BULLETIN

OK THK

MUSEUM OF COMPARATIVE ZOOLOGY

AT

HARVARD COLLEGE, IN CAMBRIDGE.

VOL. XXVI.

CAMBRIDGE, MASS., U. S. A.
1894-95.

Reprinted with the permission of the original publisher

KRAUS REPRINT CORPORATION

New York

1967

Printed in U.S.A.

CONTENTS.

Page
No 1. — A Reconnoissance of the Bahamas and of the Elevated Reefs of Cuba

in the Steam Yacht " Wild Duck," January to April, 1893. By Alexan-
der Agassiz. (47 Plates.) December, 1894 1

No. 2. — A Visit to the Bermudas in March, 1894. By Alexander Agassiz.
(30 Plates.) April, 1895 205

Bulletin of the Museum of Comparative Zoology

AT HARVAR-D COLLEGE.

Vol. XXVI. No. 1.

A RECONNOISSANCE OF THE BAHAMAS AND OF THE
ELEVATED REEFS OF CUBA IN THE STEAM YACHT
"WILD DUCK," JANUARY TO APRIL, 1893.

By Alexander Agassiz.

With Forty-Seven Plates.

CAMBRIDGE, MASS., U.S.A.:

PRINTED FOR THE MUSEUM.

December, 1891.

TABLE OF CONTENTS.

PAGE

Introductory 3

The Gjieat Bahama Bank. Plate 1 17

New Providence. Plate X, Figs. 2, 3 ; Plates XV. to XX 18

New Providence to Northern Eleuthera. Plate X. Figs. 2, 3 ; Plate X.

Fig. 5 ; Plate XXXII 25

Eleuthera. Plate I. ; Plate X. Fig. 2 ; Plates XXVII. to XXX 29

Little San Salvador and Cat Island. Plates I. and XXXIV 34

Nassau to Harvey Cay. Plate I.; Plate X. Fig. 2 ; Plate XXXVI. ... 35
From Harvey Cay to Great Ragged Island and Columbus Bank. — The Sand

Bores of the Bank. Plate I.; Plate XI. Figs. 3, 5; Plate XII. Figs. 1,

2, 4 ; Plate XXXV 37

Ocean-Holes. Plate III 41

Great Ragged Island. Plate XI. Fig. 5 43

Columbus Bank 43

Long Island. Plate X. Fig. 6 ; Plate XI. Fig. 2 44

The Exuma Islands. Plate X. Fig. 5 ; Plate XI. Fig. 1 47

From Conch Cut to Green Cay. Plate XXXVI 48

Green Cay. Plate XXII 49

Andros and the Western Part of Great Bahama Bank. Plate XI. Figs. 1, 3 ;

Plate XII. Fig. 3 ; Plate XXXIII 50

From Andros to Orange Cay 55

From Orange Cay to Great Isaac. Plate I. ; Plate XII. Fig. 2 ; Plates XXIV.

to XXVI 56

Gun Cay. Plate XII. Fig. 2 58

The Beminis. Plate XIL Fig. 2 59

Great Isaac. Plates XXIV. to XXVI 60

From Great Isaac to the Berry Islands. Plate I. ; Plate XII. Fig. 4 . , . 61

The Berry Islands. Plate XII. Fig. 4 62

The Little Bahama Bank. Plate I. ; Plate X. Fig. 1 64

Great Abaco Island. Plate X. Fig. 1 ; Plate XI 65

The Bank from Great Abaco to Bahama Island. Plate X. Fig. 1 . . . . 68

From Bahama Island to Memory Rock. Plate X. Fig. 1 ; Plate XXXIX, . 73

From Memory Rock to Geeen Turtle Cay. Plate X. Fig. 1 ; Plate XXI. . 76
"Whale Cay Channel and the Eastern Face of the Little Bahama Bank.

Plate X. Fig. 1 ; Plate XI. Fig. 7 80

Salt Cay Bank. Plates I. and XXXI 81

VOL. XXVI. — NO. 1. 1

2 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

PAGE

The Eastern Bahama Islands. Tlate I.; Plate VI. Figs. 1 to 4 ; Plate IX. 85

Watliiig Island. Plate IX. Fig. 12 ; Plate XXIII 86

Pium Cay and Conception Island. Plate IX. Fig. 11 90

The Crooked Island Bank. Plate IX. Figs. 1,9; Plate X. Fig. 4 ; Plates

XXXVII. and XXXVIII 91

Mariguana, the Plana Cays, and Saniana. Plate i. ; Plate IX. Fig. 4 . . 94

The Caicos Bank. Plate L; Plate IX. Fig. 2 95

Great Inagua. Plates I. and VIII. ; Plate VI. Fig. 2 ; Plate IX. Fig. 3 . 97

The Turk's Islands. Plates I. and VIII.; Plate IX. Figs. 5, 6 100

Navidad, Silver, and Mouchoir Banks. Plate IX. Figs. 7, 8, 10 . . . . 102

Hogsty Reef. Plates I. and II 103

The Coast of Cuba. Plates I., XIII., XIV., XLI. to XLVII 108

Cape Maysi to Santiago de Cuba. Plate I. ; Plate XIV. Fig. 5 ; Plate XLVII. 110

Santiago to Saboney 112

Baracoa to Banes. Plate I.; Plate XIV. Figs. 3, 6 ; Plate XLI 118

Banes to Padre. Plate I.; Plate XIV. Figs. 6 and 7 121

Padre to Nuevitas. Plate I.; Plate XIII. Fig. 1 ; Plate XIV. Fig. 7 . . 123
The Cays from Nuevitas to Caidenas. Plate I.; Plate XIII. Figs. 1 to 5;

Plate XIV. Figs. 1, 2 123

CayConfites. Plate L; Plate XIII. Fig. 2 ; Plate XIV. Fig. 2 .... 124

Cay Lobos. Plate 1 125

Paredon Grande to Cay Frances. Plate I.; Plate XIII. Fig. 2 125

Sagua la Grande. Plate 1.; Plate XIII. Fig. 4 ; Plate XIV. Fig. I. ... 127

Cardenas to Matanzas. Plate I.; Plate XIII. Fig. 4 128

Matanzas. Plates I., XLIL, and XLIII 129

Matanzas to Havana. Plate I. and Plates XLIV. to XLVI 130

The Coral Reefs of Cuba. Plates I. and XIII.; Plate XIV. Fig. 2 ; Plates

XLIV. to XLVII 133

Distribution of Corals in the Bahamas. Plate I. and Plates IX. to XII. 136

Hydrography of the Bahamas. Plates I. -VIII., and Plate X. Figs. 4, 5 . 139

The Coral Reefs and Banks of the Caribbean District. Plate VIII. 145

Santa Cruz, the Virgin Islands, and the Greater Antilles 153

The Banks of the Caribbean Sea 156

The Coral Reefs along the North Shore of South America 158

The Coral Reefs and Banks of the West Coast of the Caribbean Sea ... 160

The Reefs of the Yucatan Bank 161

The Reefs and Banks of British Honduras 162

The Mosquito Bank 163

The Bottom and Rocks on the Bahama Banks 166

The Bottom 166

The Rocks of the Bahamas 170

Some Recent Views on the Theory of the Formation of Coral Reefs 170

Index 189

List of Figures inserted in the Text 195

Explanation of the Plates 197

No. ]. — A Reconnoissance of the Bahamas and of the Elevated
Beefs of Cuba in the Steam Yacht ''Wild Duck," Jaimary to
April, 1893. By Alexander Agassiz.

The fullest accoimt we have of the structure of the Bahamas is that
given by Captaiu Nelson/ who clearly recognized the great part played
by the "seolian" rock in the formation of the islands. His examination
has formed the basis of the resumes of their structure subsequently
published by Darwin and by Dana,

Most accurate surveys of the Bahamas have been made by Captains
Owen and Bamet, and published by the Admiralty. These charts have
immensely facilitated the examination of the banks, and with such ex-
cellent guides a great deal of time could naturally be devoted to an
examination of the rocks, which otherwise might have to be spent in
making surveys of the ground covered. The hydrography of the deep sea
has been developed by the "r)]ake," "Albatross," and other vessels, so
that with the aid of the many deep-sea soundings taken in the region of
the Bahamas and in the Caribbean many problems can now be discussed
for which no data existed as long as the soundings were limited by the
100 fathom line. In a letter to Professor James D. Danii,'-^ written at
the time, was given a short account of my explorations. This is repro-
duced here, as the most convenient method of showing the track of the
" Wild Duck " among the Bahamas and along the Cuban coast.

Passing out of Nassau, we entered the Bahama Bank at Douglas
Channel and crossed the bank to Northern Eleuthera, where we exam-
ined the " Glass Window " and the northern extremity of Eleuthera.
We then sailed along the west shore of the island close enough to get
a good view of its characteristic features as far as Eock Harlior at the
southern end. We steamed out into Exutna Sound tlirough Powell
Point Channel, and round tlie southern end of Eleuthera past Little 8an
Salvador and to the northwest end of Cat Island, where are the hiirh-

1 On tlie Geology of tlie Baham.is, and on Coral Formations fjonorally, liy Cap-
tain K. J. Nelson, R. E. Quart. .Jour. Geol. Soc. of London, Vol. IX. p. 200, 1853.

- Observations in the West Indies. By Alexander Agassiz. (In a letter to J. D.
Dana, dated Steam Yacht " Wild Duck," Nassau, ^March, 1S93.) American Journal
of Science, Vol XLV. p. .1.-jO, April, 1893.

VOL. XXVI. — NO. 1.

4 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

est hills of the Bahamas. We then skirted Cat Island along its west-
ern face, rounded the southern extremity, and made for Riding Rocks
on the western side of AYatling Island. We circumnavigated Watling,
passed over to Rum Cay, then to the northern part of Long Island, visit-
ing Clarence Harbor ; we next crossed to Fortune Island, rounded the
southern extremity, and passed to the east side, near the northern end of
the island, on the Crooked Island Bank. Leaving the bank by the same
track, we skirted its southern face, and steamed to Caicos Bank, crossing
that bank from French Cay to Long Island, passed by Cockburn Harbor,
and ended our eastern route at Turk's Island. From Turk's Island we
made Cape Maysi, skirting the southern shore of Cuba as far as Santiago
de Cuba.

After coaling there we visited Inagua, and next steamed to Hogsty
Reef, a regular horseshoe-shaped atoll with two small cays at the western
entrance, where we passed three days studying the atoll. This to me
was an entirely novel experience ; we were at anchor in three fathoms
of water, surrounded by a wall of heavy breakers pounding upon the
narrow annular reef which sheltered us, forty-five miles from any land,
with a depth of nine hundred fathoms only three miles outside our quiet
harbor. I made some soundings in the lagoon and of the slope of the
reef outside. From Hogsty Reef we returned to Crooked Island Bank,
to the westward of which I also made some soundings to determine its
slope. Next we again visited Long Island, taking in the southern and
northern ends, which I had not before examined. From there we passed
to Great Exuma, stopping at Exuma Harbor and sounding into deep
water on our way out. We skirted the line of cays fringing Exuma
Sound to Conch Cut, where we entered the bank and sailed west, cross-
ing to Green Cay From there we made the southwest end of New
Providence, and returned to Nassau.

On our first cruise we were fortunate enough to strike Cape Maysi a
short time after daylight, and thus had a capital chance to oliserve the
magnificent elevated terraces and raised coral reef which skirt tlie whole
of the southern shore of Cuba from Cape Maysi to Cape Cruz, and make
so prominent a part of the landscape as seen from the sea. We were
never more than three miles from shore, and had ample opportunity to
trace the course of some of the terraces as far as Santingo, and to note
the great clianges in the aspect of the shores as we passed westward,
due to the greater denudation and erosion of the limestone hills and ter-
races to the west of Cape Maysi, which seems to be the only point where
five terraces are distinctlv to be seen. The hei<rht of the hills back of

1/ o

AGASSIZ : BAHAMAS. 5

Point Caleta, where the terraces are most clearly defined, I should esti-
mate at nine hundred to one thousand feet ; though the hills behind the
terraces, which, judging from their faces, are also limestone, reach a
somewhat greater height, perhaps eleven to twelve hundred feet.

During our second cruise we steamed fi'om Nassau for Harvey Cay,
crossing the bank from north to south as far as Flamingo Cay, and from
there passing to the westward of Seal Cay, Nurse Cay, etc. to Great Kagged
Island, from which we took our departure for Baracoa.

At Baracoa I hoped to be able to ascend the Yunque ; unfortunately,
the trip was given up, owing to the desperate condition of tlie roads.
From Baracoa we steamed to the westward close to the shore, touching
at Banes, Padre, Cay Confites, Sagua, Cay Frances, Cardenas, Matanzas,
and finally ending at Havana. On this trii> we continued the observa-
tions made on the south coast of Cuba, and thus traced the gradual
disappearance of the terraces from Baracoa to Nuevitas, and their re-
appearance from Matanzas to Havana, from the same causes which evi-
dently influenced theii" state of preservation from Cape Maysi west.
We obtained a pretty clear idea of the mode of formation of the flask-
shaped harbors found not oidy on the northern coast of Cuba to tlie
eastward of Nuevitas, but all along the southern coast, and between
Matanzas and the Colorado Reefs. They give us in part the explanation
of the mode of formation of the extensive system of cays reaching from
Nuevitas to Cardenas, and wliich find their parallel on the south coast
of Cuba from Cape Cruz to Cape Corrientes, only upon a much wider
plateau, and in the chain of cays stretching towards the western extrem-
ity of Cuba behind the Colorado Reefs.

From Havana we steamed to Cay Sal Bank, visited Cay Sal, the Double
Headed Shot Cays, and the Anguila Islands, and then crossed the near-
est point of the Great Baliama Bank. The bottom of this l)ank is of a most
uniform level, three and three and a half fathoms for miles, sloping very
gradually towards the west shore of Andros, so that we anchored nearly
six miles from AVide Opening of the central part of Andros which we
visited. The bottom consists of a white marl, resembling when brought
up in the dredge newly mixed plaster of Paris, and having about its con-
sistency just as it begins to set. The same bottom extends to the shore,
and the land itself, which is low where we landed (Wide Opening), not
more than ten to fifteen inches above high-water mark, is made up of
the same material, and feels under foot as if one were trending upon
a sheet of soft India-rubber. Of course on shore the marl is drier, and
has the consistency of very thick dough. It appears to be made up

6 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

of the same material as the seoliau rock of the rest of the Bahamas,
but from want of drainage it has become thoroughly saturated with salt
water. In that condition it crumbles readily, and is then triturated
into a line impalpable powder, almost like deep-sea ooze, which covers
the bottom of the immense bank to the west of Andros. After leav-
ing Andros we crossed the bank to Orange Cay, following the east-
ern edge of the Gulf Stream to the Riding Rocks, Gun Cay, and the
Beminis. We then passed to Great Isaac, where we saw some huge
masses of seolian rock which had been thrown up along the slope of
the cay about eighty feet from high-water mark to a height of twenty
feet. One of these masses was fifteen feet six inches long by eleven
feet by six feet. We then kept on to Great Stirrup Cay, and, coasting
along the Berry Cays, crossed over to Morgan's Bluff on the east side of
Andros, running down as far as Mastic Point, and then returned to
Nassau.

We made several attempts to examine the eastern sic^e of Andros, but
only succeeded in reaching North Bight, being, invariably driven back
to Nassau by northers. We made a special expedition to the Little
Bahama Bank, running eastward from Nassau south of Rose Island to
Fleeming Channel, where we passed into Northeast Providence Channel
and anchored near Current Island Cut. We crossed to Great Abaco at
the Hole in the Wall, following the southwest shore of that island to
Gorda Cay. We entered the bank at Mores Island Channel, steamed to
the Woollendean Cays and to Rock Harbor, examining the " ]\Iarls " to tlie
westward of Great Abaco. Returning, we passed into Northwest Provi-
dence Channel, steamed north to Burrow Cay, and skirted the southern
shore of Bahama Island as far as Settlement Point. At Memory Rock
we entered the bank, steering east past Great Sale Cay for Pensacola
Cay, and from there keeping in the channel between the outlying cays
and Great Abaco as far as Green Turtle Cay. Here we passed out from
the Bank through Whale Cay Channel, and skirted the eastern side of
the Little Bahama Bank back to Hole in the Wall, and crossed to Egg
Island. We examined both it and Egg Reef on our way to Nassau.

The islands of the Bahamas from the Little Bahama Bank on the north
to Cay Sal on the west, and as far as Turk's Island on the east, are all of
peolian origin. They were formed at a time when the banks up to the
10 fathom line must have been one huge irregularly shaped mass of low
land, the coral sand beaches of which supplied the material that must
have built up the successive ranges of low hills which we still find in
New Providence, and which are so characteristic of all the ridges of the

AGASSIZ: BAHAMAS. 7

islands of the group. After the formation of the islands came an exten-
sive gradual subsidence, which can be estimated at about three hundred
feet, and during this subsidence the sea has little by little worn away
the teoliau hills, leaving only here and there narrow strips of land in the
shape of the present islands. Inagua and Little Inagua are still in the
original condition in which I imagine such banks as the Crooked Island,
Caicos, and the Turk's Islands Banks, and other parts of the Bahamas,
to have been ; while the process of disintegration going on at the western
side of Andros still shows a broad island, which will in time leave only
the narrow eastern strip of higher ajolian hills on the western edge of
the Tongue of the Ocean. Such is also the structure of Salt Cay Ba-nk;
it owes its shape to the same conditions as those which have given the
Bahamas their present configuration. My reason for assigning a subsi-
dence of three hundred feet is that some of the deep ocean-holes on the
bank have been sounded to a depth of thirty-four fathoms, and I take
them to be submarine blow-holes or canons in the seolian limestone of the
Bahama hills when they were at a greater elevation than now.^ Subsi-
sidence explains satisfactorily the present configuration of the Bahamas,
but teaches us nothing in regard to the substratum upon which the Ba-
hamas were built. Indeed, the present reefs form but an insignificant
part of the topography of the islands, and they have taken only a
secondary part in filling here and there a bight or a cove with more
modern reef rock, thrown up against the shores so as to form coral
reef beaches such as we find in the Florida Reef.

We steamed in the " Wild Duck " nearly forty-five hundred miles
among the Bahamas^ visiting all the more important points, and made
an extensive collection of the rocks of the group,

I had on board a Tanner sounding machine, kindly lent me for this
trip by Colonel McDonald of the Fish Commission, and some deep-sea

1 Dr. Jolin L. Northrop, ("Notes on the Geology of the Bahamas," Trans. N. Y.
Acad, of Sci., Oct. 1-3, 1890,) who passed considerable time in New Providence and
in Andros, has given an excellent account of the characteristic features of Andros
He considers the evidence he has collected as conclusive of the recent elevation of
Andros and of New Providence. During my extended examination of the Bahamas
I did not meet anywhere with deposits either of corals or of mollusks, the position of
which could not be satisfactorily accounted for as resulting from the action of winds
and waves, or hurricanes. On the contrary, the verj' facts Dr. Northrop brings forth
regarding the configuration of the western coast of Andros seem to me to lead to
the opposite conclusions from those arrived at by him. All the evidence I have
shows that the Bahamas owe their present configuration to subsidence and erosion,
and that they are not rising.

8 bulletin: museum of comparative zoology.

thermometers were also supplied by him and by Professor Mendenhall,
the Superintendent of the United States Coast Survey. I provided myself
with a number of self-closing Tanner deep-sea tow-nets, with a supply of
dredges and surface tow-nets, and carried on board a Yale and Towne
patent winch for winding the wire rope used in dredging and towing in
deep water. The yacht was provided with a steam capstan ; by increas-
ing its diameter with lagging we found no difficulty in hauling in our
wire rope at the rate of a hundred fathoms in eight minutes. We carried
six hundred fathoms of steel wire dredging rope of the same dimensions
which I had used on the " Blake," and which has also been adopted on
the " Albatross."

Both on going into Havana and on leaving we spent the greater part
of a day in towing with the Tanner net. I thought I could not select a
better spot for finally settling the vertical distribution of pelagic life than
off Havana, which is in deep water — nine hundred fathoms — close to
land, on the track of a great oceanic current, the Gulf Stream, noted for
the mass of pelagic life it carries along its course. "We towed in one
hundred, one hundred and fifty, two hundred and fifty, and thi-ee hun-
dred fathoms, and on the surface at or near the same locality, and I have
found nothing to cause me to change the views expressed in the Prelim-
inary Reports of the " Albatross " expedition of 1891.-^ At no depth did I
obtain with the Tanner net any species which were not also at some time
found at the surface. Even at one hundred fathoms the amount of
animal life was much less than in the belt from the surface to that
depth. At one hundred and fifty fathoms there was still less, and
at two hundred and fifty fathoms and three hundred fathoms the
closed part of the Tanner net contained nothing} At each of these
depths we towed fully as long as was required to bring the net to the
surface again. Thus we insured, before the messenger was sent to close

1 Alexander Agassiz: General Sketch of the Expedition of the "Albatross,"
from February to May, 180]. Bull. Mus. Comp. Znul., Vol. XXIII. No. 1, 1892.

'-' Tliis is fully in accordance with the observations of the Plankton Expedition,
as far as they have been published by Apstein, Ortmann, Giesbrecht, and others,
relative to the bathymetrical range of the pelagic fauna. The diminution and final
disappearance at sea of pelagic animal and vegetable life below a comparatively
narrow limit seems general. That there are local conditions near shore, or in com-
paratively closed or shallow areas, or in districts adjoining submarine banks near the
500 fathom line, which modify these conclusions, I have already stated. Many of
the observations which form the basis of statements proving the indefinite exten-
sion in depth of the pelagic fauna and flora, are of little value, owing to the imper-
fect working of the apparatus in use.

AGASSIZ: BAHAMAS. 9

the lower part of the bag of the Tanner net, as long a haul through
water as the open part of the net would have to travel till it reached
the surface. This gave for comparison of their respective richness the
fauna of a liorizontal column of water obtained in the closed part of the
Tanner net at one hundred, one hundred and fifty, two hundred and
fifty, and thi-ee hundred fathoms, of the same or of greater length than
the fauna of a vertical column from those points to the surface obtained
in the sweep of the open part of the Tanner net.

In all our tows with tiie deep-sea self-closing Tanner net we took the
usual precautions of carefully filtering the sea water into which the con-
tents of the closed part of the net were emptied. We also made some
slight modifications in the construction of the Tanner net. Iron rods
were substituted for the rope guides of the pulleys, and one side was
loaded so as to cant the net while it was towed. Oil Clifton, New Provi-
dence, we made some trials in the Tongue of the Ocean, at a distance of
not more than a mile from the edge of the bank, the depth being seven
hundred fathoms. AVe towed at 9.30 a. m. in from one hundred to one
hundred and ten fathoms for about twenty minutes : the net closed suc-
cessfully. Only one Copepod was brought up from that depth, while in
the open part of the net we obtained several specimens of Eucope, many
bells of Diphyes, numerous Copepods, Alciope, Schizopods, larvfe of Bra-
chinrans, Macrurans, Doliolum, Appendicularia, Gasteropod larvse, and
CoUozoum. A fine Rhyzophysa came up attached to the wire after
hauling in forty fathoms. At 10.30 we took a second tow at three
hundred fathoms ; the closed part of the net contained nothing, and a
preliminary examination of the contents of the open part of the net,
which remained open from three hundred fathoms to the surface, showed
that it contained nothing we had not obtained from the shallower
depths between the surface and one hundred fathoms.

In a haul with the deep-sea Tanner net, made at 1.50 p. m., five miles
off Havana, in a depth of seven hundred fathoms, we towed for twenty
minutes at a depth of three hundred fathoms. There was nothing found
in the closed part of the net. There was a strong southeast breeze, so
that we obtained comparativelj' little in the surface tow-net except a few
pelagic algse, Diphyes bells, and Copepods, while in the part of the
to\v-ne+ which remained open all the way from the 300 fathom line to
the surface we brought up a mass of pelagic stuff, consisting of Cteno-
phores (probably species of Idyopsis and Eucharis), of two species of
Diphyes, of floats and tasters of Khyzophysa, of masses of Copepods, of
Schizopods, three species of iSalpae, many specimens of Doliolum, a few

10 BULLETIN : MUSEUM OF COMPAKATIVE ZOOLOGY.

Alciope, annelid and crustacean larvae, Brachiurans, as well as Macru-
rans, and several species of Pteropods, the whole tow mixed in a broth
of pelagic alga;, in striliing contrast with the paucity of the pelagic fauna
and flora at tlie very surface.

On another day, at 2.10 P.M., we made some hauls near the same
locality, olf Morro Castle, in seven hundred and forty-nine fathoms. Wc
towed at two hundred and fifty fathoms for twenty minutes. The open
part of the net contained many Sagitt^, Copepods, bells of Diphyes,
pelagic algte, Thysanopoda, Hyperiae, Limacina, and crustacean larvae.
There was neither animal nor vegetable life in the lower closed part of
the net. The water was smooth, a very light wind only having slightly
ruffled the surface. The surface tow-net contained many fragments
of Ctenophores, masses of Copepods, of Schizopods, of Macrurans and
Macruran larvte, as well as annelid larvte, Sagittee, Salpse, Doliolum,
Autolytus, Collozoum, and great numbers of pelagic algee.

A second haul was made about an hour later at the same locality. We
towed for twenty minutes at a depth of one hundred and fifty fathoms.
This time the closed part of the net contained a few annelid larvae,
some Schizopods and Copepods, all of the same species we had found
on the surface. It also contained specimens of Limacina and Styliola.
None of these were in the surface tow-net, but they had been obtained
from the surface on another occasion. In the open part of the net,
which had remained open from one hundred and fifty fathoms to the
surface, we obtained the same Pteropods which were found in the closed
part of the net ; also the same Copepods and Schizopods, as well as the
same annelid larvae. In addition, the open part of the net contained
bells of Diphyes, fragments of Ctenophores, Salpae, Doliolum, crustacean
larvae, Collozoum colonies, and pelagic algae.

On comparing the amount of material obtained in the closed part of
the bag from towing twenty minutes at a depth of one hundred and fifty
fathoms with the amount obtained at the surface in a tow-net dragged
during the same length of time, we cannot fail to be struck with the
poverty of the deeper haul as compared with the abundance of the haul
made at the surface. After hauling in about thirty fathoms of the wire,
we brought up a fine specimen of Pihizophysa. At six in the afternoon
we towed the Tanner net for twenty-five minutes in one hundred fath-
oms near the same locality. The wind had sprung up somewhat, so that
the surface tow-net contained practically nothing ; everything had been
driven froni the surface. In the closed part of the bag (towed at one
hundred fathoms) we found Doliolum, Copepods, Sagitta, and Schizo-

AGASSIZ : BAHAMAS. 11

pods, all of species identical with those collected in the open part of the
bag dragged tli rough the belt extending from one hundred fathoms to
the surface. It contained in addition some large Schizopods and Sa-
gittte, Tomopteris, Salpa3, Diphyes, Pteropods, tasters of Rhizophysa,
and crustacean larvce.

We hoped also to have made a large number of deep-sea soundings ;
unfortunately tlie trades were unusually heavy during the greater part of
our visit to the Bahamas, greatly interfering with such work on a vessel
no larger than tiie "Wild Duck," — one hundred and twenty-seven feet
on the water line. For the same reason, the number of deep-water pe-
lagic hauls was also much smaller than we had expected to make, as in a
heavy sea the apparatus would have been greatly endangered. This was
something of a disappointment, as the yacht Avas specially equipped for
this work, and I had hoped, with the assistance of such skilled draughts-
men as were my assistants on this trip, Messrs. J. H. Emerton and A. G.
Mayer, to accumulate a large series of sketches of pelagic types. It is
a very different thing working at sea in a small yacht like the "Wild
Duck," from working in such vessels as the "Blake" and the "Alba-
tross," fitted up w'ith every possible requirement for deep-sea work.
The "Wild Duck," on the other hand, was admirably adapted for cruis-
ing on the Bahama Banks, her liglit draught enabling her to go to every
point of interest, and to cross and recross the banks where a larger ves-
sel could not follow. I am under the greatest obligations to my friend,
Mr. John M. Forbes, for having so kindly placed his yacht at my dis-
posal for this exploration.

The Bahamas (Plates I., VIII. , IX., and X.) naturally divide them-
selves, first, into sunken banks like the Navidad, Silver, and Mouchoir
Banks; next, islands occupying the whole or nearly the whole summit
of tlie banks from which they rise, like Watling, Rum Cay, Conception,^
Samana, ]\Iariguana, the Plana Cays, Inagua, Little Inagua, and the atoll
of Hogsty ; then, banks having the semblance of atolls, like the Crooked
Island and Caicos Banks, which are fringed Jiy low islands forming a cres-
cent witli an open lagoon or flat between its horns ; next, Salt Cay Bank,
which from its structure holds a position intermediate between the group
of sunken banks like Navidad and that resembling the Caicos Bank;
and finally, such composite banks as the Little Bahama and Great Ba-
hama Banks, with the characteristics of a combination of banks resem-
bling all the others.

Darwin, in his account of the reefs of the West Indies,* assumes "that

1 Coral Reefs, 3d edition, London, 1889, Appendix, p. 266.

12 bulletin: museum of compaeative zoology.

large accumulations of sediment are in progress on the West Indian
shores, . . . especially of the portion north ^ of a line joining Yucatan
and Florida. The area of deposition seems less intimately connected
with the debouchement of the great rivers than with the course of the
sea currents, as is evident from the vast extension of the banks from the
promontories of Yucatan and Mosquito." Darwin considers the iso-
lated banks, such as the Misteriosa Bank, the bank off the northern point
of Old Providence, Thunder Knoll, and others of various dimensions, to
be " composed of sand firmly agglutinated, with little or no coral." He
considers their steep slopes as characteristic of similar banks " in all
parts of the world where sediment is accumulating, , . . the banks
shelving very gently far out to sea, and then terminate abruptly. . . .
The form and composition of the banks in the middle parts of the West
Indian sea clearly show that their origin must be chiefly attributed to
the accumulation of sediment ; and the only obvious explanation of their
isolated position is the presence of a nucleus round which tlie currents
have collected fine drift matter." Further he says, " There cannot be
any doubt that the' Mosquito Bank has been formed by the accumulation
of sediment round the promontory of the same name." And, finall}', he
says that the origin of the Bahama banks " is easily explained by the
elevation of banks fringed on their windward side by coral reefs." But
he modities this assertion by stating that the Bahamas, as well as many
of the submerged banks of the West Indian sea, " have been woi'u down
by the currents and waves of the sea during their elevation."

From what we have learned of the geology of the West Indian sea-
shores, there is nothing to confirm Darwin's views of the formation of
great accumulations of sediment against the promontories of Yucatan or
of Mosquito. The great submarine plateaus off those coasts consist of
the seaward continuation of their shore strata, and not of accumulations
of sediment. Whatever loose particles compose the bottom on these
plateaus are due to the remains of the animals and plants thriving upon
them, and flourishing to an unusual degree from the mass of pelagic
life brought to serve them as food by the prevailing currents and winds.
The currents, while they bear but little sediment in suspension, on the
contrary carry along a pelagic fauna and flora unsurpassed in richness,
but which stipplies a comparatively small amount of material towards the
building up of the submerged West Indian banks compared to that fur-
nished by the carcasses of the animals and plants fed by this pelagic
material, and the remains of which supply the great bulk of the deposits
which go to build up these banks.

* This must be a misprint for " east."

AGASSIZ: BAHAMAS. 13

Substituting older limestones in place of the oceanic deposits which
Darwin imagines to have been made on the shores of the West Indian
sea, and it seems to me that he has himself in his account of the coral
reefs of the West Indies, the substance of which I have quoted above,
given an admirable summary of the possible condition of the substruc-
ture of coral reefs in areas where it did not appear that his theory fur-
nished a satisfactory explanation of the facts.

The following report must decide how far the explanation given by me
satisfies the conditions of the coral reefs existing in Cuba, the Bahamas,
Florida, the Bermudas, and other West Indian areas.

The plateau upon which the Bahama Islands are situated is connected
with the shore plateau (the "Blake Plateau") of the east coast of the
United States, extending in a triangular shape from Cape Hatteras to
the Little Bahama Bank.^ Its western face is separated by compara-
tively shallow straits from Florida ; these become deeper as we proceed
south, and their depth increases regularly towards the westward to the
deepest points between Cape San Antonio and Yucatan.

The Bahama Plateau is separated from Cuba by the Old Bahama
Channel. It gradually increases in depth eastward as we go from
three hundred fathoms off Cay Frances to its greatest depth, over one
thousand fathoms, off Diamond Point, between it and Boca Guajaba.
(Plates 1. and VIII.) The plateau itself slopes to the eastward at its
northern termination, as is well shown also from the line of soundings
to the north of the Little Bahama Bank parallel with it, and along the
axis of the Northwest Providence Channel separating it from the Great
Bahama Bank.

The relations of the banks to the eastward of the Great Bahama Bank
are best seen from an examination of the hydrographic sections on Plates
IV. and v., and of the charts, Plates I. and VIII.

In attempting to explain the formation of the Straits of Florida, we
should remember that, in addition to wliatever part the Gulf Stream
may have played in cutting them out, on the Florida^ side of the straits
the land was nearly stationary during the time of the formation of the
Florida reefs, while on the Cuban side the coral reefs have been elevated,
and on the Bahamas the reefs now flourishing are in a region where there
has been considerable subsidence, increasing as we pass eastward. If
these movements of the Bahama Banks and of the land on their sides

1 See cliarts, Figures 56 and 176, " Tliree Cruises of the Blake "

2 See Suess, Antlitz der Erde, Vols. I. and II., Chapters X. and XVII.

14 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

were synchronous, they must have produced considerable warping of the
surfaces enclosed within the area explored by the " Wild Duck."

Mr. William H. Tillinghast has published some " Notes " ^ on the
historical hydrogi'aphy of the Handkerchief Shoal, " thinking that an
examination of old maps might reveal a change of condition in the shoals
since the time of the discovery of the Bahamas. . . . They throw more
light on the condition of the cartography of the West Indies than on
any physical changes among the islands."

An examination of the earliest charts of the Mouchoir Bank enu-
merated in the " Notes " which Mr. Tillinghast was kind enough to make
with me at once showed the impossibility of their having been based upon
actual surveys. On the chart of Thomas Jeffreys (1775), the Windward
Passage from the east end of Cuba and the north part of Saint Domingo
there is a legend regarding the Banc du Mouchoir Quarre : " This bank
is very little known. The soundings are taken from an English chart."
It contains nine islands. On writing to Captain W. J. Wharton, R. N.,
Hydrographer of the Admiralty, regarding this English chart, I was
kindly informed by Staff Commander Tizard that tliere appeared to
be no earlier chart at the Hydrographic Office than that of 1775 by
Thomas Jeffreys.

It is rkiost natural that in those days, when computations for longitude
were comparatively inaccurate, that in a region where the currents are
most variable and often quite strong, the positions assigiied to shoals and
islands should be very inaccurate.

Even at tlie present time it is most difficult, owing to the varying
strength of the currents, to pick up the position of well known banks,
and still more difficult to find banks like that on which the "Superb"
and " Severn " anchored, the position of which never was accurately
fixed. But little importance can be attached to islands as shown on an
old chart. On banks like those to the eastward of Turk's Islands (Mou-
choir and Silver Banks), where there are many rocks awash, it is not
impossible that patches of rocks awash should have been mistaken for
islands by ordinary observers, and so marked on the charts.

But when we come to actual surveys, such as those of Count de Chas-
tenet-Puysegur (1787), who cruised over the Silver Bank and Mouchoir
Quarre and took a number of soundings, we get a fairly accurate account
of the aspect of the shifting bottom, of the rocks awash, and a series of
quite characteristic soundings, but no islands are reported or plotted on
the shoals.

1 Library of Harvard University, Bibliographical Contributions, No. 14, 1881.

AGASSIZ: BAHAMAS. 15

"With the exception of the changes in the distribution of the sand
bores there are only some unimportant modifications of the topography
of the islands recorded. Since the careful surveys of 1834, no changes of
any consequence have been detected in the configuration of the land ; so
that we are warranted in assuming that the configuration of the Bahamas
as we now know them does not differ materially from that of the Y" de
los Lucayos as they were first discovered by Columbus. The char-
acter of the forests alone has been greatly modified since the advent of
Europeans.

The large scale charts of the region we explored in the " Wild Duck "
are the British Admiralty Charts, Nos. 393, 659, 1256, 2579, and 2580,
and the five Charts of the Bahamas, Nos. 399, 1496, 2009, 2075, and 2077.
The deep-sea soundings are in the main those taken by the officers of
the United States Coast Survey and United States Fish Commission, in
addition to a few lines derived from soundings taken hy the United States
Navy Department and the steamers of private companies engaged in the
laying of submarine cables.

The charts of the United States Hydrographic Bureau covering the
same field are Nos. 373, 944, 946, and 947, and for the Bahamas, Nos,
26 a, b, c, d. Also the United States Coast Survey Chart of the Atlantic
Coast from Cape Hatteras to Key West, scale Ymm' ^^s well as the charts
of the Florida Eeefs, Nos. 166, 167, 168, and 169. For the smaller
charts and plans, and other charts of the W^est Indian and Caribbean
districts, referred to in this Eeport, see the Index, and the chapters on
the Hydrography of the Bahamas, page 139, and on the Coral Reefs and
Banks of the West Indian and Caribbean, pages 145 and 160.

THE GREAT BAHAMA BANK.

Plate I.

The Great Bahama Bank, by far the largest of the group, is irregularly
V-shaped ; it extends four hundred miles from northwest to southeast,
and its greatest width is about two hundred and fifty miles. The west-
ern face, swept by the Gulf Stream, is slightly convex, and curves round
to the southeast and south, forming the northern edge of the Old Ba-
hama Channel. It then forms the edge of a great bay at the eastern
extremity of the channel, extending northward and eastward, and ter-
minating in a blunt projection, the Columbus Bank. From the eastern
extremity of this bank the edge of the Great Bank runs north, and is
protected by a line of low islands making a sweep which terminates in
the southern point of Long Island. The island is the easternmost pro-
jection to the west of Exuma Sound, a deep gulf separating the eastern
face of Great Bahama from the edge of the bank which extends north-
ward from Great Exuma to the western spit of the southern part of
Eleuthera.

The northern face of the western shank of the Great Bahama Bank
runs nearly east from the Gulf Stream. It has, like the western side,
but few fringing islands. Upon the eastern edge of the western shank
of the bank are the Berry Islands. From their southern line the edge
of the bank runs in a curve to the west edged by the Jonlter Cays, which
run into the northern extremity of Andros. This island, the largest of
the Bahamas, forms the western edge of the Tongue of the Ocean, which
separates the eastern and western shanks of the bank. On the eastern
edge of the Tongue of the Ocean there are bo islands except Green Cay
and a few insignificant islets. The Tongue of the Ocean is a deep
pockat ; the extension of its eastern shore forms a shai'p angle at New
T'rovidence, running slightly north of east to the extremity of Eleuthera.
This edge of the bank is flanked by New Providence and a series ot
narrow islands which separate the Northeast Providence Channel from
the inner bank. Eleuthera forms the northern part of the eastern face
of the Great Bahama Bank ; its southeastern spit is united to Little San

VOL XXVI. — NO. 1. 2

18 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Salvador by a shallow narrow bank, which in its turn is connected with
the northern extremity of Cat Island. The latter island forms the east-
ern side of Exuma Sound and the most easterly jDoint of the Great
Bahama Bank. AVe may now proceed with the description of the islands,
beginning at New Providence and taking the eastern shank of the Great
Bahama Bank first.

New Providence.

Plate X. Figs. 3, 3 ; Plates XV. to XX.

Although the island of New Providence is not one of the largest of the
Bahamas, yet the character of its surface varies greatly, and it is of suf-
ficient size to give an excellent epitome of the features which probably
characterized the land which once must have covered the greater part of
the Great Bahama and Little Bahama Banks.

New Providence occupies the northwest corner of that part of the
bank lying between the Tongue of the Ocean and Exuma Sound, its
westei'n extremity probably extending nearly as far out toward the
edge of the bank, as is indicated by the outlying islands to the west,
■which were perhaps once a part of it. The sea has encroached but little
upon its northern shore, except near the entrance to Nassau Harl)or, wliere
a few small islands to the west of the entrance show its former northern
extension, and indicate that the same causes, together with the subsidence
of the bank, have separated it from Hog Island.

The 100 fathom line is quite close to the north shore of the island,
as well as to the western extremity, a depth of one hundred and fifty
fathoms being reached within five hundred feet from the southwestern
extremity of the island.

Hog Island, Athol, and Pose Island to the east, as well as the small
outlying islands to the west of the entrance of Nassau, constitute the
outer line of n?olian hills wliich were separated from the ridge upon
wliich Nassau is built by a valley, now the harbor of Nassau. It is
evident, from an examination of the Nassau ridge, that it is of fcolian
formation, consisting of sand dunes closely packed together and heaped
up one upon another. The outer hill, called Hog Island, tlie land which
makes the breakwater of Nassau Harbor, is nuieh lower, and runs par-
allel with this in a general way. The valley which separated the two
outer ranges now forms Nassau Harbor, and its continuation to the
eastward is the channel which leads to Cochrane Anchoi-ago and to the
parts of the Great Bank lying between Nassau and FJeuthora. The

AGASSIZ : BAHAMAS. 19

harbor owes its existence in part to the general subsidence of this part
of the Bahamas, which has sunk the valley between Hog Island and
Nassau Hill as much as seventeen feet below low-water mark ; while the
Grantstown plain to the south of Nassau Hill (Plate XVI.), extending
to the next range, is still a few feet above high-water mark. A slight
subsidence would separate Nassau Hill from the rest of New Provi-
dence and form a second range like Hog Island to the north of a large
bay (Grantstown plain), cutting off the southern part of the island.

This ridge extends westward beyond Fort Charlotte from the point on
which Fort Monta^aie is built. From the Caves to the western entrance
of Nassau Harbor the north shore of New Providence is formed by a line
of hills running nearly parallel with it, and making the north side of the
basin of Lake Cunningham. A broad valley stretches between it and
the range of hills extending from Fort Charlotte to Fort Montague,
the valley gradually opening out as we go westwai'd from the Cave
Point. The shore line itself is a long series of seolian rock sand beaches
(Plate XVIII.), separated by an occasional rocky projection forming low
cliifs.

The part of the island adjoining the north shore is covered principally
by shrubs and trees. The flats at the foot of the hills are cultivated,
and covered with sisal or cocoanut plantations.

The Queen's Stairway presents one of the finest sections of the feolian
rocks of Nassau (Plate XV.) ; not only does the road leading up to its
base pass tlirough cuts of a3olian rocks ranging from five to fifteen feet,
but the so called Stairway itself is a succession of walls showing an ex-
posure of fully sixty feet of seolian rock. This aeolian rock when first
exposed to the air is comparatively soft, but becomes quite hard after
exposure to the action of the atmosphere.

Immediately behind the Queen's Stairway, on the southern slope of
the Nassau Ilange, there is a fine quarr}', one of a series which begins at
Nassau Street and runs along the top of the range to the eastward of the
Stairway. About one third of the way up the hill there is also a large
abandoned quarry, showing the structure of the a^olian rock. A few
smaller abandoned quarries are found near the base of the hill, in which
the exposed faces differ in no wise in character from those of the other
parts of the hill. The same was found to be the case in the street
leading from the main street to the Queen's Quarry, the nature of the
exposed faces being everywhere the same.

The a}olian rock faces as exposed in the quarries are in striking con-
trast with the harder rock surfaces exposed in the low ground between

20 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

the Nassau Range and the next range on the Grantstown road. There
we find the seolian rocks of the surface coated with a hard ringing layer,
and modified to a great extent by the intercalation of amygdules of
red earth, which give to the rocks the appearance of an indistinct
conglomerate.

At the foot of Nassau Street, twelve to fifteen feet above high -water
mark, the vertical wall (ten feet high) of the back of the quarry also
plainly showed lines of seolian deposit. On the way up the street there
are other quarries which have been abandoned, and excavations for
buildings showing the same structure. The quarries at the top of
Nassau Street are being worked, and fine vertical faces could be exam-
ined there fully twenty feet in height, which showed on the whole face
the same aeolian structure so plainly seen at the foot of the same street,
and at some other points of a still lower level, below high-water mark.
The whole of the ridge forming the first hill back of the har]Dor of
Nassau is evidently of seolian formation. Its highest point is about one
hundred feet.

To the eastward of Fort Fincastle there are other old quarries, the
exposures showing the irregular lines of wind-blown deposits. Wher-
ever streets have been cut through the hillside as trenches, the same
lines are visible on the faces of the side slopes. The seolian rock is used
as building material throughout the Bahama Islands. It is either quar-
ried from the surface in rough blocks from accessible faces, or regularly
quarried, as at Nassau and Green Turtle Cay, by cutting narrow vertical
channels five to six feet in the face of the vertical cliff' exposui'es. A
similar cut is made parallel to the face. The blocks thus laid off are
then sawed in columns, and these in their turn sawed into blocks of
the size wanted for building purposes. Imbedded in the rock faces of
this quarry, and in the fragments scattered about, we found a number
of land shells, quite similar to those still living. Mr. Dall has kindly
prepared for the Bulletin a short account of the species collected.

Through the mass of the exposures tubes were scattered irregularly,
formed from the decayed roots of bushes, grasses, and other plants which
covered parts of the sand dunes while they were forming.

To the westward of Nassau the shore of New Providence is marked
by a succession of beaches and seolian rock outcrops, and outlying caA's,
which once must have been short lines of hills more or less parallel to
the principal hills on the north side of New Providence. They were
separated from one another by low wooded flats, very similar to those
now stretching at the base of the existing hills, but which by subsidence

AGASSIZ: BAHAMAS. 21

and erosion have been changed into the narrow channels separating the
outlying cays from the main island of New Providence. In some in-
stances there has been an accumulation of recent shore coral rocks,
Hanking the hills and overlying the lower part of the older eeolian rocks.
Beyond the shore flat we cross Prospect Hill, then a second and third
range uf low eeolian hills, and then we come upon the sink which forms
Lake Cunningham (Plate X. Fig. 3). This lake fills a long valley, with
its sides flanked by mangroves and with a few mangrove islands scat-
tered on its surface. The water is slightly brackish. The vegetation out-
side of the mangrove belt runs into the characteristic Bahamian plants,
most conspicuous among which is the so called grape tree (Plate XX.).

