shore line is mark-
edly irregular. The island is evidently of composite
origin; the rather large embayment on which Fort

74 THE LESSER ANTILLES
de France is situated apparently occupies a reentrant
between two maturely dissected volcanic masses; but
the smaller bays which diversify the sides of this
embayment are plainly due to the submergence of
valleys of erosion. La Montagne Pelee at the north
end of the island, famous for its terrific eruption in
1902, is a relatively young cone, not greatly changed
from its constructional form by the incision of its
numerous but narrow and shallow radial valleys.
The geological structure of the island as a whole has
been mapped and described by Giraud,19 and the
description includes a chapter on its geological his-
tory, in which the sequence of eruptions by which the
island mass has been built up, earlier in the south,
later in the north, is fully treated; but nothing is said
of its erosion and subsidence.
The small part of the island's coast, seen by me,
being selected for inspection, as it were, by the
accidental time of a steamer's passage along it, may
perhaps be taken as a fair sample of other parts of
the southern coast. Its leading features, outlined
in Figures 22, 23, and 24, are simple enough. Ma-
turely dissected mountainous masses, bearing no
clear indication of their initial forms, rise in the
interior; well opened valleys with smoothly graded
side slopes descend to the coast; and the shore line
continually varies from the small bays of partly sub-
merged valley mouths to the steep headland cliffs of
truncated spur ends. The irregularity of the south-
1 J. L. Giraud: Esquisse g~ologique de la Martinique avec carte
g6ologique, Hanoi-Haiphong, 1918.

FIRST-CYCLE ISLANDS

western and eastern coasts is striking and is well
shown in Figures 25 and 26 reproduced from B. A.
Chart 371. Although some of the headland cliffs rise

Fig. 22-Clffed headlands of southwestern Martinique.

Fig. 23-Cliffed headlands of southern Martinique.

Fig. 24-Cliff and stack of southernmost Martinique.

several hundred feet above the shore line, they seem
low in comparison with the summits that rise a
thousand feet or more not far inland. The relation
of the cliffs to the valley-mouth bays makes it clear
that the cliff base lies below present sea level and that
the cliffs were cut back either at normal ocean level
before the island subsided to its present altitude or
while the ocean was lowered beneath its present

75

THE LESSER ANTILLES

IL _r Zi

Fig. 25-The embayed and cliffed coast of southwestern Martinique
(from British Admiralty Chart No. 371).

P nG
L a

FIRST-CYCLE ISLANDS 77

U

Fig. 26-The embayed coast and coral reefs of eastern Martinique
(from British Admiralty Chart No. 371).