After passing the range of Prospect Hill, we come upon large tracts of
pines, and in the hollows and valleys between the hills, where there is
more moisture, we find a richer soil, there being more red earth in the
decayed amygduloidal eeolian rocks. In these low marshes and flats we
find a peculiar flora, — a mixture of pines and groves of palmettos quite
similar in character to .the woods in some parts of Florida (Plate XIX.).
In fact, with the exception that the country is more rolling, one might
imagine the scene,. by its clusters of palmettos and tall pines, to be laid
tliere ; while in the more open spaces, where the forests have been
burned off, whole tracts are covered with bayonet palms and magnificent
brushes of young pines.

The road running west as far as Cave Point is more or less parallel
with the coast, and is flanked on each side with low bushes ; but the
principal shrub is the grape tree. Further inland the slopes of the hills
are covered with larger growths, and on the flats between the prolonga-
tion of the hills, especially in the space formed by the extension of
the valley in which Lake Cunningham lies, we meet fine clusters of pal-
mettos and tracts of pine.

The spurs of the sliore hills extending in a northwesterly direction
form more or less prominent spits nearly perpendicular to the shore line,
according to the hardness of the rocks and their exposure to the action
of the sea. The cave from which Cave Point derives its name is an old
vertical cliff", which is now separated from the shore by a narrow belt of
coral shore rocks which have been thrown against the eeolian cliffs, and
have formed also the small flat flanking the southerly extension of the
spur of which Cave Point is the northern end. Tiie cave diff"ers in no
way from the many similar caves which exist throughout the Bahamas.
The cave is from twenty to twenty-five feet wide, and perhaps ten to
fifteen deep in places.

22

BULLETIN : MUSEUM OF COMPMIATIVE ZOOLOGY^

From *;he top of Cave Point, about twenty feet above high-water mark,
a road leads at right angles to the shore line to Lake Killarney (Plate
X. Fig. 2). The road runs nearly level the whole way to the lalie ; the
country is covered with pines and clumps of palmettos until we strike
the mangrove swamp which forms the edge of the lake. Lake Killarney
is full of mangrove islands, forming very pretty vistas between their
headlands, and occupies a shallow sink between two short ranges of
eeolian hills, similar to the one forming Lake Cunninghan).

The blutfs at Clifton, which form the vertical cliffs of the southwestern
extremity of the island, are the termination of the a3olian range of hills
extending nearly parallel to the south shore. This range gradually re-
cedes from the shore, and the cliffs die out and are replaced by a long
beach line, which extends nearly unbroken to the southeastern end of

. ~-~^-C::^-- ■•-:

^

"^J^^

^^

—-.^-^

— = — =r--^— - =4miM^i ' ' '

^'^-^^ ,

■ ^5!^sa^^ii==

~=:^^^^i^^^^^^^=^^^ rL^-:^^--

-=■— --=-'°7:r-.- — '^^

T= ^ . 1-:=^ i^vi ^

■ n

— — — ■- ,.

■.=_:

- , ^^-^

cow AND BULL.

the island. At Clifton these cliffs are eaten in more or less by cav-
erns and fissures both to the east and west of Moss Hill, and are admi-
rable examples of the effect of the action of the sea upon reolian rock.
The constant pounding of the ordinary swell breaks off large blocks,
■which in their turn break into smaller blocks and are thrown up in the
season of the hurricanes above high-water mark, forming a stone wall
along the coast line similar to those to be seen in other parts of the
Bahamas.

Along this part of the shore there are man}- sea-holes and pot-holes,
extending from the surface to low-water mark, or even below. ]\Iany of
these irregular wells or caverns are from twenty to twenty-five feet deep,
and when occurring in high cliffs, or hillsides farther inland, as on Long
Island, from a hundred and fifty to two hundred or more feet in heiglit.

AGASSIZ : BAHAMAS.

23

would readily account for the deep ocean-holes, one of which has been
sounded to a depth of thirty-eight fathoms. These caverns would, if the
islands sauli to their full height, appear like ocean-holes in the general
levels of the bank.

Moss Hill Blutf shows also the manner in which huge isolated rocks,
sometimes twenty to thirty feet in height, like Cow and Bull on Eleu-
thera, near the Glass Window, may be left as fragments of hills formerly
much higher, but gradually eaten away on the sides where parts of the
rocks are not protected on the surface by a coating somewhat harder
tlian tliat of the surface rock immediately surrounding them.

Tliere are two bluifs to the north of Moss Hill, which now stand
out like rounded knobs above tlie sun-ounding country, and which in
time may appear almost like erratic boulders left high and dry, or like
huge masses thrown up perhaps during a great hurricane.

The seolian rock sand from the bottom where we anchored is coarse
and of a grayish color.

s

a

M

a

Ji

d

s

>

o

cc

ai

m

O

n

QQ

■Ji
<a

iz;

e

em
o

SECTION ACROSS NEW PROVIDENCE.

A section of the island of jSTev/ Providence in a nearly north and south
line from Nassau to the south shore is most instructive.

Leaving the shore at Nassau, we follow up Nassau Street, and at the
quarries cross the crest of the ridge. The slope falls rapidly to the level
of the extensive plain of Grantstown, which separates the shore range
from tlie Blue Hills range. On the flat the exposed seolian rocks are
greatly honeycombed and weather-worn, the whole surface occupied by
pits, banana-holes, pot-holes, and cavities of all sizes, which have formed
in the more or less rotten rocks of tlie Grantstown flat-. On both sides of
the road there are marshes and pools filling the depressions. It would
require but slightly deeper depression to change the greater part of the
mangrove marshes of this flat into lakes like Killarney, Cunningham,
and other ponds which lie in deeper sinks. The greater part of the
Grantstown flat is covered by low bushes (Plate XA'L), with here and
there patches of mangroves.

At the foot of tlie Blue Hills we come upon the more or less amygdu-

24 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

loidal rocks which usually characterize the aeolian rocks a little inland,
away from the immediate action of the sea, where the red earth wliich
once filled the cavities has become a part of the rock itself. As we
rise we pass into the regular stratified seolian rocks, still retaining their
original structures, and only coated with a thin ringing crust at the sur-
face. On reaching the summit and descending the hills, we come upon
a low flat,^ which separates the Blue Hills from the undulating plateau
district known as the Seven Hills. These are six or seven irregular
ranges, slight undulations merely, running more or less parallel with the
Blue Hill range, which occupy the country between it and the south
shore. These hills follow in rapid succession, and are perhaps twenty
to twenty-five feet above the general line of the slope, which ends in the
low flat behind the south beach. These hills become less and less promi-
nent as we go south, and the last scarcely rises above the flat just referred
to. The shore plateau imm.ediately behind the beach is low and swampy,
filled with clusters of mangroves, and when we reach the beach there are
extensive mangrove flats extending some distance out (for about half a
mile) from the shore line, and islets of mangroves running parallel with
the line of the beach (Plate XVII.). This must evidently have been
formerly a mangrove swamp similar to the one to the rear of the beach,
which occupied undoubtedly a position similar to it at the foot of a low
seolian hill m the valley between it and the next range, or between it
and the old shore line, which by subsidence and erosion has been brought
to its present level. All the way from the Blue Hills to the mangrove
swamp the vegetation consisted of small pines thickly crowded together.
In the more open spaces young pine shoots were starting up in all
directions, and in the lower and more swampy districts between the
successive ranges of hills groves of palmettos of all sizes varied some-
what the monotony of the pine barrens. Intermingling with the pines
and palmettos are large tracts covered by bushes, and near the shore
often prominent patches of larger trees. The pines and palmettos are
affected by the proximity of the sea, being of a smaller size and growing
less vigorously ; while on the beach itself, and immediately behind it,
we come upon the common plants characteristic of the shore line of the
Bahamas.

1 The rock exposed upon this flat resembles the more or less rotten aeolian rock
characteristic of Grantstown Flat.

AGASSIZ. BAHAMAS.

New Providence to Northern Eleuthera.

Plate X. Figs. 2, 3 ; Plate XI. Fig. 5 ; Plate XXXII.

On passing out of Nassau Harbor, making toward Douglas Channel,
we obtained the first view of the peculiar aspect presented by the low
seolian hill ranges forming the outer islands of the Bahama Bank. As
we passed to the seaward of Hog Island we obtained an excellent view
also of the sand beaches which form the western part of the island.
Similar longer or shorter sand beaches extend throughout the Bahamas
on the sea faces of such islands as are more or less protected by outside
rocks, outlying islets, or banks or bars, from which the material for the
inner beaches is worn away by the action of the sea and thrown up on
the surface of the inner island. The sand beaches may also be formed
from the disintegration of the more gentle slopes on the sea face of the
seolian rocks of the inner islands.

As seen from the sea, the outline of the outer islands is most charac-
teristic. Seen slightly obliquely, such an island as Rose Island, for
instance, seems made up of a series of solidified dunes falling one over
another. Rose Island is a long narrow island, the highest point of which
is not more than thirty to forty feet above the sea level, of about twelve
miles in length. One might readily imagine from the distant outline
that the whole mass of the island was moving bodily to the westward, in
a series of low sand waves, under the influence of the prevailing winds.
Rose Island is parallel to the range of islands of which Hog Island may
be called the centre. Long Cay and North Cay form the western exten-
sion, while Athol Island makes the eastern extension. Parallel to Rose
Island are Salt Cay to the west and Booby Island to the east, — the
remains, perhaps, of the range of seolian hills which formed the very edge
of the northern part of the bank to the eastward of Nassau, very mu«h
as Egg and Royal Islands are the remnants of the ranges of the north-
eastern points of the bank.

Rose Island is perhaps one of the most characteristic of the Bahama
Islands. It consists of a single narrow ridge of seolian rocks, .extend-
ing for twelve miles along the edge of the bank. In continuation of it,
on the other side of Douglas Channel, is Booby Island, a very simi-
lar ridge. Rose Island is a low range of seolian hills, similar to those
now running across New Providence ; it has become isolated by the sub-
sidence, erosion, and destruction of the land of which it once formed a

26 bulJjEtin: museum of compakative zoology.

part, little of which is left. At the eastern end of the island, where it
is wider, a salt pond exists.

Passing into Douglas Channel, Booby Island (Plate XXXII.), a long
barren line of rocks, is seen to the eastward, presenting very inuch the
same appearance as Rose Island, with low asolian hills, but without a
beach to either the sea face or the bank side of the range. The contrast
of the appearance of Booby Island, as seen from the bank or from the
northern side, is most characteristic. On the sea face it is one series of
seoliau hill slopes, piled one on top of another, while on the bank side
the steeper slope of the hills has been cut away from the base nearly to
the summit, leaving only a series of rounded tops on the horizontal lint,
of the summits of the vertical shore clifis. Booby Island is bare, with
rounded summit, the whole surface of the island honeycombed more or
less by the action of the sea. On both sides of the island the waves
have made greater or less inroads, and have cut away the lower parts
of the island, especially on the southeastern face, leaving round them a
shallow bank formed of disintegrated ai'olian rock, on which animal life
is not plentiful — having on it a few patches of Gorgonians and of coral-
Ime algge — either on the sea or the bank face of the shoal. But as we
go into deeper water, four to six fathoms, toward the steep edge of the
bank, in the direction of the Providence Channel, we come upon more
or less extensive patches of coral heads, separated by clear sand, which
form an irregular belt of corals to twelve or to sixteen fathoms in depth
along the sea face of the bank, from its northeastern extremity to the
western end of New Providence.

Samphire and Upper Samphire Cays have the same general structure
as Booby Island. There is perhaps a little more low vegetation on these
rocks, and they appear a little greener. They are the remnants of cays
which were undoubtedly similar to Hog, Rose, and Boob}- Islands, and
which once formed long narrow seolian ranges on tliat part of the bank.
The disintegrating action of the sea has, however, left only parts of these
ranges, either as low cays like the Samphires, or the patches of small
banks rising nearly to the surface on the two sides of the Fleeming Chan-
nel. In the distance we passed the western end of Current Island, a
low point covered by a few trees. The point is formed of low undulat-
ing hills, of about the same height as Hog Island, and presenting the
same structural features.

On our way to the Little Bahama Bank we passed again over the bank
lying between Nassau and the Fleeming Channel. Through this we
passed, and steered for our anchorage on the northwestern extremity of

AGASSIZ: BAHAMAS. 27

Northern Eleutbera off Current Cut. The Six ShiHing Cays and other
rocks forming the northwestern edge of the banli to the eastward of
Nassau are all low, bare, eaten away at the base (the remnant of the
Eleuthera-Nassau land), formed of seolian rocks greatly weathered and
worn ; Fleeming as well as Douglas Channel being formed by the subsi-
dence of the bank, of which the islets and cays and rocks are the higher
summits and the channels the lower parts of the intervening space. As
we came up to Pimlico Cay we saw an extensive belt of angular a;olian
rocks thrown up along tlie north shore of the cay above high-water mark.
The nortli face of the cays between Fleeming Channel and Current Cut
has been greatly eroded, and the rocks rounded and honeycombed by
the action of the sea.

From Pimlico Cay to Egg Island round to Bridge Point there are a
number of cays which have formed an outside line as it were upon
the triangular extension of the bank to the northwest of Eleuthera.
These cays are the outliers of the former Eleuthera land, having become
disconnected from the main island. Of these caj's we examined Egg
Island, Northern Eleuthera, and the north end of Current Island adja-
cent to Current Cut. The constitution of the islands was everywhere
the same, of aeolian origin. The bottom on the bank where we anchored
was made up of fine' coral sand, with many broken shells, mainly bi-
valves. A strong current sweeps north or south through Current Cut
according to the stage of the tides.

From the western face of Egg Island an extensive reef runs parallel
to Eoyal and EusscU Islands, from two to three miles distant, as far as
Bridge Point on Eleuthera, where it sweeps aci'oss the northern extrem-
ity of Eleuthera and skirts the cays as far as the southern end of Harbor
Island. The 100 fathom line runs so near the bluffs of Eleuthera, near
the Glass Window, that the reef is interrupted, but reappears near
James Point. From there it skirts the eastern face of Eleuthera as a
fringing and barrier reef, and it extends south to meet the reef which
connects Eleuthera, Little San Salvador, and Cat Island. Fronr Eleu-
thera Point it extends on the west coast, but rather irregularly, as far
as Bamboo Point. It appears again in patches on the west side of Powell
Point Channel, to the south of the Schooner Caj's, on the northern edge
of the Exuma Sound Bank. From Ship Channel Cay to Wide Opening
the corals on the edge of the bank are in from three to five fatlioms of
water, and the gaps between the cays are one mass of coral heads, with
such passages as Ship Channel, Highborn Cut, Wax Cay Cut, and Wide
Opening leading from Exuma Sound on to the bank.

28 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

In the passages — such as Douglas Chaunel and Fleeming Channel
— leading from the Northeast Providence Channel on to the bank,
we find that coral heads extend into three to four fathoms, and mav
be found even at less depth. But the heads are not vigorous, and
do not compare with those growing on the sea face of the banks.
They occur everywhere on the banks in proximity to the open waters
of the deep channels, forming more or less extensive patches, such
as are known near Nassau, between Hog and Athol, as the Marine
Gardens.

On striking the open bank itself, beyond the influence of the fringing
cays, one cannot fail to be struck with the povferty of the fauna and flora
upon this great expanse of disintegrated seolian rocks. ' We found almost
nothing upon the bank in the line extending in a general way parallel
to the northern edge, from Douglas Cliannel to the Glass Window on
Eleuthera ; here and there were to be met patches of Millepores or of
Gorgonians, or a few sponges, very few mollusks, small clusters of Zos-
tera and of coralline algaD, and a few broken sea-urchins and shells, and
here and there a Crustacean or a Holothurian. The bottom was plainly
visible as we steamed along, the ground occupied by animal or vegetable
life presenting a different coloration.

The long cays which are scattered upon the northeastern extremity
of the Bahama Bank to the northwest of Eleuthera are edged, as seen
coming from the north, by very low clifts of eeolian rocks breaking in
here and there on the otherwise continuous shore line. The north line
of cays are ftiirly sheltered from the violent action of the swell, owing
to the width of the shallow bank which extends beyond them, and also
by the line of the coral reef which fringes nearly the whole width of
the bank from a depth of five or six fathoms to ten or fifteen; so
that the sea is eating away but slowly the base of the shore hills of
these cays.

Egg Island, on which we landed, does not differ from the typical
Bahama cay. The base of the seolian hills on the south face has been
beaten into diminutive vertical cliffs, which pass gradually into the
rounded hill slopes of the cay. The surface of the rocks is more or
less water worn, according to its distance from high-water mark and its
exposure to the action of the sea.

AGASSIZ: BAHAMAS. 29

Eleuthera.

Plate I. ; Plate X. Fig. 2 ; Plates XXVII. to XXX.

On both sides of the passage separating Northern Eleuthera from Eleu-
thera the island is so narrow that tlie whole surface is washed by the
spray from both the eastern and western sides. The surface is free from
vegetable growth, and the incessant action of the sea-water and of the
rains has honeycombed the rocks from the water up to the highest ridge
of the cliffs forming the eastern edge of the island. These clifls are per-
pendicular, the heavy trade swells having broken off large masses of the
eastern sides of the a^olian hills and gradually washed them off into deep
water, the 100 fathom line not being more than about a mile from the
shore, and often much less, except at three or four points where the
island must formerly have extended from five to six miles farther east-
ward, as off Savannah Sound, James Point, and the northeast bank to
the eastward of jNIan Island. These perpendicular cliti's extend north-
ward and southward, and form the sea face of Eleuthera. There is no
shelf of any width at the base, and only along those stretches of the
island where the coast shelf has any width do we find extensive patches
of coral reefs forming a sort of barrier reef; as, for instance, at irregu-
lar intervals all the way between Eleuthera Point, Savannah Sound, and
James Point, and again from Harbor Island to Egg Island, on the north-
east and northern faces of the outside bank.

The photographs (Plates XXVIIL, XXX.) will give a better idea than
could any detailed description of the honeycombed surface extending
north and south of the Glass Window Passage. In some cases the
height of these little Gothic pinnacles is nearly three feet, and walking
along the surface is very difficult. We met with similarly eroded sur-
faces at many other places in the Bahamas where the islands are narrow
and their surface subjected to the action of the spray or surf, in addition
to that of the rains in districts with a scant vegetable growth.

In the hollows formed by the disintegration of the surrounding surface
there are accumulations of red earth, and there a few isolated stunted
bushes begin to make their appearance. They gradually pass into
larger patches as we come to wider parts of tlie i.slands not influenced
by the action of the spray, where the decomposition of the surface rocks
has been limited to the action of the rains. From three quarters of a
mile to a mile both north and south of this barren tract, which looks
more like an area covered with volcanic scoria^, the vegetation begins

30

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

again, and going south we soon pass into regions thickly covered with
the same underbrush which is so characteristic of the shores of all the
Bahama Islands ; and inland we get patches of aloes, of palmettos, and
of tlie forest trees characteristic of the Jiahamas.

The arch of the Glass Window was evidently formed by the undermin-
ing action of the sea, which little by little disintegrated the underlying
rock ; finally the overlying arch gave way in part, leaving an opening
known as the Glass Window, and remnants of the rocks forming the
arch lining the western sides of the opening.

THE GLASS WINDOW.

P)0th the northeast and northwest winds act upon the southern shores
of Northci-n l'>lcutliera, and are gradually eating away the outer shore
shelf, forming low vertical cliffs, and leaving here and there numerous
isolated rocks in from two to three feet of water, more or less water
worn, and ci-odcd or diminutive islets formed by the general subsidence
and liv erosion, whi<:'h will gradually disappear and finally leave only the
uniform flat level of the parfs of the bank adjacent to the islands. An
examination c>f the large-scaled charts of the ])ahaiiias will show a large
luu'iibcr of such islets and islands anywhere in this group, either on the
windward or lee faces of the larger islands (see I'latcs IX., XIL). "I'liese
diifereut islets are now prominent, according to the height of the original
hills of which tliey once made a part, or form shallow j)ortions of the
orincijial banks. The long island to the south of Savannah Sound, on
:ho east face of l"Ieuthei-a, is an excellent example of such sloughing oil
(I'lateX. Fi- '-')•

The west face of Xortliern Eleuthera, as seen from our anchorage off

AGASSIZ : BAHAMAS. 31

the Glass Window, is very much eaten away into low perpendicular cliffs
behind a nearly iiat area of considerable extent, eastward of which rise
the hills forming the eastern face of the island. As we proceed south,
the western cliffs become somewhat higher (Plate XXYH.), as the west-
ern liills composing that part of the island, instead of being cut away at
the base like the northern hiils, have been cut away at a considerable
distance inland. To the eastward of the outer western range are piled
thiee to four, or even five, irregular rows of pcolian hills, tljo sides of
which encroach one upon another, and form more or less irregular
ranges of hills according to the width of the island.

Even a casual examination of Eleuthera and of the adjacent islands to
the northward and westward cannot fail to give one a very fiiir impres-
sion of the forces which have been at work on the Bahamas to bring them
to their present condition.

Nortliern Eleuthera is nearly disconnected fi'om Eleuthera by the gap
forming the passage to the south of the (ilass Window. Through this the
sea breaks at high tide, and during a heavy swell the wash covers the path
leading to Hock Harbor. The eeolian hills which form the westward
limits of the bay are protected from tlie sea by Harbor Island and Man
Island. They are the inner range of hills which was once separated by
a low flat from the outer range of which these islands are the remnants.
With the general snbsidence this flat has been changed to the shallow
bay forming Harbor Island Bay. The action of tlie sea conpled with
subsidence has in its turn formed the passage between these islands,
as well as between Current Island and the western spit of Eleuthera.
Similarly, George, Russell, Royal, and Egg Islands, with the adjoining
rocks and islets, have become separated, and are left as tlie outcrops of
a greater Northern Eleuthera which occupied in comparatively recent
times pcrliajis the whole of the northeastern extremity of the Great
Bahama Bank.

The action of the nortli westerly and sonthwesterl}^ winds along the
western face of tiie islands has com|)letely washed away tlie western
ranges of hills which undoubtedly once formed a ])art of Eleuthera
proper. The more southerly extensions now forming that part of the
island from Balmetto Point to Eleuthera Point are tlie remnants of
these ranges, tlie more prominent being tlic hills extending from Elen-
thera Point to Powell Point, tlie outlyers to the eastward of which are
the Schooner Cays, I'inlay Cay, and the great sandbore bank to the
eastward of the so called Middle Gronntl. This shil'ling bunk we can
readily reconstruct as a part of the greater Eleuthera, New Pj-ovidenco
Island, to which 1 have already referred.

32 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

The condition of the western face of Eleuthera shows admirably the
method of erosion, disintegration, and denudation which has taken place
on the inner face of an island facing the great bank of which the part now
worn away must once have been the summit surface.

Going south from the Glass Window, we keep sufficiently within sight
of the land to read, as we pass along, its former history. About two
miles south from tlie " Cove," where we anchored, begin a series of ver-
tical cliffs, which continue almost uninterruptedly as far as Hatchet
Point. Tiiey are full of holes and of small caverns, ribbed with stalac-
tites, giving the face of the cliffs, not only here Init everywhere nearly
in the Ijahamas, the appearance of basaltic rocks, more or less eaten
away at tlie base. The cliff's are the remnants of the headlands, which
have been worn away first, leaving only here and there a slope reaching
from the inner hills to the shore. All along the west coast there is excel-
lent sponging and fishing for large conchs. We found conch-shells and
fragments of corals thrown up fully twenty-five feet above high-water
mark, and lighter fragments of shells and dried stenr^ of Gorgonia^p blown
by the winds to the highest points of the hills on each side of the' Glass
Window.

Dana describes the surface of Metia, an elevated coral island which
presents, I should say, much the same honeycombed appearance so char-
acteristic of the more exposed and weathered islands and islets of tlie
Bahamas, especially as seen at the Glass AVindv:»w (see Plate XXX.). Its
shore cliffs and rounded summits present a striking resemblance to some
of the Bahamas. Compare the figure given by Dana (Corals and Coral
Islands, page 193) with the figures I have given of tlie Bahama a^olian
hills and cliffs. ,

All the way from Hatchet Point to Governor Harbor (Plate X. Fig. 2)
the same succession of vertical clili's continues, with the same undulat-
ing, rolling teolian lulls, perhaps a little lower behind the shore cliffs than
farther north. To the south of Governor Harbor the ground falls off" a
good deal to Savanna SotukI, and tliere are fewer vertical clili's along the
re.st of the shore of Eleuthera extending to I'owell Point, ^yhile coast-
ing along Eleuthera we were taken at a good rate by a fresh northwester,
which stirred up the bottt>m very extensively, and the whole sea was one
mass of milky water carrying a very perceptible amount of particles of
lime in suspension, derived both from the bottom and from tlie shores.
There is to the westward of the island extending from tlie Cove to a few
miles north of Tari)on Point a marked depression in the general level of
the bank, varying from four to six fathoms in depth, with an average

AGASSI2 : BAHAMAS. 33

width of more than eight miles (Plate X. Fig. 2). This must at some
time ill the history of the bank have formed an extensive lagoon or flat
very similar to the lakes and flats now existing on a smaller scale at New
Providence.

Although from abreast of Savanna Sound to Powell Point the general
aspect of the island continues the same, yet from the altered trend of the
shore its outline is greatly changed, and the island widens out between
Powell Point and Eleutliera Point and extends eastward to form a long,
irregular triangle, deeply indented by Tarpon Bay and Kock Harbor an-
chorage. The low spit forming the outer barrier of the harbor seems to
be of recent shoj-e sand origin, and not to be the remnant of some wider
promontory. It is similar in structure to the few beaches and spits of
recent origin which here and there have been thrown up to form small
lagoons or barriers across headlands in favorable localities, where they
are exposed to any length of reach of wind, as this point is to the north-
west winds, which often blow here with great violence.

We leave the bank by a shifting channel leading from Tarpon Point
by the oand spit to the north of Powell Point, to the westward of which
lie an extensive flat formed by sand bars, and a few low isolated cays,
like the Schooner Cays (Plate X. Fig. 2). This flat separates Exuma
Sound from the deeper water on the bank to the northward of the
sound. The shifting character of this channel is well shown from the
fact that, drawing eight feet of water, we passed safely over a bar marked
one foot on the chart as corrected to 1882. This extensive flat, filled
with numerous dry sand bores, extending from Schooner Cays on the
east to Finley Cay on the north and the Sail Rocks on the south, may
mark the position once occupied by the westward extension of Eleuthera
Island.

Entering Exuma Sound we skirted the Sound shore of Eleuthera past
Bamboo Point to Eleuthera Point. The whole of this face of the island
is low, the seolian hills not rising more than twenty-five or thirty feet in
height. Near Eleuthera Point the rolling hills become slightly higher,
some of them reaching forty to forty-five feet, — as Miller Hill, for in-
stance, which is noted on the charts as forty-five feet high. There are
numerous disconnected beaches on the Soimd face of Eleu^lhera. The
100 fathom line is not more than two miles distant from the shore line,
falling abruptly from the 10 to the 15 fathom line. Between Powell
Point and Bamboo Point it is less than a mile off" shore. The eastern
extremity of Eleuthera Island at Eleuthera Point shows admirably the
action of the sea in .breaking through the long and narrow spit which

VOL. XXVI, — NO. 1. 3

34 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

once must have connected Little San Salvador with Eleuthera Island,
and gradually reducing it to the bank which now alone forms the con-
necting linlv, after having worn away the islands first formed, the rem-
nants of which still exist as a series of isolated islets and rocks extending
to the line of breakers to the east of Eleuthera Point, while here and
there a single rock may still be detected standing in the breakers.

Little San Salvador and Cat Island.

Plates I. and XXXIV.

Little San Salvador, and the islets and islands between it and the
northern end of Cat Island, are the last vestiges of the former land
extension of Cat Island, when it must have covered nearly all the space
now limited by the 10 fathom line to the eastward of the island, from
Hawk's Nest Point to Little San Salvador. Little San Salvador will
eventually disappear. It is now low, not more than twenty feet high,
and a long beach broken by six or seven rocky bluffs forms its easterly
face.

Cat Island, where we anchored off Orange Creek, is interesting as
having the highest land of any of the islands of the Bahamas. The

HIGHEST HILLS OF CAT ISLAND.

feolian hills to the north of Orange Creek are marked on the charts of
the islands as being nearly four hundred feet high. Dunes of this height
are not unknown ; there are at the present day in many parts of the
world high dunes covering extensive tracts, as along the Atlantic coast
of the United States, both inland and along the sea border. They often
rise to heights fully as great as those observed in the Bahamas, and
that from comparatively narrow beaches. On the coast of the Baltic
there are long stretches of unbroken dunes for many miles, the crest
of which averages from ninety to one hundred and fifty feet in height,
the summits rising to one hundred and eighty feet. On the west coast
of Africa, near Cape Bojador, sand dunes are said to reach a height of
over five hundred feet.

The west face of Cat Island is a series of low bluffs and beaches. The

AGASSIZ ; BAHAMAS. 35

hills become lower as we proceed south, and then rise again to form the
rectangular shank of Cat Island extending from Hawk's Nest to Colum-
bus Point. The highest hills of this part of the island are to the
northwest of Columbus Point, on the western shore, near Fernandez
Cay. There are a few isolated patches of reefs on the northeastern ex-
tremity of Cat Island, but none along the narrow bank formed by the
100 fathom line, extending from there to Columbus Point. To the east-
ward of Hawk's Nest there are patches and stretches of coral heads,
formins: an excellent reef harbor, Port Howe, between the Devil's
Point and the Bluff.

Nassau to Harvey Cay.

Plate I, ; Plate X. Fig. 2 ; Plate XXXVI.

Passing out from Nassau to the eastward. Potter Cay divides the
main channel into two parts, one of which is quite shallow. The a3olian
character of the low hills forming the base of the promontory at the ex-
tremity of which stands Fort Montague is well marked, and as we steam
out we can see their continuation to East Point. On the other side of
the channel we pass a series of low islands,^ Hog, Salt Cay, and Athol,
on the last of which stands the lighthouse. Their southern face is formed
of low vertical cliffs ; the vegetation upon these islands is scanty.

Dredging occasionally beyond the eastern channel, we brought up
nothing except algfe and corallines ; we passed a few patches of coral
heads and of Gorgonians, generally opposite to openings between outlying
islands which give a freer access to the water from Providence Channel.
Our course lay southeast from the Porgee Pocks, and for a considerable
distance we passed over a bottom nearly barren of animal and vegetable
life ; it consisted mainly of clean sand, with here and there a coral patch
or a cluster of Gorgonians ; but the farther south we steamed on the

1 It seems to me tliat the expl.'inntion fjiven bv Dr. Xortlirop (Trnns. New York
Acad, of Sci., Oct. 1.3, 1800) retranling tlie formation of the cays nortli of Nassau
Harbor is not tlie correct one. Tliey owe tlieir existence, not to the coral snnd wliich
has been thrown up from the outlyinrr coral reef, but to the denudation and erosion
of an outer line of asolian hills parallel to the Nassau rannjc wliich forms their base-
ment. It is true that upon their sea face coral sand has been lieaped up between
the headlands of the a?olian hills composiuir the cays, and sometimes blown to a
considerable height to form broad and hi^h beaches, as on tlie sea face of Hog
Island, but the ajolian rock underlying; them crops out in all directions, especially
on the south side of the cays.

36 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

bank, the less numerous they became. The same was the case also
when we crossed the bank from Nassau to the Glass Window; the
patches of Gorgonians were more numerous as long as we kept in the
vicinity of the line of cays extending from New Providence toward
the Fleeniing Channel.

On reaching the so called INIiddle Ground (Plate X, Fig. 2), we came
upon a mass of Gorgonians and coral heads, througli which our pilot
picked his way by sight. This stretch of coral and Gorgonia ground
did nut extend far ; we soon passed out of it and were back again upon
the usual sand bottom, nearly bare of animal and vegetable life, with
only an occasional patch of corals or Gorgonians. We next came upon
the line of cays extending south of the Ship Channel, and were in plain
sight of the tcolian hills and cliifs of Saddle Cay and Norman Cay. The
bank as we go south continues bare of animal life, no patches even of
Gorgonians or clusters of heads of coral to be seen in any direction. Later
we could see the vertical faces so characteristic of the western sides of the
cays in the vicinity of Conch Cut, and to the south of that the seolian
hills belonging to the chain of cays to the north of Harvey Cay.

Coming upon Harvey Cay, P>itter Guana Cay, and Great Guana Cay
from the west, we were able to foiin an excellent idea of the character
of the shores of the larger cays which are somewhat protected from the
action of the trade wind swells. Their structure is the more striking
in contrast with the outlying small low and water-worn caj's which char-
acterize Conch Cut and Rudder Cut. These low islets are mere ledges
of rocks rising but a few feet above the water line, and likely to stand
but a comparatively sliort time as a barrier to the inroads of the sea
upon tlie larger inner cays.

Harvey Cay is itself greath^ disintegrated ; it is attacked by the spray
and rain. There is but little vegetation upon it, and the limit of action
of the sea upon its eastern face is very plain, at mau}^ ])laces bare rocks
extending; close to the top of the cay. The east face of the cay is pitted
fi'dui tho action of the sea as far as it can reach. The clifls of the lee
shore are full of cavities and caverns, which have been exposed by the
gi-adu;d inroads of tlie sea on the base of tlie a^olian hills. As we steamed
south, we coulil sec the breakers throwing the sea over many of the
smaller outer islets, or dashing its spray high above the summits of
the higli cays of the outer chain on the eastern edge of tlie bank. On
ITarvc}' Cay a little wild sisal grows ; the more sheltered j)arts are
covered with hard-wood shrubs, and the ordinary plants arc found near
luu sliore line. The bottom where we anchored was composed nf a

AGASSIZ: BAHAMAS. 37

kind of sticky marl, — greatly decomposed aeolian sand. The nature
of the bottom is accurately noted on tiie charts, and from tlie examina-
tion of the many samples collected it will be quite possible to give an
excellent account of its characteristics.

On our way south we attempted to reach some of the ocean-holes
marked as existing near Blossom Channel. Unfortunately the sea was
too rough, and it was impossible in the disturbed condition of tlie bottom
to recognize their position.

Harvey Cay is connected with the outer cays to the south of it by a
number of small low cays, half eaten away by the action of the sea, which
reach to Exuma Sound. They form a series of narrow parallel lines
of rocks, with passages like those of liudder Cut between some of the
cays. These islets are in every respect similar to the numerous small
islets we saw as we passed through Conch Cut (Plate XXXVl.). They
extend across the passage between Harvey Cay and Great Cuana Cay.

From Harvey Cay to Great Ragged Island and Columbus
Bank. — The Sand Bores of the Bank.

Plate I.; Plate XI. Figs. 3. 5; Plate XII. Figs. 1, 3, 4; Plate XXXV.

The passage round Galliot Bank gave us an excellent idea of the alter-
nating channels which exist between the different sand ridges running in
a westerly direction from Galliot Island to the Barracouta Rocks. The
navigation is entirely by the eye, and the boat is forced across the deep-
est part of a terminal ridge or of a lateral spur into the nearest channel.
To the windward and leeward we could see on each side three or tour
additional bores, parallel in a general way to the two between which
we were steaming. The channels and their probable deptli were fairly
indicated by the color of the band of water separating tiiem. This
bank is very similar to the bores to be neen on many other parts of the
bank. The sand bores and sand spits separating the channels are
plainly indicated by the liglit emerald-green color of the water, which
is in marked contrast to the darker purplish color of the belts of deeper
water forming the navigable channels.

Of course, where there is such constant aisturbance over the bottom,
and shifting of these large sand masses in accordance with the action of
the winds prevailing for any length of time, as on the Galliot Bank, we
wore not surprised at finding th'e bottom bare of all animal or vegetable
life.

38 bulletin: museum of comparative zoology.

Some of the shallower parts of the bank are practically impassable
even for the smallest boats, owing to the existence of extensive stretches
covered by shifting sand bores more or less exposed at low water, when
they drift like ajolian sand in the direction of the prevailing winds, or are
run into more or less broken ridges parallel with the direction of the
short seas breaking over the Huts. Such tracts on the Great Bahama
Bank are formed on the northeast part of the bank extending ten miles
to the westward of the Berry Islands, where we find the water on the
bank varying from a quarter fathom to a fathom to the westward of the
cays, and extending to a line running diagonally across the bank from
the Northwest Channel to Great Stirrup Cay (Plate Xll. Fig. 4). An-
other track extends to the north and northwest of Andros, about fifteen
miles to the westward of the Joulter Cays (Plate XI. Fig. 3). A similar
small patch runs parallel with the southeast end of Andros for a distance
of nearly ten miles, extending westward from Curley Cut Cays and rising
from the great sand fiat, which increases very gradually in depth as we
go west, having a depth of only two fathoms fifteen miles southwest of
the cays and three fathoms at a distance of twenty miles in the same
direction. Extensive bores also occur to the eastward of the Beminis,
and a belt of sand bores varying in width from one to five miles extends
from eastward of Gun Cay to South Riding Rock, a distance of more
than twenty-five miles.

There are also a number of such bores on the Mackie Bank, to the
eastward of the Beminis (Plate XII. Fig. 2), and between that and the
Northwest Passage another extensive tract, forming a bank of about ten
miles by eight, traversed by numerous sand ridges, carrying from one
and three quarters to two fathoms and from two and a half to three
fathoms between them. From Orange Cay in a norther!}- direction as
far as the latitude of the Beminis we find a number of isolated patches or
banks of sand, held togetlier by the masses of a species of Thalassia grow-
ing u])on them. Aliout twenty-five miles to the south of Orange Cay
there is a narrow belt of sand ridges running nearly parallel with the
100 fathom line for a length of about eighteen miles. These hanks and
bores are limited to the area north of the great marl deposit to the west
of Andros, which extends from a line running west a few miles north of
Billy Island to from five to ten miles south of South Bight, the western
limit of which varies from three to five miles from the 100 fathom line.
Outside of these limits we again find the seolian sand, more or less modi-
fied by the fragments of coralline algfe or of Invertebrates which once
lived upon the banks. Algee also flourish upon the white clay or marl

AGASSIZ : BAHAMAS. 39

district ; but the number of species is not so varied as upon the other
kinds of bottom.

By far the most striking of the sand bore districts is the one which
forms the great sweep of the southern cul-de-sac of the Tongue of the Ocean
(Pkxte XII. Fig. 1). South of tlie cays to the southward of Washer-
woman's Cut (Plate XL Fig. 3) is a tract of about ten by hfteen miles which
is a mass of sand bores, many of them dry. To the eastward for a dis-
tance of tiiirty-five miles they run in a southwesterly direction, at a sharp
angle with the course of the 100 fathom line. In this part of the bank
they are broad ridges, more or less undulating, some of them half a mile
in width and sometimes twelve miles in length, often nearly dry in places,
and with from one to three fothoms on the ridges, separated by broad
channels with from four to six fathoms of water. Some of the wider of
these cliannels are regularly used as approaches to the interior of the
bank, and are known as Queen's, Blossom, Thunder, and Lark Channels,
through which vessels bound for Cuba cross the bank, coming out either
through the Man-of-War Channel, south of Flamingo Cay, or running
west of the Ragged Islands and crossing the Columbus Bank. The east-
ern extremity of these sand bores is formed by a tract of narrow sand
ridges, with deep -water between them, extending some fifteen miles
along the edge of the bank on the eastern face of the Tongue of the
Ocean.

To the eastward of Hawk's Bill Rock there is a line of sand bores to
the south of the line of small cays reaching to the centre of the west
shore of Great Exuma. To the west of the north end of Great Exuma
occur a series of dry sand bores, with from one to two fathoms of water
between them. They trend in a westerly direction, and run north, whei'e
they join the southern bores of the Galliot Bank. An extensive series
of sand bores, many of them dry, runs east from Green Cay across the
bank to within about ten miles of Conch Cut. During one whole day's
sailing from Harvey Cay south nearly to our anchorage off Flamingo
Cay we did not come across any patches of Gorgonians or of coral
heads.

The Brigantine Cays, the Barracouta Rocks, and Hawk's Bill Rocks,
low cays and patches of teolian rock, are the fragments of the western
extension of the northern extremity of Great Exuma.

"When we reached our anchorage off Flamingo Cay, we found the bot-
tom a mass of broken shells, of fragments of corals and Gorgonians, and
covered by Nullipores. As we approached Flamingo Cay we came in
sight of some of the small islands forming a part of the chain of cays

40 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

which extends iu an easterly direction from Flamingo Cay along the
eastern curve of the bank towards the western extension of the central
spit of Long Island. These cays are well rounded and weathered, pre-
senting the usual features of cays forming the broken wall which now
denotes the former extension of the Kagged Cay and Flamingo Land
to the westward till it perhaps once formed part of the Andros, Lobos,
Orange, and Bernini Land.

Some of the cliffs of the western face of Flamingo are high, separated
by small coral and reolian sand beaches. The high coral sand beach
opposite our anchorage formed the sea-wall of a small lagoon. The
southern end of the beach was formed by ft-agments of conchs of all
sizes up to nearly perfect shells, cemented together with itolian sand,
fragments of corals, and broken shells, forming a splendid breccia. The
shells in the lagoon of Flamingo Cay were very much smaller than
specimens of the same species thrown up on the sea face of the beach.

After passing Hawk's Bill Bank we came upon patches of corals and
Gorgonians, which became more extensive as we approached fTamingo
Cay, the water at the same time becoming somewhat deeper and clearer,
the bottom being less affected by the action of the seas due to the pre-
vailing winds.

Flamingo Cay (Plate XXXV.) is comparatively well covered by vege-
tation ; there are many wild guava bushes and shrubs. The aiolian
rocks, as seen on the two sides of the landing beach, are greatly eroded
by the action of both fresh and salt water, the whole surface of the
rounded seolian hills being pitted and honeycombed. The seolian rocks
of Flamingo Cay are much harder than is usually the case with the
jBolian rocks so close to the sea-shore. The cays to the south of Fla-
mingo are excellent specimens of ceolian cays, with rounded summits
almost bare of vegetation, and with surfaces pitted and worn by the
action of the sea and rain upon them. Heavy seas could be seen break-
ing over those nearest tlie outer edge of tlie bank. The vertical cliffs
of Flamingo were riddled with cavities and fissures, as were also the
cliffs of the islands to the south of Flamingo, between it and Man-of-
War Channel.

Steaming south from Flamingo Cay, we cross a series of bars running
apparently at right angles to the trend of the cays, having from two to
two and a half fathoms over the ridges, with numerous patches of coral
heads and of Gorgonians. Some of the cays to the south of Man-of-War
Channel were formerly inlialiited ; but since the destruction of the salt
trade on Great Ragged Cay they have been abandoned. All except

AGASSIZ: BAHAMAS. 41

Nurse Cay, which is quite fertile, are bare, or support at best a very
scanty vegetatiou.

From Nurse Channel the chain of cays continues toward the Great
Ragged Cays. We pass the white cliffs of Nurse Cay, with its low
aiolian hills forming a slightly midulating line on the horizon, separated
by a narrow channel from Buenavista Cay. Next come Racoon Cay and
Double-breasted Cay, all very bare of vegetation, with here and there a
long coral sand beach separating the low cliffs formed by the eating away
of the base of the aeolian hills which form this chain of cays.

The disintegration of the formerly existing land masses, and their
breaking up into smaller masses or islands or islets, and finally rocks
and sunken banks, is also in great part a process not entirely due to the
mechanical action of the waves. Both the seolian and coral shore rocks
become most friable when saturated with sea water, so that large masses
are constantly sloughed off from the base of the hills which project into
the sea. These fragments of greater or less size are themselves rapidly
disintegrated by the same process, resulting in a coarse sand consisting
of oeolian or shore coi'al sand, which helps to form the small beaches so
frequently separating indistinct headlands marked by vertical cliffs. By
this process, beaches, small bights, or diminutive harbors may be formed
in the midst of faces of high or low cliffs, parts of which have been
affected more than others by the action of salt water.

Ocean-Holes.

Plate III.

May we not to a great extent measure the amount of subsidence which
must have taken place at certain points of the Bahamas by the depth
attained in some of the so called ocean-holes, as marked on the charts'?
Of course we assume that they were due in the a3olian strata to the
same process which has on the shores of many islands formed pot-
holes, boiling holes, banana-holes, sea-holes, caverns, caves, sinks, cav-
ities, blow-holes, and other openings in the seolian rocks. Tliey are all
due more or less to the action of rain percolating througli tlie fcolian
rocks and becoming charged with carbonic acid, or rendered acid by the
fermentation of decomposed vegetable or animal matter or by the action
upon the limestone of sea water or spray under the most varying condi-
tions of elevation and of exposure. None of them have their upper
openings below low-water mark, though some of them may reach many

42 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

feet below low-water level. Ocean-holes were formed in a similar way at
a time when that part of the bank where they exist was above high-water
mark, and at a sufficient height above that point to include its deepest
part. Tlie subsidence of the bank has carried the level of the mouth
and of the bottom of the hole below high-water mark.

From the description of the strata which crop out upon the banks in
the vicinity of some of the ocean-holes at Blue Hole Point, there seems
to be little doubt tliat the stratification characteristic of the aeolian rocks
has been observed.

The principal ocean-holes, Blue Holes, are the following : one five to
six miles from Hawk's Bill Rock ; three, of eighteen, twenty-four, and
thirteen fathoms, a little east of north of Blue Hole Point, each about
five miles apart on a northerly line; and two, of seventeen and thirty-
eight fathoms, in the extension of the line of Blossom Channel leading
from the Tongue of the Ocean upon the bank.

I am able, thanks to the kindness of Captain Wharton, the Hydrog-
rapher to the Admiralty, to give three sketches of these Blue Holes,
showing the character of the soundings around them (Plate III.). They
are such as we should expect to obtain from any part of the cays where
there are many light-holes, if sunken below the level of the sea.

At other places on the banks ocean-holes are said to exist. Among
those not on the charts, I may mention a fifteen fathom hole at High
Point, Andros, and a twenty fathom hole in the Middle Bight, between
Gibson Cay and ]>ig Wood Cay. Dr. Northrop has examined some of
the ocean-holes of Andros, and has given a description of those he
visited.^

Except in the case of some wells at Nassau, there has been no obser-
vation of geolian rocks at any great depth below the surfoce. At the Ber-
mudas the a'olian rocks have been traced in situ during the building of
the dry dock to a depth of over fifty feet. The presence there of trunks
of trees would imlicate tlie invasion of sand dunes at so^ne time, much as
they invade the gardens of the Bermudas at the present day.

From tlie desci'iption of the Bermudas given by Pein,. Thomson,
Fewkes. and lloil[irin, there appears to me little doubt that we have
there repeated on the bank of the Bermudas the identical processes
which have been described in this paper, and tliat the Bermudas and
the Bahamas owe their present configuration to the same process of
waste which lias l)ecn going on during their subsidence ; tliat the so
called diminutive lagoons we find there are not lagoons iu the ordinary

1 Trans. N. Y. Acad, of Sci., Oct., 1890.

AGASSIZ : BAHAMAS. 43

sense of atoll lagoons, but are merely small pot-holes or former banana-
holes which have come to be below high-water mark. Certainly there
seems to have been nothing written to prove that the present configura-
tion of the so called Bermuda atoll is directly due to the formation of
the coral reefs which are still growing "upon the bank, and their increase
in thickness owing to the subsidence of the bank. I shall return to
this subject in the description of my visit to the Bermudas.

In the shallower places on the Bahama Bank the whole body of water
was discolored by the presence of sand stirred up by the action of the
waves, and it is only natural that in all those parts of the banks
where the water is constantly rendered turbid by moderate winds we
should meet with so little animal life ; for Gorgonians and corals can
only flourish in clear water, and even corallines cannot obtain a foot-
hold where the ripple-marks are too frequently changed or the bottom
sand is in constant motion. Between Nurse Cut and Kacoon Cut very
little animal life was to be seen on the bank. After passing Racoon Cut
we came upon a good many patches of coral heads and Corgonians. At
the same time the water was gradually deepening to the westward, and
we seemed, steaming parallel with the general trend of the cays, to be
cutting across a number of sand spits about at right angles to the shore
line of the cays, and trending in the general direction of the prevailing
winds.

Great Ragged Island.

Plate XI. Fig. 5.

Great Ragged Island, the most southerly of the cays on the Great
Bahama Bank, does not differ in structure or appearance from the other
cays of the group. The rocks are seolian, hard, full of caverns and
cavities in the cliffs of the west side, with here and there fine sand
beaches between the spits of projecting rocks. At our anchorage the
coral sand was much coarser. After leaving the anchorage we passed
Hobsim's Breakers, to the south of Ragged Island, which are all that is
left of what must once have been an island of considerable size.

Columbus Bank.

Columbus Bank, which lies to the southeast of Ragged Island, is fully
exposed to the swell of the prevailing trades. Whatever cays onre ex-
isted upon it have disappeared, with the exception of Cay Verde and Cay

44 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

San Domingo. The depth of water upon this bauli. is considerably
greater than upon similar stretches of the Bahamas. Here and there
patches of shallow water are dotted over its surface, but the average
depth varies from seven to eleven fathoms. The sea at the time we
crossed it was quite rough, and the action of the heavy swell upon that
part of the bank must be quite effective.

Long Island.

Plate X. Fig. 6 ; Platef XI. Fig. 3.

From the southwestern side of Exuma Sound extends Long Island,
which projects far out from the angle of the bank. The eastern shore of
Long Island is made up of a series of aeolian hills, with more or less
steep faces to the sea. Its highest point is one hundred and fifty
feet. The island is narrow, often not more than a mile in width,
varying from that to little more than three. It flanks the eastern
face of the most easterly extension of the Oreat Bahama Bank. The
100 fathom line is generally within two miles or so of the eastern shore
line. South of Clarence Harbor the island forms a long narrow spit,
running southeast with the 100 fathom line within a cou[)le of miles on

^OLIAN HILLS, CLARENCE HAKBOR.

both sides, while off the eastern face of the northern part of the island an
extensive flat, with less than three fathoms, connects Long Island with
the chain of tlie Exumas. This flat is the continuation of the great flat
Avhich extends to the westward of the Exumas and the line of innumer-
able small cays forming the western bank of Exuma Sound. On the
west side of the island, opposite Clarence Harbor, the bank makes a
great sweep, the outer edge, protected by the Junientos Cays, forming
a half-circle nearly to Ragged Island and the northern edge of Co-
lumbus Bank. A few small salt ponds, formed in the valleys of the
seoliau hills, are found near Clarence Harbor. The general trend of the
seolian hills, both north and south of our anchorage, could be readily
traced till the low hills disappeared below the horizon. Patches of coral
are growing in the harbor, which is protected by small cays from the
prevailing winds.

AGASSIZ: BAHAMAS. 45

From Clarence Harbor we crossed to the west shore of the island.
The salt ponds near Clarence Harbor are separated from the sea by an
outer range of seolian hills. From the level of the ponds we gradually
rose to the top of the second range, perhaps eighty feet above the sea
level, and then gradually passed down the sea face of the slope of the last
teolian range to the low western sliore. Here we found a long, narrow
lagoon, Salina Flat, formed by the throwing up of a low outer bank of
feolian sand recently washed up and inclined at a slight angle to the
sea. The lagoon, which skirts the western shore of the southern part of
Long Island, is about twenty-five miles in length, extending to the south-
ern extremity of the island. To the northward, what is now a flat must
have been a wide lagoon ; only a part of the southern end of the bank
which once separated it from the sea is now left. In fact, as will be
shown later, Long Island, with the string of islands extending north of
Exuma as far as Ship Channel at the north end of Exuma Sound,
and the line of cays forming the curve of the southeastern face of the
Great Bahama Bank as far as Columbus Bank, are the remnants of what
must once have been an extensive island. This island gradually became
separated, first into a series of islands closely packed together, and later,
by greater subsidence, into the innumerable islands now forming the east-
ern edge of the Great Bank, the islands and the Great Bank itself being
all that now attest the existence of the island or islands which must once
have covered the bank inside the 10 fathom limit.

Long Island is noted for its many caverns. On the road across from
Clarence Harbor to the west shore we could not fail to be struck with
the many pot-holes, banana-holes, sinks, and other signs of the extensive
denudation to which the seolian limestone rocks forming the hills had
been subjected. The accumulations of red earth here and there add
their testimony to the extensive action of rains, which must have carried
otf the surface of the hills as they percolated through the fissures of the
rocks and forced their way, little by little, through the porous mass, to
form the numerous and often extensive caverns which are met with in
every direction.^

In many parts of the island we passed through forests with trees of
quite a respectable size, such as lignum-vitae, pidgeon plum, tamarinds,
and the like, giving us an idea of the fine forests which must have
covered some of the islands at the time of their discovery by Columbus,

1 Perliaps the most extensive caves of the Bahamas are those of East Caicos,
described by Sharpies (Proc. Boston Soc.-Nat Hist., XXII. 247, 1883).

46

BULLETIN : MUSEUM OF COMIAKATIVE ZOOLOGY.

and in marked contrast to the pine-covered cays of Andros, Ne>v Provi-
dence, and the Little Bahama Bank.-^

From Cape Verd north the coast of Long Island is formed by low
rounded seolian hills with gentle slopes to the eastward ; it then passes

^OLIAN HILLS AND CLIFFS, SOUTHERN PART OF LONG ISLAND.

into much higher hills, the base of which is formed by vertical cliffs of
aeolian rock, extending to the southern extremity of the opening of
Clarence Harbor. To the north ajolian hills of varying heights succeed
one another, flanked to the eastward for almost the whole lensth of the
island to its northern extremity by vertical cliff's full of holes and cav-
erns. At a short distance south of Cape Santa Maria these eroded cliff's
are quite striking.

^OLIAN CLIFFS SODTH OF CAPE SANTA MARIA.

At the landing place the shores of Clarence Harbor consist of recent
coral sand strata, dipping slightly to the sea. The summit of the ridge
of one of the islands forming the outer barrier of the harbor, say twenty-
five feet above high water, was formed of a?olian rocks ; on the inside,
round the base of it, shore coral deposits had collected, which were ex-
posed in the flats between the feolian hills forming the outer line of the
harbor and the shore of the island. On the summit of the outlying
islands we observed many huge angular blocks similar to those which
are thrown up by hurricanes and line the outer shores of so many of
the islands of the Bahamas.

We were able on this island easily to observe how the sea undermines
the seolian rocks. Huge rocks are thus broken off" from the sea foce, fall
into the sea, and are in turn broken into smaller blocks, either as they
crumble from the fall, or by the subsequent action of the sea as they
lie piled up just as if the base of the hill had been blasted ; and sub-
sequently, at times of violent storms or of hurricanes, these huge

1 See List of Plants collected in the Bahamas by A. S. Hitchcock (4tli Ann.
Rep. Mo. Bot. Garden, 1893).

AGASSIZ: BAHAMAS. 47

masses are hurled inland and deposited far above the ordinary high-
water mark. All the way along the sea face of Long Island we passed
large masses of gulf-weed, many of these patches more than twice the
size of the yacht.

The difference in the aspect of the vertical cliffs in different parts of
the Bahamas is very marked. If formed at the base of the gentle wind-
ward slope of the asolian hills, they are low and of very uniform height,
especially if that face is the bank face of the island ; but where the lee
face of the hills — lee wlien formed — is also the sea face, the cliffs
are on the steeper slope of the teolian hills, are far more irregular, and
their height varies greatly, according to the height of the hill which
is attacked and the distance inland to which the action of the sea has
reached.

The corals do not form a regular reef off the east face of Long Island.

The Exuma Islands.

Plate X. Fig. 5 ; Plate XI. Fig. 1.

To the west of Cape Santa Maria, Long Island has disappeared, leav-
ing only a shallow bank flanking it for its whole length and connecting it
with the Exuma Islands. From the central part of the western line of
these islands extends a long, narrow spit, the only mark of the former
extension of Long Island in that direction. It must have been a con-
tinuous shore, forming a great curve, indicated now by the row of islets
and rocks which flank that part of the bank, extending in an unbroken
line to Flamingo Cay, and from there by Seal Cay and the islets run-
ning to the north of Columbus Bank from the Ragged Cays as far as
Nurse Channel.

Great Exuma is the largest fragment remaining of the land which once
formed the eastern edge of the bank flanking the west side of Exuma
Sound. Like the many islands, islets, and rocks extending northward as
far as Ship Channel, it is built up of aeolian rocks. Georgetown, the prin-
cipal port of Exuma, lies upon a long inland sound studded with islands,
and sheltered from the outer sound by a series of low outlying islands
which form a barrier against the force of the trade swell, which gains
considerable force over the forty miles or more of sea-way from the east-
ern side of Exuma Sound. These outlying islands are themselves gradu-
ally being eaten away, and were once also a part of the greater Exuma
land, which has little by little become dismembered by the action of the
sea upon the subsiding seolian hills. In the inner sound, which extends

48 bulletin: museum of comparative zoology.

for many miles southward and northward of Exuma Harbor, and forms
one of the prettiest stretches in the Bahamas, there are many diminutive
cays, isolated rocks of all sizes, with rounded ridges undermined at the
water's edge and ready to topple over and disappear. The seolian rock,
kept constantly wet, becomes soft and readily crumbles, and is washed into
ajolian rock sand. The whole floor of the harbor is covered with fine
sand of this kind, and the harbor is gradually filling with material de-
rived from the wash of the windward row of outer islands. Off Exuma,
between it and the northern end of Long Island, the edge of the bank
is comparatively wide. The British Admiralty Chart, No. 393, shows
admirably the gradual wasting of the land which has taken place to form
the inner sound. ^ There was comparatively little animal or vegetable
life on the floor of the harbor where we examined it. On leaving Ex-
uma we sounded at short intervals, and found the sea slope of the bank
much less steep and more gradual than that of other faces we had
explored.

From Conch Cut to Green Cay.

Plate XXXVI.

From Exuma Harbor we steamed northward to Conch Cut, keeping as
close as was prudent to the outer line of cays, which all showed indica-
tions of great erosion. On the northern extremity of Groat Exuma the
ajolian hills are closely packed, and reach a height of one hundred and
fifty feet. The islands of Lee Stocking and Great Guana present the
usual features of seolian hills attacked by the sea at the base of the longer
slope. The smaller cays are bare, while even on Exuma and the larger
cays the vegetation is far less luxuriant than we might expect from the
size of the islands.

Passing through Conch Cut to enter upon the bank on our way to
Green Cay, we were struck with the number of small islets, which
form a wide protecting belt against the encroachments of the sea. The
islets are all exposed on the lesser slope of the ajolian hills to the action
of the trade swells, while the steeper face has been attacked by the
shorter, sharper waves reaching across the bank, and undermining the
western faces of the islets to an extraordinary degree. The solvent action
of the sea water has also undoubtedly played an important part in pro-
ducing the fantastic shapes which some of the islets and isolated rocks
have assumed ; so that by this combined action a wide shallow bank

1 See also U. S. Hydrographical Charts, Nos. 26'> and 26".

AGASSIZ: BAHAMAS, 49

has been formed which connects all these islets together, leaving only
such entrances as Conch Cut, Wide Opening, and Wax Cay Cut for the
passage of craft of shallow draft. The light green color of the bank

WATER-WORN ISLET, CONCH CUT.

connecting the islands indicates very markedly its position as contrasted
with the deeper waters of the passages and of the outer sea edge of the
bank. The larger cays are flanked on the western face by white sand
beaches, formed by the rapid disintegration of that side of the cays.

All the way across the bank, from Conch Cut to Green Cay, we found
but little life upon the bottom ; there were no patches of weeds or of Gor-
gonians, the bottom being everywhere composed of coarse seolian sand.
But when about two to three miles ofl' Green Cay, patches of coral heads
and of Gorgonians begin to appear, and become more numerous as we
approach the western edge of the bank forming the eastern side of the
Tongue of the Ocean. The absence of animal and vegetable life upon
the bottom of the interior of the banks is undoubtedly due to the con-
stant shifting of the coral sand from the action of the sea. At moderate
depths of one to three fathoms we could everywhere see that action
plainly indicated by the presence of ripple marks. In the shallower
parts of the banks this action forms great sand bores, which, exposed to
the action of the winds, also tends to increase them in size in the direc-
tion of the prevailing winds. To the eastward of Green Cay we could
see such a great sand bore, seven feet high, forming as it were a cay
consisting of nothing but a constantly shifting tract. In many locali.ties
on the banks these great sand bores have assumed quite definite posi-
tions, whicli they retain, merely shifting north or south or advancing
eastward or westward within narrow limits. At our anchorage off Green
Cay the bottom consisted of fine hard coral sand, fairly well covered with
coralline algtB.

Green Cay.

Plate XXII.

The western extremity of Green Cay itself is the termination of one
of these shifting bores, and its frequent changes of position and dimen-

VOL XXVI. — NO. 1. 4

50 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

sion are known to every skipper who anchors off its shores. Green
Cay has been extended by this sand bore to the very edge of the plane
where the sudden drop in depth beyond the 20 fothom line takes place.
The seolian rocks of Green Cay crop out at the eastern edge of this sand
bore. Against these outcrops the bore rests, and in some places has com-
pletely overwhelmed them. At the eastern extremity of the western
spit a mass of blocks of teolian rock has been piled \ip by some hurri-
cane. On Green Cay we find the same pup<T3 so characteristic of all the
Bahamas. The vegetation reminds us of that found on Eleuthera along
the shore line. The principal mass of Green Cay, which rises to a height
of sixty feet, is of seolian origin ; but the extent of the cay is greatly
modified by its being a buttress for the sand bores which extend from it
both to the westward and eastward. Green Cay is marked among the
Bahamas for the combination it presents of sand bores and seolian rocks,
and shows how some pai'ts of the banks themselves may have been
formed of movable bores at certain stages of their subsidence, when
masses of sand large enough to become afl'ected in their movements by
the action of the sea and of the winds had accumulated from the dis-
integration of fixed nuclei of teolian rocks. It is very easy to imagine
such a process as we saw going on at Green Cay to liave been an impor-
tant fiictor in connecting adjoining hills or in filling low valleys between
neighboring islets. Green Cay is the only cay left to indicate the former
existence of the land which once extended round the edge of the Tongue
of the Ocean all the way from New Providence to the islets to the south
of Andres, parallel with the line of the 100 fathom soundings, which
now alone indicate its probable original outline.

Andros and the Western Part of Great Bahama Bank.

Plate XI. Figs. 1, 3; Plate XII. Fig. 3; Plate XXXIII.

From Anguila Island we crossed the Santaren Channel during the
night and made for Wide Opening early in the morning. When in sight
of Andros we made a haul for a specimen of the bottom, and obtained our
first sample of the so called " white marl," wliich is marked on the charts
as extending many miles west of tlie Wide Opening of Andros, as well as
to the north and south, forming a liuge triangular patch characterized
by this peculiar bottom. It consists of a very fine chalky ooze, resem-
bling plaster of Paris which has just been mixed for setting.

From our anchorajre to the shore of Andros at the mouth of Wide

AGASSIZ: BAHAMAS.

51

Opening, a distance of about eight miles, we found the bottom sloping
gradually, with here and there a few blades of Thalassia. It was every-
where composed of the same marly substance, and sustained but little
animal life. A few Lupas, a couple of small sharks, and the holes of
a Squilla seemed to be the limit of the fauna and liora of this marly
waste. On reaching the shore, we found the land, as far as we exam-
ined it, to consist of the same white marl as the bottom, but still
sufficiently solidified to enable us to walk upon it. It was somewhat
elastic, giving us the sensation of walking upon a sheet of India-rubber.
In some places tlie marl was covered with a l)lack alga, forming a thin
crust, or it was often coated with a harder material composed of minute
fragments of shells and of sand, giving the surface a gray appearance.
On digging into the soil, we found the same white marl, more or less
mixed with vegetable matter. The shore upon which we landed was at
no point more than from twelve to fourteen inches above the high-water

s;::;,^^;^.- • " -->- mI' "J' o

WEST SHORE OF ANDROS, WIDE OPENING.

mark. The country inland seemed well covered with low vegetation,
and mangroves flourished in every direction. Here and there a ridge of
sand had been blown up, composed of fragments of shells and of crabs.
These ridges, rising a few inches higher than the general dead level
around them, formed the high ground, as it were, upon which was
growing a somewhat richer vegetation, composed mainly of the same

52 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

plants so characteristic of the shore lines of all the Bahamas. Upon
these saud ridges also flourished small groves of palmettos.^

As we steamed northward, the low, dark line of Audros stood out in
marked contrast with the intensely light-green water, extending as far as
the eye could reach over the shallow bank to the westward of the island.
Here and there the low line was broken by a high mangrove tree loom-
ing up like a cliff in the distance, or was interrupted' by a stray line of
high palmettos oh some sand ridge near the shore. The edge of the
island near Wide Opening is occupied by a lagoon inside the shore
line^ and the deep bights which characterize the island show clearly
how it has little by little been eroded, then cut into halves and
thirds, next into smaller cays, and finally, wearing away having com-
menced, has left here and there a small cay (Plate XII. Fig. 3, and
U. S. Hydrographical Chart, No. 26*) to the westward of the main
island. As the water is quite shallow all along the west shore of An-
dros,its action is most feeble, and it must have triturated and ground
very slowly all the shore material into the very fine and minute particles
composing the " white marl " covering so large an area of the bank be-
tween the island and the Santaren Channel. This white marl looks
almost like deep-sea chalky ooze ; it has about the same consistency, is
made up of the same nearly impalpable fine particles, and is of the same
whitish color. A similar white marl deposit is also found, only on a rhost
limited scale, in protected pools exposed to the very lightest wash of the
sea in the recesses of the shore line.

The shores of the inner lagoons, and in many places the main shore
of the island, is lined with mangroves, many of which are large, and form
the most prominent landmarks available in the navigation off the west
shore of Andros.

As far as Williams Island, about twenty-five miles to the northwest of
Wide Opening, the bottom on the bank is composed of the same white
marl. All the way to Billy Island we found no change in the character
of the bottom, which seemed fully as barren of animal life as it proved
while rowing to the shore of Wide Opening from our anchorage. On
our way north from Wide Opening we steamed in water so shallow that
we left in our wake a broad belt of white water, stirring up the white
marl with the wash of the screw.

On going ashore at Billy Island we found it to consist of the same
finely triturated seolian rocks. The shores were formed by miniature

^ See Northrop's description of the " Swash " (" The West Coast of Andros,"
Trans. N. Y. Acad, of Sci., Oct., 1890).

AGASSIZ: BAHAMAS.

vertical cliffs of about eighteeu inches in height, presenting all the
appearances of water action so characteristic of the higher limestone
clifts on other islands of the Bahamas. The diminutive seas had eaten
out small bays, formed promontories, and indented the coast in a man-

TOUNG MANGROVES, WIDE OPENING.

ner no less characteristic tlian the shore lines of higher islands when
exposed to the full action of the trades. Masses of dead shells are found
blown up or thrown up on the diminutive beaches of the recesses cut out of
the shore line. Billy and Williams Islands must at one time have formed
a part of the northwestern extension of Andros. There is no part of the
Bahama Bank which is so instructive as that now occupied by Andros.
Nowhere else do we find so large an island undergoinij; all the processes
of disintegration, division, and ernsinn wliich have on other parts of the
bank ended in forming the submarine shoals, and leaving liere and there
traces of the former extension of the larger islands of which tlie bank
was composed. Andros still occupies a romparativoly large part of the
Great P)ahama Bank to the west of the Tongue of the Ocean ; yet it is
cut into three islands by the so called bights wliicli connect the Tongue
of the Ocean with the shallow waters of the bank to the westward of
Andros. Its former soutlicrn extension is marked by tlie numerous
small islands, isolated rocks, sand banks, and ridges reaching southward
and eastward from the southern end of Andros. Its former western

54 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

extension, perhaps close to the edge of the Santaren Channel, is well
marked by the great belt of white marl, which we may see formed under
our eyes on the shores of Wide Opening. The northern limit is marked
by the spread westward and northward of the white marl, with here and
there a I'emnaut of the main island as an isolated cay, like Billy or
Williams Island, reaching perhaps to the northern extremity of the
bank, where the only traces of its former existence are the rocks, islets,
and shallow bars extending from Great Isaac to Great Stirrup Cay,
forming to the eastward of the northern extension of Andrus a broad
promontory, of which the Berry Islands and the Joulter Cays are the
only remnants. On the west, along the line of the Old Bahama Chan-
nel, the Santaren Channel, and the Straits of Florida, we are reminded,
by Lobos Cay, Guinchos Cay, and the disconnected flats and patches edg-
ing the channel as far as Orange Cay, of the western extension of Audros ;
while the Riding Rocks, Gun Cay, and the Bemini Inlands, lying be-
tween Orange Cay and the Great Isaac, attest the former existence of
the shore line of a large island — probably the west coast of Andros —
along the eastern edge of the Gulf Stream.

The appearance of the east coast of Andros, the high aaolian bluff of
which flanks nearly the whole of the western edge of the Tongue of the
Ocean, is in striking contrast to its low western shore. The east face of
Andros is a series of alternating beaches and bluff's, extending from the
northern extremity to High Point. The island is comparatively well
wooded, large tracts being covered by pine and by hard-wood forests.
To the south of Morgan's Bluff" near the shore begins a magnificent
coral reef, extending the whole length of the island, and running paral-
lel with and distant from it half a mile to a mile and a quarter, with
openings in the outer reef to allow the passage of spongers. This reef,
though narrow, is one of the finest reefs I liave seen, and the patches
of corals and Gorgonians which flourish between the outer reef and the
shore are not surpassed in beauty by the corals of any district known
to me. Tliese patches form an intricate network, rendering naviga-
tion inside the reef very difficult. In fact, it can be followed only by
the eye. The inside patches of Gorgonians are in many places most lux-
uriant, while the outer reef is mainly made up of masses of Astrreans
and Madrepores; the patches between tlie outer reef and the shore
consist of Porites, Mpeaudrina?, and Millepores, growing upon the Kolian
rocks. These corals, when broken up by the action of the sea, supplied
the sand forming the beaches which cover the underlying aeolian rocks.
I saw but few stretches of the shore coral rock formation.

AGASSIZ : BAHAMAS. 55

The vegetation of Andres consists mainly, as I was informed by Mr.
Cliamberlain, of pine, mahogany, mastic, and the heavy undergrowths.

The ridge which runs parallel or nearly so to the coast is followed by
a second ridge from forty to sixty feet high, separated from the first by
a wide flat plateau, beyond which we come, by a rapid descent about five
miles inland, upon the low flat land which extends to the western edge
of the island and forms the marl shores we visited at Wide Opening.

Morgan's Bluff' (Plate XXXIIL) has the finest limestone cliff's on the
eastern face of the island. The outlying islands and rocks and islets off"
the east coast of Andros, of which High Cay is a good example, as far as
the South Bight, are all of fEolian origin, and have been separated from
the main island by the same agencies which have been at work in other
parts of the group. The a^olian rocks of Andros itself differ in no way
from those of other islands. The eastern edge of Andros is separated
from the deep water of the Tongue of the Ocean only by the narrow
shelf of the bank of which the eastern edge is occupied by the long bar-
rier reef which protects the eastern edge of Andros, leaving an excellent
well prutected inside channel for small boats along the whole length of
the island.

From Andros to Orange Cay.

From Billy Island we steamed across the northern end of the great
white marl belt forming the southern edge of an excellent sponging
ground which extends northward towards the edge of the Bahama
Bank, south of the Northwest Providence Channel. The white marl,
as we stopped to dredge or to sound, seemed very barren of animal
life. Here and there an occasional sponge could be seen. On our
way to Orange Cay we found that the white marl as we got farther
west gradually contained more coral sand, which became coarser and
more abundant as we approached the western edge of the bank, where
the bottom was again composed of the characteristic coral and a;olian
sand found upon other parts of the Bahama Bank. With the in-
crease of the coral sand we came upon a species of Thalassia with huge
roots, by which it anchored in the fine marl. The great development of
the roots is very characteristic of the coralline algfe, which thrive upon
the coral sand bottoms. A macroscopic examination of the marl from
the shore of Billy Island showed that it contains a good deal of vegetable
matter and a few F<iramiiiifera, an indication most probably that the marl
was formed in a lagoon with free access to the sea, perhaps a sink much

56 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

like those on Nassau and on other islands, in the shape of salt ponds.
The presence of Foraminifera in this marl, so like deep-sea ooze, yet
known to be formed only by the slow trituration of aiolian rocks in very
shallow waters, raises the question of the depth at which Radiolarian
earth, as at the Barbados and elsewhere (Baracoa), niay have been
formed. It is no more difficult to imagine that the formations round
the shores of such a shallow water lagoon have been elevated, than it is
to believe the same of an elevated coral reef. In the Billy Island marl
the presence of Lobulina and Globigerina would merely indicate that the
open sea had free access to the lagoon. Similarly, the Radiolaria of the
Barbados earth may have floated into the shallow waters in which it was
formed, were it not that the presence of Cystechinus, by analogy with
our recent species, indicates deep water. Tlie marl of the shore of
teilly Island contained Lobulinag and Globigerina^, while the marl from
nearer the edge of the bank, in about four to five fathoms, and within
four miles south of Orange Cay, which contained a greater percentage of
coarse coral sand, also contained Lobulina^ and Globigerina, as well as
fragments of silicious Radiolarians. When within five miles from the
edge of the bank the bottom was quite clear of mnd and marl, and was
composed mainly of nicely rounded particles of coral sand.

From Orange Cay to Great Isaac.

Plate I. ; Plate XII. Fig. 3 ; Plates XXIV\ to XXVI.

When approaching the western edge of the bank off the Racolas
Rocks to the south of Orange Cay, we began to see bars of corals in
from five to six fathoms of water. This had now become much clearer,
and, compared to the water on the midway of the bank, apparently car-
ried nothing in suspension. Still farther to the westward could be dis-
tinctly seen the dark blue line indicating the deep water of the Gulf
Stream off the edge of tlie bank. After passing the Racolas Rocks, we
came upon Orange Cay, a low ridge of a^olian rocks absolutely bare,
pitted and honeycombed over its whole surface. Vessels crossing
the banks from Stirrup Cay sail an almost straight course from
there to Orange Cay, and next strike Salt Cay Bank, thus avoiding a
great part of the strength of the Gidf Stream, and making for Havana
from off Matanzas close t(^ the north shore of Cuba. From Orange Cay
as far as Riding Rocks bars of corals run parallel to the edge of the
bank on an approximate north and south line. The corals on these

AGASSIZ: BAHAMAS. 57

bars are all flourishing, and consist mainly of heads of Astrseans and
of MsEaudringe.

The shore lines of Orange Cay, as well as the adjoining isolated rocks,
are all eaten away at the base, the surface being completely pitted and
honeycombed; this small cay, with the scattered isolated rocks extending
to Riding liocks, being all that is left of the western part of the great
Andros Island land. Riding Rocks are found, like Orange Cay, to be
composed of the same kind of aeolian rock. Masses of corals are flour-
ishing, forming great patches more or less disconnected on the face of
the bank towards the Gulf Stream, extending to fifteen or seventeen
fathoms of water. Astra^ans, Maeandriuce, Millepores, and Madrepores
extend all the way to Gun Cay. Nowhere perhaps oji the Bahamas is
it more obvious that the existing bars of corals found along the sea face
of the nearly continuous bank throughout the Bahamas have had noth-
ing to do with the formation of the cays and rocks on the west side of
the Andros Bank. The coral masses are merely surface growths as it
were upon the banks, which owe their building up to entirely different
causes and to a state of things which has long passed away.

From Riding Rocks to Gun Cay a series of isolated patches, rocks,
and islets, all of the same character, sometimes in two or more parallel
rows, run north on the edge of the bank ; these are flanked on the west-
ward by very flourishing patches of corals growing close together, and
extending on the sea face of the bank into about seventeen fathoms
of water. These rocks and islets are comparatively less barren as
we go north. The remnants of the former land also increase in num-
ber as we approach Gun Cay, and become still more numerous and large
with the Beminis and the adjacent rocks and islets. On the southern
end of South Cat Cay there is a considerable peolian hill, covered with
vegetation similar to that of other islands on the east face of the
bank. North Cat Cay has a grove of cocoanut palms and low bushes.
The sea face shore lines of this and South Cat Cay are protected against
the action of the sea by lines of outlying rocks, and are thus less affected
by the wearing action of the prevailing winds. South Cat Cay is high
enough to have, near the middle of the north end, a line of low vertical
cliffs, both the Cat Cays being somewhat higher than Gun Cay.

58

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Gun Cay.

Plate XII. Fig. 2.

On the south point of Gun Cay a lot of loose rocks are thrown
up well above high-water mark. The cliffs of the southern shore are
eaten away, and on the western face of the cay the rocks up to the line

^OLIAN ROCKS THROWN UP ABOVE HIGH-WATER MARK, GUN CAY.

of scanty vegetation are deeply pitted and honeycombed. All along the
west face of Gun Cay rectangular masses of rocks are thrown up close to
the line of vegetation. These masses of rocks are the fragments of the
shores torn off from the vertical cliffs by undermining, and then thrown
up by the action of the heavy seas or hurricanes. Picket Rocks derive
their name from such a mass of angular rocks thrown up on the very
crest of the ridge of the cay, looking a little way off like an irregular
picket fence. Holm Cay is another island of similar character, but
smaller.

Running paranel with the line of outer rocks which make the extension
of Gun Cay to the north, we crossed magnificent banks and patches of
corals, some of which run close up to Gun Cay, as well as to the rocks
in the vicinity. Off the little sandy reach inside of the north end of
Gun Cay we brought up large masses of Thalassia in the dredge.

On the east face of Gun Cay the shore rocks are so eaten away as to
form numerous caverns and cavities in the face of the shore line.

AGASSIZ : BAHAMAS. 59

Going on shore, we found it composed of the ordinary aiolian rock of
tlie Bahamas, its character admirably shown in the cuts made for the
boat landintj and for tlie foundation of the derrick used in handlinj?
the sni)phes. Tiie cay, wherever exposed to the strong waves of the
Gulf Stream, is deeply pitted and honeycombed. We found a few speci-

-r':<f^'piTSSit^.

GUN CAY.

mens of the same pupa so common on other islands of the group.
The vegetation was very scanty, — a few scattered sjiecimens of cac-
tus, wild grape, wild geraniums, and verbena, — with tlie usual plants
characteristic of the shore belt throughout the Baliamas. Among the
most prominent rocks north of Gun Cay, between it and the Beminis,
are Turtle Rocks, forming a low rocky icolian ridge.

The Beminis.

Plate XII. Fig. 3.

The Gulf Stream skirts close to the western edge of the two Beminis.
They are separated by a picturesque opening, forming an excellent har-
bor even for boats of a considerable size. At the south end of both
South and North Bemini we tind a long line of angular a;olian rocks
thrown up by hurricanes well above high-water mark, often fully twenty
feet. From here to the northern end extends a long C(M-al sand beach,
disconnected in many places by extensive patches of shore built coral
rocks, covering in many places the underlying reolian formation. The
greater part of the shores of the Beminis is made up of this shore
deposit, piled up above the older reolian rocks wherever they have been
washed away, and forming small cliffs along the northern face of the
western shore line of North Bemini. Here and there an occasional
cliff of the older feolian is seen, but greatly weathered and worn, as
near the northern extremity of North Bemini, some distance above the
village. The sea has thrown up a high beach and formed a lagoon at
that end, which is filled with mangroves and small mangrove islands.
The outlying rocks of North Bemini are all teolian, the surface pitted

60 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

and honeycombed, and eaten away at the base. On the gulf side of the
Beminis, and all the way north to Great Isaac, we meet tine patches of
corals in from four to six fathoms of water. These patches and bars
extend but little way eastward on the bank, and flourish most luxuri-
antly as we approach the Gulf Stream edge of the bank out to fifteen
or sixteen fathoms, where they often cover extensive tracts separated
by irregular sandy lanes.

Great Isaac.

Plates XXIV. to XXVI.

Great Isaac, as it is approached from the south (Plate XXV.), reminds
one of Double Headed Shot Cay, with its rounded outline and outlying
islets and rocks presenting a most barren and desolate appearance. The
southern faces of the island, and of its adjoining islets and rocks, are all
cut off by low vertical cliffs, full of caverns and cavities. The surface
of Great Isaac is an admirable example, perhaps one of the best, of the
forces which have been at work during the subsidence of the Bahamas.
The aeolian structure of the rocks is everywhere most apparent. The
action of the sea in undermining the shore slopes and forming low ver-
tical cliffs by the breaking off and falling into the sea of huge angular
masses of rocks is seen on all sides. The action of the sea during hur-
ricanes in throwing up huge masses far above high-water mark is not
better shown on any other island of the group. Finally, the action of
the salt-water spray and waves, as well as of the tropical rains, in pit-
ting and honeycombing the surface of the island, the former near the
shores, the latter everywhere else over the cay, cannot fail to strike the
eye of the most casual observer. As we passed the west end of Great
Isaac we observed a number of huge isolated irregular rocks, left far
above the high-water mark, passing into an irregular mass of angular
blocks forming a rude broad wall, left as a witness of the fury of the
hurricane of 1870. ]iut on the north face of the island we come upon
another tract covered by still larger blocks, thrown up by an older hur-
ricane, the date of which is not known. Two of the largest of these
blocks we measured, and found them to be, the one twelve feet eight
inclies long by four feet wide and six feet high, at a distance of one
hundred and twenty-five feet from the shore and fully twenty-five feet
above high-water mark ; the other was fifteen feet six inches long eleven
feet wide and six feet liigh, at least eighty feet from the shore and about
eighteen feet above high-water mark (Plate XXIV.). On the highest

AGASSIZ : BAHAMAS. 61

part of the island, to the eastward of the lighthouse, we found a number
of fantastic isolated pinnacles of ceolian rock left standing high above the
present general level of the cay by the denuding action of the trojjical
rains. Many of these pinnacles were limestone tables of varying tops
and heights, with a slender pedestal still attached to the ground. They
owe their origin undoubtedly to the harder character of the surface
limestone at certain points.

On the north side the slope of the island is very gentle. There are
many pot-holes and sinks, as well as ocean-holes, on the northern face of
the cay, and at the south side the sea has vvashed the cliffs into the most
fantastic pinnacles (Plate XXVL), some of them several feet in height,
so that a j)art of this shore recalls very vividly the aspect of the surface
of Eleuthera near the Glass Window. Only the most scanty vegetation
is found at certain points of the island, not enough for tlie food of a goat.
A cactus, a few stunted bushes and shore plants, compose the flora of this
island. Gun Cay, the Beminis, Great Isaac, and the long line of rocks
and sand bars to the eastward of it extending to the Berry Islands, such
as North Eock, the East and West Brothers, and the Eastern Isaacs, are
probably the remnants of the northern and western edges of a larger
island, now only barely apparent above tlie level of the sea, perhaps itself
formerly a part of Andros. Off Little Isaac the bottom samples became
quite coarse, consisting principally of broken shells and masses of coral-
line algfc. When approaching Great Stirrup Cay, about thirty miles
southeast of Great Isaac, the bottom was coral sand, with many broken
shells, small whitish Mellita hexapora, and light brownish Clypeaster.
In fact, the prevailing tone of the whole fauna of the banks proper is
of light hue, except where coral patches are found close to the edge of
the bank, and nothing is more apparent than the bleached aspect of
the shells, echini, or coralline algae characteristic of the bank itself. The
darker patches of sponges, corals, and Gorgonians alone apparently form
an exception to the general monotony of color characteristic of the bottom
on the inner parts of the banks.

From Great Isaac to the Berry Islands.

Plate I. ; Plate XII. Fig 4.

To the eastward of Great Isaac run a series of rocky ledges, the
remnants of former cays, running parallel with the trend of the 100
fathon> line, distant from it three to seven miles, known as the North-
east Rock, the Brothers, Little, Middle, and East Isaacs, Rockawash, and

62 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

other rocky banks, ending witli the Gingerbread Ground in the east.
There we strike the extensive bight between it and Little Stirrup Cay,
where the bank slopes most gradually from the 3 fathom line to the
12 or 15 fathom line, and then drops suddenly to the 100 fathom line,
the ledges named above forming the northwestern edge of the bank.

The Berry Islands.

Plate XII. Fig. 4.

We passed the islands to the westward of Great Stirrup Cay in the
dark, but that island and those of the Berry Islands which we saw dur-
ing the day were all of aeolian rock. Great Stirrup Cay is full of pot-
holes and sinks ; it is comparatively well wooded, as well as many of the
Berry Islands and of the islands to the northwest visible from our an-
chorage in Great Harbor.

All the way from Gi-eat Harbor south we found irregular patches and
bars of corals on the shelf to the eastward of the Berry Islands, extend-
ing from three to four fathoms or less to fifteeua or sixteen near the
eastern edge of the bank. Coming from the west, I am informed that
from Gingerbread Ground similar coral reefs extend as far as Great
Harbor Cay, on the shelf of the banks, outside of the cays.

The islands, islets, and rocks known as the Berrv Islands, extending'
along the Providence Channel and forming the northeastern edge of that
part of the bank of wliich Andros is the principal land, are both as
to numljer and size in marked contrast with the small and insignificant
islets and rocks which occupy the edge of the bank flanked by the Gulf
Stream. The principal cays of the Berry Islands, such as Haines Cay,
Little Harbor Cay, Alder Cay, Bond's Cay, as far as "Whale Cay, present
no features of special interest. Their surface appears well pitted and
honeycombed, as their eastern face gets the full force of the northeast
trades. Near their highest point they are covered by a very scanty
vegetation. The outer line of cays protect a beautiful sheet of water of
a brilliant light green.

They show the usual variation in height of from twenty-five to thirty
feet, occasionall}' rising, as at Haines Bluff and Devil's BhifT, to fifty or
sixty feet, their eastern faces presenting the ordinary variation of low
vertical cliffs Avhere the headlands have been cut off by the action of
the sea, or the more or less extended coral sand beaches stretching
between those promontories, with here and there extensive walls of
ajolian blocks thrown up above high-water mark, as on the Market Fish

AGASSIZ : BAHAMAS.

Cays. The cays in this vicinity are moi'e exposed to the action of the
sea, the eastern shelf of the bank being quite narrow, and in many cases
the eastern faces are formed by low vertical cliffs of a;olian rocks, thirty
to forty feet high, separated by short sti'ctches of sand beaches. The

DEVIL S BLUFF.

north end of Bond's Cay and the sea face of Alder Cay are both very
mucli eaten by caverns, with rows and patches of loose angular blocks
thrown up above high-water mark. There is but little vegetation on
either Cay. Beaches and low cliffs alternate along the southern end of
Bond's Cay. Along the whole length of the nurth end of Whale Cay
the action of the sea is well marked in the undermining of the low
cliffs forming the sea face of the cay.

The effect of the shallow wide shore shelf to the east of the Berry
Islands is very marked on the swell, which is far less powerful than on
the Atlantic or Gulf face of the bank, wherever deep water comes close
to the sea face of the cays.

An extensive tract of sand bores, dry at low water, runs from the
Northwest Channel to Great Harbor Cay. In the whole of the track to
the eastward of them as far as the Berry Islands there is only a very
limited area with a depth of one fathom. To the westward of the cays
from Great Harbor Cay to Whale Cay the shores run into broad sand
flats. The westward extension of these sand flats forms the southern
edge of the bank from Whale Cay and the Chub Cays to the entrance
of the Northwest Channel and to the Joulter Cays. From that edge of
the bank extensive sand bores run diagonally across in the diiection of
Haines Cay. (Hydrographic Chart, No. 2G^)

Fine patches and bars of corals follow us south, with lanes of sand
separating them and extending to the edge of the bank from three or
four fathoms into fifteen or sixteen. Toward the outer edge the corals
grow most vigorously. The same kind of bottom followed us as far as
Whale Cay. Corals begin to grow in from three to four fathoms, where
they are less disturbed by the constant movement of the coarse coral
sand of the bank, and hence the corals have not assisted in building
up the shores. Very few coral patches cnme close to the surface, as
they do in the Florida Cays, where the corals play an important pait in
the formation of tlie outer ones.

64 BULLETIN : MUSEUM OF COMPAIIATIVE ZOOLOGY.

From Whale Cay we crossed over to Mastic Point on Andros. As
soon as we got into deep water after leaving the bank we began to en-
counter again flying lishes, not one of which we had seen while steaming
on the outer shallow part of the bank.

THE LITTLE BAHAMA BANK.

Plate I. ; Plate X. Fig. 1.

The Little Bahama Bank is dumbbell-shaped, one of its shanks running
in a northwesterly direction, the other irregularly north and south. The
northern edge of the bank from Walker Cay Channel to Elbow Cay, and
the eastern edge from that point to Cheroki Sound, is skirted by numer-
ous cays running nearly parallel with the 100 fathom line, at a distance
varying from one to four miles. The rest of the southeastern edge of
the bank is flanked by the high clifls of Great Abaco as far as Hole in
the Wall ; from that point the shore trends to the southwest, and extends
northwest from Southwest Point to Eocky Point. Of the outer line of
eastern cays the principal ones are Walker Cay, the Double Breasted
Cays, Stranger, Carter, Fish, Pensacola, Spanish, Munjack, Green Tur-
tle, Great Guana, Man-of-War, Elbow, Tilloo, and Lynj'ard Cays. A
broken line of reef extends from Cheroki Sound along the eastern edge
of the bank to Matanilla Reef. This reef becomes specially prominent
north of Elbow Cay, leaving only here and there a passage through the
reef for small boats to gain admittance into the large sheet of water,
from two to six miles wide, and twelve feet deep, which separates the
outer cays from the eastern shores of Great and Little Abaco. Large
craft can enter this inner sheet of water through Man-of-War or Whale
Channel Ca}'.

As far as we saw the outer line of cays from the Stranger Cays to
Lynyard Cay, they are all low, generally bare, Green Turtle .Cay being
an exception. They probably represent the outer line of peolian rock
hills which once formed the eastern shore of Great Abaco, but which,
owing to the subsidence and to the wearing action of the sea, has sepa-
rated these hills into numerous islets and formed the wide and navigable
channel intervening between them and Great and Little Abaco.

In tlie vicinity of Little Harbor the outer range of a'olian hills is still
connected with the main island, Ocean Hill and the adjoining promon-

AGASSIZ : BAHAMAS.

65

tories forming the southern edge of tlie harbor and the northern boun-
dary of Cheroki Sound, while tlie promontory of Sweeting protects the
northern side of Little Harbor. At Marsh Harbor settlement and at
Black Point, both on Great Abaco, promontories nearly two miles in
length are left as monuments of the former eastern extensions and
connections of Abaco with the outer cays. In fact, very little more
wearing away would separate the promontories forming Little Harbor
from Abaco, and turn them into cays similar to the outer line, leaving
perhaps at the same time a second inner row of cays like those now in
formation off so many parts of the eastern shore of Abaco, traces of which
farther north are seen in the numerous banks of all sizes and shapes
which exist both to the westward and to the eastward of the outer cays,
being parallel to them in a general way.

Great Abaco Island.

Plate X. Fig. 1 ; Plate XL..

We made the lighthouse on Abaco Island at Hole in the Wall, the
southern extremity of the Little Bahama Bank. That part of Abaco is
low, with rounded outlines (Plate XL.). The rocky surface is bare, all
that part of the island being exposed to the full action of the prevailing
trade winds and swell. In consequence the shore is formed of low cliffs,
having the peculiar basaltic appearance so characteristic of the darker
limestone cliffs of the Bahamas. The surface of the island within reach
of the action of the sea is pitted, honeycombed, and full of pot-holes.
This is specially seen on the east face of the Lighthouse Hill and the
spit to the south of it.

A well marked narrow rocky promontory, covered near the shore by
pinnacles, runs out from the southern end of the island, at the extrem-
ity of which are situated two islets undermined at the base. Near the

HOLE IN THE WALL.

extremity of this narrow promontory the sea has broken through and
formed a large hole, which has given to the locality its name of Hole
in the Wall. The outlying islands have perhaps been formed in part by

VOL XXVI.

NO 1.

66 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

the same process which will in time separate the extremity of the prom-
ontory as a distinct islet, now connected with the inner part of it only
by a comparatively thin arch. From its exposed position we naturally
find many fragments of rock thrown up on the shore. A few at the
base of tlie clilf leading to the lighthouse are quite large. The largest
rock train is thrown up on the beach near the spit to the west of Abaco
Lighthouse. The cliffs to the westward, to the eastwai-d of Southwest
Spit, are the remnants of a range of seolian hills existing at one time to
the south of those which form the extension of Liglithouse Hill. These
western hills must have reached very close to the edge of the bank before
the general subsidence of the Baljaaias had produced any marked cliange
in the topography of the land. The valley which separated them from
the northern range of hills is still well marked.

After turning the Southwest Spit we ran parallel to a long coral sand
beach, the shore line of which was flanked at intervals by stretches of
angular rocks thrown ujj by southeastern hurricanes or gales above
high-water mark and extending inland. On landing we found rocks,
sponges, Gorgonians, and large masses of corals, thrown up to a con-
siderable distance on the lowland, forming an extensive plateau ex-
tending inland behind the beach. Soon after passing Southwest Spit
we began to see fine stretches of pines, which are so characteristic a
growth on the larger islands of the Little Bahamas.

The edge of the narrow bank Hanking western Abaco is covered with
fine coral heads, growing abundantly in from three to ten fathoms close
towards the shore, and separated by wide lanes or patches of sand. On
the western face of the Little Bahama Bank this edging of corals is
nearly continuous, extending all the way from Southwest Spit on Abaco
to Memory Rock, where we entered on the bank to cross it en route for
Green Turtle Cay.

Great Abaco from Southwest Spit westward is low% and covered with
a fine forest of pines, wdiieh extends into the interior. The Lighthouse
Hill range gradually falls to the westward, but extends to the noi-theast
fif tlie lighthouse, and forms the eastei'u face of the island, which is
edged by vertical cliffs from forty to one hundred and fifty feet in
height, forming a nearly continuous wall, upon which pounds the heavy
trade swell. These vertical cliffs extend north as far as Guineaman's
Bay, where the outer row of rocks, islets, and cays begins, extending in
a nearly unbroken breakwater off the main island from Cheroki Sound
as far as Matanilla Reef. On the west side the lowland extends as far
as Rocky Point, and then disappears to form the extensive shallow bank

AGASSIZ : BAHAMAS. 67

which stretches to the southeastern extremity of Bahama Island on
one side, and to Little Abaco on the other. On the interior of the
bank the west face of Great Abaco runs at many points close to the east-
ern shore of the island, leaving only low, narrow ridges connecting the
various parts of tliis singular island (Plate X. Fig. 1). On examining
tlie cliart, one cannot fail to be strucii. witii the endless islets and pas-
sages which have been left on the east coast of Abaco as records of the
subsidence of the bank, and the numerous cays which flank the western
edge of the bank between Eocky Point and the southeastern end of
Bahama Island, while the many cays found upon the shallow interior
bank attest the former extension of the Abaco Bahama Island Land.
The Little Bahamas are perhaps a tiner exainijle than even Andros
can be of the former greater extension of the land, and of the causes
which have resulted in the present configuration of the group.

The Aliaco Bahama Island Land, which once covered the gi-eo.ter
part of the Little Bahama Bank, and probably corresponded in outline
approximately with the line of ten fathoms, was exposed at its north-
western face to the violent action of the northers. They have eaten
away the whole of the northern face of Bahama Island, leaving only
Memory Rock and the banks to the north as witnesses of its former
extension. On the north face of tlie Little liahama Bank the patches
forming Middle Shoal, Matanilla Beef, and the long line of outer cays,
give us approximately the outline of the former Little Bahama Land,
of which Little Abaco and the cays extending to the westward to the
Centre of the World are remnants, these remnants being in turn the
outliers of Great Abaco before it became disintegrated by the action of
the northers, when it was perhaps only se[)arated by a narrow channel
from Bahama Island. Of course, as soon as a wide channel was formed
to the north of Bahama Island or to the westward of Abaco, the action
of the northeast trades also came into play to cut away the low shores
of these islands, thus helping to increase rapidly the dimensions of the
bank.

But what has made the shape of the banks such as they are, and what
has shaped the outline of the old land in so absolute conformity to what
we may reasonably assume to have liccn their original outline]

As far as the sliape of the "\\'indward Islands is concerned, we can still
see the action of the volcanic forces which have elevated islands of very
different shapes and different sizes above the bottom of the surrounding
ocean, — islands which are separated one from the t)ther by channels of
very varying depths, and round which have been formed on one face or

6S bulletin: museum of comparative zoology.

the other narrow fringes of limestone by the reefs surrounding them.
As we pass nortJiwards from South America we find that tliese ishmds
become larger, or rather that they are the summits of larger plateaus,
forming banks of greater or less extent and separated by channels of
vai-ious depths. We also see that on some of the banks there are no
volcanic islands, the whole surface of the plateau being covered by coral
sand, in }>art the renmants of limestone tracts wliich had a greater ex-
tension, (ir in part perhaps of limestone banks which during periods of
great volcanic activity have gradually formed upon the folds of the bot-
tom of the ocean. Even granting for the Bahamas the greatest possible
subsidence as indicated by the deepest ocean-holes, the outlines of the
banks at the time of their greatest elevation could hardly have been
materially ditl'erent from that of tlie present charts, as the whole change
of level is taken to be well inside the 100 fathom line, at not more than
fifty to sixty fathoms.' Take the sea face slope as we find it to-day, it
would nut have changed materially the position of the coral reefs which
must have been growing there, perhaps as barrier reefs exposed to the
disintegrating action of the sea, and supplying by their own disintegra-
tion the material needed for the formation of the teolian hills from which
the Bahama Land was built. Or, more probably, these reefs existed as
frino'ing reefs, mucli as thev do in our davs alonir some parts of the
coast of the Sandwich Islands, and from them were formed the immense
stretches of coral sand beaches which, swept alternately by the trades
and the winds prevailing at other seasons, sup})lied the sand to build up
the gigantic dunes of former days. These formed the highest hills of the
Bahamas, and they in their turn have, from varion?- causes mentioned
in this account of the Bahamas, been reduced to their present limits.

The Bank from Great Abaco to Bahama Island.

Plate X. Fig. 1.

Between Bocky Boint (Great Abaco) and the southeastern, extremity
of Bahama Island only a few cays exist, — CJorda, Channel Cay, Black
Bock, Lily, and liurrow Caj's. The reef can be crossed at Burrow
Cay and at Mores Island t'hannel. Between Burrow Cay and Ba-
hama Island we found numerous sand bores extending five or six miles
from the edge of the bank. Inside of Gorda Cay and to the east of
Southern Cay and of Mores Island extends a great tract filled with
banks and numerous sand ridges reaching towards the ill defined low

AGA.SSIZ: BAHAMAS. 69

swampy shores of the western side of Great Abaco. These sand bores
also extend in a northerly direction from Mores Island to the sand bores
east of Burrow Cay. In crossing the bank from Mores Island to the
Woollendean Cays, we carried from one and a half to two fathoms of
water, and before reaching the cays had struck the territory of the so
called marl of the Little Bahamas. It seems to have been formed under
very much the same conditions as those which have formed the great
white marl flats to the west of Andros ; but the Woollendean Cays and
the Joe Downer Cays being the remnant of a land rather higher than
that on the west shore of Andros, this marl is not so pure, and contains
a greater amount of vegetable matter derived from the decomposition
of a larger amount of soil.

When off Rocky Point we could easily follow with tlie eye the changes
which hail perhaps taken place in tiie configuration of the west side of
Abaco. To the east stretched the low coast of the island itself, covered
with a dense forest of pines, and deeply indented by channels which
seemed, as seen from the rigging, to cut the shore line into numerous
islets. To the westward extends a low rocky spit, and still farther west
rocky cays are found on the edge of the bank, — the outliei's of the
former Abaco. Corda Cay, the summit of which is covered Viv a regular
picket line of angular a^olian rocks, attests the strength of the hurri-
canes which have gradually eaten away the greater part of the west
shore of Abaco ; while farther inland Mores Island and the Woollendean
Cays, reolian islands rising upon the shallow interior bank, indicate the
action which has gradually reduced the western pai't of Great Abaco to
its present dimensions. Long Cay Rocks and other small cays south of
Mores Island are "bare, like Gorda Cay, though often topped here and
there like the latter by a wall of loose rocks thrown up during tl)e hurri-
cane season. There is excellent sponging on the interior of the bank.
The bank is entered by a good passage a little north of Mores Island, and
in that way the west coast of Great Abaco can be reached. "We anchored
in Rock Harbor, on the west coast of Abaco, after liaving visited the
Woollendean Islands and the marl district of the Little Bahamas.

Burrow Cay is low, not more than twenty feet in height, with vertical
cliffs on the channel side. A long row of angular seolian blocks is thrown
up on the western face of the island to a 'height of about fifteen feet
above high-water mark. On the western extremity a small low rocky
cay protects the northern shore of the island from the action of the
northers. On the northern shore there is a short stretch of recent
shore coral rock, masses of Nullipores, of Gorgonians, of corals, and

70

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

an immense accumulation of large conch shells covering an extensive
flat, and also forming a wall of at least five feet iu height above low-
water mark. These conchs are of course of all sizes, and the fragments
into which they have been broken are in all stages of wearing by the
action of the sea. Such masses of conchs, forming almost as great an
accumulation as might be due to an Indian shell heap, we have not
found anywhere else in the Bahamas, although in many places we met
with smaller heaps. All along the line of the bank, in from four to six
fathoms of water, we found the continuation of the coral reef wliich we
struck soon after reaching the lighthouse near Hole in the ^yall.

Owing to threatening indications of a norther, we turned back, and
passing in by Channel Cay entered on the Little Bahama Bank. We
soon came upon the northera extremity of Mores Island, which can
readily be distinguished by the greater height of the cay and the
rounded hills of which it is formed. In this vicinity the bank is thickly
covered with conchs, which are collected by the inhabitants and buraed
to make lime. Mores Island is composed of teolian rock, and is fairly

MORES ISLAND.

wooded. Here and there the promontories of some of the hills have
been washed away by the sea, so that the shore is composed of patches
of low limestone cliffs alternating with sandy beaches. On the west
shore tlie modern coral sand beach formation hides from view the
underlyinsi- reolian rock. From Mores Island towards the north end
of Ahaco there is excellent sponging ground, and all the way from Black
Rock to Mores Island the bank is covered with extensive patches of
sponges and Gorgonians. The extensive coral reef to the westward of
I'.lack Rock and Cliannel Cay reaches some little way over the bank,
and shows special vigor in the vicinity of the Channel.

Off the northeast point of Mores Island the dredge came up filled with
several species of coralline algae and sponges. The bottom samples con-
tained but few specimens of Foraminifera, aiid the sand at this point was
already much finer and more sticky, as compared with the coarse coral
sand of the entrance of the Channel. We found an abundant surface

AGASSIZ : BAHAMAS. 71

fauna off Burrow Cay, — Leptocephalus, Squillro, pelagic flounders,
Sagitte, Doliolum, Diphyes, Copepods, floating algae, and many larvae of
Crustaceans. After we passed Mores Island the specimens of the bot-
tom became more and more sticky, and contained a greater number of
Foraminifera, changing also to a grayish color. As seen from the north-
west, the northern extremity of Mores Island, with its rounded hill-tops,
is very characteristic, and in striking contrast to the low seolian hills
which form the southei-n part of the island. Ou the horizon to the
south is seen the low line of Great Abaco, covered with its pine forests,
and to the eastward the line of the two Woollendean Cays, the outliers
of the former extension of Great Abaco to the west of the marls. These
cays are partly rocky and partly sandy. The sea, although shallow, has
evidently considerable force here, especially during the northers, and
low walls of seolian rocks are thrown up here and there on these cays
just at high-water mark. We landed on one of the cays to the north of
Cambridge Harbor. It consisted of ajolian rocks in an interesting stage
of decomposition, nearly marl, the holes of the rocks full of red earth and
of vegetable matter. The sand on the beaches was made up of the same
material, a little less compact and quite marly. This stage of the a^olian
rock seemed to be the condition immediately preceding that of tlie
locality which is marked "The Marls" on the charts extending from the
Woollendean Cays to the eastward towards Abaco and to the northward
to Little Abaco. The so called marl which we obtained just inside of
the cays differed greatly in its darker color from the whitish marl west
of Andros. It also differed materially in being made up of far coarser
materials, though it seemed to be fully as tough and sticky as the white
marl from Andros.

As we approached the northern extremity of Great Abaco, near Nor-
man's Castle, we could see the Kolian cliffs on the south of it, formed
from the hill slopes cut away at the base. At many points huge blocks,
eaten aivay at their base, had been broken off, and looked now like huge
white sails scattered along the coast line. It is interesting to follow
to the westward of Norman's Castle the continuation of the a^olian
hills, which as small cavs extend in a line outside of the main shore of
Abaco and form Rock Harbor, the lowland lying between them and the
mainland having all been washed away. The rocks were, as at Woollen-
dean Cays, nearly changed into marl, mixed with m'ore or less vegetable
matter and red earth, so that very little additional disintegration would
change it into the sticky and half sandy bottom so characteristic of this
part of the Little Bahama Bank.

72 BULLETIN : MUSEUM OF COMPAEATIVE ZOOLOGY.

The chain of small cays in the vicinity of Rock Harbor, to the west-
ward of Great Abaco, affords one of the best examples of the evidence
we have of the former continuity of the many cays scattered all over
that part ©f the bank. The cliffs of the cays are eaten away at the base
by the slow action of the sea, which here has a for more limited range
than when acting upon the WooUendean Cays to the southward. The
bottom here is sticky, and, though still of the characteristic gray color,
is made up of much finer particles than the samples of bottom we ob-
tained on our way to this point from Mores Island.

The few large pines still left upon some of the smaller islands near
Rock Harbor indicate clearly their former connection with the pine
forests whicli are seen upon the main island (Great Abaco) to the
eastward. Rock Harbor Cay was interesting as showing us the rem-
nants of the inner western line of hills which form the Black Point of
Little Abaco, appear again on Randall's Cay and on Norman's Castle on
Great Abaco, and which may have formed the line of hills connecting
Great and Little Abaco.

Little Abaco is separated by a narrow shallow channel from Great
Abaco, and is only a narrow spit, the remnant of a line of hills running
westward from the northern extremity of Great Abaco.

Coming back to Channel Cay, we pushed rapidly north to Bahama
Island, steaming all the way from Channel Cay over the fine coral
reef which fringes the western edge of the bank in from four to ten
fathoms. The reef is bare at many points, especially in the extension of
some of the ledges of rocks or low cays in the channels formed between
them. The western slope of the bank is often very steep ; it was not an
uncommon occurrence while steaming over the reef to see the patches of
the great coral heads in from five to eight fathoms on one side of the
yacht, while on the port side we could not see bottom. From Burrow
Cay to Carrion Crow Harbor there is a continuous stretch of coral sand
banks separated by shallow channels, leaving passages for small boats to
enter the bank. In the channels, or flanking these sand banks or their
extensions as rocky ledges, thriving patches of large heads of corals
could be seen whenever we came near enough to the outer line of
cays. A few miles north of Carrion Crow Harbor the line of the
gi*eat barrier reef, which runs parallel with Bahama Island, makes a
sharp angle. As we ran parallel with this beautiful reef, we could fol-
low the spurs of the main reef striking toward the shore of the island,
and becoming changed at many places for a short distance into a
fringing reef.

AGASSIZ: BAHAMAS. 73

From Bahama Island to Memory Rock.

Plate X. Fig. 1 ; Plate XXXIX.

Bahama Island forms the southern face of tlie northwestern shank of
the Little Bahama Bank. It is sixty-tive miles in length, but not more
than six to seven miles in width. It is low, covered with a thick forest
of pine. It may be fifty to sixty feet at its highest point, to the east
of High Rock. The high ridge runs close to the south sliore, and on

HIGH ROCK, BAHAMA ISLAND.

crossing this we come upon the level stretch sloping gently north to the
northern shore. Its southeastern extremity is broken into numerous
low cays, and the north shore, which has not been surveyed, appears,
as far as could be seen from the rigging, to be made up of innumerable
low wooded (pine) cays running east of Settlement Point.

The highest cliffs we saw were at High Rock, where there is a small
settlement, but even there the cliffs are not more than from twelve to
twenty feet. In the vicinity of High Rock, to the eastward of Gold
Rock, where we anchored, we had an excellent oi)portunity to see the
barrier reef growing upon the submarine extension of tlie sliore eeolian
rocks. Tlie reef where we anchored is in about four fathoms of water.
We dropped our anchor in an open space between patches of fine heads
of Porites, of magnificent huge clusters of Madrepora palmata and colos-
sal heads of Micandrinre and Astrpeans, many of them overgrown by
splendid Millepores. Inside the reef, towards th<.' shore, the sheltered
waters were filled with patches of large Gorgonians and isolated coral
heads. The distance from the 5 fatliom line to the sliore line is nearly
a mile. Inside of this the coral heads were not very numerous, except
in the lines where they formed spurs reaching to tlic shore of Ba-
hama Island, constituting an incipient fringing reef The shore teolian
rocks are well protected, except at such places as High Rock, by the
barrier reef running parallel with the shore line of the island. I'ack
of the beach the highest poiut cannot be more than ton to twelve feet.
The surface of the exposed rocks is more or loss water-worn from the

74 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

combined action of the rain and sea. Behind the shore Hne of rocks
extend vast flats, the pine barrens of the island. The tcolian rocks are
everywhere fully exposed, presenting the peculiar characteristics of the
great expanses of level or nearly level surfaces which in other islands
frequently separate parallel ranges of aiolian hills. In the sinks and
pot-iioles, or depressions of greater dimensions, pools and ponds, often
of considerable size, have accumulated, many of which are separated from
the sea only by the narrow wall forming the low line of aioliau hills im-
mediately back of the shore.

The shores of Bahama Island, all the way from Carrion Crow Har-
bor to our anchorage at Turtle lieef near High Eock, have once been a
succession of coral sand beaches and of low clift's along the edge of the
low line of hills, forming a sort of dam between the pine tract levels and
the edge of the island. Bej'ond Turtle Reef the extension of this line of
hills forms a few insignificant cays to the west of Gold Rock. The reef is
outside of this line of cays, which represents a part of the ancient shore
line of Bahama Island. The reef dies out at Southwest Point, where
the shore is clear close np to the beach. Beyond Southwest Point, at
Barnard's Point, the low shore hills with vertical cliffs are again char-
acteristic of the shore line, the pine barren plains appearing to be from
one half to three quarters of a mile beliind the beach mound. "Wherever
there is any uutcrnp[)ing of rocks between the stretches of coral sand
beaches, many blocks of a'olian rock are thrown up abtive high-water
mark. There are a number of these rocky outcrops, and as we go
north past Southwest Point the sandy beaches become shorter and
are much more frequently interrupted by considerable lengths of out-
crops. Hawk's Bill Creek is an estuary which has cut Bahama Island
in two, and which comes out on the north side. The shore line df
cliffs leaves a wide opening flanked with mangroves and shrubs; in
the distance are the pine barrens aliout one mile inland. Soon after
leaving Hawk's P>ill Creek the pine l>ari-cns recede fai'ther from tlie
shore, and towaids tiie narrower part of the northern extremity of the
island the pines tliiiiinish gradually in size and in thickness, becoming
(juito sctittered. The low range of shore hills increases in width, ex-
tendiiiL,' farther inland. The continuation of the northern extrem-
ity (if flic island consists of fiuir or five small cays, the remains
or the f)rnirr northern extension of I'ahama Island. These small
cays are of the usual tvpe ; the rocks are a'olian, with vertical faces
more or less undermined, and the surface of the islets pitted and honey-
combed. As the })ines diminish north of Hawk's Bill Creek, their

AGASSIZ : BAHAMAS. 75

place Is taken by scrub vegetation. The island becomes quite flat ;
there ate but few rock exposures, and longer reaches of sand beaches.
We still tind flourishing patches of coral heads all along the coast, but
they are disconnected, and the patches are often far apart. When ofl'
the northern extremity of IJahania Island we could see from the rig-
ging the east shore of the island cut up into numerous small islands,
many of which are still well covered with pines, far better than is the
western part of the northern extremity of the island. These numerous
cays and estuaries are not marked on the charts ; in fact, the eastern
and northern coasts of Bahama Island have not as yet been thoroughly
examined and mapped out. Here and there on the west coast, close to
the shore, are left a few pine trees, and the mangroves are in many
places large and most flourishing.

All along Bahama Lsland, as well as along Andros, where the bar-
rier reefs are perhaps better developed than elsewhere along the sea
face of the bank, wherever there are breaks in the barrier reef so that
the shore line is not protected l)y it from the action of the sea, we find
stretches of sand beaciies corresponding to the openings left in the bar-
rier reef; while opposite the unbroken reaches of the barrier reef the
shores are rocky, exposing in full view the underlying a^olian rocks, which
are not covered up in part by the reef sand, as in the shore opposite the
breaks.

After leaving Settlement Point we came upon Indian and Wood Cays,
the rocky remnants of the western side of Bahama Island. Sandy Cay
is low, not more than fourteen feet in height, with no exposed trace of
the underlying a'olian rock of the bank.

Memory Rock (Plate XXXIX.), close to which we passed as we turned
eastwaul to cross the bank, is perhaps one of the most characteristic of
the outlying sentinels of what once formed a part of the greater Little
Bahama Land. The gradual disappearance of this land as we pass north
is most characteristic. We can as it were follow the disintegration
which has taken place about the northern part of Bahama Island,
the character of which changes radically as we leave the pine barren
flats and pass to the bare rounded rocky hills, pitted, honeycombed, and
worn, which form its western end, and to the numerous cays extending
eastwanl, thickly wooded, covered with pines, which are the continua-
tion northward of the pine barrens of the main island.

Nearer the western edge of the bank, Indian and Wood Cays are
more or less exposed to the same agencies which have acted upon the
Bahama Islands in so uniform a manner. On Wood Cay we find large

/6 bulletin: museum of comparative zoology.

blocks of seolian rock thrown up above high-water mark, and above this
belt it is covered with low shrubs. When we finally come to Memory
Rock, we find nothing but a few pinnacles of feolian rocks, pitted and
honeycombed, and worn into fantastic shapes, the only land still visible
of the northern pai't of the older Bahama Abaco Land.

The northwestern extremity of the bank is fringed by patches of coral
reefs and sand bars. The reef which extends nearly unbroken from
Southwest Point on Great Abaco along the western face of the Little
Bahama Bank beyond Memory Rock runs to within about fifteen miles
from Matanilla Shoal, the northernmost coral patch on the bank.

From Memory Rock to Green Turtle Cay.

Plate X. Fig. 1 ; Plate XXI.

The character of the bottom on the bank from IMemory Rock for about
thirty miles eastward is that of grayish coral sand, rather finer than the
bottom samples near the edge of the bank, which are clear coral sand
and broken shells. At each haul of the dredge masses of Thalassia
were brought up.

The Barracouta Rocks, five to six isolated little rocky patches, twelve
to fourteen feet high, are pitted and honeycombed, and water-worn
at the base. The seolian lamination in some places is most distinct,
dipping at times thirty-five to forty degrees! The cavities of these
rocks were in part filled by very peculiar rounded lumps of a cellular
mass of red earth, sometimes also arranged in ridges on the dividing
edges of adjoining depressions.

Little Sale Cay rises about twenty feet above high-water mark, and is
devoid of vegetation. It is, like the Barracouta Rocks, the remnant
of a larger island, which must have covered the greater part of the
bank to the north of the eastern end of Bah.ima Island. Little Sale

LITTLE SALE CAT ROCKS.

Cay, the Barracouta Rocks, and the Centre of the World, standing as
they do as the outposts most exposed to the agency of the waves, are
cays and patches of rocks the outlines of which have been far more
affected by the action of the sea than the more eastern range of inner
cays, which protect the eastern extremity of Little Abaco.

AGASSIZ: BAHAMAS. 77

Grand, Double Breasted, Stranger, Carter, Fish, Pensacola, and the
Hog Cays are, in the order named, the westernmost of the long line
extending from Walker Cay Channel on the northern edge of the bank
alonrf its eastern edije as far as Cheroki Sound near the southeastern
extremity of the bank. This line of cays protects the eastern face of
Little Abaco and of Abaco from the action of the heavy trade wind
swell which pounds upon the eastern face of the cays ; they will iu
time disappear completely, as the intervening channels become wider,
thus exposing their western face to the long reach of the sea, which
has already removed nearly all traces of the former line of outer cays,
islets, and sheltered cays which must once have formed the continu-
ous Little Bahama Bank, — then a bank on which were many islands,
and which was fringed on its sea face by cays of which no trace is now
left except the shallower patches on the banks to indicate their former
existence. The outer line of cays is flanked on the sea side by the
narrow flat of the bank between them and the 18 or 20 fathom line,
where the bank drops suddenly to a hundred or a hundred and twenty
fathoms. On this flat coral reefs flourish, as more or less extended
patches of heads or clusters of heads, in a depth of four to six fiithoms.
Inside of this line the corals do not flourish well, being too much ex-
posed to the full force of the Atlantic swell. This belt of corals we
crossed twice, on going off" the bank at Green Turtle Cay, and again on
attempting to enter Little Harbor.

The inner shore line of the outer row of cays is generally formed by
low vertical cliff's, behind which rise the rounded summit lines of the
cays, scarcely reaching at any point a greater height than fifty feet.
All these islands are of peolian origin, the rocks composing them differ-
ing in no wise from those of the other parts of the Bahamas. The char-
acter of the bottom, however, changes somewhat after we approach the
Barracouta Rocks. There it begins to be somewhat more marly, and
soon after going eastward we enter a district the bottom of which is
characterized on the charts as marly ; finally, when we get off" Green
Turtle Cay, we find that the marl closely resembles the peculiar white
ooze covering so great a part of the bank to the westward of Andros.
This white marl fills the channel all the way from Little Abaco to
Man-of-War Channel. The samples of the bottom taken off" West End
Rock of Little Abaco, in three fathoms of water, are fine marl of a
light gray color, and of a consistency almost like plaster of Paris, but
of a V)luish tint. Off" the eastern face of the outer cays the bottom is
covered by the coarse sand formed of coral debris and of seolian rock,

78 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

or even of large pieces, according to the conditions due to the outer line
of protecting reefs. The outer line of cays forms a sort of sieve through
which the action of the sea outside due to the prevailing winds is greatly
modified and tempered both upon the western face of the cays and the
eastern channel face of the Abacos.

Steaming along in an easterly direction, keeping Little Abaco to the
south, the island is covered with low growth, the cliffs of the eastern
face are not prominent, and the aeolian rocks can be seen at all exposed
places on the sides of the cay which are bare of vegetation.

To the eastward are seen the outline of the low rounded hills of
Spanish and Powell Cays, the latter about eighty feet in height, off Great
Abaco. South of Spanish Cay the bottom consists of very fine marl,
and is covered with sponges, coralline algae, Thalassia, and several species
of Penicillus.

After passing the spit at the north end of Great Abaco the island
widens somewhat, so that we get two lines of aeolian hills parallel to the
shore. The western range, which is the highest, is covered with low
vegetation. Nearer the shore the trees are taller, seemingly mastic and
mahogany woods, while from the south the pines come in again, and the
southern part of the island, which falls off rapidly to a low flat from
Mango Hill, is covered with a. thick pine forest, which extends unbroken
to Rocky Point. About opposite Munjack Cay the shore Eeolian hills,
which run south from Angel Fish Point, are not more than from four to
six feet in height. Farther south they rise to twenty or twenty-five feet,
and pass into the flats upon which the pine forest extends, as we could
see when steaming from the Woollendean Cays on the inside of the bank
towards the western shore of the island.

After Powell Cay comes the narrow line of Munjack Cay, and next
Green Turtle Cay, the most important settlement of the Bahamas after
Nassau. Seen from the west all the cays appear to be low seolian
hills with rounded outlines. Green Turtle Cay is a little more tlian
seventy feet in height. It is somewhat broader than any of the other
outer cays. There is a fine bluff of white aeolian rock at the southern
extremity of the island. The base of the cliff is constantly eaten away
by the action of the sea, and supplies the material from which is derived
the whitish marly bottom which extends over the channel to the east of
Great Abaco. No Name Cay protects the anchorage of Green Turtle
Cay from the outside swell to the south, and Munjack and Crab Cays
protect it from the north. Owing to the short distance between the
outer cays and Great Abaco the sea in the channel acts upon its shores
with little force.

AGASSIZ: BAHAMAS, 79

An examination of the larger charts of this part of the Bahamas will
show the effect which subsidence has had in forming a number of boat
harbors and inlets on the shores of this cay. Some of them have twenty
feet of water, and are more or less protected by the patches of coral
reef growing between the cay and the edge of the bank. Upon these
patches the sea beats violently, forming two or three lines of breakers,
through which it is often impossible even for small boats to find their
way out. Similar sinks exist upon many parts of the Bahamas, one of
the most characteristic of which is Hurricane Hole, south of Andros
(Plate XI. Fig. 4).

After leaving Green Turtle Cay, immediately along the sea face of the
Little Bahama Bank on our way south, we passed through great masses of
gulf-weed. There is little doubt that much of the Sargassum we meet
witli in tlie Bahamas comes from the shores of the cays. It is found in
considerable profusion in many localities, and, while growing vigorously
during the comparatively short season of calms, immense quantities are
torn off during gales or heavy seas, so that a fresh stock of Sargassum
must at all times be increasing the mass floating in the range of
the Gulf Stream and of the northern extension of the equatorial set.
How long it can thus float, and perhaps propagate by budding, is not
known. It is, however, a well known fact, that Sargassum torn off" from
the rocks can be kept thriving for months in tanks or dishes.

The slopes of the low shore hills of Great Abaco, as seen extending
southward from the anchorage at Green Turtle Cay, show in a most
characteristic way their seolian structure. The base of these hills is cut
off by low vertical cliffs which have been formed by the gradual wearing
away of the rocks by the sea. We examined the shore of Great Abaco
at Black Point, pulling round it and- landing inside on the sand beach
formed beiiind the spit. The rocks present the same seolian structure,
and are greatly worn by the breakers. The material back of the beach
is more in the amygduloid stage than in the marl stage. In three and
a half fathoms off the beach the bottom is somewhat more sticky than
the cleaner samples taken immediately on the beach.

In our way out through Whale Channel Cay we found all the islets
and rocks rounded off on the summits, the surface pitted and honey-
combed and eaten away at the base of the rocks.

From the summit of the white cliff to the eastward of the landing on
Green Turtle Cay we could see i\)e breakers on the outer coral reef,
which extends more or less parallel with the edge of the bank all the way
round its eastern edge in a nearly unbroken line.

80 bulletin: museum of COMPAKATIVE ZOOLOfJY.

Whale Cay Channel and the Eastern Face of the Little

Bahama Bank.

Plate X. Fig. 1 ; Plate XI. Fig. 7.

Passing out through the channel and crossing the opening left in
the reef through the breakers, we had a fine exhibition of the skill and
coolness of the pilot as he steered the " Wild Duck " into deep water.

Whale Cay Channel Rock is but a small outlier of a sunken patch.
Whale Cay, as seen end on, shows the seolian hills to have their longest
slope on the east face, and the same structure is admirably shown for
Great Abaco. Great Guana Cay displays the same feature, though here
and there short stretches of the base of the shore hills are eaten away
into low vertical cliffs, probably opposite smaller or greater gaps in the
belt of coral heads protecting the outer islands from the terrific pounding
of the trade wind swell. Where not cut away the lower part of the
shore hills is pretty well covered with coral sand torn off from the reef
and thrown up on the shore. Great Guana Cay is covered by a most
scanty vegetation near the shore, but is a little better wooded on the
hill face near the summit.

Between Great Guana and Elbow Cays is a series of low cays, rocky
or sandy, with very little vegetation near the summit ridge. The line
of breakers forms a continuous wall with the islets whenever there is a
heavy swell rurning, as was the case when we steamed past.

Man-of-War Cay has the same characteristics as Great Guana Cay.
It has a high sandy beach with an occasional rocky outcrop covering the
underlying rock, and the usual scanty vegetation just above high-water
mark becoming somewhat thicker near the ridge of the island.

The succession of aeolian hills piled up one by the side of another
and sloping up to the westward is clearly seen in the line of Man-of-
War Cay and of Elbow Cay. The settlement on Elbow Cay is protected
by an outlying line of rocks, a part of the former eastern extension of
that island, which now forms an outside line of islets connecting it with
Man-of-War Cay.

As we steamed down towards the south end of Elbow Cay the face of
the northern ridge above the settlement was seen to be white seolian rock
cropping out between the bushes and scrub vegetation. All the way from
Whale Cay we were running parallel with the belts of coral reef and the
outer line of rocky cays. Near the south end of one of the small cays,
between Tilloo Cay and Lynyard Cay, there is an accumulation of large

AGASSIZ : BAHAMAS. 81

rocks opposite one of the gaps in the reef and in the outer line of
rocky cays, giving the sea full swing to throw up the fragments of the
seolian cliffs as they become broken off.

The east face of Lynyard Cay is low and worn into low cliffs. In
the background can be seen the pine forests of Great Abaco rising be-
hind Tilloo Cay Sound. The last of the outer islets we were near enough
to examine made the northern spit of the entrance to Cheroki Sound,
while Ocean Hill, a promontory of Great Abaco, formed the opening to
Little Harbor to the south of Lynyard Cay. As we lost Great Abaco,
the eastern face of which is no longer protected by outlying islands, we
could see the comparatively high vertical cliffs extending southward in
a nearly unbroken line to the point which we had seen on our tirst
reaching the island north of the Hole in the Wall.

SALT CAY BANK.

Plates I. and XXXI.

Before describing the banks to the eastward of the Great Bahama
Bank I will give a short account of Salt Cay Bank, the westernmost of
the tln-ee banks lying on the eastern edge of the Gulf Stream. Salt
Cay Bank is triangular, with rounded angles, its greatest width being
about forty miles and its length nearly sixty. It lies at the western
opening of the Old Bahama Channel, the forks of which, separating the
])ank from the Great Bahama Bank and Cuba, are known as the Santaren
and the Nicholas Cliannels. In general, Salt Cay Bank resembles more
the Crooked Island and Caicos Banks, and has reached a condition in-
termediate between them and the Mouchoir and Silver Banks. There
remain on the edge of Salt Cay Bank fewer islets and rocks than along
the Crooked Island and Caicos Banks, but they have not disappeared
so as to leave mere rocks awash, as upon the edge of Mouchoir and Silver
Banks.

Salt Cay Bank has been described by Professor Agassiz in his Eeport
to the Supei-intendent of the United States Coast Survey for 1851, and
republished in the Memoirs of the Museum of Comparative Zoology.
Having visited the Florida Keys first and Salt Cay next, he naturally
extended his explanation of the formation of the Florida Keys to all the
cays of Salt Cay Bank. The structure of Salt Cay is, I believe, however,
radically different from that of the other cays of the bank which I visited.

VOL. XXVI. — NO. 1. 6

82

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

It is, like Cay Lobos, built of recent coral sand rock accumulated after
the foraiatiou and disintegration of the cays of the bank. It is on
the very edge of the bank, and owes its existence to the throwing up at
that point of coral sand so as to make an island, the underlying base
of which there is every reason to suppose does not differ from that of
the other banks of the Bahamas. This we may safel}' infer after an
examination of the remaining outcropping islands forming the Double
Headed Shot Cays, the Muertos Cays, the isolated islets of the eastern
edge of the Salt Cay Bank (Dog Eocks and Damas Cays, etc.), and the

30

■j-a.

IB 12 '■"

Miuertos CavsSVt

Doj? RockB V;\

1 ^'30 fi! ~^

NElbo-wCajj

■A

--'Vfater Cajre
316,.' ,-^y- ■;■'>■. '

\^iy^ SAL r ® c

JO^ 7

Pf2 .., s si

■; 7 -si

Sli

(.7

4)

■■■J 4 3i
212 .

3i

B A N

4j

325

,rh-S

L4e

JV

<.i

534.

A^. Y

9
3
« 9

36

3»\

r.,rK\ h..,f1a

6i

&0

-*tA

•si

•*
V
■S

•A

Ano"\iila Isles ^-;. "'-■..

6 ' ^ 4\Kl'

""■■.. 7

■^^

'. 340
H/t Co S

328

Anguila Islands. The shores of Salt Cay are edged with strata inclined
to the sea, composed of recent coral rock sand of different degrees of fine-
ness, including in some cases excellent examples of coarse heavy breccia.
At onother point on the shore we found a mass of broken fragments of
Strombus and Turbo of all sizes and shapes, rounded, elliptical, or an-
gular, often more or less worn, forming a regular coarse conglomerate.
Mixed with this are numerous pieces of bivalves. At other places the
fragments of shells have been broken to very small pieces, forming a
very fine breccia.

The Salt Lagoon, which was once the basis of a thriving industry,
destroyed by the hurricane of 1866, is separated from the sea by a long

AGASSIZ: BAHAMAS. 83

steep beach, the summit of which is from twelve to fifteen feet above the
level of the sea. The sea face is not more than twenty to thirty feet wide,
■while on the lagoon side the beach is from three to five hundred feet
wide. As has been noted by other observers, tlie water of the lagoon is
intensely salt. We saw nothing living in it, but many dead conchs and
the shells of other large mollusks, as well as those of innumerable
smaller mollusks were found scattered all along the lagoon beach. These
mollusks must liave lived in the lagoon during a higher stage of water,
when its salinity did not differ materially from that of the sea. It is
surrounded by mangroves, many of which run far into the lagoon. Tlie
hills to the east are not more than twenty-iive feet in height. I could
not satisfy myself of their true character, and could not decide from what
I observed whether the rocks composing them were seolian, or whether
they had been thrown up during hurricanes, as was undoubtedly the
case with a bluff which forms a part of the same low range separating
the lagoon from the sea. This bluff was built up of large rectangular
blocks of coral rock similar in structure to that of the inclined strata
along the shores. Some of the blocks had been thrown up to a height
of fully twenty-five feet above the level of the sea during the hurricane
of 18G6. At least our pilot says that is the general report.

Between Salt Cay and Double Headed Shot Cays as far as Rompidas
Rocks we steamed on the edge of the bank in about seven to eight
fathoms over masses of coral heads closely clustered together, separated
by narrow sand bars consisting of Astrteans and other corals. At our
anchorage off Salt Cay the coral heads were growing luxuriantly in from
three to five fathoms. The great abundance of living corals on the edge
of the Salt Cay Bank fully accounts for the masses of fragments of
corals in all stages of comminution which we found on the beach of Salt
Cay, and for the formation of an island like Salt Cay, which is probably
wholly of reef-rock sand origin, and not of fpolian origin like the other
cays on the northern and eastern edges of the bank. After making the
shallower water of the bank, the coral heads are fewer, there are longer
stretches of sand between the heads, and large spaces entirely devoid of
corals.

Salt Cay Bank is, like the other banks of the Bahamas, a triangular
plateau fringed on the eastern and northern edges with islands and islets
or isolated rocks all of seolian structure, while on the western edge of the
bank, with the exception of the Rompidas and Lavanderas Rocks, they
have disappeared. The subsidence of the bank has probably been
fully as great as that of many of tli« other banks, if we are to judge

84 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

by the condition of the Double Headed Shot Cays and of the Anguila
Islands.

The Double Headed Shot Cays, as approached from the south, present
one mass of isolated rocks, islets, and islands without vegetation. Elbow
Cay (Plate XXXI.), the largest island of the group, is a splendid specimen
of the transformations to which the seolian rocks of the Bahamas have been
subjected. The surface of the island is deeply pitted and honeycombed
in all directions. Near the edge there are many blow-holes, through
which the spray is sent in all directions, and on the north side, where the
water is deepest, the action of the northers must have been one of the
principal causes of the wearing away of that part of the island. The
seoliau structure of the rocks of the island could be most plainly seen.
At the landing to the south of the lighthouse, as well as in the deep
channel for landing stores cut into the rock on the north side, the strati-
fication so characteristic of aeolian rocks was most distinct. Steaming
across the bank towards the Anguila Islands we found the bottom on
Salt Cay Bank much like that of the other Bahama Banks, with little
animal life but many coralline algae. We dredged a few times on the
way, bringing up many fragments of shells and broken Clypeastroids
showing the effect of the waves on this shallow plateau. The outlying
rocks to the northwest of the Anguila Islands could plainly be seen to

.S:OLIAN CLIFFS, SOUTH ANGUILA.

De of seolian origin. The islands are comparatively low, the north-
ern and southern extremities being somewhat higher than the central
mass of the islands. The long stretches of low vertical cliffs forming
the western shore of the islands are separated by sand beaches. The
hills near the southern extremity on the west shore mark the highest
part of the islands. Opposite our anchorage were fine seolian^hills from
forty to fifty feet in height, full of cavities and deeply honeycombed.

AGASSIZ: BAHAMAS. 85

THE EASTERN BAHAMA ISLANDS.

Plate I. ; Plate VI. Figs. 1 to 4 ; Plate IX.

To the eastward and southeastward, and separated from the Great
Bahama Bank by very deep channels, are a series of islands and banks
extending from Watling Island to Navidad Bank, which in their turn
rise sharply from the bottom of the ocean. These islands and banks
are of all shapes, either elliptical, or circular, or more or less rectangular,
or sometimes irregularly triangular. They are interesting as showing
the different stages through which the Bahamas as a whole have passed,
from the time when they covered a far greater area than that now
indicated by the islands, which in some cases merely form the sea fringe
of the banks of which they represent the summits. The former isl-
ands have been eroded and eaten away, and have wholly disappeared
on the western feces of the banks, or have left only here and there a
small island or isolated rock to testify to the former existence of the
same seolian hills which form the summits of the present islands.

The smaller islands — like Watling, Conception, Rum Cay, Atwood
Cay, the Plana Cays, Mariguana, and the Inaguas — still occupy nearly
the whole area of the banks upon which they rise. The 100 fathom
line of Great and Little Inagua, of Samana, and of Mariguana, is but a
little distance beyond the shore line of these islands. On Rum Cay and
Watling the 100 fothom line bank is somewhat larger, and Conception
Island is a small part of the submerged bank upon which it rises. But
except Conception these islands have at no time differed very materially
from their present outline. When, however, we come to such banks as
the Crooked Island Bank, Caicos Bank, and Turk's Islands, we find upon
them a series of islands which have been greatly modified by the action
of the sea. The islands, such as Crooked Island, Fortune Island, 'Bird
Rock, the Fish Cays, Acklin, and Castle Island, which, with the exception
of the westward face, nearly surround the Crooked Island Bank, are all
that remain of the one large island which undoubtedly once occupied
the whole of this bank even somewhat beyond the 10 fathom line. In
fact, we may well imagine the time when the Crooked Island Bank
presented much the same appearance as Great Inagua, when it' had like
the latter its fringing reef a short distance from the shore line at a depth
of four to fifteen fathoms, and formed perhaps here and there a reef
harbor like that of Alfred Sound at the northwest extremity of Inagua.

86 BULLETIN : MUSEUM OF COMPARA.TIVE ZOOLOGY.

The cBolian hills of the western face of the former Crooked Island laud
were probably quite low, and were soon eaten away. As the bank sub-
sided, the great lagoon or flat occupying once the central part of the isl-
and, as at Inagua, sank below the level of the sea to become the bottom
of the bank, while tlie low land to the westward was all washed away
and disintegrated by the action of the sea and rains.

We can readily trace a similar course of events on Caicos Bank, and
find no difficulty in reconstructing the Caicos Eank land. This is now split
lip into West Caicos, Providenciales, and North, Grand, and East Caicos,
which form the northern and northwestern outline of the bank, while
South Caicos, Long Cay, Ambergris Cays, and the Swimmer Rock fringe
the eastern face of the bank, and Molasses Reef, French Cay, and West
Sandspit are the remnants of the land fringing the southwest line of
the Caicos Bank. Here and there on the southeastern part of the bank
rise the Seal Cays, Pear Cay, WTiite Cay, and the small rocks which
were once a part of the greater Caicos land. The bank itself, like
Crooked Island Bank, the Turk's Islands, in fact all the banks, not ex-
cepting the Great Bahama Bank, show a dip to the westward, and the
same is the case with the sunken Mouchoir, Silver, and Kavidad Banks.
Great and Little Inagua evidently have not been suliject to the same
amount of subsidence which has so materially affected the islands and
banks to the seaward.

In addition to these larger banks we have Mira por vos Bank with a iew
insignificant cays, Diana and Brown Banks, both of which are sunken,
the one with ten, the other with nine fathoms in the shallowest part.
To the southeast of Turk's Islands are Mouchoir and Silver Banks, on
both of which there are rocks awash and coral heads on the northern
faces of the banks, with a few isolated rocks irregularly scattered, the
depths on their banks ranging from nine to twenty fiithoms ; and finally
comes the last sunken bank to the eastward, Navidad Bank, with an
average depth of about sixteen fathoms and a least depth of eleven
fathoms.

"Watling Island.

Plate IX. Fig. 12 ; Plate XXIII.

Watling, as is seen from the chart, is twelve miles l^ng and from five
to seven wide. Its shores are but slightly indented. The greater part
of the surface of the island is occupied by salt water lagoons of very
irregular outline, formed in the valleys and sinks of the cross ranges

AGASSIZ: BAHAMAS. 87

of seoliau hills covering the island. These hills rise to a height of from
a hundred to a hundred and forty feet, and are covered with woods.

An examination of the map of Watling Island is interesting as show-
ing the great number of lagoons which occupy so large a part of its
surface. On the western side is found the largest of these lagoons.
It is quite shallow, the part we examined varying between one and a
half and three feet in depth. An artificial cut has been made through
a low ridge separating one of the western lagoons, about half a mile from
Eiding Rock beach, so as to make a boat passage to the larger lagoon
and reach by water the vicinity of the lighthouse on the northeastern
extremity of the island. The cut shows the same seolian structure of
the rock so characteristic of the islands we had so far visited: In the
distance, on the opposite side of the lagoon, could be seen rising the
same solidified eeolian hills which characterize the structure of all the
islands on the larger banks. The examination of this side of Watling
Island plainly shows its structure to be similar to that of the islands
to the westward, and also shows that Watling owes its presfent config-
uration and the existence of its many lagoons to the subsidence which
has caused the gradual disappearance of the extensive tracts of seolian
land which once covered the greater part of the Bahama Banks.

The bottom of the great lagoon is thickly covered with algae (Ace-
tabularia), and the shores of its beaches are lined with diminutive speci-
mens of the same species of shells found on the open sea beaches. The
water of the lagoon is intensely salt. It connects evidently with the sea,
as our guide mentioned several blow-holes through which the tide is
forced into the lagoons.

The shores of the lagoon are lined with mangroves. Parts of the
lagoon have been separated from the sea by a high narrow beach thrown
up by the incessant Atlantic swell ; but by far the greater number of
the lagoons of Watling Island are due to the general subsidence of
the island, forming drainage areas, and allowing the sea to cover the
flats intervening between the ranges of feolian hills and then to the
closing of these openings by coral sand beaches thrown up by the sea.
Many of tlie ponds are disconnected, are salt, and are supplied by per-
colation through the barriers separating them from the sea, or by blow-
holes connecting them directly with it. This is especially the case
with some of the ponds only separated from the sea by narrow beaches,
Many of the lagoons are in underground communication with the sea.
and in some of them the position of the blow-holes through which the
sea water is forced up can be traced by the commotion of the water of
parts of the lagoon.

88

bulletin: museum of comparative zoology.

The subsidence and erosion of the island have formed an extensive har-
bor at the northeastern end, hemmed in on all sides by islands and islets,
leaving a couple of passages to the northeast of Green Cay into Graham
Harbor. Fringing this harbor, and following its outline in from three to six
fathoms of water, extensive patches of corals are met with which form to
the eastward a more or less continuous coral reef, with an inner protected
passage for boats and smaller vessels along the whole of that face of the
island. These independent patches of corals are also found on the west
coast, as at Riding Rocks, but are not so continuous as on the east coast.
The corals forming these patches are the common West Indian species
of Madrepores, Mseandriua, Astrseans, and Orbicellas, with Flabellum as
the most common of the Gorgonians. Sailing round Graham Harbor, we
followed a course parallel to the eastern shore. The outer reef shelters
Ihe long sand beaches fairly well, and the former outline of the island
was such that there are but few vertical bluffs on the windward side of
the island. The highest hills are on this eastern side, one rising to the
southward of the lighthouse, its sea face forming the white bluffs which,
according to Captain Becher of the Royal Kavy, iianked the beach where

LANDING PLACE OF COLUMBDS, ACCORDING TO CAPTAIN BECHER, R. N.

Columbus first landed, while the spurs from Fortune Hill nearer the
southern end of the island extend to the sea and form the white cliffs
which according to Sir Henry Blake flanked to the north of Columbus
Bight the spot where Columbus first landed in the New World. The

LANDFALL OF COLUMBUS, ACCORDING TO SIR HENRY BLAKE.

beach at the nortlieast end of the island is so entirely shut out by
patches of coral reefs that it would have been impossible for Columbus to
have anchored at the spot which Becher assigns as his anchorage. Our
pilot says no sponger ever dares to anchor there ; while farther south,
at Columbus Bight, the spot selected by Sir Henry Blake, tliere is a
small reef harbor where boats of the size of the caravels could readily

AGASSIZ: BAHAMAS. 89

fiud an anchorage, and from which the boats of the fleet might have
rowed in safety to examine the " other side," and return the same day
to their ships.

I may here quote from a translation, by Mr. H. L. Thomas of the U. S.
State Department, of the Journal of Columbus,^ the following points of
interest regarding the so called barrier reef off the east face of Watling
Island : " Went along the island, in a northeasterly direction, to see the
other side, which was on the other side of the east. ... I was afraid of
a reef of rocks which entirely surrounds that island,^ although there is
within it depth enough and ample harbor for all the vessels of Chris-
tendom, but the entrance is very narrow. It is true that the interior of
that belt contains some rocks, but the sea is there as still as the water in
a well."

From Columbus Bight on the east coast there is an excellent boat
passage inside the reef leading to Graham Harbor, and it is within this
passage that the boats of Columbus probably rowed to explore the
other side of the island. It is undoubtedly to this passage, sheltered by
the reef, and to the reef harbor, that he refers as capable of holding the
navies of the world. Such a wonder as a reef harbor or a passage inside
a barrier reef could not fail to strike him, and it seems strange that it
is not better described in his journal, and that so little is said by him of
the striking contrast of the light green color of the water inside the bar-
rier reef and in Graham Harbor with the dark blue of the ocean beyond
the 100 fathom line.

The 100 fathom line extends about a mile beyond the reef. Inside
the reef, the passage is on an average about a third of a mile wide. The
100 fathom line runs at a fairly uniform distance from the shore all
round the island, except off the Hinchinbroke Rocks at the southeast
end, where the bank extends seaward more than two miles, and comes
in again along the south shore to Southwest Point. .

Along the west coast disconnected patches of coral heads extend from
a depth of three or four fathoms towards the edge of the bank. We carae
to anchor that same day at Southwest Point on the steep slope of a sand
beach. Here we also found, in from four to twelve fathoms of water,
outside the action of the breakers, fine patches of Madrepores and other
masses of corals furnishing material for the recent coral beaches which
form a low plateau flanking the seolian hills. This plateau lies at inter-
vals between headlands along the western face of the island.

1 G. V. Fox, Appendix 18, U. S. Coast and Geodetic Survey, Report for 1880, p. 14.

2 Not the west side.

90 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Rum Cay and Conception Island.

Plate IX. Dig. 11.

From the southwest poiut of Watliug Island we steamed to Rum
Cay, the structure of which is very similar to that of Watling. The
salt ponds are limited to two small areas, from which formerly a consid-
erable amount of salt was exported.

On our passage from Watling Island to Rum Cay, and while sailing
round it and on our way to Clarence Harbor, we passed through many
streaks of gulf-weed.

The bank of v.hich Rum Cay forms the summit extends northeast
about five miles to the 100 fathom line. The bank is narrow on the west
face. On the north side the 100 fathom line is not quite two miles from
shore, and varies from three miles to three and a half on the west and
south sides. On the eastern side the salt ponds are separated from the
sea by seolian hills varying from fifty to ninety feet in height, and on the
south by walls of recent coral sand. The northern and western ends of
the island are also capped by seolian hillocks.

Rum Cay is about nine and a half miles in length by five in breadth
at the east end and two at the west. The southern side of the cay is
edged by a coral reef nearly a mile off shore, growing upon sunken
patches of seolian rocks. A wide entrance through the reef forms Port
Nelson inside the reef with four fathoms of water. There are patches
of corals all along the north shore about half a mile from it. The
corals are thriving, and consist, as far as we examined them, mainly of
huge masses of Astrseans, Orbicellas, Madrepora palmata, and Millepores.

The low southeastern part of the island is formed of debris of the reefs
now growing on the southern edge of Port Nelson. A low bank of coral
sand is thrown up, forming strata slightly inclined and protecting the
coral sand flat extending inland to the base of the a^olian hills. The
north face is bold, steep, with low blufi's and no anchorage along it.
The east face is flanked by low seolian hills. The highest hills are
along the northern side near the northwest extremity of the island.

Conception Island we did not visit. The adjacent cays rise to one
hundred and thirty feet above the circular bank, which carries from six
to fourteen fathoms of water. Reefs extend in a nearly unbroken line
round the southeastern and western edges of the bank. It is probable
that the 1,000 fathom line connects Conception Island Bank with the
southeastern side of Cat, and that Rum Cay also lies within that line.

AGASSIZ: BAHAMAS. 91

The Crooked Island Bank.

Plate IX. Figs. 1, 9; Plate X. Fig. 4; Plates XXXVII. and XXXTIII.

The most westerly of the smaller outlying banks of the Bahamas is
the Crooked Island Bank, which is triangular in shape. Its northern
side is about thirty miles long, its eastern face forty, and the westei'n
forty-eight miles. The bank slopes very gradually towards its western
and southern edge. The 100 fathom line runs close to the edge of the
bank along the sea face of both Crooked and Acklin Islands, as well as to
the west face of Fortune. This and Caicos Bank are, as it were, epitomes
of the Great Bahama Bank, representing on a small scale the charac-
teristic physical features of the Bahamas. Fortune Island forms the
western edge of the bank, Crooked Island the northern, and Acklin
and Castle Island part of the northern and eastern face of the trian-
gular bank, which is open to the sea for the greater part of its western
side. The islands forming the outside edge of the bank are all narrow •
both Crooked and Acklin Islands are somewhat wider than Fortune
Island, spreading out on the northern and western sides very gradually,
and passing into the shallow waters of the inner northeastern part of the
bank. From the western extremity of Crooked Island extends a wide
range of seolian hills, the Blue Hills, occupying the central line of the
greater part of the island ; the summits reach a height of about two
hundred feet. Near the eastern end Mount Pisgah rises to two hundred
feet close to the sliore.

The whole northern face of the bank is edged by a coral reef extend-
ing from Northeast Breaker on Acklin Island to Bird Kock, where the
reef forms a well sheltered basin. About a mile from the northwest
point of Crooked Island is Portland Harbor, with three to four fathoms
of water.

Acklin Island is separated from Crooked Island by a wide passage of
about two miles, but very shallow. The eastern face of the island is
skirted by a reef nearly continuous from Northeast Point to Castle
Island, beyond the southern extremity of Acklin Island. Castle Island
stands on the southern end of the Crooked Island Bank, and is about
two miles in length. On the northern and eastern sides of Acklin Isl-
and are a series of seolian hills lying along the eastern face, which
rise to a hundred and fifty or two hundred feet. Near the southern
extremity they are somewhat higher than those of Fortune Island.

The west shore of Acklin, like the southern shore of Crooked Island,

92 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

is low, ruuning into the shallow water of the bank. Nowhere, except
near the southern end of Fortune Island and to the west of the Fish
Cays, and a small patch to the eastward of them where there are two
fathoms, is this shallow water of greater depth than one to one and a
half fathoms, and a great stretch of the bank carries even less than
that.

Acklin Island and Castle Island, as seen from the southeast, pre-
sented no features differing in any way from those of the other Bahama
Islands. Vertical bluffs of aiolian rock, of greater or less height, char-
acterize their sea face. The trend and outline of the eeoliau hills of
Acklin plainly indicate their origin.

Fortune Island is nine miles long, comma-shaped, barely a quarter of a
mile wide at its northern extremity. Near the south end a hill rising to
a hundred and ten feet slopes gradually towards it. Off the east coast
near the south end there is a deeper belt of water running rapidly
into six or seven fathoms. From the southwest end of Fortune Island
a narrow reef extends along the whole western face, in from four to
twelve or fifteen fathoms, towards the edge of the 100 fathom line
bank, which drops off abruptly from the outer edge of the reef. The
Fish Cays are the only remnants of the land once skirting the south-
ern part of the western edge of the bank, or perliaps of the land which
once covered the whole bank, and of which tlie larger islands are
the vanishing tops. To the eastward they are surrounded by a series
of sand bores which do not quite reach the surface at low water. Off
the west coast of Fortune Island the 100 fathom line runs close to the
shore, leaving but a narrow belt of soundings. This belt widens out
somewhat about half-way north along the west shore, and from that
point an irregular coral reef extends, in from three to five fathoms,
almost to the northern end of Crooked Island. The western end
of Crooked Island is cut up by narrow lagoons opening on the bank
side. The inner one opens by two channels edged with mangroves into
a large inland bay nearly ten miles across. The outer one is separated
from the Crooked Island passage by a narrow band of recent coral sand.
This gradually disappears as one goes farther on the bank, until finally
the bottom is made up of the debris of aiolian rocks mixed with frag-
ments of shells, and other Invertebi'ates, the whole kept more or less in
place by the numerous calcareous algse which flourish on the limestone
bottom.

On entering the bank off the southern extremity of Fortune Island
we found the bottom to be dotted with coralline akw, fine sand, and

AGASSIZ: BAHAMAS. 93

with patches of the common West Indian Gorgonians. The corals on
the bank exist in irregular patches, and do not extend any distance
eastward. They flourish along a narrow belt on the edge of the bank,
between three to six and fifteen or sixteen fathoms.

On the bank side of Fortune Island the slope of the hills runs insen-
sibly into shallow water ; but, steaming on the western face, the shores
all tlie way from the South Spit of Fortune to its northern extremity
are flanked by low ceolian cliff's interrupted by steep coral sand beaches.
Above the low cliff's extends for a considerable distance a line of an-
gular rocks thrown up above high-water mark by an unusual swell or
by hurricanes.

The lagoon which extends south from the settlement nearly to the
southern end of Fortune Island is separated from the sea by a high
steep coral sand beach (Plate X. Fig. 4). The beach becomes gradu-
ally much lower north of tlie settlement ; it extends all the way
to the northern extremity of Fortune Island, and is again seen on
Crooked Island, being nearly unbroken along the whole shore line as
far as Portlaud Harbor at the northern end of the island. Crooked
Island is also built of a^olian hills with gentle slopes to the west except
near its northern spit, whore the shores consist of low vertical cliffs.

From Bird liock lighthouse the beautiful sweep of a coral reef con-
necting it with Crooked Island forms a fine bay sheltered from the north-
east trades, tiie green waters of which stand out in marked contrast with
the dark blue of the deep water off the bank, while the white wall of
breakers marks the dividing line between it and the deep water to the
east.

If Columbus visited Bird Rock, or Cape Beautiful, as some writers call
the western end of Crooked Island, he could not have failed to notice
a physical phenomenon so strange to him as that of a coral reef forming
a bay well protected from the prevailing winds, nor could such a sharp
observer have failed to describe at length, or to note at least, this pecu-
liar feature of the sea.

To the westward of Castle Island is the Mira por vos Tknk, with
soundings near the edge of from five to nine fiithoms. It is pear-
shaped, witli a cluster of low barren rocky cays, the highest of which
is but thirty feet. At the north end they are connected by a reef; and
there are several disconnected patches of corals, with three fathoms of
water over them.

Two other small banks rise in the Crooked Island passage. Diana
Bank, twenty miles to the west of Fortune Island, is about four miles

94 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

in length, in from nine to fifteen fathoms of water. The bank is said to
consist of sand and of coral. South of Mira por vos Bank rises the
peak of Brown Bank (Kansas Bank), which has from ten to twenty-
three fathoms of water on it. Both these banks drop abruptly into deep
water from the 100 fathom line, close to the soundings indicated.

A number of other small banks have also been reported, but their
position is doubtful. They are Cuidado Bank, south of Mariguana, the
Clarion Bank, south of Mathew Town (Inagua), the Fawn Shoal, south
of Turk's Island, and the Severn Shoal, to the northeast of Silver Bank.
It is not astonishing that, in a district where the currents are so strong
and so variable, it should be difficult to find the position of reported
shoals, the original positions of which are generally very inaccurate.

Mariguana, the Plana Cays, and Sanaana.

Plate I. ; Plate IX. Fig. 4.

On account of the strength of the trades we did not visit Mariguana,
Samana Island, or the Plana Cays. The highest point of Mariguana is
stated to be a hundred and one feet, and at the east end there are several
hummocks of from forty to sixty feet. The hills, I am informed, consist
of the reofular Bahamian aeolian rocks, and the same rocks also occur on
the Plana Cays. The north shore of Mariguana is skirted for its whole
length by a reef. The west shore is bold, and the 100 fathom line of the
bank is close to .the land. The south shore, with the exception of a reef
which protects Abraham Bay, is clear; the bottom is covered with clean
sand. At Southeast Point a reef of nearly nine miles in length extends
to the easternmost extremity of the spit formed by the bank off the east-

- ■•, Pleuna Cays

-■%^>= 5 r\

Lat. 22° 40' N. ; Long. 73° 34' W.

ern face of the island. The reef is close to the 100 fathom line on the
south side of the spit, but on the north side it is half a mile distant.

The eastern side of the West Plana Cay is fringed by a reef, while both
the north and south sides of East Cay are skirted by a reef about a third
of a mile from the shore, and extending nearly five miles off the east end.

\im'^

AGASSIZ: BAHAMAS. 95

The highest point of East Plana Cay is about seventy feet. On the north
side of West Cay the bank limited by the 100 fathom line extends about
two miles.

Samana Cay is nine miles long and about a mile and a half broad.
Its surface is uneven, stated by Sir Henry Blake to be perfectly barren,

Samana
or At^ vo o d C aj^

<M'"''' '

Lat. 23° G' N. ; Long. 73° 40' W.

and the most inhospitable looking of the islands. It attains a height
of one hundred feet. The narrow cay running nearly due east and
west is practically fringed by a coral reef, occasionally broken for its
whole length both on the north and south shores.

The Caicos Bank.

Plate I.; Plate IX. Fig. 3.

Caicos. Bank is the largest of the eastern outlying Bahamas. Its out-
line is most irregular, extending sixty-four miles east and west and fifty-
eight north and south. The northern and the greater part of the eastern
face are edged with islands, all consisting of eeolian rocks.

The wearing away of the Caicos Bank has taken place on a propor-
tionally larger scale than on the Crooked Island Bank. The Caicos
Bank is protected from the northeast trades by a line of islands extend-
ing from West Caicos east in a great curve, namely, the Providenciales,
North Caicos, Grand Caicos, and East Caicos, South Caicos, Long Island,
the Ambergris Cays, and Swimmer Rock, forming the only land on the
east face of the bank. A few isolated patches of rocks of a^olian origin
rise a few feet above the surface on the southern spit of the bank, and
on its west face French Cay and other shallow patches mark the out-
line of the former Caicos Land. On the sides facing the interior of the
bank the shores of the northern islands are all low, passing into lowlands
scarcely rising above the level of the sea, and thus forming most indis-
tinct coast lines, which gradually pass into the shallow soundings char-
acteristic of the northern part of the bank.

From the northern end of West Caicos as far as the south end of
South Caicos there is on the outer edge of the bank a coral reef

96 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

extending nearly unbroken along its whole length in from three to
twelve fathoms of water. The reef forms a wide belt between West
Caicos and Providenciales, with a narrow passage of from two to five
fathoms of clear water between the reef and Providenciales. It skirts,
both as a fringing and as a barrier reef, the northern shore of Provi-
denciales, with but a narrow channel between the barrier reef and the
shore. It extends along North Caicos, at a distance varying from
half a mile to two miles off shore. There is a break in the reef at
Juniper Hole, between the North Caicos and Grand Caicos reef. The
latter extends nearly unbroken along the north side of Grand Caicos
and East Caicos to Cape Coraete, and thence south along the east faces
of East and South Caicos. Along Long Cay the 100 fathom line is close
to its eastern face, from its southern extremity the reef runs in isolated
patches towards x\.mbergris Cays. We were unable to observe its course
along the southern edge of the bank. The edging reef reappears again
about twenty miles southeast of West Sand Spit, extending all the way
to that point and westerly to French Cay, beyond which it forms exten-
sive and nearly continuous reefs along the edge of the bank, such as
Molasses and the Southwest Reefs east of West Caicos.

A little to the south of French Cay we entered upon the bank, passing
through an opening in the edging reef, and steamed nearl}' east across
the bank, carrying from two and a half to three fathoms of water, to
Long Cay, passing in sight of Middleton and Six Hill Cays, and out

SIX HILL CAY.

into Turk's Island Passage south of South Caicos. We were in sight of
Ambergris Cays, beyond which are dotted the Seal Cays, running across
the southern part of the bank. The highest point on the islands of the
Caicos Bank are the hills on North Caicos and Ambergris, which are
said to be one hundred feet high. West Caicos is about sixty -feet high.
North of the track we followed in crossing the bank, as the pilot in-
formed me, the water shallowed very rapidly, w^iile in the direction of
the Seal Cays the water is deeper.

The peculiar light green tint assumed by the western edge of the bank
follows the belt of the edging reef, and is in striking contrast to the white
bottom inside of the reef and the adjacent dark blue color of the water
outside of the 100 fathom line. This color, when reflected on the edge

AGASSIZ : BAHAMAS. 97

of the clouds, is often seen at great distances ; it is known to the bank
fishermen as the bank bhuk.

Our pilot informed me that in from six to ten fathoms of water on
the southern part of the east face and the southern face of the bank there
are large tracts covered by a species of Thalassia, judging from his de-
scription of the weed.

We examined the bottom on the Caicos Bank on our way across from
French Cay to Long Cay, As we steamed along we came upon patclies
of Gorgonians and of large heads of corals (Astrseans) separated by long
irregular lanes of sand, or nearly choked by it. The coral belt of the
western face of the bank was, where we followed it, in from four to ten
fathoms, although limited in extent, of a rather more luxuriant growth
than on the east face of the bank. At Molasses Reef it has reached the
surface, and isolated patches are exposed at low tide. The bottom con-
sisted mainly of broken shells and of ajolian sand, mox'e or less tied
together by the roots of calcareous alg?e.

On leaving Caicos Bank the fine white aeolian cliffs of the southeast
end of Long Cay formed a prominent landmark, and gave ns an excel-
lent idea of the extent of the disintegi'ation of all that side of the Caicos
Bank land. The Turk's Island Passage is remarkable for the abundance

jimIJM^

LONG CAT.

of its pelagic life. Ifi the first place, gulf-weed was most abundant ;
Appendicularise in their houses swarmed ; a species of Mertensia was a
common Ctenophore ; Liriope, annelid and crustacean larvae, embryos
of corals, Pneumodermon-like Pteropods, large and small Arachnactis ; —
in fact, a finer display of pelagic types than we found anywhere in our
cruise outside of the track of the Gulf Stream between the Cuban coast
and Florida.

Great Inagua.

Plates I. and VIII.; Plate TI. Fig. 3; Plate IX. Fig. 3.

Inagua is about forty-five miles long ; its outline is most irregular, its
greatest breadth being only eighteen miles. East Hill, the highest point
on the island, stands on its east face, and is one hundred and thirty-two

VOL. XXVI. — NO. 1. 7

98 bulletin: museum of comparative zoology.

feet high. North of East Hill a number of hills rather higher than those
of the south shore extend to Northeast Point. A number of low hills
from thirty to ninety feet in height skirt the southern face of the island ;
when approaching them from the south, as we did in coming from Cape
Maysi, they appear like distinct islands. The hills are covered with
stunted bushes and palmettos. The west shore is comparatively low,
Mortimer Hill north of Mathew Town being its highest point. Beyond
Alfred Sound on the north shore rises James Hill, about ninety feet
high, the only elevation west of Carmichael Point ; flanking Ocean Bight
are a number of hills from seventy to ninety feet in height. The interior
of the island is low, wooded, and intersected by many salt water lagoons.
A recent coral sand beach, quite steep, is the landing place of Mathew
Town to the north of the lighthouse. The material for the beach comes
in part from the wearing of shore coral sand rocks which extend all the
way along the coast from the Lighthouse Point northward, and in part
also from the fragments of the belt of coral reef which grows upon the
narrow coast shelf of the island. The low plain forming the greater
part of Inagua Island apparently consists of shore coral rock rising but
a few feet above high-water mark, which has been deposited in the sinks
between the seolian hills flanking the northern and southern shores. The
aeolian hills rise from the northern and western ends of the island, and
increase in height to the eastward. This plain is fairly covered with
vegetation and with coarse grass, and aff"ords excellent pasturage for
cattle and horses, which run wild over the island.

The eastern face of the island is skirted by a reef from six hundred
feet to half a mile from. the shore, the 100 fathom line being about half
a mile from the reef Nearly the whole of the south shore is also sim-
ilarly skirted by a reef, with the exception of a short distance round
Southwest Point on the west coast. The coral growth is in patches, the
water off" shore being bold and the 100 fathom line close to the land.

Alfred Sound on the east side of Northwest Point in Inagua is an
anchorage for small vessels, protected by a reef nearly dry at low water,
which runs out from Palmetto Point at a distance of a mile and a half
from the shore. Standing at the western opening of the reef harbor
are two low cays covered by a little vegetation, and connected by a reef
or patches of corals and Gorgonia3 growing in the shallower water. To
the eastward a long spit extends, formed by a coral reef, inside of which
no less than seven fathoms are found in places. At the east end of the
spit the reef is quite awash, and the course of the reef is clearly indicated
by the long horseshoe-shaped line of breakers. The reef which extends

AGASSIZ: BAHAMAS. 99

east from Palmetto Point also skirts the western part of Ocean Bight.
All along the shore from Palmetto Point to Northeast Point the bank
formed by the 100 fathom line is very narrow. The coast is steep, and
is exposed to the full action of the northerly winds.^

We did not visit Little Inagua. It is quadrangular in form, eight
miles long by five miles broad (Plate IX. Fig. 3). Its highest point
on the north side is sixty feet, and on the south side there are a few
hills of nearly the same height. The northeast and northwest shores
are bold, with a clear sandy bottom to a distance of about one third
of a mile. There is a coral growth off the south and east sides ; on the
west side the bank to the 100 fathom line is perhaps a mile wide.

We may readily imagine the various banks on the Bahamas to have
been somewhat in the condition of Inagua. An extensive and flourishing
coral reef on the edge of the bank may have supplied the material for the
formation of the great shallow bank which once existed at the foot of
the seolian hills, which in their turn were formed from dry material,
supplied from the shallower portions of the bank, blown into dunes and
consolidated.

The outlying eastern islands, vSamana, the Plana Cays, and Mariguana,
resemble in their structure Watling and Rum Cay, occupying nearly
the whole of the bank of which they form a part.

Neither Inagua nor little Inagua has changed its outline greatly
since the subsidence took place whicli so materially altered the face
of the line of banks from Crooked Island to Navidad Bank. Judoing
from an examination of the charts, it appears that, as we pass eastward
from the Crooked Island Bank, the banks have been worn away generally
in proportion to their easterly position. The Crooked Island Bank has
in proportion to its size a greater amount of dry land fringing its bor-
der than Caicos Bank, that in its turn more than the Turk's Islands
Bank, while the Mouchoir, Silver, and Navidad Banks carry more water
as we pass to the east.

1 Dr. Henry Bryant visited Inagua in 1859, and gave a short account of the
island in 1806. (Proc. Bost. Soc. Nat. Hist., Vol. XI. p. 63.) lie considers the salt
lake "the remains of the original lagonn which once occupied tiie greater part of
the island, and wliich has heen gradually filled up so evenly that the portion which
still remains is nowhere more than a few feet in depth." As Dr, Bryant well says,
the fertile plains are merely salt plains covered with a coarse grass dotted here and
there with clumps of stunted trees, and scarcely elevated above the level of the salt
lake. Dr. Bryant considered both Inagua and Watling Islands as " merely a nar-
row margin of an interior lagoon" ; the plains as "the remains of lagoons formed
by fringing reefs filled up by the action of the winds and waves."

100 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

It is worthy of notice tliat the position of all the important islands
of tlie Great Bahama, Crooked Island, and Caicos Banks — as well as
of Little Bahama Bank with the exception of Bahama Island — is on
the weather side of the hanks, either on the eastern or northern face.
On the western side we hnd only small islets or rocks, the remnants of
the western islands which were the first to disappear on the subsidence
or the tilting of the western shore. This must have resulted in leaving
the northwestern and eastern faces of the banks at a higher level, and
must also have given to all the banks of tlie Bahamas a very gradual
dip to the westward. Tliis is found to be the case whether we examine
the section of the bank from Andros to tlie Santaren Channel, or from
Eleuthera to the Tongue of the Ocean, or from Long Island to the Old
Bahama Channel, or that of the Little Bahama Bank from Great Abaco
to Burrow Cay, or a line parallel to this from Spanish Cay to INIemory
Eock. The configuration of these wide shallow banks, flanked by isl-
ands on the windward face and by islets or rocks on the western face,
plainly shows that the gi'owth of the recent corals, which flourish in
from three to fifteen or more fathoms almost around the whole outer
face of the banks, has added very little to the amount of land, which con-
sists wholly of a?olian rock. It is only here and there that bights in the
disintegrating islands Iiave been filled as it were by the material thrown
up from adjoining coral reefs, and thus coral sand flats have been formed
abutting upon older scolian hills. These flats, however, form but an
insignificant part of the laud which remains, and play no important
part in the configuration of the islands of the Bahamas.

While crossiuiT the Caicos Bank we encountered large masses of
gulf- weed.

The Turk's Islands.

Plates I. and VIII. ; Plate IX. Figs. .'>, 6.

The easternmost of the Baliama Islands are the Turk's Islands, which
rise from a narrow bank running about thirty-five miles in a north and
south direction. From the middle of the bank a wide tongue extends
for seven miles eastward, on which the soundings are from eight to ten
fathoms. Grand Turk Island is five and a half miles long and about
one mile broad. The eastern f;ice is formed by a narrow ridge of a3oli;ui
hills about seventy feet high, witli their steep face on the east. The
western slope of the ridge is flanked i)y recent shore coral rocks, which
have formed the flats between the landing and the hills, and have iso-

AGASSIZ : BAHAMAS. 101

lated the salt punds once the source of a profitable industry for the
inhabitants. On the west side of the island near the southern end
the 100 fathom line is less than eight hundred feet fi'om the shore,
and the reef which skirts that part of the island forms an exceed-
ingly narrow belt from the 4 or 5 fathom line to the 12 or 15 fathom
line. It fringes the southeastern end of the island, and on the east
face extends in a mass of disconnected heads irregularly scattered
over the 3 fathom line reach which stretches eastward from the shore,
gradually passing into the reef which flourishes along tlie 4 or 5
fathom line, and which extends from Northeast Reqf to Tucker Rock,
along the centre of Turk's Bank, connecting all the cays of that side.
On the northern part of the west face the I'eef is broader and has a
width of from an eighth to a quarter of a mile. It increases in width
on the north side, and there connects with the broad reef on the east of
the island. To the southwest of the southern spit of the island a reef
runs for nearly four miles which protects Hawk's Nest Anchorage. The
whole of the space included between Gibbs, Long, and East Cays on
one side, ;lnd from the south spit of Turk's Island to Salt Cay on the
other, is filled with huge heads of corals, some of which have l)een de-
scribed as having attained an exceptional size.^ I am more inclined
to look upon these gigantic coral heads as coi-al blocks of the usual
size growing upon isolated pinnacles of seolian rock such as are so
common in the Bahamas and the Bermudas, and which here represent
the remnants of the disintegrated greater Turk's Island, which has all
disappeared towards Endymion Rock.

East Cay is the highest of the Turk's Islands, the feolian hills being
ninety-six feet high. Cotton Cay and Salt Cay are both formed, as is
Grand Turk Island, in part of shore coral rock and in part of ceolian
rock. Sand Cay, the southern of the Turk's Islands, also consists of
seolian rock. A coral reef runs off from the island in a northerly direc-
tion for nearly two miles.

The southernmost trace of the former Turk's Islands land is Endymion
Rock, which stands isolated near the southwestern spit of the Turk's
Islands Bank. From Salt Cay to the southwest spit of the bank the
soundings vary from six to thirteen fathoms. The whole bank has a
remarkably uniform depth, and a very abrupt slope from ten or twelve
fathoms to one hundred fathoms. The bank is covered with decom-
posed seolian rock sand, mixed here and there with corallines.

1 See A. E. Verrill, Proc. Bost. Soc. Nat. Hist., 1862.

102 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Navidad, Silver, and Mouchoir Banks.

Plate IX. Figs. 7, 8, 10.

The most southerly of the Bahama banks is Navidad Bank. It has
a least depth of eleven fathoms near the southeastern extremity. From
the soundings it appears to vary little in depth. The southeastern part
of the bank slopes off to the 100 fathom line somewhat less abruptly
than either the eastern or western face. The depth on the main body
of the bank varies from eleven to seventeen fathoms. It is covered with
aeolian sand modified by Nullipores. Navidad Bank is oval-shaped,
twenty-two miles long in a n^vtli and south direction, and in the centre
about eleven miles wide.

Between Navidad and Silver Banks there are three small submarine
banks rising to a depth of ten, twelve, and seventeen fathoms.

Silver Bank, the next bank to the westward, is an irregularly rectan-
gular bank. Its eastern somewhat concave side is twenty-nine miles
long. The northern side is thirty-eight miles. The central part of that
side, at a distance of about eight miles from the southeastern extremity,
is occupied by a triangular patch of coral heads which are awash and
extend nearly five miles inland, towards the northwest angle of the
bank.

The west side of the bank runs nearly north and south thirty-seven
miles. Coral heads crop out only at one point of this side of the bank,
and also on the western part of the southern line of the bank, which is
twenty-four miles long.

The soundings on the bank vary from six to twenty fathoms. Tne
deeper soundings occur nearer the western side of the bank.

The last of the larger submerged banks is Mouchoir Bank, the outline
of which is more irregular than that of Silver Bank. Its greatest length
from east to west is about thirty-one miles. Its breadth north and
south varies greatly. The eastern face is twenty-five miles, extending
into a long narrow spit. The depth on the bank varies from seven to
fifteen fixthoms. On the north side there are two large patches of coral
heads awash, both extending some distance south on the^bank.

From what I can learn, the coral heads on Mouchoir and Silver
Banks are growing upon pinnacles of aeolian rock.

AGASSIZ: BAHAMAS. 103

Hogsty Reef.

Plates I. and II,

Hogsty Eeef is a small atoll, irregularly elliptical ; its longer axis is
something more than five miles and its shorter about three at its widest
part, measured approximately from tlie limits of the 20 fathom line.
As will be seen from the deep soundings to the northwest and south-
west of the reef, there is a depth of nine hundred and five fathoms not
more than four miles and a quarter from the 100 fathom line to the
northwest of Northwest Cay, and as great a depth as twelve hundred
and eighty-one fathoms somewhat less than four miles to the south of
Southwest Cay. The accompanying sketch of the reef (Plate II. Fig. 1)
I owe to the kindness of Captain Wharton of the British Admiralty. It
has been sliglitly modified from the soundings taken by the " Wild
Duck " off the northeastern face of the reef. Quite an extensive pla-
teau was developed on that side, extending the 100 fathom line more
than a mile to the eastward from the position formerly assigned to it.
With that exception, the shelf between the 3 fathom line and the 100
fathom line is quite narrow, and the slope most abrupt between seven-
teen and a hundred fathoms. It is only on the eastern edge that suc-
cessive soundings were taken of ten, eleven, and twenty fathoms before
reaching the 100 fathom line. Usually we might strike ten or twelve
or perhaps fifteen fathoms, and the next sounding would show no bot-
tom with the 50 fathom hand-line. The slope to the south of the atoll
is slightly greater than I : 2.7, and that on the northern side not quite
so steep, 1 : 3.1. (Plate II. Fig. 3.)

With the exception of the broad triangular shelf on the eastward side
of the reef, where the corals extend out fully one mile from the breakers,
the annular ring of growing corals is less than a thousand feet wide. In
the lagoon itself no heads are growing except those which are found scat-
tered between one and a half and two fathoms ; these grow more luxu-
riantly as they get within reach of the effects of the last inner line of
breakers. Here and there a head is also found in the lagoon in some-
what deeper water, but none were seen inside at any distance from the
shallower parts of the reef. On the eastern face there is a stretch of
corals perhaps a hundred yards in width, of which here and there a mass
is exposed to the air at low tide, but as a rule there is a foot to a foot
and a half of water at low tide over the shallowest part of the reef. On
this narrow shelf and on its lee side are thrown up fragments of corals,

104 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

or small heads, Nullipores, Gorgoiiians, and debris of Mollusks torn off
from the outside of the reef by the incessant swell of the trade winds.
Owing to the steepness of the weather shelf there are not more than two
or three lines of breakers usually pounding on the reef. The corals
growing to seaward are almost entirely made up of large masses of As-
trseans and of a few heads of Madrepora palmata. There was a greater
variety of species on the lee side of the shelf, — Madrepora cervicornis,
small heads of Maeandrinas, Mauicina, and clusters of Millepora, as well
as large patches of Gorgonians. Nullipores are most abundant on the
summit of the reef, growing upon the smaller fragments of broken corals,
which they also often cement together, when they are forced inward into
deeper parts of the lagoon, where the cemented masses frequently form
heads of considerable size.

Longitudinal and cross sections of the lagoon show that its bottom
is uniformly covered with coarse sand and broken shell material, or fine
sand, according to the distance from the action of the breakers. Upon
this looser material algae and corallines thrive and grow abundantly,
generally in large patches. As will be seen from the survey of the atoll
(Plate II. Fig. 1) the annular ring of corals within the 3 fathom line is
of nearly uniform width, except that on the northern edge the belt is
slightly narrower. On the shelf of the atoll, both seaward and on the
lagoon side, alga?, corallines, and Nullipores are most abundant. On the
sandy bottom of the lagoon the patches of vegetation consisted mainly
of masses of Thalassia, Caulerpa, Penicillus, Halimeda, and Udotea, while
on the lagoon coral shelf are found masses of a species of Sargassum, of
Padina, Blodgettia, Laurencia, Digenea, and several species of incrusting
Nullipores.-' The heads exposed at low tide are also more abundant on
the eastern face of the atoll, where there are no boat passages, while
on the south face, as well as on the northwestern face near Northwest
Cay, there are a number of points where the reef can safely be crossed.
There and on the spit to the west of South Cay the reef does not come
as near the surface, that end of the reef being lower both on the north
and south side of the atoll.

The entrance to the atoll is on the west side. It is flanked on th'fe
north side by Northwest Cay, and by the smaller South Cay about two

^ For a list of the West Inrlian marine al^^fe, tlie greater number of which huve
been collected also at the Bahamas, see G. Dickie, " On the Marine Algae of Bnr-
badoes," Journ. Linn. Soc, Botany, Vol. XIV. p. 146; and "Contributions to the
Botany of H. M. S. Challenger," communicated by Sir J. D. Hooker, "I. On the
Marine Algae of St Thomas and the Bermudas," Ibid., p. 311.

AGASSIZ: BAHAMAS. 105

miles off in a southeasterly direction. The entrance is steep to directly
soutli of Northwest Cay, while a long sunken spit extends from South
Cay to the northwest, encroaching upon neai'ly half the opening to the
lagoon. Coral heads are scattered all the way across the entrance, in
six to nine fiithoms of water; they become quite abundant and luxuriant
as we approach the Northwest Cay, and are continued in sliallower water
on the outer rim of the atoll all the way to the eastern face, seemingly
finer and more thriving as they extend to the eastward.

The same is the case with the heads on the outer face of the southern
and eastern face of the atoll; the coral heads increase in size and num-
ber as they extend towards the eastern extremity, and take their fullest
development on the comparatively broad shelf stretching to the eastward.
As far as I could ascertain with the water glass and lead the coral heads
do not extend beyond eighteen or nineteen fathoms, but owing to 'the
steepness of the outer slope of the atoll it was difficult to decide this.
When we examined the eastern shelf the sea was too rough to enable me
to detect the presence of heads beyond seventeen fathoms, and at that
depth they were more distant, and frequently separated by large patches
of coral sand, while towards the atoll they gradually became more nu-
merous and more thickly crowded, attaining apparently their maximum
of development in the belt inside of ten fathoms.

The poverty of the coral fauna of the interior of the lagoon is very
striking. As I have stated above, the bottom is nearly level, sloping
as seen in the longitudinal section (Plate II. Figs. 2, 4) very gradually
from a deptli of three fathoms to five and a half fathoms, when the slope
increases more rapidly to seven, ten, and then suddenly drops down to
over a hundred (140) fathoms. Transvei'se sections of the atoll show
well its gradual deepening as we proceed from the inner eastern edge to
the section across it near the western entrance of the lagoon. The floor
of the lagoon is of a very uniform depth transversely, as all the sections
taken from north to south readily show ; but the depth increases as they
are taken more to the westward (Plate II. Figs. 5-7), with the exception
of one or two small shallower patches, on which some Gorgonians and
algae are growing. On the western face, the coral heads are more devel-
oped near the two little islets firming the gate posts as it were to the
lagoon. They form a large broad patch, gradually passing on the one
hand into the broad ring of scattered coral heads below tlie 10 fathom
line on the western face of the atoll, and on the other hand into the
two narrow belts of scattered coral heads and patches of Gorgonians
found running along the interior of the lagoon more or less parallel

106

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

with the shallowest part of the reef. This belt of heads rarely extends
to a greater depth than one and a half to two fathoms, and has a width
of usually less than one hundred yards, thus leaving the whole interior
of the lagoon free from coral heads. Judging from the descriptions of
Dana, Darwin, and others of the growth of coral heads in the interior
of other lagoons, their absence in a lagoon as open as that of Hogsty
Reef seems to be an exception. This is the more remarkable as the
position of this atoll is such that from its constant exposure to the
action of the prevailing trades a great mass of fresh sea water is con-
stantly poured into the lagoon for more than half its circumference.
This mass of water of course washes constantly the scattered heads par-
allel to the line of least depth, but also throws into the inner lagoon
water loaded with particles of sand, which cover its floor and leave ap-
parently no chance for the growth of anything except algaj and corallines
in the main basin of the lagoon. It is quite evident, when watching the

•v^^

HOGSTY KEEF.

huge trade wind breakers following one another in rapid succession, that
they must act as a very efficient force pump in driving an immense quan-
tity of water to the westward, thus constantly changing the water in
the lagoon, and rapidly removing the carbonate of lime it liolds both in
solution and in suspension.^ It is only when we reach the western part
of the atoll, where the depth increases, that a great part of the water
forced into the lagoon by the breakers finds its way out in part through
passages both on the north and south faces, but mainly through the
wide western entrance of the lagoon. The entrance is nearly a mile
wide between tlie 3 fathom lines of opposite sides, and the section
across it sliows a greater depth (five, six, and seven fathoms) than at
any other section of the lagoon. A strong westerly current is always
running out, probably due to the mass of water incessantly piled up
by the breakers into the eastern half of the comparatively closed por-
tion of the lagoon.

1 Murray and Irvine, Proc. Roy. Soc. of Edinburgli, 1889, p. 79.

AGASSIZ: BAHAMAS. 1.07

The action of the water on the bottom of the lagoon is clearly seen in
all the patches of clear sand by well defined ripple marks, showing the
unstable condition of the sand even at depths of six to seven fathoms.

Owing to the steepness of the edge of the Hogsty bank, the 100
"athom line represents very closely the width of the growing belt of
corals. The approximate centre of the inner line of breakers is within
one hundred yards of the inner 3 fathom line. I was informed by
our pilot that both Northwest and South Cays had increased in size,
the former in length on the northerly face, the latter on both its east
and west faces. Northwest Cay is mainly built up of the broken frag-
ments of corals thrown up on its northern faces derived from coral
heads to the eastward and westward of the cay, and South Cay from
the heads growing either on the spit to the northwest or to the east-
ward on the soutliern face of the atoll. The cays are nearly fiat, with
steep shores about fifteen feet high, their surfaces covered with frag-
ments of broken shells and corals, and with here and there small patches
of vegetation consisting mainly of the different plants so characteristic of
the shore line of the Bahama Islands. On parts of the shores of the
cays the recent coral rock forming the underlying base of the islets can
be seen sloping seawards. The cays at the time of our visit were occu-
pied by a rookery of boobies.

One cannot fail to be struck with the fact that on Hogsty Reef the
most luxurious growth of corals is from a depth of five or six fathoms
to ten or twelve, and that below that point, while still flourishing, the
heads are not clustered as closely together as in shallower water, while
the heads similarly diminish rapidly in number as they reach the line of
breakers. Contrary to the observations of Bourne ^ and Hickson, who
lay great stress on the growth of corals inside of the lagoons of tlie atolls
they have examined, in our atoll their growth is practically limited to a
very narrow belt close to the line of breakers, of not more than one
hundred to one hundred and fifty yards from the principal line of the
corals awash. The slope from the ring of the atoll to about ten or fifteen
fathoms is quite moderate, while beyond that it drops very fast. It is
on that edge, as has been observed by others, and as is specially marked

1 Bourne states (Proc. Royal Society, March 22, 1888, p. 440) that corals thrive
remarkably and most vigorously in that part of the Diego Garcia Atoll which is
most remote from the influence of currents, and he attributes to corals a greater
capacity as veg:etab]e feeders than has been supposed. But no decayed vegetable
masses derived from the shores of the lagoon are found in either of the atolls I
have examined, Alacran or Hogsty. In these atolls there is nothing else for the
corals to feed upon except what is brought by the currents of the sea.

108 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

everywhere in the Bahamas, that the coral heads flourish best. The ring
of coral heads is generally from one hundred and fifty to four hundred
yards wide, except at the eastern face, where the width of the reef is
nearly a mile and a quarter.

There is of course nothing to show that the substructure of Hogsty
Reef differs in any way from that of the other banks of the Bahamas.
It is fair to assume tliat, lilie them, it was built up of seolian rocks which
have gradually been worn away by the action of the sea, so that between
erosion and subsidence nothing has been left to indicate the existence of
the former Hogsty land beyond the growth of corals which now flourishes
upon the underlying seolian strata, and which they completely cover.

That the corals grow in a more or less circular belt I have attempted
to explain by the action of the breakers pounding upon the surface of the
Tl)ank, thus forming a gigantic pot-hole. Finally the water breaks through
this pot-hole on its lee side, forms an opening, and changes it into a
lagoon. This explanation of the formation of Hogsty Reef had occurred
to me before I had seen the so called Serpulee Reefs of the Bermudas, of
the structure and mode of formation of which I gave a short account in
the June number of the American Journal of Science for 1894. I shall
in my report on the Bermudas further consider this subject.

THE COAST OP CUBA.

Plates I., XIII,, XIV., XLI. to XLVII,

We were fortunate enough, both in going to Santiago de Cuba and in
leaving it for Hogsty Reef, to run along the south shore of Cuba, to and
from Cape Maysi, in daylight. This gave us an admirable opportunity
to obtain an excellent impression of the terraces and the elevated reef
extending along that coast, forming so prominent a feature of Cuba at
intervals all the way from Cape Cruz to Cape Maysi, and similarly all
along the northern coast, past Baracoa, Gibara, Matanzas, Havana, and
Mariel. But while we thus obtained an excellent idea of their extent,
so superficial an examination could only give me data for the most gen-
eral conclusions regarding the part these elevated limestones and the
reefs had played in the geological history of Cuba, far less for any views
regarding their mode of origin.

The white honeycombed limestone cliffs which are so characteristic of
the whole coast of Cuba are known locally as " soboruco," the name
applied by the natives to the honeycombed limestone rock forming the

AGASSIZ: BAHAMAS. 109

elevated terraces, though originally applied only to the elevated coral
reef forming the first terrace. These terraces are striking features on
certain i-eaches of the coast, and are so noted in the descriptions of the
coast pilots that they often serve to distinguish special points.

During my trip along the north coast of Cuba I became satisfied that
I had misinterpreted the nature of the elevated terraced limestones of
the shores of the island. With other naturalists, I had been led to
think from a first examination, made many years ago, that these elevated
limestones were coral reef rock limestones representing coral reefs of per-
haps twelve to fifteen hundred feet in thickness. The true character of
the coral reef rock, which has in general been elevated to a height of not
more than twenty-five or thirty feet, is readily made out. The exposed
surface of the platform of the first terrace is often one mass of coral heads,
and it requires but little effort to reconstruct this into a recent living
reef. When we come, however, to the flats of the second and third ter-
races, the problem is not so simple. It is evident that the limestones
are of older age than those of the raised coral reefs flanking the shores.
They underlie the elevated reefs, and, while greatly altered, yet contain
an occasional mass of coral belonging to the reef-building species, but so
few in number that we can hardly call the stone of which they form a part
a coral reef limestone. These questions were brought to mind vividly at
Santiago, while exploring the limestone terraces; also at Saboney, at
Baracoa, where, the thickness of the elevated reef is not more than thirty-
five fe6t, at Banes, at Matanzas, and at Havana. It became important
to determine exactly the age of the limestones forming the second, third,
and higher terraces, and to ascertain how fiir the presence in older lime-
stones of an occasional mass of a species of reef-building coral, at different
heights all the way up to two hundred and fifty or even to over four hun-
dred feet from the level of the first terrace, justified the conclusion that
these deposits of older limestones belong to the group of coral reef lime-
stones. Such an inference from the presence of a few corals in the older
limestones seems to me no more reasonable than to speak of a few iso-
lated heads of reef-building species found along the shore lines of any
coast or island as constituting a recent coral reef.

My observations are here given as noted during the cruise ot the
"Wild Duck." Not having myself the time to make the necessary ex-
plorations, I gladly accepted the proposition of Professor Robert T. Hill,
of the United States Geological Survey, to carry on this exploration, and-
to determine the character of the elevated terraced limestones, and their
relation to the overlaying elevated coral reef, as well as the greatest alti-

no

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

tude to which true coral reef rocks could be traced. His Report will
be printed in the Museum Bulletin, and I shall refer to this paper for
such illustration as will supplement the observations of Professor Hill.^

Cape Maysi to Santiago de Cuba.

Plate I. ; Plate XIV. Fig. 5 ; Plate XL, VII.

From the westward of Cape Maysi to Caleta Point there are four very
distinct terraces, and signs of a fifth, the summit of the hill forming
Caleta Point being probably nearly twelve hundred feet in height. The
accompanying slietch will give better than any description of mine an
idea of the appearance of the shore as seen from the yacht's deck. As

TERRACES AT CALETA POINT.

we go westwai'd these terraces become more and more indistinct.
They are no longer continuous, as we saw them for a long stretch
to the eastward of Caleta Point. There has been great erosive action,
and many valleys have been cut from the shore through the lime-
stone hills back tcJ the older formations upon which the shore hills rest.
During the erosion, the hills parallel to the coast line have been cut into
disconnected patches, and often reduced to a height less than thfCt of the
second or third terrace on the shore line near Caleta Point. In the
process the lines of the terraces have also frequently been obliterated, so
that it is impossible to detect their continuity. Here and there iso-
lated remnants, generally of the third terrace judging from their com-
parative altitude, can be traced at intervals all along the shore line to
Santiago de Cuba. The first terrace formed by the line of the elevated
shore reef is most persistent ; it can often be traced for miles where the

* A preliminary note on the Geology of Cuba, by Professor R. T. Hill, will- be
found in the September number of the American Journal of Science, 1894, p. 196.

AGASSIZ: BAHAMAS. Ill

existence of the others cannot be detected. The faces of the terrace
show the exposed limestone to be greatly honeycombed. It is often
columnar in appiearance, and large caverns and cavities of all sizes open
into the exposed vertical faces of the terraces. Some of the ravines run-
ning vpestward from the shore line expose, in their section of the shore
hills, high cliffs of limestone similar in structure to those of the terraces.
This is particularly well shown in two of the ravines beyond Caleta
Point. Some of the hills near the shore, which as seen from the ship
seemed to be limestone, must have reached an elevation of at least
eleven or twelve huudred feet. To the west of Jauco River the hills in
the rear rise to at least sixteen or perhaps eighteen hundred feet, but no
shore limestone hills could be detected of a probable height greater than
one thousand to twelve hundred feet, judging from the exposures visible
in the ravines.

Near Imia Bay another great ravine has cut through the coast range
limestone hills. The second terrace is obliterated by the talus of the
third terrace, and the hill behind this rises to about the height of the
fourth terrace. This part of the coast must have jutted out somewhat,
and that portion of the shore which formerly formed the first and second
terraces has been washed into the sea, and left only the steep cliffs form-
ing the base of the third terrace, which was further inland.

Between Sabana la Mar and Baitiqueri there are traces of the second
terrace, the third terrace is specially well marked, and indications perhaps
of the fourth are visible near the summit of the shore hill.

To the eastward of Port Escondido the shore hills show the lines of at
least four of the terraces. But the breaks in continuity of the terraces
due to the elevation of the coast are, as I have stated, very numerous, and
by far the greater mass of the shore limestone hills must have been
carried into the sea. Formerly their characteristic terraces undoubtedly
extended at intervals, if not continuously, all along the southern shore,
and probably were as well marked as we have seen them to the eastward
near Caleta Point.

When we come to the entrance of the harbor of Guantanamo, the
shore hills- show still more j)lainly the effects of the great erosion which
has taken place all along the southern coast. We find this erosion to
be greater in proportion as we go northwestward from Cape Maysi.
To the westward of the harbor of Guantanamo we have fragments of
the first and second terraces. To the east of the harbor the shore is
flanked by low more or less conical hills, which from their shape com-
pletely conceal the terraces of which they are only the remnants.

112 bulletin: museum of comparative zoology.

From Guantanamo the shore hills are not terraced, and to the east-
ward of Hatibonico the first shore terrace again becomes prominent.
From the pier to the westward of Morillo Chico the shore limestone hills
are again seen. They rise there to six hundred or eight hundred feet
perhaps, and behind and above them are seen the dioritic mountains
upon the flanks of which the shore limestone ranges have been depos-
ited and have been raised during the periods of elevation. Near Baiti-
queri River an isolated shore limestone hill shows plainly four terraces.

Guantanamo (Plate XIV. Fig. 5), like Santiago de Cuba and Port
Escondido, is a flask-shaped harbor which has been formed by erosion
in the limestone during the elevation of the belt which flanks the whole
of the southern shore of Cuba from Cape Maysi to Cape Cruz. The
■same formation is said to extend to Cape San Antonio from Cape Cruz.
Its limits on the south shore are not known to me, but I have followed
it from Cape Maysi to Cape San Antonio on the north shore. The har-
bors of the north shore from Baracoa to Bahia Honda, like Livisa,
Banes (Plate XIV. Fig. 6), Padre (Plate XIV. Fig. 7), Nuevitas, Matan-
zas, Havana, and others of Ji. similar character, have all been formed
by the erosion of their drainage area across the shore limestone hills
during the rising of the shore line.

Santiago to Saboney.

During a trip we made to the Juragua iron mines I was able to
examine more in detail a considerable stretch of the shore limestone
hills and of the elevated reef. The narrow gage railroad which runs
from the dock of the company in the harbor of Santiago de Cuba rises
rapidly upon the hills which surround the bay, and attains a height of
about two hundred and fifty feet. It then drops gradually to the shore,
and runs for a distance of about ten miles, to Saboney, along the shore
line, keeping all the way on the top of the first terrace, the surface of
the elevated reef. At the greatest height reached by the railroad behind
Santiago, in the cut about three kilometers from the company's wharf,
we found a number of fossil shells, embedded in a sort of marl formed of
the decomposed limestone of which all the exposures of the hills consist.
So much is this the character of the limestone in the vicinity of the
harbor of Santiago that nowhere except in some of the cuts to the east-
ward of the city, and after passing over the sea face of the hills, are we
able to obtain characteristic specimens of the limestone rock of the shore
hills.

AGASSIZ: BAHAMAS. 113

According to Mr. Dall, who has been kind enough to examine these
fossils, they " are of old Miocene age, the same as the age of Bovvdeu
in Jamaica and those of the Isthmus of Panama. They are also repre-
sented in Florida by what has been called the Chipola epoch, to which
the beds of Ballast Point, Tampa Bay, belong. Some of them appear
identical with fossils from the Chipola River, Florida. Fossils of the
same age were obtained on the Cuban coast by Professor R. T. Hill and
Dr. J. W. Spencer at various points."

After passing the summit and descending to a point a hundred
and sixty feet above the level of the sea, we came upon a cut in the
road, where we found hard limestone and a small isolated head of a
fine Astraea, but no coral rock. From there down, the rock is hard
limestone greatly honeycombed, as are all the faces of the terraces
along the coast parallel to which the railroad runs. The first terrace,
upon which the railroad runs, is about twenty feet above the level of
the sea. It is nowhere more than three hundred feet wide, and forms an
ideal railroad bed. At Saboney the road enters into a gap of the shore
hills and soon passes out of the region of the limestone shore hills to
meet the dioritic rocks which compose the second range of mountains.
At tlie gap of Saboney, and thence to the westward, the second and third
terraces are very prominent. The second terrace is about fifty feet above
the first, and the third not more than sixty feet above that. The third
terrace is often obliterated by erosion. The first terrace in spots has
been carried away by the sea, and cliffs have been formed which reach
the height of the second or even of the third terrace.

We were indebted to the courtesy of Mr. Shuman, the agent of the
Juragua mines, for an excellent opportunity to examine a part of this ele-
vated coral reef, forming the surface of the first terrace at Saboney, —
the old surface of the shore reef. We made good collections of speci-
mens of Mteandrin?e, two species of Astrajans, a large Colpophyllia, and
a fine Allopora, all belonging to species now found living.^ The corals
were clustered very closely, much as they would now be found growing
on a living coral reef, and quite as thickly packed.

The examination of the terraces along the line of the Juragua Railroad
shows plainly that the vertical distance between the second and third

1 The species of corals characteristic of the elevated reef (the " soboruco") be-
long to tlie following genera : Colpophyllia, Orbicella, Poritcs, Maeandrina, Agaricia,
Favia, Stephanoccpnia, Manicina, Madrepora, Isophyllia, and Siderastrea ; wherever
tlius far collected in the Caribbean, they all belong to recent species of reef builders
characteristic of the West Indian fauna.

VOL. XXVI. — NO. 1. 8

114 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

terraces, not being more than fifty to sixty feet, is not of greater height
than the depth at which corals would readily grow up from a new base.
The flat surface which forms the upper part of the elevated reef, the first
terrace, is covered with several species of corals apparently killed in situ
by the gradual elevation of the base upon which they were growing, and
of which they formed the living top at the time the elevation began.
The lower portions of the reef have grown up at greater depths, and,
as they died or were covered by more flourishing or by growing corals,
have formed the base upon which the newer crop built their way little
by little to the surface. The terraces as we see them form a series of
steps, each of considerable width, so that the fifth terrace, for instance,
might be fully twelve to fifteen hundred feet horizontally removed from
the shore line. Of course the first terrace may have been much wider
than where I measured it, and much of its sea face may have been worn
away. Still, as the depth off the south shore of Cuba is very great, and
the limit of coral reef builders is at almost every point of the coast
reached within a comparatively short distance, it is not probable that
the width of the first terrace was very considerable. The examination
which I made does not settle the question that each of the terraces cor-
responds to a separate growth of corals laterally ; it neither proves or
disproves the former continuity of the coral reef, nor its division through
the agency of the waves into perhaps one or more distinct terraces. I
can only say that tlie character of the rock exposed on the faces of the
second and third terraces showed quite plainly that they consisted of
older rocks with very marked differences, — differences which would in-
dicate a greater age in the face of the second terrace than in the first,
and a still greater age in the third, but not necessarily, as the terraces
niav have been cut successivelv in limestone of the same age. The con-
dition of the fossil corals was specially marked, and the differences
noted in fossilization quite agreed with the different conditions of the
limestones of the second and third terraces as contrasted with each
other, or as compared with that of the first and youngest elevated coral
reef terrace. We may conclude that at Saboney at least tlie fii-st ter-
race only is composed of an elevated coral reef, while the limestones
of the second and third terraces are much older, and, although containing
a few corals, yet are not coral i-eef limestones.

It may be that the isolated corals identical in species with those now
living or with those of the elevated reef forming the first terrace, which
are found on the second and third terraces near Santiago, at Baracoa, and
to' heights of nearly five hundred feet near Banes, are the remnants of

AGASSIZ: BAHAMAS. 115

the former extension of the veneer of coral reef which covered the shore
of Cuba during the time of elevation of each terrace ; that is, there
probaWy existed at the time when the older terraces were formed a
frino-ing reef, much like the present elevated fringing reef forming the
first terrace, each reef corresponding to its terrace and forming on the
sea face of the older underlying limestones a veneer of about the thick-
ness of the present " soboruco " reef, say twenty-five to thirty feet. The
patches or isolated heads upon the faces of the older limestones are prob-
ably the remnant of these successive reefs, the one upon the first terrace
being the youngest. The existence of these scattered patches has led
many observers to assume that the elevated limestones are of the same
age as the elevated reef (the first terrace).

The sheet of coral, which may have extended to a height of nearly five
hundred feet along some parts of the Cuban shore, has been eroded and
swept away by the same agencies which have formed the peculiarly
shaped limestone hills of the shore line. This erosion has been very
extensive ; with the exception of the first terrace, where the coral reef
is nearly continuous, it has left only an occasional patch of coral reef or
an isolated head here and there, at varying altitudes, to testify to the
former existence of the coral reef sheet, the higher inland parts of the
reef having all been eroded from the sides of the underlying older ter-
tiary rocks. Yet it is strange that no patch of recent reef rock of any
extent should have been met with thus far, unless we except the ele-
vated'reef of Matanzas, where, according to Professor Hill, it reaches
a height at least of a hundred feet.

A similar coral sheet, resting upon oceanic beds and the so called Scot-
land beds, has been carefully described by Harrison and Jukes-Browne ^
as extending at the Barbados from the sea level to nearly eleven hun-
dred feet. This sheet consists of a series of " platforms built up one
around the other as the island slowly rose from the sea," and this I im-
agine to have been the case with the successive elevated reefs of the
Cuban coast. The erosion at the Barbados seems to have been much
less than along the Cuban coast, leaving the bases of the successive plat-
forms of coral reef continuous. According to Browne and Harrison,*
there are six distinct terraces between the sea level and a height of five
hundred feet, and between that height and one thousand feet they indi-

1 The Geology of Barbados, by J. B. Harrison and A. J. Jukes-Browne. Pub-
lished by Authority of the Barbadian Legislature, 1890.

2 The Geology of Barbados, by A. J. Jukes-Browne and J. B. Harrison. Quart
Journ. Geol. Soc, XLVII. 197.

116 bulletin: museum of comparative zoology,

cate at least five additional terraces. They say that there are not many
places where the thickness of the coral rock is greater than two hundred
feet. In some places it appears to reach two hundred and thirty feet, and
in one place two hundred and sixty feet.

Giippy^ describes the elevated reefs of the Solonion Islands, where
extensive degradation of the surface has also taken place, as being of
very moderate thickness. This agrees with the structure and thickness
of the elevated reefs of Barbados and of Cuba, as I imagine them to have
been.

Professor Dana says : " The atoll of the Tortugas, and others in the
West Indies, are regarded by Mr. Agassiz as having a basement up to
the coral growing limit of pelagic limestone or of some other material.
It may be so ; but there is as yet no proof of it." ^ Surely the exten-
sion of the miocene rocks of the peninsula of Florida south under the
district of the belt where coral reefs are found may fairly be assumed.
The existence of the modern limestone of the Pourtales Plateau has been
proved by dredging to run close to the seaward limit of the coral reefs.
The Alacran atoll, as well as all the coral reefs of the Yucatan Plateau,
are underlain by a marine basement (not composed of corals), into the
extension of which one can penetrate nearly three hundred feet in depth.
And, finally, the recent explorations of Professor Hill and myself along
the northern coast of Cuba have proved beyond doubt the existence of a
miocene basement underlying the elevated coral reefs.

While not denying that subsidence is necessary to account in many
cases for the formation of deep lagoons and deep channels, it must be
left to others to prove that these depressions and channels have been cut
through the thickness of the coral reefs, and not through that of the
basement. It is quite possible to imagine lagoons to be formed here-
after in some of the Bahamas and in the Bermudas by subsidence, which
might have a depth of nearly sixty fathoms and yet have a thickness of
corals of not more than a few feet on the upper margin, extending along
the inner and outer slopes to a depth of not more than twelve or fifteen
fathoms.

The vertical distance between the lower terraces near Saboney was far
less than between those observed at Caleta Point. It is quite possible
that at the eastern extremity of the island the terraces which are still
so plainly marked are not the only terraces once existing. The erosion
which made, as we have seen, such great changes in the physical aspect

1 The Solomon Islands, by H. P. Guppy. London, 1887.

2 Corals and Coral Islands, p. 292.

AGASSIZ: BAHA.MAS.

117

of the shore line to the west may also have acted there to a less extent,
but perhaps quite enough to have obliterated the lines of some of the
terraces intermediate between those which at Caleta Point have been
designated as one to five. From the Juragua mines one gets an excel-
lent distant view of the shore limestone hills stretching to the west and
east, and well separated from the inner range which forms a part of
the mass of Gran Piedra. At Saboney, and at several places on the
way from Santiago along the outside of the line of the railroad, small
boat harbors have been formed on a diminutive scale by erosion, very
similar to the larger ones so characteristic of the Cuban coast, and
of which Guantanamo (Plate XIV. Fig. 5) and Santiago are the two
finest examples on the southern coast. Many bights on the coast have
been produced in a similar way. Neither at Santiago nor at Guan-
tanamo do we find any trace of terraces along the sides of the eroded
limestone hills which surround these harbors. They have been obliter-
ated by the wash along their sides into the drainage basin fringing the
bays. On our way east from Santiago we so timed our start as to see by
daylight that part of the coast near Santiago which we had passed after
dark. To the westward of the entrance of Santiago the terraces are most
indistinct, the third terrace being alone fairly defined, while to the east-
ward a fine line of terraces can be seen from the sea. We were able to
get a distant view of the second and third terraces running along the
hill slopes above the first terrace, forming the road-bed of the railroad
leading from Santiago to Saboney. Through the gap at Saboney we
could see the line of the Juragua mines on the foothills of the Gran

GRAN PIEDRA AND TERRACES NEAR SABONET.

Piedra. To the eastward of the Saboney Gap, before reaching the pier
of the Spanish American Mining Company, one of the foothills shows
four terraces quite plainly, the same number which on our way to
Santiago we had seen on the isolated hill near Baitiqueri River.

118

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY,

Baracoa to Banes.

Plate I. ; Plate XIV. Figs. 3, 6 ; Plate XL,I.

We reconnoitred the north shore of Cuba from Baracoa to Havana,
reaching the former place from Ragged Island. Baracoa is a small har-
bor, eroded out of the first terrace in a gap of the shore hills caused
by the drainage of the country behind it (Plate XIV. Fig. 3). The en-
trance is narrow, flanked on either side by well worn and honeycombed

ELEVATED REEF (SOBORUOO), ENTRANCE OF BARACOA.

reef rocks. To the south of the entrance the platform of the first ter-
race extends for nearly a mile, and upon the inner edge, at the foot of
the cliffs of the second terrace, around Miel Bay, the town of Baracoa
is situated. The second and third terraces can be seen plainly, and the
small fort which commands the harbor is built upon the second terrace.
Traces of the third terrace can also be seen on the hill to the north of
the entrance to the harbor. A river is banked off from the bay by a
beach thrown up by surf, so that it does not empty directly into the bay,
but runs parallel with the shore for the greater part of its circumference,
emptying finally into the soutlieast corner of the port, where it is pro-
tected from the swell of the sea.

The first terrace of the elevated reef of Baracoa presented the same
features which had so greatly interested us when examining the reef at
Saboney. "VVe found huge masses of Astrreans, of Mseandrinae, of Mad-
repores, of Allopora, apparently all in place as they grew, and likewise

AGASSIZ: BAHAMAS. 119

we found in the section of the terrace on the harbor side the same species
of corals composing the bulk of the reef rock. On going up to the flat
of the second terrace, we found isolated masses of corals and shells,
which evidently form a considerable part of the limestone of the cliff
extending from the first to the second terrace. To the south of the foi*t
one can reach the base of the cliffs of the third terrace, in which we found
again isolated corals of species characterizing the limestone of the sec-
ond terrace. On the second terrace we found considerable red earth.
The whole thickness of the first terrace of the elevated reef at Baracoa,
of not more than thirty-five to forty feet, wherever there was a section
showing its character, proved to be one mass of heads of corals closely
packed together, — huge masses of Astraeans, of Allopora, of Mseandrinee,
etc., all cemented together just as they must have grown on the surface
of the reef when it was living.

Not more than five miles from us in a straight line rose Yuuqne, an ele-
vated mass of limestone underlaid, according to Professor 0. W. Crosby,
by older metamorphic rocks, and stated by him to be one continuous
mass of solid reef rock,^ the upper part perhaps being of a greater age
than the lower limestone cliffs. Unfortunately the condition of the roads
was such as to make it impossible for. us to reach the mountain, which I
was most anxious to visit in order to ascertain whether in this great
mass of limestone any trace of the successive terraces of the coast reefs
could be found. Judging from what could be seen of the base of the
mountain from the sea, the talus at its foot, built up of the disintegrated
limestone which for years has been falling from the upper layers, must
make it wellnigh impossible to ascertain the thickness of the successive
terraces which compose the Yunque. Still more difficult would it be to
determine their number, and whether they were caused by the erosive
action of the sea or were built out successively seaward during the dif-
ferent periods of elevation which must have followed one another with
(geologically) considerable rapidity in order to bring about such a suc-
cession of terraces as can still be recognized to the south of Baracoa.
The apparent columnar structure of the flanks of Yunque, especially as
they approach the summit, due to the breaking away of huge masses of
limestone, gives no great hope that we shall ever be able to make out
the history and succession of the terraces of which this mass of limestone
is the monument (Plate XLI.).

Passing to the westward, the ranges of shore hills show markedly the

1 Professor Hill in liis report to me distinctly states that the Yunque limestone
is not a coral reef limestone.

120

BULLETIN : MUSEUM OF COMPAEATIVE ZOOLOGY.

effect of the extensive erosion to vvuich they have been subjected. The
terraces becop- "udis.ct after w6 pr ,he first point beyond Baracoa.
The vegetation is very dcuse, and nothing can be seen from the sea
indicating the successive hnes of former elevation. But the shape of
the hills — vrorn into sadd'- s, into peaks, into isolated cones, forming
indistinct ranges more or less parallel to the coast — sufficiently shows
how great has been the effect of the erosion on the limestone hills. Be-
tween Baracoa and Mangle Point the whole country is broken up into
disconnected bits of hills, those nearest the shore being from six hundred
to nine hundred feet high. A careful study of the cross breaks of the
shore hills might reveal some limestone faces, and throw light on the
age and succession of the limestone belt which covers so great a part of
the coast of Cuba from Bai'acoa to Cape San Antonio. Here and there
along the coast we may recognize — as, for instance, near the Saddle of
Bay, in the range of hills nearest the shore — the second and third ter-
races. To the west of the Sierra de Moa, near Cayo Grande de Moa,
begins a series of saddles and peaks culminating in the Cuchillas del

CUCHILLAS DEL PINAL.

Pinal, which seem, as seen from the sea, a series of gigantic ant-heaps
and disconnected saddle-like ridges, showing the great eftect of erosion in
this district. We could see the eroded shore hills flanking the Sierra de
Cristal, and the low shore extending past the ports of Cabonico, Livisa,
and Nipe, all fine examples of the flask-shaped harbors of Cuba. To the
eastward of Livisa indistinct traces of the third terrace could be seen.

TERRACES AND ENTRANCE TO NIPE.

It was not till we faced the entrance to Nipe that we once more clearly
traced in the spit to the westward the first, second, and third terraces.
To the eastward of the entrance the second terrace was well defined.

AGASSIZ: BAHAMAS. 121

Banes to Padre.

Plate I. ; Plate XIV. Figs. 6 and 7.

On each side of the entrance to Banes the second terrace could be
seen, towards Alulas Point the third can be indistinctly followed for a
considerable stretch, and towards the Sama hills there appear to be
four, if not five terraces. The winding canon forming the entrance to
Banes is a fine example of the passages, more or less tortuous, leading
into the flask-shaped . harbors so characteristic of the Cuban coast, —
harbors due to the gradual cutting away of the drainage area of which
they are the sinks during the elevation of the coast. From the
wharf of the banana plantation of the Messrs. Dubois we went inland
to the village, to a height of fifty-five meters above the level of the
sea and a distance of six kilometers, passing through a region of lime-
stone similar in structure to that of the rocks through which the
canon forming the entrance of Banes had been cut. We drove up
to the summit of a hill about eighty-eight meters higher than the vil-
lage, say one hundred and forty-three meters above the level of the
sea. The top of this hill was probably a part of the fourth terrace.
We found on its highest ridge, forming a nearly level saddle, a species of
Astraea in situ. The rock in place here, and lower along the slope of the
hill, is everywhere the same limestone, greatly disintegrated, so that the
lines of the terraces are nearly everywhere obliterated.

Having become acquainted with the entrance to Banes (Plate XIV.
Fig. 6), we had, on leaving the port, an excellent opportunity to see the

OUTLET OF PORT BANES.

canon through which we passed out, and to note the deep cuts, from
seventy-five to one hundred feet in height, forming the nearly ver-
tical sides of the main channel. The channel carries from four to six
fathoms, is long enough and tortuous enough to make a perfectly land-
locked harbor, which opens out after a little over a mile into a broad
bay with deep water, from five to six fathoms, close to the shore. The

122 bulletin: museum of compakative zoology,

shores of the bay are low, consisting mainly of decomposed limestone anr
of red earth, making an exceedingly fertile soil. Harbors like Nipe,
Banes, Padre, and the like, which are so common on the north coast of
Cuba, were of little use before the days of steamers. Their entrances,
facing the full force of the prevailing trades, make it nearly impossible
for a sailing craft to weather the shores and get an offing. It was simple
enough to sail in, to tack out against the trades was quite auotl ei
matter.

After passing Point Lucrecia four terraces can be distinguished be-
tween it and Point Sama, but owing to the extensive denudation and
erosion of the district only isolated parts of the several terraces are
left, and it is often difficult to determine to which terrace they belong.
Port Sama is an indentation formed by a break in the shore hills. To
the westward of the port a long ridge extends, at the extremity of
which the four terraces composing it stand out very clearly ; and here
we have an excellent guide for the determination of the terrace to which
the isolated saddles of the country both east and west belong. We then
come upon a stretch of shore eastward of Naranjo crowded with isolated

LIMESTONE HILLS BACK OF NAEANJO.

saddles and peaks, giving a peculiar aspect to that part of the coast.
All the way to Gibara the shore hills are eroded into the most fantas-
tic shapes, leaving no trace of any of the higher terraces. Thus the
limestone hills appear from the sea to be one continuous mass from the
base to the summit. A careful examination of the slopes and valleys
would however probably reveal in favorable localities the lines of the old
terraces. To the westward of Gibara the high limestone hills, so char-
acteristic of the shore line all the way from Baracoa to Gibara, recede
from the coast, and we now only find low shore hills, the highest point
of which can scarcely be as high as the fourth ten-ace.

The coast as seen from the sea is now low as far as Sagua la Grande,
where the Sierra Morenas approaches the coast again. The spurs of
the main range which receded from the coast westward of Gibara are
again within sight from the sea, and continue to form the principal part
of the background of the shore line past Cardenas and Matanzas, and
well towards Havana.

AGASSIZ: BAHAMAS. 123

Padre to Nuevitas.

Plate I, ; Plate XIII. Fig. 1 ; Plate XIV. Fig. 7.

The entrance to Padre is very wide, fully a third of a mile. It is
straight, and has all the appearance of being the mouth of a great river.
After about two miles it opens into a Y-shaped channel formed by isl-
ands lying opposite its western end, and then spreads into a bay (Plate
XIV. Fig. 7). The shores behind the beaches are low, and wherever
there are any rock exposures the same reef rock is found which occurs
all along the coast, and the same limestone half-way up the low hills
to the westwai'd of the town. There is a great deal of red earth, and
the soil is very rich.

The shores on both sides of the channel are flanked with low mangrove
islands. We find here the beginning of the extensive system of cays
which runs uninterruptedly from Nuevitas to Cardenas ; but to the
south of Nuevitas the channel which separates the cays from the main-
land is not as yet cut through the space separating consecutive harbors
to the north of Padre. With little more erosion the condition of things
existing to the north of Nuevitas would be extended as far as Padre.
To the westward of Padre are a number of low mangrove islands within
the reef which extends from the western side of the entrance. South
of Padre an extensive reef with coral heads awash extends from Herra-
dura Point to the entrance of the harbor. The shore line of the first
terrace is now frequently obliterated by the sand formed from the debris
of the reef outside and constantly thrown against the shores by the pre-
vailing winds. The renmants of the second and third terraces form the
hills west of Manati, which may be about two hundred feet in height,
and are the only high land in this vicinity.

The Cays from Nuevitas to Cardenas.

Plate I.; Plate XIII. Figs. 1 to 5 ; Plate XIV. Figs. 1, 3.

The entrance to Nuevitas is very similar to that of Padre. A coral
reef has formed a prominent spit to the south of the entrance, enclosing
a shallow lagoon open to the north. Such elongated patches or spits,
like the reef at Padre, are found all the way from Mangle Point, and
north of Nuevitas they form a nearly continuous living reef to the west-
ward outside of the cavs, with here and there a break or passage for boats

124 bulletin: museum of comparative zoology.

and shallow draft vessels. After passing Nuevitas tlie clusters of gulf-
weed became more and more numerous.

The long, low cays which extend westward beyond Maternillos Point
are separated from the shore by a shallow lagoon, forming a continuous
passage for small boats between them and the mainland all the way to
Cardenas. The highest point on any of the cays is on Cay Eomano, the
hills on the northern extremity of which are marked on the charts as
being about two hundred and thirty feet high. The monotony of this
low shore along which we skirted after passing Maternillos Light was
relieved by-the hills of Cay Guajaba, ninety feet in height perhaps.

While we could make out the reef rock of the immediate shore line of
Cay Sabinal, Cay Guajaba, and Cay Eomano, we could also see how great
a part of the beaches concealing it was made up of sand formed from the
decomposition both of the soboruco and the fragments of the outlying
living coral reef. After passing Boca Guajaba, the plateau to the east-
ward of Cay Romano widens greatly, and upon its eastern edge are Cay
Verde, Cay Confites, Cay Cruz, Caiman and Anton Cays, and Paredon
Grande, "which must undoubtedly once have formed a part of Cay Ro-
mano, but have been separated from it by the extensive erosion which
has taken place all along this coast, and which has swept away between
Nuevitas and Cardenas nearly all traces of the second and third terraces,
and perhaps others, and in the majority of the cays has reduced the area
once occupied by them even below the level of the first terrace.

Cay Confites.

Plate I. ; Plate XIII. Fig. 3 ; Plate XIV. Fig. !}.

The rock of which Cay Confites is composed showed this clearly. We
anchored for the night to the westward of the cay, and had an opportunity
of examining its structure. It is on the very edge of the Old Bahama
Channel, holding to it on the west very much the same relation which
Cay Lobos holds to it on the east. Those parts of the cay not hidden
by sand are all made up of coral reef rock of the first terrace. We col-
lected here, as we had at so many other places, specimens of Astrsea,
Mseandrina, and Madrepora identical with those collected at Baracoa
from the reef rock of the first terrace. Between the spits of the older
reef rock were stratified beaches of coral sand dipping to the sea, and
formed from the debris of the living reef spit extending to the south of
Cay Confites, parallel to the edge of the Old Bahama Channel. On the

AGASSIZ: BAHAMAS. 125

beaches were thrown up some of the largest specimens of Madrepora
pahiiata I have seen. The branches were from twenty to twenty-eight
inches wide, fully eighteen inches thick, and some of the pieces were as
long as eight feet, and many were over six.^

Cay Coufites is just on the edge of the bank (Plate XIV. Fig. 2). The
northern end of its slope off the cay is steep, leaving no space for an
extended reef. The cay is situated at the narrowest part of the Old
Bahama Channel, where the trade wind drives into a deep funnel all
that comes floating along the equatorial drift from the northern shores
of the larger West India Islands and from the Virgin and Windward
Islands, so that, besides the blocks of recent corals thrown up by heavy
surf on the surface of the cay, there are found logs and twigs of all sizes
and of many species of wood, beans, calabashes, sugar-cane, bamboo, and
cocoanuts. With a fertile soil many of the waifs thus thrown upon this
island would soon get a foothold, and it would be interesting to make a
list of the species of plants which are thus carried by the trades and the
currents far from their origin. On the poor soil so characteristic of all
reef rock land we found growing mainly such plants as are character-
istic of the immediate shore line in the Bahamas and other West India
Islands. The Pupa so common on the Bahamas was also found on Cay
Corifites.

Cay Lobos.

Plate I.

Cay Lobos, on the opposite side of the old Bahama Channel, is a recent
coral island, at least the exposed surface was entirely made up of frag-
ments of recent corals which conceal the foundation seolian rock upon
which the lio;hthouse has been built.

o

Paredon Grande to Cay Frances.

Plate I. ; Plate XIII. Fig. 3.

The island of Paredon Grande is composed of low, rocky bluffs, re-
maining from the disintegration of the second terrace, the sea front of

1 On Enderbury's Island similar large Madrepores with tlie mode of growth of
M. palmata extend over areas of twelve to fifteen feet in diameter. These vie ii>
size with the hu'je masses of Porites one meets with on the reef flat near Honolulu.
They are nearly as large as those found in the rock of the inner reef of Tongatabu,
stated by Dana (Corals and Coral Islands, p. 180) to measure twenty-five feet in
diameter, or as the Astraeans and Ma;andrinae both there and at the Feejees meas-
uring from twelve to fifteen feet (Ibid., p. 146).

126

bulletin: museum of comparative zoology.

■which in many places is changed into extensive sand beaches. On the
mainland in the distance can be seen the outline of Mount Gunagua.
After passing Cay Coco, which only at Coco Point rises above the
first terrace, we sailed along a series of disconnected patches of reef
rock, the remnants of larger cays undoubtedly, like Cay Coco, Paredon
Grande, and the like. Along the face of Cay Guillermo are here and
there short stretches of low cliffs, parts of the first terrace, separated
by low rounded hills. The low hills of Triguano Island are now faintly
seen in the background. The plateau formed by the system of low
cays we are describing is widest in the region from Cay Verde to Cay
Frances. The terraces here were probably of great width, as the hills of

PATCHES OF SOBORUCO, CAT FRANCES.

older rocks upon which they flank do not in this part of Cuba come near
the coast line, as is the case from Santiago to Cape Maysi, and as far west
as Gibara.

On the cays to the eastward of Santa Maria Cay, as far as Cay Frances,
the reef rock cliffs are separated by sandy beaches, and the hills are so
eroded that it is impossible to determine the terraces of which they once
formed a part. Upon the outer edge of the plateau, in from four to ten
fathoms of water, an extensive nearly unbroken reef runs from Nuevitag
to Cardenas, and is the source from which are formed the long sand
beaches that reach between the low cliffs of reef rock forming the sea
faces of tlie line of cays. There are patches of living corals often ex-
tending close to low-water mark, with heads awash at low tide. From
these sand beaches, exposed to the full force of the trade winds, patches
of recent scolian rock have occasionally formed on this line of cays.
Such a patch was observed at Cay Frances. The inner cays, many of
them extending toward the shores of the mninland, are composed of lime-
stone, and have been washed away below the level of the sea and con-
verted into regular mangrove islands. Cay Frances is really the first
break admitting vessels of any depth inside the cay reef plateau, and
forming a channel leading to tlie main passage running parallel to the
coast. Between this and Sagua there are two others, the principal and
deepest one leading to Sagua la Grande.

AGASSIZ: BAHAMAS, 127

The Fragoso Cays are low, and protected on the face of the inner
channel by innumerable small cays. Here and there we find cays with
a low reef rock bluif or point, but generally they are bordered by sand
beaches. Passing Tocinero Point, we come to a wide bay, a passage for
small boats making for the mainland, and soon opened upon the mainland
an isolated saddle similar to those so frequently seen farther to the east-
ward, the remnant probably of limestones belonging to the third terrace.

To the eastward of Vela Cay there is also a wide passage, leading from
the outside of the cays to the interior channel along the main shore of
the island. We next come to the Lanzanillo Channel, a similar break
between the outer cays. To the westward of this extend Lanzanillo
Cay, the Jutias Cays, Cay Canete, and Cay Cristo, separated by the
Boca de Marillanes. All these low cays are eroded parts of the sec-
ond terrace, with low reef rock bluffs separated here and there by long
stretches of coral sand beaches derived from the outlying living reefs of
the edge of the shore plateau.

Sagua la Grande.

Plate I.; Plate XIII. Fig. 4; Plate XIV. Fig. 1.

We now reach the entrance to Sagua, a wide channel not less than
eleven miles long, carrying nineteen feet of water, fringed on all sides
with low mangrove islands. To the northwest of the anchorage opposite
the town stretches out a wide, shallow bay, the outer edge of which is
protected by the continuation of the numerous cays on the north of the
entrance to Sagua (Plate XIII. Fig. 4 ; Plate XIV. Fig. 1). In the dis-
tance behind Sagua rise the Sierra Morenas, and to the eastward the
Lomas de Sagua la Grande. Sagua is built on piles on the highest
part of a mangrove island, which can hardly be called a part of the
mainland. A causeway of limestone brought from the neishborinc hi<rh-
lands forms the main street and the sidings of the railroad station. To
the westward of Sagua and of Bahia de Cadiz extends an archipelago of
low mangrove cays, reaching to the entrance of Cardenas Bay (Plate
XIII. Fig. 4). Cay Piedras, the most westerly of the islands on the
Nuevitas-Cardenas plateau, shows a ledge of low bluffs of tlie usual reef
rock. In the background of the cays the Liraonar range becomes a
prominent feature of the landscape. The hills approach the shore of
the mainland again, and we soon make out indistinct traces of some of
the higher terraces in the horizontal lines of tlie saddle-like hills seen

1'28 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

in the distance. To the eastward the Pan de Matanzas rises above the
horizon, and the hills which surround it to the eastward and westward
make part of the shore line itself. The Bay of Cardenas is easily en-
tered, and we anchored close to the town. There is no rock in place
at (Jardenas, but the rocks brought for building purposes from the neigh-
borhood are older limestones. All the way east from Cardenas to the
inner southern extremity of Cay Romano there is an excellent water
way fur small boats. The inland navigation is protected the whole way
by an almost unbroken line of cays. The distance between the outer
cays and the cays forming part of the immediate shore of the mainland
is often considerable, though the depth of water is not great. This
wnde inland sea-way, and also the channels separating the cays on the
wider passages, like Boca de Cahete, Boca de Marcos, the entrances to
Buenavista Bay, Cardenas, Sagua, Caybarien, and others, are undoubt-
edly due to mechanical agency. The rock washed by the greater or
smaller waves of the outer sea or of the inner bays is readily changed to
sand, and thus great stretches of the low cays are converted into flats,
and the finer particles washed away, increasing little by little the depth
of the water upon them ; the currents deepen the channels between
the cays, or form the still deeper passages, giving egress to much of
the disintegrated reef rock.

Cardenas to Matanzas.

Plate I. ; Plate XIII. Fig. 4.

The peninsula of Tcacos, which forms the western protection of the
harbor of Cardenas, is low, fairly wooded in parts, and opposite Diana'
Cay there is a fine cliff of older limestones ; other outcroppings occur
also on the peninsula, farther out from Cardenas, and nearer its eastern
point.

After rounding Cay Piedras and passing Monito Cay we followed the
low coast making the continuation of the peninsula of Icacos apd stretch-
ing toward ^Matanzas. This is flanked on the sea face by low reef rock
cliffs, which become higher as we proceed westward ; behind them rise a
low range of well rounded limestone hills, perhaps two to three hundred
feet in height. In the background, to the eastward, the Limouar range
passes into the Santa Clara hills, and they in their turn are connected
by a low range of hills (the Camarioca Paps) with the Pan de Matanzas
lying to the westward of the entrance to the Bay of Matanzas.

AGASSIZ: BAHAMAS.

129

Matanzas.

Plates I., XLII., and XLIII.

As one approaches the entrance of Matanzas harbor the eroded hills
gradually pass into distinct terraces to the eastward of the mouth of the
bay. The low elevated coral reef forming the first terrace comes into
prominence, and next the line of the second can be distinctly traced,
while on the opposite shore near Sabanilla Point three of the terraces
are plainly marked. A part of the Bay of Matanzas and the Yumnri
Valley have undoubtedly been formed by erosion during the elevation of
the surrounding country. The extent of this erosion is seen on all sides
of Matanzas, from the terrace's of the eastern and western sides of the
harbor to the gently sloping hills rising to Bellamar, or to the conical
paps of Camarioca, or to the rolling foothills rising to the massive Pan

PAN DE MATANZAS.

de Matanzas, or to the vertical limestone cliffs nearly as high as the Pan
itself to the westward of it. The caves of Bellamar are the finest exam-
ples of the many caverns found in the limestone district of Cuba. Their
opening is on a comparatively level plateau from three hundred and fifty
to four hundred feet above the harbor; but to which one of the terraces
the plateau corresponds it is difficult to say. With the exception of the
terrace (perhaps the second) indicated in the railroad cutting as we as-
cend from the bay, there are no traces of other terraces. A few isolated
corals are found on the level of the entrance to the caves. The first ter-
race follows the line of the road, which runs round the bay on the east
side. At all cuts of the first terrace we find several species of corals,
apparently in place in the reef rock. From Montserrat one gets a
beautiful view of the Yumuri Valley. I found no corals anywhere in
the limestone on the way up, but a good many shells are embedded in
the limestone at an elevation of four hundred to four hundred and fifty
feet. The wide valley of Yumuri, with its terraces to the north, is shut
in by low ranges of hills, which are cut through immediately back of

VOL. XXVI. — NO. 1. 9

130

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Matanzas by a deep gorge. This gorge must at one time have been the
entrance to a flask-shaped harbor very similar to such harbors as Nipe,
Padre, and the like, which once opened through this canon into the Bay
of Matanzas, but the bottom of which has been elevated above the present
level of the bay.

We had an excellent opportunity, on going out of Matanzas on our
way to Havana, to observe the plainly marked first and second ter-
races inside of Sabanilla Point, just beyond the pilot station. Before
reaching Guanos Point the base of tlie second terrace is seen to be dis-

TERRACES, GUANOS POINT, MATANZAS.

tinctly undermined, the limestone sloping very gently to the low bluff
of the first terrace. After passing Guanos Point the sea has encroached
upon the first and second terraces, leaving vertical bluffs which probably
represent the summit of the third terrace. The fourth terrace may be
seen indistinctly to the east of Guanos Point, at a height of from five
hundred to six hundred feet. In the background there are indications
of the fifth terrace.

Matanzas to Havana.

Plate I. and Plates XLIV. to XLVI.

On leaving Matanzas we come upon a stretch of coast line which
is in marked contrast with the configuration of the island to the east-
ward. The line of shore, protected by outlying cays, which extends
from Cardenas to Nuevitas, is now replaced by a nearly continuous
vertical bluff, representing probably the summit of the third terrace,
which extends to about opposite the Managua Paps. This low shore
line belt of the third terrace is cut through by valleys following the
course of the small rivers which empty on this part of the north
shore between Matanzas and Havana ; as, for instance, the Rio Canasi,
the Santa Cruz valley, and the Jaruco River. An examination of some
of the cuts made by the rivers on this part of the coast might throw
important light on the history of the limestone forming the different

AGASSIZ : BAHAMAS.

131

terraces of the shore line. The height of the bluffs of the shore line, or
of the hills corresponding to the third terrace, is about three hundred
and fifty feet. Seen from the mouth of Canasi River Gap the limestone
hills through which the river has cut its way remind one of the flat hills
apparently damming the course of a river in the anthracite region of
Pennsylvania.

In the background, after passing westward of the gap, the Jaruco hills
come into view\ They are greatly eroded as seen from the sea, and
present the characteristic erosion of the vertical cliffs of the limestone
hills on the eastern extremity of Cuba. The same seems also to be the
case with the hills to the rear of the Santa Cruz cut. Inland behind the

^^ - ., r^^'

SANTA CRUZ GAP.

third terrace, where it can be traced near Santa Cruz, the general level
of the land appears lower than the terrace in front of it. There seems to
have been an extensive valley, parallel with the coast, formed by erosion
to the rear of the shore hills. The range of hills extending to the west-
ward towards Havana is the continuation of the Sierras do Jaruco. It
seems to trend somewhat away from the coast, and also to be considerably
lower. The Managua Paps are marked on the chart as not more than
seven hundred and thirty-two feet high. In the vicinity of Havana the
second row of hills is not of greater heiglit than that of the third ter-
race. To the westward of the Jaruco River entrance there are large
isolated blocks, the remnants of the soboruco after the general disinte-
gration of the first terrace. These indicate cleai'ly the extent of tlie
erosion whicli has taken place, and wliicli hns here widened tlie reach
of tlie slope of the first terrace towards the base of the cliffs or of the
hills forming apart of the second terrace. Professor Dana ^ has called
attention to the existence of similar huge masses of reef rock upon the
shore platform of coral islands.

i Corals and Coral Islands, p 179.

132 BULLETIN : MUSEUM OF COMPAKATIVE ZOOLOGY.

To the westward of the Jaruco River the shore bills gradually become
lower ; they are not above four to five hundred feet in height, with
rounded slopes, from which all traces of the terraces have been oblit-
erated. Only here and there can we trace the first terrace. As we
approach Havana from the eastwai'd the shore hills continue to dimin-
ish in height, showing greater marks of erosion. They are not above
two hundred or two hundred and fifty feet, and when the Morro comes
in sight they are about a hundred to a hundred and twenty feet. It is
only near tlie Morro that tlie terraces are again discernible. The first
terrace is most prominent to the westward of the entrance of the har-
bor, while some of the hills which surround the port can readily be
recognized as parts of the second and third. The higher hills to the
southeast of Havana show no traces of terraces. To the westward the
table-land of Mariel consists of several terraces, the steps of which are
plainly visible on entering the harbor of Havana. Tlie reef rock of
the first terrace (Plate XLV.), extending westward from the Castle de la
Punta to the San Lorenzo quarries, is full of fine specimens of corals, of
the same species which haV^e chai'acterized the first terrace of the ele-
vated reef wherever we have examined it, whetlier at such distant points
as Saboney, Baracoa, and Cay Confites, or at Matanzas or Havana.

To the westward of Havana the coast hills are low. They are flanked
b}' the line of the first tei'race, and cut through here and there by the
valleys of the small rivers taking their rise in the Mariel table-land, and
in the broken range extending westward to the Sierras de los Organos.
]\Iariel, Cabanas, and Bahia Hunda are the only flask-shaped harbors
found between Havana and the eastern extremity of the Colorado Reef.
The Colorado Reef forms the edge of a plateau extending outside of the
shore line from Bahia Honda to Cape San Antonio ; there are on it com-
paratively few cays (Plate Xlll. Fig. 5). That part of the coast resem-
bles only in a general way the cay protected shore of Cuba extending
from Cardenas to Nuevitas.

On leaving Havana for Caj' Sal we kept farther out to sea than when
steaming from ]\Iatanzas west. We were thus able to get a better view
of the inner range of hills rising behind the low shore hill range.

AGASSIZ: BAHAMAS. 133

THE CORAL REEFS OF CUBA.

Plates I. and XIII.; Plate XIV. Fig. 3; Plates XLIV. to XL VII,

Outside of the elevated coral reef which may be said to flank nearly
the whole coast of Cuba is found the growing reef, extending to ten or
twelve fathoms or more wherever there is an insular platform of any
width between the elevated reef and the 100 fathom line. This living
reef is most irregular, occurring often merely as patches, or as spits
forming small boat harbors, or as either short barrier or fringing reefs.
But in many localities the general line of the reef is interrupted, either
from the shifting bottom of coral sand on the vsea face of the elevated reef,
or from the great steepness of the shore slope, as along the south coast
of Cuba from Cape Cruz to Cape Maysi, and nearly as far north as Nue-
vitas. Here, and along the face of the cays protecting the main shore
of the island as fiir as Cardenas, the continental platform is much wider,
and forms a coast shelf of considerable width, upon which corals flourish,
principally in from three to seven fathoms of water, though they extend
into depths of ten or twelve fathoms, or even more. From Cardenas to
the eastern extremity of the Colorado Reef corals do not flourish, except
on the Jaruco Bank, probably from the same causes which influence their
distribution on the south coast. Inside the Colorado Reef as far as San
Antonio exist a series of cays similar to those on the submarine bank to
the eastward of Cardenas, but on a smaller scale. On the south shore,
from the Isle of Pines to Cape Cruz, we have an extensive archipelago
of similar structure upon a wide platform the edge of which is fringed
with corals. On this plateau are, to the east of Trinidad, the Great
Bank of Buena Esperanza, the Manzanillo and the Laberinto de Doce
Leguas Banks, and innumerable cays dotted over its surface. ,The reefs
between Cape Cruz, the Isle of Pines, and Cape Corrientes I have not
examined, and know them only from the description of the Coast Pilot.
The Colorado Reef I have only seen at two points.

On the south coast, between Cape Maysi and Cape Cruz, the promi-
nent stretches of reefs are those to the eastward of Jucaro Anchorage,
on each side of Damas Cay, and the belt extending eastward fronr
Brava Point, whei-e the reefs extend to the foot of the Ojo del Toro and
form a sheltered reef harbor. A small patch of corals also sti;etches
to the westward of Cape Cruz. In addition to the many cays which
stud the banks between Cape Cruz and the Isle of Pines, many parts

134 BULLETIN : MUSEUxM OF COMPAKATIVE ZOOLOGY.

jf this region are " thickly covered with mushroom-shaped coral shoals,
which suddenly rise from soundings of five to eleven fathoms to just
awash." These coral heads are said to be especially thick in the vicin-
ity of Boca Grande, and extend towards the mainland. A prominent
reef extends all the way from Breton Cay north of Boca Grande as
far as Zarza de Fuera Cay. From that point the reef is limited to a
short stretch of patches to the eastward of Trinidad. The coast there
is flanked by steep rocky cliffs. The 100 fathom line is close to the
shore as far as Cochinos Bay, the western shore of which is sandy, and
terminates in Piedras Cay, whence a bank extends to the westward
which connects with the Jardinillos Bank. The reef encircling this
bank starts from Diego Perez Cay in Cazones Bay, and extends on the
east face to East Guano Cay, the most easterly extremity of the bank.
The southern fece of the bank is "skirted by a chain of reefs," broken
to form Rosario Channel. Beyond the channel a "formidable barrier
coral reef" runs on the prolongation of the bank out towards Dry
Shingle, to the eastern end of the Isle of Pines. Within Dry Shingle,
the Calapatch Mehagon Reef is one mass of corals just awash.

The south coast of the Isle of Pines is skirted by a reef. On the
western side of the island there are, as on the shores of many of the
cays of the banks between Cape Cruz and Cape Corrientes, extensive
patches of " soboruoo," parts of the first terrace or elevated reef of
Cuba. A coral reef closely skirts the whole west coast of the Isle of
Pines. The Isle of Pines is by far the largest of the islands and cays
which are dotted over the greater part of the wide banks off" the south-
ern coast of Cuba. The eastern extension of the Isle of Pines Bank
reaches to Rosario Channel, and the western extension to Cape Frances.
The coral reef skirting the west coast of the island stretches to the
westward, bordering the San Felipe Cays, and from there to Cape
Frances there are patches of coral reefs. From Mangle Point, near
Cape Frances, to Cape Corrientes the coast is flanked by soboruco, with
here and there a short stretch of coral reef.

On the north shore the Colorado Reefs extend nearly all the way from
Cape San Antonio to Cabaiias. There are a number of openings through
the reef, with an inside channel navigable for small vessels (Plate XIII.
Fig. 5). To the east of Cape San Antonio the reef is at a consid-
erable distance from tlie coast line, extending across the mouth of the
wnde bay formed between it and Buena Vista Opening. From here the
great barrier reef runs parallel with the coast, and is separated from
it bv an irregular belt of low cays, not many in number, compared

AGASSIZ : BAHAMAS. 135

with the endless chain of cays extending eastward from Cardenas as fat-
as Nuevitas.

The entrances to the flask-shaped harbors of Bahia Honda, Cabanas,
and Mariel are made by breaks tiirough the extension of the barrier
reef, which from the Manimar Opening to Cabanas has become ahnost a
flinging reef. Beyond Cabanas, and as far east as Matanzas, there are
but few patches of corals, one to the westward of Mariel, another ofl:' Man-
gles Point to the westward of Havana, and we find corals again on the
Jaruco Bank and off Maya Point on the eastern side of the entrance to
Matanzas. Tlie absence of corals is undoubtedly due to the great depth
immediately oft' shore along that whole stretch of coast line.

From Cardenas to Nuevitas extends a shore platform varying in width
from five to twenty-five miles. The outer edge of this coast shelf is
edged for nearly its whole distance by a narrow coral reef extending from
three to seven fathoms in depth, or perhaps more. A belt of cays, many
of them of great size, like Cay Coco, Cay Romano, and Cay Sabinal (Plate
XIII. Figs. 1, 2), lie between the edging reef and the shore, leaving
large bays or channels navigable for small boats between their western
shores and the main island. Between Cardenas and Cay Coco the cays
are smaller, and are either scattered in clusters over the shore plat-
form, forming small archipelagos like those off Santa Clara Bay, Sagua
la Grande, and Caybarien, or else form a belt more or less parallel
with the 100 fathom line (Plate XIII. Figs. 3, 4).

These cays are the result of erosion combined with the wearing action
of the sea. On their sea face we find patches of soboi'uco elevated to a
height of from five to fifteen feet. Thq first terrace, formed mainly of
soboruco, has been greatly eroded in the district between Cardenas and
Nuevitas, and in many cases only small patches of the soboruco are left.

To the eastward of Cay Confites the edging reef little by little trends
toward the shore as the shore platform becomes narrower, and to the
east of Nuevitas the reef becomes a fringing reef, forming disconnected
patches separated by sandy banks between Maternillos Point and Herra-
dura Point. At some places, as oflfthe entrance of Padre and of Nuevitas,
the reefs form small reef harbors off" the shore, similar to that of Alfred
Sound on Inagua. They are readily distinguished, even at a consider-
able distance from the land, by the exquisite emerald color of the enclosea
waters.

The Jaiagua anchorage is an excellent example of the formation of
an anchorage behind a coral reef caused by the breaking through of a
channel oDPosite the mouth of a river. (Hydrographic Chart, No. 518*.)

136 bulletin: museum of comparative zoology.

Opposite the river Moa the coral reef which stretches across the port
of Cayo Moa is interrupted, and the break forms the entrance to Cayo
Moa Bay. A still better example is perhaps that of the anchorages of
Yamaniguey and Cafiete, where the outlying coral reef is broken through
in two places, opposite each of the rivers which gives the anchorage its
name. These lead into an elongated bay, the northern end of which is
the Bay of Cahete and the southern the Bay of Yamaniguey.

Between Herradura Point and Cape Maysi we find only here and there
patches and short stretches of fringing and edging coral reefs, or small
reef harbors, as between Gibara and Sama Point, and south of Mulas
Point to Banes, and between the entrances of Nipe and Port Canova.
From that point the shore platform becomes wider and the reef more
continuous to oft' Cayo Moa ; it runs close to the shore from Mangle
Point to Jaragua. From this place to Cape Maysi corals grow outside
of the elevated soboruco reef only where the deep water does not come
too close to the sea face of the elevated reef. But the living coral reefs
of this part of the coast, all the way from Cape Maysi to Nuevitas on
the west, and to Cape Cruz on the south coast, are of little importance.

DISTRIBUTION OP CORALS IN THE BAHAMAS.

Plate I. and Plates IX. to XII.

It is interesting to follow the distribution of corals upon the banks.
Excluding the edging reefs of the Great and Little Bahama Banks, as well
as those of the outer edge of the smaller outlying eastern banks which
have been referred to elsewhere, we find that upon the interior of the Great
Bahama Bank reef comls are limited to a wide belt running east from New
Providence as far as Current Island, and extending in a southerly direc-
tion. Between New Providence and the Ship Channel another extensive
tract, known as the Middle Ground, is full of coral heads. It is separated
from the New Providence ground by a barren area, and a similar area
separates it on the south from another great patch of coral heads to the
westward of Norman Cay. On examining the chart, it will be found
that these areas of reef corals are in a position to be swept by the clear
water of the Tongue of the Ocean and of the Northeast Providence
Channel in regions where the bottom is not in constant turmoil from the
action of the sea. In fact they probably grow upon the isolated patches
of seolian rock, which afford the corals a solid foundation. With the

AGASSIZ: BAHAMAS. 137

exception of the edging reef on the west side of the Great Baliama Bank,
on the east face of the Gulf Stream, we find the bank wholly bare of
coral heads. It is one wide expanse of marl and of coral or coralline
sand to the west of Andros, and to the north or northwest as well as
west of the line from Andros Island to Gay Lobos. To the southeast
of the termination of the Tongue of the Ocean tlie whole surface of the
bank is dotted over with an innumerable mass of rocky heads, forming
a wide field of seolian rock ledges, separated by narrow channels of from
two to three fiithoms of water or less, througli which only small boats
can navigate. This field extends almost to Nurse Channel on the south,
and occupies a large triangular area reaching to the southern edge of the
Jumentos Cays. Many of the heads are covered with coral growth.
There are similar but more widely separated patches of rocky heads on
Columbus Bank, and another extensive area flanking the southern edge
of the bank from north of Blue Hole Point to Lobos Gay. In the dis-
trict to the north of tlje line from Blue Hole Point to Diamond Point
these patches are met with nearly forty miles from the edge of the bank.
Another important patcli of coral heads is found to the west of the south-
ern end of the Tongue of tlie Ocean, south of Andros.

In addition to the Middle Ground southeast of New Providence, there
is a band of rocky heads occupying the centre of the bank to the west
of Norman Cay, and extending south till it reaches the sand bores to the
east of Green Cay. Over the greater part of the rest of the bank the
disintegration and reduction of the heads and islands and islets have
been so complete as to leave the surface covered only witli aeolian sand
and fragments in various stages of destruction, from the finest impal-
pable powder to fine sand, coarse sand, and flakes of coralline algee from
an eighth to a quarter of an inch in leugth.

On the smaller banks, the western and northern edge of Little Bahama
Bank north of Bahama Island is occupied in part by heads, and the edge
between Settlement Point on Great Abaco and Carrion Crow Harbor is
fairly covered with heads between the few islands and islets reaching
the surface. On the eastern edge of the bank the heads are interspersed
between the cays, and nowhere form such a peculiar ground as the Mid-
dle Ground of the Great Bahama Bank. On Crooked Island and Caicos
Banks the sunken heads are limited to a comparatively small area; in
the former to the west side of the bank between Fish Cays and Castle
Island, and in the latter to the area of the bank south of Ambergris
Cays to South Rock and Swimmer Rock, and along the western edge
of the bank. On Mouchoir and Silver Banks there is but a small area
where coral heads are found.

138 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Bat perhaps the most characteristic of the Bahama banks, showing
the disintegration of the larger a3oHan islands into small isolated heads,
is the Turk's Islands Bank, where the whole area to the south and west
of Grand Turk Island, as far as Salt Cay on the one side and East Cay
on the other, is one maze of heads with deep channels between them
(Plate IX. Figs. 5, 6).

Tlie "coral heads" described by Mr. J. A. Whipple^ in his Journal,
referred to by Professor Dana^ as standing in fifty feet near Turk's
Island, are not coral heads in the strict sense of the word. They are
mushroom-shaped rocks of seolian structure, veneered with coral growth.
Similar mushroom-shaped rocks of volcanic origin covered with corals
occur in Kaneohe Bay, on the weather side of Oahu in the Sandwich
Islands, and are common in the Bahamas and Bermudas.

These tracts of coral heads are seen to be in areas accessible to the
pi'evailing winds, and to the food supply swept upon the banks by them
and the prevailing currents. These patches, forming the Middle Ground
or " coral heads " region, are without doubt clusters of seolian rock, which
have not become entirely disintegrated, and which rise to very varying
heights above the bottom of the bank. Upon these in many places
corals and Gorgonians have obtained a foothold. Upon the smaller
banks, closed on one or more sides, the corals or coral heads extend
only a shoi't distance upon the side open to the action of the prevailing
winds.

Upon Little Bahama Bank we find no extensive tract of coral heads
outside of those areas which may be called edging or fringing reefs, or,
as on the southern fiice of Bahama Island, their extension a short dis-
tance upon the bank. The edging reefs may in general be said to exist
along the margin of all the Bahama banks in from three to ten, or even
fifteen to eigliteen fathoms of watei-.

Tliere seems to be nothing in the distribution of the coral reefs upon
the Bahamas to show that they form anything more than comparatively
thin layers of corals overlying the reolian rocks which constitute the
substructure of the islands and the smaller sunken patches, andthe foun-
dation of the banks. Wliat their thickness or extent must have been in
order to form the extensive beaches which supplied the material which
was blown into the dunes of the Bahama seolian land, is another ques-

1 "The trunk wbicli made up two tliirris of its heiglit was only fifteen feet in
diameter along its upper half, and it supported above a great tabular mass one
hundred feet in diameter, wliose top was bare at low tide."

2 Corals and Coral Islands, p. 133.

AGASSIZ: BAHAMAS. 139

tion. It may uot be far out of the way to assume that their thickness
was uo greater than that of the elevated reef of the Cuban coast on the
opposite side of the Old Bahama Channel.

The diiierence in the formation of the coral reefs of Florida, the Ba-
hamas, and Cuba is niost striking. The coast of Cuba is characterized
oy the elevated reef terrace which surrounds it. In the Bahamas the
corals are scattered over wide areas, either on the interior of the banks
or lining their margijis ; tljey have played no part in the building up of
the Bahamas. In Florida the corals have not only built up the outer
keys, but also the inner line of islands, and the outer reef reaches the
surface at many points. In the Bahamas the reefs are subsiding, in
Florida they are stationary, and in Cuba they have been elevated.

HYDROGRAPHY OP THE BAHAMAS.

Plates I.-VIII., and Plate X. Figs. 4, 5.

This brief sketch of the Hydrography of the Bahamas, of the Wind-
ward Islands, and of the Caribbean, will serve to show the relation of the
coral reefs to the great banks upon which they are found. The Con-
tour Map of the Caribbean (Plate VIII.) was prepared for my Eeports
on the " Three Cruises of the Blake," at the Hydrographic Office,
through the kindness of Commander J. R. Bartlett, U. S. N., Hydrog-
rapher, and it is for the sake of convenient reference reproduced in
this Report. For the chart of the Bahamas (Plate I.) I am indebted to
Commander C. D. Sigsbee, U. S. N., the present Hydrographer of the
Navy Department.

The sea ftice of the Bahamas indicates depths and steep slopes "fully
as remarkable as any yet observed in the Pacific, even among its equa-
torial islands," ^ and which are in great contrast to the more gentle
slopes oft' the Florida Reefs.^

The character of the slope of the bottom off" the Florida Reefs has been
well described by Pourtales ^ in the Introduction to his Monograph on

1 Dana, Corals and Coral Islands, p 216.

■■^ See Langenbeck, he. cif., p. 28. I nm surprised to find that Dr. Langrenbeck
questions this. A comparison of the West Indian and Equatorial Pacific Hydro-
graphic Charts leaves no doubt on this point.

3 Deep-Sea Corals, by L. F. de Pourtales, 111. Cat. Mas. Comp. Zool., No. IV.,
1871. See also Pourtales, in I'eterm. Geog. Mittheilungon, 1870, Heft XI. ; and A.
Agassiz, The Tortugas and Florida Reefs, Mem. Am. Acad., XI. 107, 1882.

140 bulletin: museum of compara.tive zoology.

Deep-Sea Corals. He has called attention to the great width (compar-
atively speaking), varying from three to six miles, of the slope covered
by calcareous mud extending from the outer reef to the 100 fathom
line. This is in marked contrast with the narrow plateau found outside
of any reef on the Bahamas between it and the 100 fathom line. This
plateau is nearly always a mere narrow shelf, sometimes less than a
hundred feet, and rarely more than a mile in width at its greatest
breadth.

An examination of the sections given on Plate IV. Figs. 1-3, and Plate
V. Fig. 16, will readily show the character of the slope of the bottom oft"
the Florida Keys, and the striking contrast of this slope with the steep,
almost vertical sides of the sea faces of the edges of the coral reef banks
of the Bahamas (Plate IV. Figs. 4, 5, 6, Plate V. Figs. 10, 15). Along
the Florida Reef, after passing the 100 fathom line, we find a rocky
belt of limestone built up of organic fragments of all kinds of Inverte-
brates, which extends very gradually to the 250 or 300 fathom line,
and has a greatest breadth of about eighteen miles off Sombrero Light.
Beyond this rocky belt, the Pourtales Plateau, the slope becomes
steeper towards the trough of the Gulf Stream. Off Cuba the rocky
slope is very abrupt (Plate V. Fig. 16), and this abrupt slope is char-
acteristic of the whole line of the Cuban coast, both north and south,
except along the face of the Old Bahama Channel. Figure 16 of
Plate V. is a section from Sand Key to Havana ; Figure 14 is a section
from Coffin's Patches to Elbow Cay on Salt Cay Bank. This shows with
great clearness the gentle slope off the reef to two hundred and thirty
fathoms, and the abrupt pitch of the Salt Cay Bank off Elbow Cay.

The character of the sections across the trough 'of the Gulf Stream
becomes more clearlj' indicated in the sections from Fowey Rocks to
Gun Cay (Plate IV. Fig. 3), from Hillsboro Inlet towards the North-
west Providence Channel (Plate IV. Fig. 2), and from Jupiter Inlet to
Memory Rock (Plate IV. Fig. 1). The abrupt slope on the Bahama
Bank side characterizes all the sections.

The sections of the two branches of the Old Bahama Channel passing
into the Straits of Florida on the two sides of Salt Cay Bank (Plate IV.
Fig. 5) show a more abrupt slope on the Cuban side than on the oppo-
site side off Salt Cay (Plate V. Figs. 13-15), where the slope is more
gentle ; while the section from Anguila to the opposite bank (Plate IV.
Fig. 6) shows its steepest slope off Anguila as compared with the
Bahama Bank slope.

As we proceed eastward, the section from the Santa Maria Cays to the

J

AGASSIZ: BAHAMAS. 141

opposite shore of the Great Bahama Bank indicates that beyond the 100
fathom line the extension of the wide plateau off' the main island takes a
more gentle slope to the central part of the trough, while the bank slope
is quite abrupt. (Plate V. Fig. 1.) It is the opposite in a section from
Paredon Grande to Guinchos Cay, the Cuban slope l)eing the more abrupt.
(Plate V. Fig. 2.) lu a section from Cay Lobos at right angles to the
channel across to Romano Cay, the two slopes are equally abrupt.
(Plate V. F'ig. 3.) The same is also the case in the deeper section from
Cay Romano to Diamond Point, though the Cuban slope is slightly less
abrupt. (Plate V. Fig. 4.) As we go eastward the sections across the
approaches to the Old Bahama Channel become deeper ; in that from
Maternillos Point (Cay Sabinal) to Blue Hole Point, the Cuban slope is
the steeper. (Plate V. Fig. 5.) The contrary is the case in the section
from Sama Point to Cay' San Domingo and Ragged Island. (Plate V.
Fig. 6.) The section from Cape Maysi (Point Azules) to Inagua (Plate V.
Fig. 7) shows the first eight hundred fathoms to be steeper, followed by
a gentler slope to the deepest part of the channel, while the Inagua slope
is more uniformly steep beyond the 200 fathom line. Off Haiti in aline
from Cape Isabella to Turk's Islands the Haitian slope is seen to be
steeper than that of the Turk's Islands Bank. (Plate V. Fig. 8.) The
sections across the Straits of Florida and across the Old Bahama Channel
(Plate IV. Figs. 1-3, Plate V. Figs. 1-9) show the depths of the troughs
which separate the Bahama Banks from Florida and from Cuba. The
extension of some of these sections across the banks to the Atlantic give
us an idea of their great expanse, and of the deep channels which cut into
the banks. First, the Providence Channels, which separate the Little
Bahama Bank from the Great Bank (Plate IV. Figs. 2, 3) ; others,
like the Tongue of the Ocean or Exuma Sound (Plate IV. Figs. 4-6),
which form deep bays into the bank itself; and finally the deep passages
that separate the smaller banks and connect the Atlantic directly with
the deep valley forming the approaches to the Old Bahama Chatmel.

A section run from Jupiter Inlet to Memory Rock (Plate IV. Fig. 1),
across the bank to Great Sale Cay, and rnnning off" the Little Bahama
Bank outside of Pensacola Cay, shows that the soundings on the bank
between Great Sale Cay and Memory Rock are slightly deeper than
those to the eastward towards Pensacola Cay. There seems to be little
difference perceptible in the steepness of the Atlantic slope of the bank
from the 10 fathom line to about three hundred fathoms, as compared
with the slope off Memory Rock into the channel of the Gulf Stream.
A section from Hillsboro Inlet across the Northwest Providence Channel

142 BULLETIN : MUSEUM OF COMPAKATIVE ZOOLOGY.

(Plate IV. Fig. 2) shows the southerly extension of the Little Bahama
Bank separating the Gulf Stream channel and the Atlantic bight of the
Northwest Providence Channel. This line runs across the bank from
Mores Island to Great Abaco, and shows a much steeper sea face off
Little Harbor on the Atlantic than off Mores on the Providence Channel
slope. The section from No Name Cay south of Green Turtle Cay
(Plate VI. Fig. 4) is interesting as showing the line of the outer ridge
of cays and patches, often rising to the surface, and joining the discon-
nected parts of the outer reef with the belt of deeper water between it
and the inner line of cays of which Green Turtle Cay is one.

The section from Fowey Rocks to Gun Cay (Plate IV. Fig. 3) across
the northern part of the Great Bahama Bank to the Berry Islands shows
a much steeper slope oft' these islands into the Northeast Providence
Channel than the western slope off Gun Cay in the direction of the
trough of the Gulf Stream. The line across the bank from Royal Isl-
and to Hhe Atlantic face of Eleuthera indicates a very abrupt slope both
on the channel side and on the Atlantic side, these two slopes being as
steep as the slope east of the Berry Islands.

A section made somewhat farther south (Plate IV. Fig. 4), from Orange
Cay across the northern extremity of Andros, the Tongue of the Ocean,
Nassau, and that part of the bank lying between it and Eleuthera, shows
the same featnres as the more northern section. The Orange Cay slope
is less abrupt than the east and west slopes of the Tongue of the Ocean
off Andros and off Nassan. The Atlantic slope off Elenthera is somewhat
less steep than the slope farther north. A section from Salt Cay Bank
(Plate IV. Fig. 5) across to the southern part of Andros, to Green Cay, to
Harvey Cay, and across Exnma Sonnd to Cat Island, shows a less abrupt
slope on both faces of the Tongue of the Ocean, though steeper than the
Santaren slope of the bank. Off Harvey Cay the slope into Exuma
Sound is far less abrupt than either, and the slope from the 1,000 fiithom
line to the 500 fathom line off the west side of Cat Island is nearly as
flat as that on the west side of the Straits of Florida. From the 500
fathom line the west face of Cat Island is as abrupt as the east face.
The' two slopes of Cat Island are, however, both equally abrupt, and as
steep as the slopes of the outer Atlantic faces of the bank elsewhere.

A cross section somewhat farther south (Plate IV. Fig. 6), from the lati-
tude of Anguila Island on Salt Cay Bank across the Hurricane Flats, the
southern part of the Tongue of the Ocean over to Exuma Island across
Exuma Sound, Conception Island, and Watling Island, shows the peak of
Conception Island to be one of a ridge in continuation of Cat Island, and

AGASSIZ: BAHAMAS. 143

of which Bum Cay is auother summit. The eastern slope of the Tongue
of the Ocean is les^ steep than the western face, and the slope off Exuraa
is far less abrupt than that of the peaks forming Conception Island and
Watling Island, both of which rise abruptly from a narrow base at a
depth of over twelve hundred fathoms, and finally on the line from Cape
Cabron (Haiti) to Navidad Bank the slope is equally abrupt on the two
sides of the deep valley which separates these points. (Plate V. Fig. 9.)
A line from Acklin to Hogsty Reef (Plate V. Fig. 10) and thence to,
Inagua shows very abrupt slopes in the deep channel separating Acklin
from Hogsty (over one thousand fathoms) and that between Hogsty and
Inagua,* which is more than fourteen hundred fathoms deep. The ex-
tension of that line to the eastward of Inagua runs into over seventeen
hundred fathoms, and cuts somewhat diagonally across the slope of the
northern face of the deep channel to the north of Haiti.

There are unfortunately no soundings on the Atlantic face of the
lines between Crooked Island Bank, Mariguana, the Caicos, Turk's Isl-
ands, Mouchoir, Silver, and Navidad Banks, to show definitely how far
these banks are separate elevations, or how many may be connected
together. (See Plates I. and VIII.) On the western slope the soimdings
seem to indicate that Turk's Islands and the banks to the eastward may
be the summit of a greater bank connecting them all, and that Mari-
guana, Plana Cays, and Caicos and Crooked Island Banks are the summits
of distinct banks separated by deep channels, much as Crooked Island
Bank is separated from Long Island by the Crooked Island Passage, with
nearly thirteen hundred fathoms in the deepest part. But perhaps the
Plana Cays, Mariguana, the Turk's Islands, and the other banks may
hold to one another much the same relation which the Mira por vos
Bank summit holds to Castle Island (Plate V. Fig. 11) and the Crooked
Island Bank, being separated from the last by a channel of more than
nine hundred fathoms in depth and from Cay Verde by a channel with a
depth of nearly fifteen hundred fathoms.

The soundings parallel to Turk's Islands Bank would seem to indicate a
deep channel of perhaps a thousand fathoms between it and the Caicos
Bank and the extension of the slope of Mouchoir Bank to the eastward
south of Turk's Islands Bank.

We made a line of soundings to three hundred fathoms at right angles
to Crooked Island Bank (Plate VI. Fig. 2) about one mile south of South-
west Point on Fortune Island. The bottom from three to five fathoms
was coarse nullipore and coral sand ; at from six to fifteen fathoms it
became quite fine; at one hundred and eighteen it was still finer; and

144 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

as no specimen was collected at either of the soundings beyond that, it
must have been rocky bottom. On steaming back to our starting point
for dredging, as we reached seventeen fathoms we found a few coral heads,
with large patches of fine white sand between them.

On the line of soundings we ran at right angles to the bank off
Exuma Harbor (Plate VI. Fig. 3) we brought up as far as seventeen
fathoms the coarse sand of the banks ; at eighty fathoms we found rather
finer coral sand and a few broken shells ; at one hundred and fortyrtwo
fathoms the bottom was composed of fine nummulitic and coral sand ;
the bottom was rocky at one hundred and fifty-seven and two hundred
and forty-two fathoms, and at two hundred and sixty-six fathoms we
struck fine coral ooze.

Neither of these slopes differs materially from the northerly slope of
Hogsty Reef. Off the reef the 900 fathom line (Plate II. Fig. 1) is
about four miles from the 100 fathom line. Off Fortune Island it is
at about the same distance, and off Exuma Harbor the 800 fathom
line is four miles from the 300 fathom line. (Plate IV. Fig. 6.) The
sudHen drop of these banks to the 100 fathom line occurs from the 15
to the 20 fathom line, the distance to the 100 fathom line varying from
one sixth to three fifths of a mile from these soundings, or even less at
other points on the banks, as off Green Cay, where it is not more than
five hundred feet. Off the coast of Florida the 100 fathom line is from
four to nine miles distant from the 10 fathom line, the line of the reef.
(Plate IV. Figs. 1-3, Plate V. Fig. 16.)

Off the Mosquito Bank, south of Rosalind Bank, the 600 fathom line
is nearly thirty miles from the 100 fathom line, and between the Rosa-
lind and Pedro Banks the slopes east and west from each bank are fully
as flat as that. (Plate IV. Fig. 7.) , Off the northeastern edge of the
Yucatan Bank the slope is even less, the 600 fathom line being more
than thirty-five miles from the 100 fathom line. (Plate VI. Fig. 5,
Plate VIII.)

The west face of the banks of the Windward Islands is often much
the steeper. Off Saba the 600 fathom line is about twelve miles distant.
(Plate VII. Fig. 3). The slope of the bank at the east end of Santa Cruz
is about the same. (Plate VIII.) Off the west side of Barbados the
800 fathom line is over twelve miles distant. (Plate VII. Fig. 7.) At
some points of the Grenadines Bank (Plate VII. Fig. 7, Plate VIII.) the
1,000 fathom line is within fifteen miles of the 100 fathom line. The
eastern end of the Virgin Island Bank is steep on the sea faces ; the 1,000
fathom line is in many places less than ten miles distant. (Plate VII.
Fig. 1.)

J

AGASSIZ: BAHAMAS. 145

An examination of the sections across Martinique (Plate VIT. Fig. 5),
Grenada (Plate Vll. Fig. 6), the Grenadines (Plate VII. Fig. 7), Domin-
ica (Plate VII. Fig. 8), and St. Lucia (Plate VII. Fig. 9), will show the
same character of the western slope of the islands. Off the Mont-
serrat and Antigua Bank (Plate VII. Fig. 4) this does not hold, the
eastern slope being the steepest ; nor is it the case with the ridge of
which Barbados is the summit, the eastern face being the steeper (Plate
VII. Figs. 7, 9). A section from north to south across the Virgin Island
Bank (Plate VII. Fig. 2) shows a steeper slope on the southern face
within the 500 fathom line.

That is, the continental slopes, or the slopes of such straits as the
Florida Straits, are less steep than the ocean slopes within the area
of the limestone banks.

THE CORAL REEFS AND BANKS OF THE
CARIBBEAN DISTRICT.

Plate VIII.

The following notes on the reefs and banks of the Windward Islands
and of the Greater Antilles were made during my connection with the
expeditions of the " Blake " from 1877 to 1880. As they bear upon the
general theory of the coral reefs, they are now published more in detail
than in the general account of the " Three Cruises of the Blake." The
statements concerning the reefs of the north coast of Cuba, and of the
south shore from Cape Maysi to Santiago de Cuba, are based upon per-
sonal examination, as well as those on the Florida reefs and the Yucatan
Bank. In making up my notes I was obliged to consult nearly all the
recently published charts of the Caribbean. This has led me to add a
short account of the distribution of the coral reefs of the north coast of
South America from Trinidad to Costa Rica, and also along the Mosquito,
Nicaragua, and Honduras coasts, as far as Cape Catoche on the Yucatan
Bank. With the exception of the coast near Aspinwall I have not exam-
ined any of the reefs or the outlying banks off the Mosquito coast, off
Honduras, or those between Jamaica and the Mosquito Bank. Neither
have I visited the line of shoals extending from Cayman Brae to Miste-
riosa Bank.

The reefs of the Windward Islands are almost without exception on
their weather side, and on Saba Bank, off Anegada, and on the bank east

VOL. XXVI. — NO. 1. 10

146 BULLETIN ; MUSEUM OF COMPARATIVE ZOOLOGY

of Santa Cruz we have horseshoe-shaped reefs following the outline of
the banks at depths of from six to ten fathoms. In the majority of these
cases we might look upon these curved reefs as atolls. They vary greatly
in outline, following closely the depths at which they best thrive, and
have usually taken from the shape of the basis upon which they have
developed the outline of a more or less circular belt. The summits
which have risen above the sea level are many of them more or legs
circular or irregularly elliptical, like St. Vincent, Grenada, and Domin-
ica. Others are greatly indented, like the Grenadines, Martinique,
Guadeloupe, and St. Martin ; while among the northern islands we find,
as among the smaller Grenadines, many irregularly shaped islands, as
Santa Cruz, Anguilla, St. Thomas, Cat Island, and the like, along the
shores of which we also find lines of irregular, broken reefs.

The 100 fathom line of the banks on the west face of the Windward
Islands is comparatively close to the shore, the sea face is quite abrupt,
and the mass of volcanic mud and other material which is washed down
during the rainy season is quite sufficient to prevent the growth of frin-
ging reefs, which are so flourishing on the weather side. Were the banks
of the Windward Islands fringed only by low, narrow islands, like the
Bahama Banks, we might expect to find reefs growing on their western
edge, forming thus a number of atoll-like banks, such as Crooked Island,
Caicos, and Hogsty Reef.

Horseshoe-shaped or semicircular reefs exist on isolated banks in the
Caribbean, as for instance on the Morant Cays, Serranilla, Serrana, and
Quita Sueno Banks. There are also curved reefs simulating semi-atolls
on the Mosquito and Yucatan Banks, such as Edinburgh Reef, Half-
Moon Reef, and the Hobbies on the former, and the Triangles, Sisal, and
Cay Arenas Reefs on the latter, besides the Alacran atoll.

Finally, off Belize we have the isolated banks of Chinchorro, Turneffe,
Lighthouse Reef, and Glover Bank, all more or less irregularly elliptical,
and rising from deep water beyond the 100 fathom line and inside the
five hundred fathom limit, the east face of the last running rapidly into a
depth of one thousand fathoms, while the slope of the Chinchorro Bank is
comparatively less steep, the 1,000 fathom line being at least twenty-five
miles to the eastward. Nearly the whole length of the coast off" Belize,
from Ambergris Point to Cape Three Points, is protected by a reef which
lines the edge of the bank inside of Turneff'e Island.

Let us now examine the banks of the Windward Islands formed by the
100 fathom line and compare them with the Banks of the Bahamas.

Beginning with the Grenadines Bank and throwing out of account the

AGASSIZ: BAHAMAS. 147

broad continental shelf formed by the 100 fathom line outside of Tobago
and Blanca Islands, we find the bank to extend from Bequia in a south-
westerly direction, forming a curved elongated plateau concave towards
the west, with Bequia and Grenada as its principal islands, the latter
rising to a height of over twenty-seven hundred feet, and other volcanic
summits such as Carriacou, Cannouan, and a number of smaller ones
scarcely rising above the level of the water. Corals flourish on the
Atlantic face of the bank, forming fringing reefs or barrier reefs, or often
coral ridges connecting the smaller islets. Round St. Vincent (Hydro-
graphic Chart No. 1279, Admiralty Charts Nos. 791, 956), St. Lucia
(Hydrographic Chart No. 1261, Admiralty Charts Nos. 956, 1273),
Dominica (Hydrographic Chart No. 1318, Admiralty Charts Nos. 697,
956), and Martinique the 100 fathom line forms comparatively small
banks. Only a narrow belt on the northern and western face of St.
Vincent and St. Lucia, and a somewhat wider plateau on the eastern
face of Martinique, are studded within the 20 fathom line with coral
reef patches existing either as barrier reefs or as fringing reefs. No
elevated reefs have been observed on Martinique, St. Lucia, or the
Grenadines, while Barbados, which rises to over eleven hundred feet,
is surrounded with elevated reefs forming a series of terraces round the
island. These have been carefully studied by Harrison and J. Jukes-
Browne. Barbados itself is the summit of an extensive elliptical bank
above the 500 fathom line, the greater part extending northward nearly
to the' latitude of the northern extremity of St. Lucia.

The northeast coast of Barbados (Hydrographic Chart No. 1010, Ad-
miralty Charts Nos. 956, 2485), is skirted by a coral reef from a quarter
to half a mile from shore, which encircles almost the whole island.
Off Palmetto Bay the reef is more than a mile offshore, but off the south-
east coast a bai'rier reef is found, known as Cobbler Reef, extending from
Kitridge Point to South Point, with a channel varying from one half to
three and a half fathoms, and from a quarter to half a mile in width, with
two to eight feet of water ; and parallel with this is a narrow sunken reef
with from seven to ten fathoms of water, separated from Cobbler Reef by
a belt of water from one quarter to three quarters of a mile wide, with
soundings close to the outer reef ranging from fourteen to twenty-seven
fathoms. To the seaward the 100 fathom line is not more than two
miles off.

Grenada (Hydrographic Chart No. 1316, Admiralty Charts Nos. 956,
2821) attains a height of from twenty-three to twenty-seven hundred
feet. It is volcanic, and rises from a bank on which are the Grenadines

148 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

(Hydrographic Chart No. 357, Admiralty Chart No. 1^872). The 100
fathom line is about seven miles from the east coast. The bank has twenty-
seven fathoms on the east edge, diminishing to twenty about half a mile
from the shore. On the western side of the island the edge ot the bank
is not more than two thirds of a mile from the shore.

The eastern and southern shores of Grenada are skirted by wide frin-
ging coral reefs, following the deep indentations which characterize that
part of Grenada. Between the island and the 100 fatliom line a number
of irregular banks rise from the main bank, with from seventeen to
eighteen fathoms upon them, from near the 20 or 30 fathom line, the
summits probably of islets and islands no longer reaching the surface.

The Grenadines consist of about one hundred islands and rocks ex-
tending for sixty miles between Grenada and St. Vincent, none of them
attaining a height of more than eleven hundred feet. The smaller islets
and islands are many of them fringed with reefs, or protected on the eastern
faces by barrier reefs.

Carriacou ^ (Hydrographic Chart No. 357, Admiralty Chart No. 2872),
Little Martinique, Little St. Vincent, Union, and Prune Islands are pro-
tected on their eastern faces by curved barrier reefs with deep water
between the reef and the shore (one to seven fathoms), or by fringing
reefs. The Tobago Cays are the centre of a complicated system of coral
reefs, forming the fringing reefs of Tobago and of the smaller adjacent
cays ; between the islands extend broad patches of coral heads. The
Tobago Cays (Hydrographjc Chart No. 357, Admiralty Chart No. 2872)
are in the centre of a great narrow horseshoe-shaped barrier reef, the
horns of which are nearly two and a half miles apart. Outside to the
eastward are the isolated reef patches known as Egg Reef and World
End Reef, which rise within the 6 fathom line and are similar to the
isolated reefs so common upon the bank, but most of which are covered
with coralline and coral sand too deep for the vigorous growth of corals.

Cannouan (Hydrographic Chart 357, Admiralty Chart No. 2872) is
flanked on the east face by a barrier reef with a belt of water from one
to six fathoms inside of it. On the islands to the north the coral reefs
are less vigorous. From- Moustique to Carrfacou there are a series ot
narrow, elongated ridges with from fifteea to twenty fathoms of water
upon them, rising from the 20 to the 30 fathom line and following in
general the 100 fathom curve.

The ridge of which the Grenadines, St. Lucia, and other Windward
Islands form the summits broadens out gradually as we pass northward

1 The largest of the Grenadines, rising to a height of nearly one thousand fee*^-

AGASSIZ : BAHAMAS. 149

from Grenada. It becomes materially wider uorth of Martinique. This
is indicated by the two small banks lying northeastward of Martinique
between it and Dominica. North of Dominica, Marie Galante, and Gua-
deloupe with its outliers of Desirade and Petite Terre, occupy nearly the
whole width of the ridge to the 500 fathom line, on either side. It again
increases in width at the line from Nevis' to Barbuda, and attains its
greatest lateral dimensions on the line from the western edge of Saba
Bank to the edge of the bank east of St. Martin, one fold of the ridge
extending from Saba Bank to Santa Cruz and another to Sombrero, the
valley between them falling off rapidly to the Anegada Passage.

The northwest end of St. Vincent rises abruptly to three thousand feet.
On the east side of Calliaqua Bay there is a lagoon protected by a reef.

The Pitons of St. Lucia are between three and four thousand feet
high. At Laboise the shore is skirted by a reef.

The depths of the passages between the banks of the Windward Isl-
ands are quite moderate. Only in the channels between Dominica and
Martinique and between Martinique and St. Lucia do we get a depth of
more than five hundred fathoms (in one case 548 fathoms, in the other
575), the 500 fathom line joining St. Lucia and St. Vincent to the Grena-
dines Bank, and the 500 fathom line uniting Dominica with all the banks
to the south of Sombrero with the exception of Santa Cruz and the two
submarine banks between it and Sombrero.

Dominica (Hydi'ographic Chart No. 1318), which rises to over forty-
seven hundred feet, is, like Martinique, flanked on the east by a plateau
of considerable width limited by the TOO fathom line. Between it and
Martinique to the eastward rises a small bank, with a depth of forty
fathoms. The northeast end rises less abruptly from the sea, the sound-
ings off the coast showing a continuation of this gradual slope, the 100
fathom line being three and a half miles off shore. It is interesting to
compare the submarine '^slopes of Dominica, Martinique, St. Lucia, and
St. Vincent with the steep slopes of the Bahama Banks, which have been
considered by many writers as due to the continuous growth of coral
reefs during subsidence, so as eventually to form walls of limestone
rising abruptly from nearly two thousand fathoms in depth. It is sur-
prising to find that it is the western faces of these islands which give us
the abrupt slopes, while the eastern faces, on which the coral reefs are
found growing upon a limestone bank of considerable width, show a
comparatively gentle slope.

The 100 fathom line surrounding Guadeloupe does not include within
its limits the volcanic summits of the Saintes, or the elevated limestone

150 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

islands of Desirade, Petite Terre, and Marie Galante. They are separated
from Guadeloupe by narrow but deep channels, having a depth of more
than two hundred fathoms.

The western part of Guadeloupe (Hydrographic Chart No. 363, Admi-
ralty Charts Nos. 885, 956) is volcanic, and rises to a height of nearly
five thousand feet. Grande Tei-re, the eastern division of Guadeloupe,
is almost a level plain. A part of the northern coast of the island is
skirted by reefs with channels through it. A coral bank extends along
the southern part of the northeast coast of Grande Terre and of the
eastern side of Guadeloupe.

The Saintes rise to more than one thousand feet, and are skirted
with patches of reefs. Marie Galante is of moderate elevation, its gen-
eral appearance being flat and low. It rises gradually from south to
'north. The east and northeast coasts are dangerous, and skirted by
coral reefs. Desirade and Petite Terre are joined by a ten-fathom bank
of coral reef, the south and east sides fringed by corals.

The French surveys of Martinique (Hydrographic, Chart No. 1009,
Admiralty Charts Nos. 371, 956) show that the eastern face of the isl-
and from Point Laboussaye is bordered by a nearly continuous fringing
reef, following the many indentations of its shores and of the small islets
off the coast. Outside of this, at a distance of from one to two miles
from the coast, extend a series of elongated coral banks, most of them
awash, rising from the 7 to the 10 fathom line, and forming a discon-
nected barrier reef with wide passages between them varying in depth
from four to fifteen fathoms, reaching from Caracoli Point on the north
to the Vauclin Channel on the south. Martinique rises to a height of
nearly four thousand feet.

St. Eustatius (Hydrographic Chart No. 1011, Admiralty Charts Nos.
487, 2600) is volcanic, and rises to a height of over nineteen hun-
dred feet. The greater part of the eastern coast is fringed with corals ;
on the southeast coast coral patches are found. Saba, also volcanic,
rises perpendicularly more than twenty-eight hundred feet from the sea.
The 100 fathom line is not more than half a mile from the west side,
and only three cables from the west face. It is so abrupt that no corals
grow upon its sides.

The eastern edge of Saba Bank (Admiralty Charts Nos. 130, 487, Hy-
drographic Chart No. 1002) is fringed with a narrow ledge of corals
nearly thirty miles in length, with six to ten fathoms of water. The
bank is about thirty by twenty miles. To the westward of the coral
reef, with the exception of a few coral patches, there arc nine to ten

AGASSIZ: BAHAMAS. 151

fathoms of water ; the bottom is clean white calcareous sand, with a
depth of from twelve to fifteen or twenty fathoms, gradually increas-
ing to thirty fiithorns, where the slope becomes very abrupt.

St. Bartholomew is nearly one thousand feet high, and rises from a
flat bank of considerable extent, composed chiefly of broken shells, white
sand, and corallines.

St. Martin, and also the smaller islets near St. Bartholomew (Hydro-
graphic Charts No. 1002, Admiralty Chart Nos. 130, 2038), are fringed
with reefs on the eastern and greater part of the northern faces. On
the shores of St. Bartholomew itself there are but few coral patches.
St. Martin is nearly thirteen hundred feet high. Tintammare is bor-
dered on all sides except the west by coral reefs extending one quarter
of a mile from shore.

A small inverted-comma-shaped bank is formed by the 100 fathom
line uniting Moiitserrat (Hydrographic Chart No. 1011, Admiralty Chart
No. 254) and Redonda, both of which are volcanic rocks. The isolated
peaks of Moiitserrat rise to three thousand feet above the sea. It is
fringed here and there with corals on the east end. Eedonda reaches a
height of six hundred feet.

The volcanic island of Saba (Admiralty Charts Nos. 130, 487, Hydro-
graphic Chart No. 1002) stands by itself, and to the westward of it
rises within the 100 fathom line the Saba Bank, on the southern and
eastern face of which a belt of coral reef is found.

An elongated bank is formed by the 100 fathom line uniting Nevis,
St. Kitts, and Eustatins.

Nevis (Admiralty Chart No. 487, Hydrographic Chart No. 1011) is
also volcanic, and over thirty-five hundred feet above the sea. With St.
Kitts and St. Eustatius (Hydrographic Chart No. 1011, Admiralty
Charts Nos. 487, 2G00), it stands upon a bank entirely detached from
the adjoining islands, separated, by channels of more than two hundred
fathoms in depth. From the south end of the bank a remarkable coral
ledge about a mile in breadth extends along its southwest edge for a dis-
tance of five miles, and has from eight to ten fathoms on it, with deeper
water within. Except on the west side, the shore is fringed with a coral
reef extending three quarters of a mile out to a depth of ten fathoms.

St. Kitts (Admiralty Chart No. 487) rises to a still greater height,
viz. over forty-three hundred feet ; a great part of the eastern face of
the island is flanked by a fringing reef.

Antigua (Hydrographic Chart No. 1004, Admiralty Charts Nos. 917,
2600), which rises to a height of thirteen hundred feet, is near the

152 bulletin: museum of comparative zoology.

southern extremity of a hank formed by the 100 fathom hue, at the
northeastern end of which is placed Barbuda, an island composed of
limestone, rising to a heigl of two hundred feet. The shores of
Antigua are deeply indented, especially the no'thern and eastern faces
of the island. It lies u' the uiiddle of the t. ra edge of the same
bank as Barbuda, the general depth of which is from seven to eleven
fathoms, with a coralline and coral sand bottom. It is surrounded by
many small islands and islets, and edgv.d on the east, south, and north
sides with fringing and barrier reefs,^ leaving only intricate passages
leading into the deep sounds of the island. From the entrance to North
Sound the patches of reefs and banks extend in an arc to the westward
as far as Diamond Bank, the most westerly of these coral reefs. The
Highlands are apparently composed of three terraces, the highest of
which is two hundred feet.

From the southeast end of the eastern shore of Barbuda (Hydro-
graphic Chart No. 367, Admiralty Charts Nos. 1997, 2G00) a coral
ledge skirts the shore at a distance of half a mile ; a reef also extends
one and a half miles off the north shore, the depth being thirty fathoms
within a mile of it ; and reefs run off the south coast nearly seven miles
in a southwest direction. Barbuda is connected with Antigua by a bank
with seven to thirty fathouis of water.

Between Antigua and Guadeloupe an isolated bank rises to a depth of
forty-three fathoms, and a similar bank is found to the southeast of the
bank limited by the 100 fathom line, of which the principal summits
are St. Bartholomew (992 feet), St. Martin (1360 feet), and Anguilla
(213 feet). With the exception of Anguilla, they are volcanic sum-
mits. They are surrounded by a number of smaller islets of the same
structure.

Sombrero is steep and rises to forty feet from a small bank about four
miles long. It is composed of recent limestone (Julieu), very much honey-
combed, and rises abruptly within the 500 fathom line. In a line between
it and Santa Cruz rise two submarine banks, the one with a minimum
depth of three hundred and seventy-six fathoms, the other with a depth
of three hundred and eighty-three fathoms.

Anguilla (Admiralty Charts Nos. 130, 2038, Hydrographic Chart No.
1002) is not more than two hundred and thirteen feet high. The
south shore and the eastern end of the north shore are fringed with
coral reefs. An extensive reef runs westward from Scrub Island within
the 10 fathom line for over twenty miles, with passages to the west of
the Pear Cays and eastward of Dog Island.

AGASSIZ: BAHAMAS. 153

Santa Cruz, the Virgin Islands, and the Greater Antilles.^

Santa Cruz is the broad summit of the bank which extends east of the
island for a distance of nearly twelve miles, and forms a narrow belt round
it. With the exception of a small part of the western end of the north
shore and of the western side, Santa Cruz is fringed along its whole
coast with corals, forming either fringing or barrier reefs off" the north
side, and a broad belt of fringing and barrier reefs skirting the south
side, of the island. Round the eastern extremity of the bank to the east
of the island an extensive reef (Lang Bank) follows the outline of the
100 fathom curve. On this we find six to nine fathoms. It extends on
the south face of the bank from the east end to opposite the eastern
point of Santa Cruz (Admiralty Charts Nos. 130, 485, and Hydrographic
Chart No. 1002).

The Virgin Island Bank, which includes within the 100 fathom line the
Virgin Islands and Porto Rico, is separated from the banks to the south-
east of it by a deep canon of over one thousand fathoms, connecting the
Atlantic with the Santa Cruz deep of more tlian two thousand fathoms,
separating that island from the Virgin Islands. Porto Rico is the east-
ernmost of the greater West India Islands, and occupies nearly the whole
width of the western end oi the Virgin Island Bank.

The Virgin Island Bank (Admiralty Charts Nos. 106, a, b, c, 130, and
Hydrographic Charts Nos. 1002, 2008) presents very much the same
characteristics as the Grenadines Bank. It is the summit of an elongated
bank seventy-five miles in length, reaching from Porto Rico to Anegada,
and from twenty-three to thirty miles wide. Some of the larger islands
which crop out upon the southern edge of the bank, like Crab Island,
Tortola, Culebra, St. Thomas, and Virgin Gorda, attain a height of more
than fifteen hundred feet, while there are a great number of smaller
islands, islets, and rocks barely reaching the surface, all of which are of
volcanic origin, Anegada alone being composed of recent reef limestones.

The sea face of Anegada (Admiralty Chart No. 130, Hydrographic
Charts Nos. 1002, 2008) is fringed with a coral reef extending to a
depth of about thirteen fothoms. Off the east point the reef spreads out
upon a broad bank over three miles in width, which also extends along
the western face of the island and is studded with coral heads. A wide
ban-d of coral heads extends from East Point, known as the Horseshoe
Reef, following in a general way the line of the 100 fathom curve. The

1 The reefs of Cuba liave already been described in a separate chapter.

154 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

reef is broken to form the Necker Island Passage, and opposite Virgin
Gorda a long narrow belt of corals extends nearly the length of the bank
as far as Crab Island, with from twelve to seventeen fathoms of water,
forming a regular wall with a belt of deeper water on the inside of from
twenty to thirty fathoms, and from two to five miles wide.

Nearly all the islands of the Virgin Bank are skirted on their eastern
and southern faces with patches and stretches of coral heads, often form-
ing extensive fringing reefs. This is the case with Virgin Gorda, Tortola,
St. John, St. Thomas, Culebra, and the eastern end of Crab Island.

The Virgin Islands lie on the southern edge of a vast bank of sound-
ings extending from Porto Rico, which rises to a height of thirty-seven
hundred feet, to the east of Anegada, which has an elevation of not more
than thirty feet. The contract between the soundings oflF the north and
the south face of the bank is quite marked. The southern edge of the
bank is not more than seven miles from the cays, while the northern
edge extends nearly thirty miles to the north of the islands. The gen-
eral depth of the bank varies from seven to thirty fathoms. Tortola
rises to a height of nearly thirteen hundred feet, and Virgin Gorda to a
little less. St. Thomas (Admiralty Charts Nos. 130, 2183, Hydrographic
Chart No. 1002) rises to a height of over fifteen hundred feet, and is
surrounded by many small islands and cays fringed with corals.

The 500 fathom line forms an elongated bank roimd Aves or Bird
Island, which is the summit of a wide fold at less than one thousand
fathoms running nearly north and south from Saba Bank to Bird
Island in the direction of Venezuela to about the latitude of Bequia,
a bay with a minimum depth of two thousand fathoms running between
it and the Windward Islands to the latitude of the southern extremity
of Dominica.

Aves Island (Hydrographic Chart No. 40, Admiralty Chart No. 2600)
IS only ten feet above the level of the sea ; it is composed of coral rock
skirted all round by a reef except on the west side. The islet rises
from a bank of considerable extent, which has not been defined by
soundings.

The south shore of Po.rto Rico (Admiralty Charts Nos. 130, 2600,
Hydrographic Chart No. 40) is fringed with reefs about four miles ofiF
shore. Mona Island is formed of white perpendicular cliffs about one
hundred and seventy-five feet hitrh full of holes, with numbers of
grottos and caves.

A part of the east coast of San Domingo (see General Charts of the
West Indies, Hydrographic Charts Nos. 40, 373, Admiralty Chart No.

AGASSIZ : BAHAMAS. 155

393) is formed of remarkable limestone cliffs from one hundred and fifty
to two hundred feet high. The projecting parts of the island in the
vicinity of Aquin Bay consist of very remarkable bold white cliffs and
hills resembling chalk, which occur both on the south and on the north
coasts. There are coral reefs all along the shores of the island. The
100 fatliom line adds but little to the extent of San Domingo, which is
separated from Porto Rico by a channel, the greatest depth of which is
only two lumdred and sixty fathoms. The area of Jamaica would be
greatly extended on its southern face by the addition of the plateau
formed by the 100 fathom line. Extensive areas of elevated tertiary
limestones have been traced along the coasts of Jamaica and of Sau
Domingo.

Between the eastern point of Haiti and Jamaica are the small island
of Navassa and the Formigas Bank. Navassa (Hydrograpliic Chart No.
379, Admiralty Charts Nos. 461, 486) is about three hundred feet high.
It rises from a small bank of soundings one third to three quarters of
a mile broad. Tliere is a patch of corals on the north end of the isl-
and. Formigas Bank (Hydrographic Chart No. 373, Admiralty Cliart
No. 486) is eight miles in length ; the bottom is rocky, with a general
depth of from five to seven fathoms. There are narrow veins of sand
near the edge, in from nine to eighteen fathoms.

To the southeast of Jamaica are the Movant Cays, a group of small
cays seven to ten feet high, forming with the adjacent reefs a crescent
bank convex to tlie southeast of about three by one and a half miles
wide. These cays rise from an elongated bank about twelve miles by
four within the 100 fathom line, with depths close to it ranging outside
of the reef from seven to eighteen fathoms, and thirty to fifty near the
edge of the bank. A terrace is said to rise seventy-five feet all round
the island, composed of ringing and honeycombed limestone. The Alba-
tross Bank lies to the northeast of the Morant Caj^s, and is a similar
bank ; it carries, however, from nineteen to thirty fathoms close to the
100 fathom line, and is marked on the cliarts as coral sand.

Fringing and detached reefs abound on the south sliores of Jamaica.
They are especially numerous from Morant Point to the west of Kings-
ton Harbor as far as Portland Point, although they extend, irregularly
distributed in favorable localities, to the western extremity of the island
off Alligator Point, and from Pedro Bay to St. John's Point (Admiralty
Chart No. 255, Hydrographic Charts Nos. 347, 373).

156 bulletin: museum of compakative zoology.

The Banks of the Caribbean Sea.

Between Jamaica and Honduras and off the Mosquito Bank rise a
number of banlis, ^the most easterly of which is the Pedro Bank (Hydro-
graphic Chart No. 373, Admii-alty Charts Nos. 450, 486), which rises
abruptly from depths of three hundred and hfty to five hundred fathoms.
It is irregular in shape, nearly a hundred miles long east and west, quite
narrow at the eastern end, not more than nine miles, while at the western
end it is fifty-five miles broad. The surface of the bank is level, ranging
from nine to fifteen fathoms except near the eastern and the southeast-
ern edge of the bank, where there ai'e many shoals and cays said to
consist of white limestone. The bottom on the bank is generally white
sand and fragments of corals, but there are patches of coral heads and
of coralline algae. Portland Rock is thirty-two feet high ; Plover Rock
is two to three feet high. Off the Pedro Cays, about twelve feet high, reefs
extend parallel with the 100 fathom line between the cays. Banner Reef
is just awash.

Next comes to the westward the Rosalind Bank (Hydrographic Charts
Nos. 373, 394). It is pear-shaped, nearly sixty-three miles long, and
ranging in width from fifteen to thirty-five miles. The only shallow
ground upon it is near the southeast end, where there is a wide coi-al belt
nearly three miles wide, and trending in a northeasterly dii-ection for
fourteen miles, with depths upon it varying from four to ten fathoms.
On the eastern face there is also a second ledge about thirteen miles
north of the first, eight by four miles, on which we find eight fathoms.
The general depths on the bank vary little from ten to twenty fathoms ;
the bottom is coarse coral, coralline sand, and broken shells.

Between Rosalind and Mosquito Bank lies a narrow bank, thirty-seven
miles long, not more than ten n:»iles wide, with from thirteen to thirty-
five fathoms close to the 100 fathom line. This bank is separated from
the Mosquito Bank by a narrow channel with a depth of about one hun-
dred and seventy-five fathomg, and by a somewhat deeper channel from
the Rosalind Bank.

To the south of Pedro Bank rises the small, isolated Bajo Nuevo Bank
(Hydrographic Chart No. 379, Admiralty Chart No. 391) on which
there are two remarkable hook and bow shaped coral reefs rising from
the 5 fathom line on the eastern and southern face of the bank. On the
bank the soundings vary from six to fifteen fathoms close to the 100
fathom line.

To the south of Rosalind are the smaller banks, the most northerly of

AGASSIZ: BAHAMAS. 157

which is the Serranilla Bank. It is somewhat circular in outhne, and
about twenty-four miles in greatest length. There is a general depth
of from five to fifteen fathoms upon it ; the bottom on the bank is coarse
coralline and coral sand. The bank is steep, the slope increasing rapidly
in depth from twenty to one hundred fathoms and more. On the east
face of the bank there are three small cays, from which extensive reefs
run along the eastern face more or less parallel with the 100 fathom
curve from Northeast Breaker to Beacon Cay. Farther south, east of
the Mosquito coast, are the Serrana and Roncador Banks to the east-
ward ; half-way between them and the Mosquito Bank comes the line of
the Quita Sueno Bank, of Old Providence, of St. Andrews, and of the
Courtown and Albuquerque Cays (Admiralty Charts Nos. 379, 1498,
Hydrographic Charts Nos. 373, 394).

Both Old Providence (Hydrographic Charts Nos. 395, 1372, Admi-
ralty Chart No. 1334) and St. Andrews (Hydrographic Charts Nos. 391,
94.5, Admiralty Chart No. 1511) resemble in their general features,
but on a smaller scale, some of the volcanic Windward Islands, upon the
weather side of which coral reefs flourish. The greatest height of the
volcanic peaks of Old Providence is nearly twelve hundred feet ; from
the central peak's spurs run off terminating in shore peaks of from three
to seven hundred feet in height. Besides the main island, there are
upon its bank the small island of Santa Catalina and a couple of smaller
cays ; the bank carries from two to ten fathoms of water. A barrier
and fringing reef extends round the island, and a long reef stretches
out north. St. Andrews is only three hundred feet high. A few small
cays are scattered upon its bank ; its northeastern end is flanked by a
reef. Upon Courtown Bank (Hydrogaphic Charts Nos. 391, 945, Admi-
ralty Chart No. 1511) a barrier reef extends across it from the northern
end to within half a mile of the southern extremity. The noi'thern and
eastern sides of Albuquerque Cays Bank are skirted by a reef.

The eastern face of Quita Sueiio Bank (Hydrographic Chart Nos. 394,
945) is flanked by a reef for twenty-five miles, leaving only a small part
of the bank clear at each end. The north and east edges of Roncador
Bank (Hydrographic Charts Nos. 373, 395, 945, 1374, Admiralty Chart
No. 1478) are fringed by a coral reef. The cays upon it are composed
of broken coral blocks and coral sand.

The extreme east end of Serrana Bank presents a solid crescent-shaped
reef. The space within the horns is crowded with detached shoals, with
from five to ten fathoms alongside. The east face of the bank is abrupt,
no bottom at one hundred fathoms one quarter of a mile from the 10
fathom line J the south edge of the hank is edged by a reef.

158 bulletin: museum of comparative zoology.

Misteriosa Bank (Hydrographic Chart No. 966) is the western extrem-
ity of a submarine chain flanking the northern side of Bartlett Deep
to the east of Cape Cruz, of which Cayman Brae and Little and Grand
Cayman are the summits, and to the north of which the isolated Pickle
Bank rises, upon which are from ten to forty fathoms of water, with a
coral sand bottom. Misteriosa Bank is twenty-four miles long, oval in
shape, with general depths of froni ten to twelve fathoms and a coral
bottom. To the south of it are four other small banks with from thir-
teen to nineteen fathoms of water. Still farther south are the Swan
Islands, but the limits of the bank from which they rise have not been
determined. The islands are separated by a coral reef; they are flat,
sixty feet in height, and bounded by a bold, rocky shore. The sound-
ings upon the bank are from six to thirteen fathoms, increasing rapidly
to forty fathoms on the western side of the islands.

Grand Cayman (Hydrographic Charts Nos. 43, 373, 947, Admiralty
Chart No. 4G2) is about forty feet high. At the southwest end a reef
runs out half a mile. All but the west side is skirted by a fringing or
barrier reef, which is steep to for a distance of a quarter of a mile to a
mile. At the east end it forms a solid barrier for the distance of a mile ;
on the south side there are several small cuts through the reef. At the
western end there is a large circular lagoon, Great Sound, with from one
to two fathoms, and a passage leading through the reef with from two to
two and a half fathoms. Cayman Bank, about ten miles west of Grand
Cayman, is a ridge of coral five miles'long and half a mile broad, with a
depth of from fifteen to twenty fathoms. The edge drops very suddenly
tc more than one hundred fathoms.

The highest point of Little Cayman (Hydrographic Charts Nos. 43,
373, 947, Admiralty Chart No. 462) is forty-eight feet. The south
coast is skirted by a fringing and barrier reef with a wide passage on
that side. The north coast is only fringed in part. The 100 fathom
line of the bank is generally half a mile from the shore.

Cayman Brae (Hydrographic Charts Nos. 43, 947, Admiralty Chart
No. 462) rises to fully one hundred and thirty feet. The^ north and
south coasts are skirted by a broken reef; the 100 fathom line is about
half a mile off".

The Coral Reefs along the North Shore of South America.

Along the north shore of South America (Hydrographic Charts Nos.
40, 964, Admiralty Charts, West India Sheets VIII. to XL, Nos. 394,

AGASSIZ: BAHAMAS. 159

395, 39G) and the adjacent islands there are many points where coral
reefs take a great development. With the exception of the central part
of the north coast of Tobago, which rises from a bank of about forty
fathoms, the island is fringed with disconnected fringing and barrier
reefs, growing most luxuriantly off the northeast end of the island and
near the southwest exti:emity (see Hydrographic Chart No. 354, Admi-
ralty Chart No. 505). The Testigos are surrounded by coral reefs
on the eastern face of the bank from which they rise; they are within
the broad belt of the 100 fathom line, and carry from twenty-three to
thirty-eight fathoms near them. Detached patclies of coral reefs extend
along the coast from the Gulf of Paria to the Gulf of Cariaco. Mar-
garita Island is flanked by fringing and barrier reefs on the eastern and
western coasts, as well as along its whole southern shore. The islands
of Cubagua and Coche, whicli lie between it and the mainland, are like-
wise skirted on the eastern faces with reefs. Orchila is low and flat,
surrounded by reefs on the eastern face. Los Eoques are a group of
cays rising from a bank of twenty-three miles in length ; the greater part
of it is covered by coral heads, and its sea face is edged by coral reefs.
The Islas de Aves are banks similar to Los Eoques. Buen Ayre Island
is fringed by a reef on the eastern and northwestern extremities of the
island. Little Curagoa is a coral islet with steep sides. The south side
of Cura^oa is bordered by a bank of coral reefs.

From Cape Codera to Cape St. Roman tlie coast is fringed with long
stretches of coral reefs, from Codei-a to Caracas, and on both sides of
Punta Tucacas, east of Zamuro Point. The whole eastern face of the
peninsula of Paraguana is fringed by corals. West of the Gulf of Vene-
zuela, with the exception of the fringing reefs along the eastern and
northern coasts of the Guajira peninsula, the reefs along the north shoi'e
of South America as far as the Mulatas archipelago are limited in ex-
tent, occurring on comparatively short stretches of shore interrupted by
long reaches of sand beaches (Hydrographic Chart No. 9G4). The Mu-
latas archipelago, a number of islets and islands from two to ten miles
off shore, extends from Pajaros Island to off Point San Bias, a distance
of eighty miles : they are fringed by reefs rising from seven to ten fath-
oms of water (Hydrographic Chart No. 945). The north shore of the
coast of Panama and of Costa Rica is, with that exception, free of exte-
rior reefs. The north side of the interior of Chiriqui Lagoon is thickly
studded with detached shoals of corals and coral patches.

160 bulletin: museum of comparative zoology.

The Coral Reefs and Banks of the "West Coast of the

Caribbean Sea.^

It is to be noted that on the western Florida Bank, as well as on the
Yucatan Bank, where the conditions of depth would cseem to be most
favorable for the growth of corals, in the one case the coral reefs are
limited to a few isolated patches like Arenas Cays (Hydrographic Charts
Nos. 403, 1233, Admiralty Chart No. 1209), Sisal Reef (Hydrographic
Charts Nos. 403, 1235, Admiralty Chart No. 1206), Madagascar Eeef
(Hydrographic Chai-t No. 1235), the Triangles (Hydrographic Charts
Nos. 403, 1239), Obispo Shoals (Hydrographic Charts Nos. 403, 1239,
Admiralty Chart No. 1830), the Areas Cays (Hydrographic Charts
Nos. 403, 1234, Admiralty Chart No. 1829), and the Alacran atoll (Hy-
drographic Charts Nos. 403, 966, 1234, 1235, 1240, Admiralty Chart
No. 1203) ; while in the other case they form a narrow thread parallel
with the course of the Gulf Stream outside of the Florida Keys from
north of Key Biscayne to the Tortugas.^

Along the west coast of the Gulf of Mexico (see Coast Survey Chart
of Gulf of Mexico, ^j-2Woijb) the corals are wanting, with the exception of
a few isolated patches of coral reefs off Vera Cruz coming np from six-
teen fathoms of water (Hydrographic Chart No. 967, Admiralty Chart
No. 523), which, although noted in the A.dmiralty West Indian Pilot,

1 Tlie reefs of Cuba liave been described in a separate chapter.

- Since nij' explorations of the Bahamas and Bermudas I am convinced that a
more detailed examination is needed to ascertain how far south the older niiocene
rocks of Florida extend, and to what depth they are covered by the recent coral
growths from Key Biscayne to the Tortugas. An exploration of the inner line of
keys may show perhaps how far the configuration of the western part of the Florida
Bank to the south of Cape Sable may be due to a slight subsidence, and how far
this has been concealed by the growth of corals to form the outer reef, and by the
deposition upon these older rocks of coral material which has formed the inner line
of cays, or how far erosion has been the main factor in shaping the topography of
the submerged southern part of Florida and of the adjacent keys. Borings alone
can determine this question, owing to the quantity of coral silt which Iws been de-
posited upon the mud flats north of the keys, unless we may succeed in finding
somewhere among the inner keys an outcrop of the older miocene or perhaps later
tertiary rocks.

For an account of the Florida Reefs see L. Agassiz, Ann. Rep. U. S. Coast Survey,
1851, Mem. Mus. Comp. Zo51., Vol. VIL No. 1, 1880; A. Agassiz, The Tortugas
and Florida Reefs, Mem. Am. Acad., Vol. XI., 1882 ; Three Cruises of the " Blake,"
Bull. Mus. Comp. Zool., Vol. XIV., 1888; and Coast Survey Charts, Nos. 166, 167,
168, 169.

AGASSIZ : BAHAMAS. 161

had failed to come to the notice of scientific men, have been described
by Professor Heilpriu.^ Is their absence due to the limited amount of
food brouglit by the current across the Yucatan Bank, sweeping as it
does northward towards tlie Tortugas and washing the shores of the
Florida Keys? The exposure of the coast of British Honduras and of
the Yucatan and Moscpiito Banks to the influence of the prevailing-
winds and currents, bringing with them a vast supply of pelagic food, is
probably the principal cause for the abundant growth of coral reefs in
all favorable localities on the Yucatan and Mosquito Banks and along
the British Hondui-as coast, as well aa upon the weather faces of the
Serrana and Quita Sueno Banks, the Roncador Reef, Old Providence,
St. Andrews, and the Courtown and the Albuquerque Cays.

The Reefs of the Yucatan Bank.

On the Yucatan Bank (Hydrographic Charts Nos. 1234, 1 235), with the
exception of the Alacran atoll, the reefs are merely more or less irregu-
larly shaped patches, either surrounding the cays wholly, or reaching
out from them, and rising generally from the 3 to 5 fathom line of the
bank itself. This is the case with the Areas Cays and the Triangles
(Hydrographic Chart No. 1239). The Obispo Shoals (Hydrographic
Charts Nos. 403, 1239, Admiralty Chart No. 1830) and the Sisal Reefs
are small irregular patches rising from the 5 or 6 fathom line, while
Madagascar Reef is a narrow coral ridge about two miles long, rising
from the 7 to 11 fathom line with from one and a half to three fathoms
of water upon it. On the extension of the Yucatan Bank plateau to the
westward there are also a few patches of reefs off' the Barra de Tuxpan
and scattered patches off Champoton, but neither these nor the reefs off
Vera Cruz, along the inshore edge of the narrow shore platform of the
west coast of the Gulf of Mexico, can be considered as reefs of any im-
portance. And indeed they are so unimportant that it was not unnat-
ural for Darwin and Dana to speak of the Mexican coast of the Gulf of
Mexico as free from coral reefs.

Along the Yucatan coast to the south of Cape Catoche, Mugcres Har-
bor (Hydrographic Charts Nos. 402, 1379) is formed by the extension
northward of the two fringing reefs which run along the east and west
face of the island and pass into the shifting ledges forming the Blancas
Lagoon. Corals are reported upon Arrowsmith Bank in twelve fathoms,

Cozumel (Hydrographic Charts Nos. 402, 1380) is flat, not more

1 Proc. Phil. Acad. Nat. Sci., 1890, p. 303.

VOL. XXVI. — NO 1, 11

162 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

than forty to fifty feet in height. It is sandy, with rocky ledges sepa-
rating the beaches. It is separated fi'om the mainland by deep straits.
From the northern end of the island a sand bank exteuds eleven miles.
The west side has deep water close to the shore.

Espiritu Santo and Ascension Bays (Hydrograpliic Charts Nos. 402,
1380, Admiralty Chart No. 1795) are both protected by narrow reefs
extending across the ■ mouth of the harbors. These reefs rise from a
depth of from four to six fathoms.

The Reefs and Banks of British Honduras.

The extensive barrier reef running along the edge of the Honduras
Bank reaches from Ambergris Cay to the southern part of the Gulf of
Honduras, and forms a broad stretch of reef building patches separated
from the coast by a wide channel parallel with it and carrying from ten
to twenty fathoms. It resembles, on a much smaller scale, the great
Austi'alian barrier reef The Honduras Reef rises in from six to seven
fathoms, and the depth gradually' increases as we go westward towards
the deeper parts of the channel which form the inland sea separated from
the Caribbean by the wide baiTier reef composed of numerous detached
patches and cays. The reef varies in width from one or one and a half
miles at the southern end to sixteen and even twenty miles in the latitude
of Gladden Spit, where the channel is divided by an irregular area of
banks, cays, and patches of coral heads, with many narrow but deep veins
of water between them. These banks are often awash, or rise from six
to seven fothoms, and carry fi-om one and a half to three fathoms. One
of the channels leading across the bank, the English Cay Channel, carries
more than twenty fathoms. The principal passages through the reef
carry from four to six fathoms, and rapidly open into the deeper water
leading to the main inside channel. Chetumal Bay and the adjoining
territory to the north and east seem to indicate the manner in which the
cays and barrier reef of Honduras have been formed. The outer sea
face, as well as the inner eastern line of the bay, is edged by a reef. If
a general subsidence or erosion of the Honduras coast is going on, that
part of it must soon be changed into a bank like the banks flank-
ing the present coast, or into something similar to the barrier reef near
Point Gladden, the small banks and cays indicating the position of the
former low ranges of hills and valleys such as we find on the lowlands
to the east of Chetumal Bay. Subsidence or erosion, or both, have formed
the outline of the bank ; coral reefs have edged it, but there is nothing

AGASSIZ: BAHAMAS, 163

to prove that they have formed the land intervening between the exist-
ing lines of reefs. The bank limited by the 100 fathom line off British
Honduras is narrow compared to the wide bank off the Mosquito coast,
the width of which at its northern extremity seems to a certain extent
to influence the growth of coral reefs and patches, for they are by no
means so flourishing as upon the narrow bank off the Honduras coast.

In addition to the barrier reef off British Honduras there are a num-
ber of outlying banks which fi-om their description resemble the smaller
of the- Bahama Banks. They are the Chinchorro Bank, the Turnefte
Islands, and Lighthouse and Glover Reefs, which in a certain degree
resemble irregular atolls. They all lie within the 500 fathom line. The
greater part of Chinchorro Bank (Hydrographic Charts Nos. 394, 1072,
Admiralty Chart No, 1796) is bounded by a reef. There are several
openings into the interior which are navigable for small fishing vessels.
Lighthouse Reef (Hydrographic Charts Nos, 394, 1120) is bounded by
an edging reef; the north end forms a half-moon curve about four miles
in diameter ; the east side of the reef is steep to. The Turnefte Islands
(Hydrographic Charts Nos. 394, 399, 1120) are an extensive cluster of
mangrove islands on a bank of coral and sand thirty miles long and from
four to nine wide. The cays are all skirted, at a short distance from both
the east and west sides, by a barrier reef with an inside channel having
over two fathoms. Glover Reef (Hydrograpliic Charts Nos. 394, 1120),
with the exception of a small part of the east side^ is fringed by a barrier
,reef.

The Mosquito Bank.

Upon the Mosquito Bank (Hydrographic Charts Nos. 394, 945), as
upon the Yucatan and Honduras Baulks, there are patches and reefs
ritiing upon the surface, not entirely covering it, but scattered, as we
find the Areas, Triangles, Madagascar, Alacran, and other small reefs.
They are more numerous upon the Honduras Bank, and get the full
benefit of the prevailing winds, and of the sweep of the currents against
their sides and over their surface.

There are a great number of irregular patches of coral reefs from
Monkey Point north of the Bluefield Lagoon (Hydrographic Charts Nos.
391, 1292, Admiralty Chart No. 1504), within four or five miles from
the coast. From Pearl Cay Lagoon to Man-of-War Cay there are a
number of small irregularly shaped banks rising from a depth of eight
to ten fathoms, the largest of which are the reef patches of the Pearl

164 BULLETIN-: MUSEUM OF COMPARATIVE ZOOLOGY.

Cays and of Tjra Bank. The distance of these reefs from the shore
varies from one to fifteen miles.

The small banks from which the Pearl Cays (Hydrographic Chart
No. 392j Admiralty Chart No. 1503) rise, which themselves rise from
the Mosquito Bank at depths varying from eight to twelve fathoms, are
fringed and covered by irregularly shaped reef patches, generally placed
on the easitern and northern sides of the banks adjoining the cays or
connecting them. Long Eeef Bank has a continuous reef on its eastern
edge rising from a depth of three to four fathoms.

Great Corn Island (Hydrographic Chart No. 392, Admiralty Chart No.
1476), twenty-nine miles east of the Bluefield River, rises to a height
of three hundred and seventy feet. It is probably of volcanic origin. The
northern and northeastern shores of the island are protected by an ellip-
tical coral reef forming a great curve at a distance of about a mile from
the shore, with numerous heads and patches between the outer reef and the
shore. Off the island the bank has a depth of from thirteen to twenty
fathoms, the reef rising from a depth of four to six fathoms. Off Little
Corn Island there is a similar reef, about a quarter to half a mile from
the shore, with two to three fathoms upon it, which skirts the eastern
and northern face of the island. Great Corn and Little Corn Islands are
seventeen or eighteen miles from the 100 fathom line, the bank sloping
seaward very gradually from twenty or thirty fathoms.

Upon the eastern half of the Mosquito Bank, between the Corn Isl-
ands and the cluster of reefs and banks to the south of Edinburgh Reef,
there are a few isolated patches of coral in from six to ten fathoms.
The Mosquito Shoals and the Morrison Cay Bank are the centre of an
extensive area of rocky heads and of coral patches south of Edinburgh
Channel. Both banks are skirted by an irregular undulating line of
reefs. Edinburgh Reef is crescent-shaped, convex to the east, rising
from thirteen fathoms, with fourteen fathoms to the westward of the reef
(Hydrographic Chart No. 94.5). North of Edinburgh Reef there is also
a cluster of small isolated coral patches and banks, of which the most
prominent are the Main Cape Bank, the Half-Moon Reef and Cays, and
the Alargate Bank.

The Mosquito Bank extends as a narrow belt varying from four to forty
miles along the north shore of Honduras from Patook Point to the Gulf
of Honduras. That part of the bank is free of shoals, with the exception
of a few patches to the east of Cape Honduras. On the outer edge of the
bank, to the eastward of Truxillo, twenty miles from the coast, is Utilla
Island. Utilla with the cays and islets at its western end may be con-

AGASSIZ: BAHAMAS. 165

sidered as the remnant of the former extension of the north coast of
Honduras : it is remarkably clear of coral patches. With the exception
of a narrow ridge rising to a height of two hundred and ninety feet at
the eastward, the greater part of the island is low and swampy. To the
northeast of Utilla are the islands of Roatan and Bonacca, both within
the 500 fathom line. Bonacca (Hydrographic Chart No. 396, Admi-
ralty Chart No. 1718) reaches to a height of over twelve hundred feet :
it rises from the 5 fathom line upon the eastern end of an elongated
bank formed by the 100 fathom line. It is surrounded by clusters
of coral bars and cays, which are most abundant on the southeast
face of the island. Eoatan (Hydrographic Chart No. 394) has an average
height of three to five hundred feet : its highest point is nearly eight
hundred feet. Its northern shore and that of the eastern islets are all
skirted by coral reefs, which sweep round the eastern face of the Roatan
Bank and extend in irregular stretches along the southeast coast, where
the shore is steep to and the reef disappears, the 100 fathom line running
close to the western spit of the island.

The northern part of the Mosquito Bank (the 100 fathom line), which
extends nearly one hundred and forty miles to the east of Cape Gracias
a Dios, is remarkably free from coral patches and banks. North of Half-
Moon Reef there are only insignificant clusters of coral reefs and bars,
forming the Caxones and Vivorilla Cays, to the eastward of which there
are a small reef and some insignificant cays designated as the Gorda
Bank.

The submarine scenery of some parts of the West Indies must be most
striking. We have first the long line of steep eastern slopes of the Ba-
hamas, with the valleys separating them, which rise abruptly from a great
submarine plain, extending far to the eastward and constituting the floor
of the ocean, from a depth of twenty-five hundred fathoms. This plain
extends all the way from the Little Bahama Bank to Navidad, a distance
of over seven hundred miles. Then follows the long, unbroken submarine
slope extending off Porto Rico to the eastward of the Anegada Channel
for a distance of over four hundred miles, with the remarkable sink of
more than forty-five hundred fathoms in depth north of Porto Rico, one
of the greatest depths as yet discovered in any ocean. Then the vftlley,
over twelve thousand feet deep, which extends south of the smaller east-
ern Bahama banks, gradually lessening in depth toward the Old Bahama
Channel and flanked on the southern face by the mountains of San
Domingo. And finally, perhaps the most striking topographic feature
of the West Indian area, the deep valley skirting the southern shore of

166 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Cuba and extendiug eastward as far as Misteriosa Bank and into the
Gulf of Honduras. This valley reaches to Swan Island ; it is flanked on
the south by the gentler slope of the eastern end of Jamaica, and the
steeper lines of the western end of that island ; on the northern edge rise
the higli mountains of the southern coast of Cuba, some of the peaks of
which attain a height of over eight thousand feet. This range is con-
tinued to the westward of Cape Cruz, and its summits are the Caymans
and the shoals of Misteriosa Bank. This valley is known as Bartlett's
Deep : it is in many places more than three thousand fathoms in depth.
The folds, islands, and banks which give the hydrography of the north-
ern, western, and eastern boundaries of the Caribbean so varied an ap-
pearance are in striking contrast to the great sea-sunken plains which
form the basin of the Caribbean and of the Gulf of Mexico.

The steep slopes off the southern coast of Cuba due to the elevation
of that part of the island, as well as the slopes off the northern and south-
western shores of San Domingo, are fully as steep as aiiy of the sea faces
of the Bahama banks. If in the one case, where the coral limestones
are of limited extent, we find no difficulty in interpreting the origin of
this slope, we should not in the other, where the reef limestones merely
form a capping to submarine folds which owe their origin to the same
causes, attempt to explain its existence in any other manner.

THE BOTTOM AND ROCKS ON THE BAHAMA BANKS.

The Bottom.

The nature of the bottom ^ on the banks varies greatly according to the
locality. In proximity to the shores of the islands, where the sea is con-
stantly wearing away and disintegrating the seolian rocks, the bottom
consists of seolian sand, varying in size according to the character of the
adjacent rocks from coarse composite grains to the finest, almost impal-
pable powder. As we pass into deeper water,^ away from the influences

1 Judi^ing from tlie specimens of the bottom brought up in the dredge, the action
of the sea has still considerable force at a depth of nearly fifteen fathoms.

2 By an unfortunate oversight, the coral ooze which extends only a comparatively
short distance from the 100 fathom line along the line of the Florida Reefs and of
the coast of Cuba is marked on the colored map of bottoms of the " Three Cruises of
the Blake" (Fig. 191) as extending all the way across the Straits of Florida. The
central part of tiie channel of tlie Gulf Stream should be colored to indicate the ex-
tension of tlie modified Pteropod and Globigerina ooze which covers so great a part of

AGASSIZ: BAHAMAS. 167

of the shore sand Kept in more or less constant motion, the grains of
aeoliau sand become mixed with fragments of corallines, or of Inverte-
brates which have lived upon its surface, and where corallines occur in
great abundance the flat fragments of their stems often constitute the
greater mass of the bottom samples. In the vicinity of coral heads the
bottom is composed of aeolian sand of different degrees of coarseness,
more or less mixed with comminuted fragments of Gorgonians, of Mille-
pores, and of other corals. On such parts of the banks there also occur
a very considerable number of Orbitolites and other Foraminifera.. Of
course, in the proximity of the edging i-eefs and along their faces frag-
ments of corals, of corallines, of Nullipores, of alga^, and of Invertebrates
are an important component of the bottom, the pieces of coral and of
corallines often constituting by far the greater mass of the sand.

The beaches are, according to their locality, composed either of very
fine coral sand, or of coarser coral sand mixed with corallines and aeolian
sand, or of large pieces of corals and of other Invertebrates which become
cemented together to form a breccia or a conglomerate, so that we may
have bottoms composed of particles varying from an impalpable powder
to coarse fragments, the components of which can readily be recognized.
When a considerable amount of vegetable matter has become mixed with
the seolian or coralline sand, the bottom becomes more or less sticky and
marl-like, and assumes a grayish color. This is the character of a great
part of the bottom on the interior of the Little Bahama Bank ; while on
the eastern shank of tlie Great Bahama Bank, and on the smaller banks
extending to the eastward of it, the bottom varies greatly according to
the local conditions ; on the western shank of the Great Bahama Bank a
large area off the west coast of Andros is covered with the " white marl"
already described.

The principal coralline algse are species of Penicillus, Acetabularia,
Halimeda, and Udotea, fragments of which can readily be recognized as
components of the bottom specimens wherever they grow in numbers.
The decaying of other algae which compose the flora of the bottom merely
adds more or less vegetable matter to its calcareous components ; the
most abundant of these algae are species of Thalassia, of Padina, Cau-

the Gulf of Mexico and the Caribbean Sea, and which also extends over the bottom
of the wider passages between the Bahamas. The so called coral ooze is in reality a
calcareous mud formed of decomposed corals and corallines, consisting sometimes of
very fine particles, which often accumulate along the shore line in sheltered localities,
or are carried to considerable distance when held in suspension in the water along
the line of currents.

168 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

lerpa, Laurencia, Blodgettia, Liguria, and Sargassnm. Professor Farlow,
who examined the algse collected, notes as an interesting acquisition from
the Bahamas specimens of Caulerpa Webbiana, thus far known only from
the Canaries.

Dr. T. W. Richards was kind enougn to have some analyses made of
bottom specimens selected with regard to their locality on the banks:
one specimen off the west shore of Eleuthera from two fathoms of water
to represent the bottom on the interior edge of the bank ; one on the
middle of Salt Cay Bank, half-way between Elbow Cay and Anguila from
three fathoms, and one from Exuma Harbor in two fathoms, to represent
the bottom on the outer weather edge of the bank ; one specimen from
Bernini, in two fathoms of water, to represent the bottom on the lee
side ; and a sample of the white marl ofl' Billy Island, in one and a half
fathoms.

The analyses, made by Mr. Donald Churchill, show a variation of the
percentage of CaO between 52.38 and 53.43, with the exception of the
sample from the west coast of Eleuthera, in which the percentage is only
45.94. This bottom specimen was composed in great part of fragments
of Orbitolites, with comparatively little aiolian sand, while the other sam-
ples varied from the impalpable powder of which the white marl is com-
posed to finer and coarser seolian sand, and to a bottom consisting of
fragments of broken shells of NuUi pores, corallines, Millepores, corals,
Gorgonians, and coarse seolian sand.

The percentage of CaO (average of two analyses) in the specimens
of bottom from off the west side of Eleuthera, in two fathoms, is 45.95 ;
from the outer edge of the bank, Exuma Harbor, in two fathoms, 53.38 ;
from the western edge of the bank, North Bemini, in two fathoms,
52.46 ; half-way across Salt Cay Bank, in three fathoms, 52.39.

A qualitative examination of the white marl showed a very large quan-
tity of CaCOs, a little magnesic oxide, a little ferric oxide, and a very little
silica and organic matter. Five grams of the powdered CaCOg bottom lost
7± milligrams upon being washed with 200 c.c. of water. Five grams
of the same were digested with 250 c.c. of sea water (from the Bahamas)
for a week. Upon washing upon a Gooch crucible with 200 ± c.c. of
water, the CaCOg was found to have lost 3± milligrams. The carbonate
was tested qualitatively for sulphuric acid, which was found in small quan-
tities. A saturated solution of gypsum in water at 20° contains about one
gram of CaO per litre. Nearly four times as much dissolves in a saturated
salt solution. Dr. Richards hopes to make some additional exj)erimental
trials with reference to the solubility of gypsum in sea water.

AGASSIZ: BAHAMAS. 169

The amount of CaO in sea water from the Bahamas appears to be quite
constant, as far as shown by the analyses kindly made for me by Dr.
Richards. Water taken from the Tongue of the Ocean, No. 1, or from
the central part of the Great Bahama Bank half-way between Green Cay
and Exuma Sound, No. 2, or close to Exuma Sound on the eastern edge
of the bank from the west side of Elamingo Cay, No. 3, or near the
western edge of the bank south of New Providence, No. 4, showed only
slight variations in the composition of the sea water. The mean of two
measurements of 1,000 c.c. of sea water at 4° were as follows : No. 1
contained 0.585 gram CaO; No. 2, 0.585 gram; No. 3, 0.580 gram;
No. 4, 0.588 gram.

The percentage of carbonate of lime present in the bottoms surround-
ing coral reefs depends in part upon the presence of magnesia or upon
the admixture of decomposed vegetable or animal matter, and in part
upon the amount of moisture. In districts where the coral reefs occur
on shores consisting of stratified or volcanic rocks, the coral sand is also
more or less mixed with particles derived from the rocks composing the
basement or abutments of the reefs. In volcanic regions the limestones
are often greatly modified, either by heat or by their mixture with vol-
canic ash. The same action was also effective in modifying the reef
limestones of the dolomitic region of Tyrol. We also find in volcanic
regions a proportionally larger percentage of magnesia.

A number of specimens of coral, of coral rock, and of coral beach sand,
from different localities, have been kindly analyzed for me by Prof. F. W.
Clark. The greatest variations in the percentage of carbonate of lime
are found in the specimens from the volcanic region of the Sandwich
Islands, and in the deposits of Sombrero (West Indies). The coralline
bottoms and coral sands from the Sandwich Islands also contained a
lai'ger percentage of MgO than any similar bottoms from the Florida
Reefs or Bahama Banks. The elevated limestones of Cuba from the
Yunque are not dolomitic, and show no trace of magnesia. In the coral
sand mixed with volcanic ash the amount of magnesia in one case rose
to over eleven per cent.

Piece of Siderastrea

Verj' coarse sand, coral fragments, Tortugas

Finer sand,

Tortugas, fairly coarse sand

" finer sand

Kej' West, very fine beach coral sand
Loggerhead Key (stratified shore coral rock)

CaO

COj

MgO

Moisture

55.16

43.74

0.20

0.54

51.52

41.58

2.08

3.18

52.24

41.46

1.53

3.27

51.24

41.07

2.09

3.57

51.95

41.53

1.44

3.07

49.38

40.39

1.93

5.12

53.54

43.71

0.78

0.81

170

BULLETIN: MUSEUM OF COMPAEATIVE ZOOLOGY.

Salt Key Bank, oolite

Sombrero ileposits (P2O5 39.12)

Coralline Bottom, Tortugas

Senote limestone near Merida

Diamond Head, Sandwich Islands, coral rock

Kohuku Bluff,

Kohuku Beach, " " sand

Honolulu Eeef, coral rock

Prison Knob, Honolulu (elevated coral)

Waianea (hard ringing limestone)

Modern chalk, Oahu 1

( SiOa 33.25, Al.Pg 19.13, Fe.^ 10-71 ) f
Lava soil bedded with coral near Pearl River )

(SiO. 46.22, ALOg 19.16, FeA 12.94) \
Same from Diamond Head I

(SiO.. 32.83, AI2O3 12.12, FeaOg 11.52) J

CaO

53.54
51.15
46.45
54.87
44.82
51.09
49.34
46.52
52.67
50.69

11.37

1.72

12.20

CO2

43.87
3.47
43.40
43.89
40.81
43.64 1
44.33
40.59
42.81
43.96

MgO

0.71

trace

5.15

0.64

5.32

2,50

4.60

2.45

0.42

2.98

11.09 —

Moisture
1.13
7.55
2.73
0.11
1.86
0.79"
0.40
2.75
1.24
0.46

9.84

0.00 2.24 16.84
11.41 11.70 5.30

The Rocks of the Bahamas.

The characteristics of the seoliaii rocks of the Bahamas have been noted
in the description of the different islands of the group. They vary from a
fine-grained compact ooUte to a coarse-grained rock composed either of
rounded or of small angular fragments, according to the condition of the
fragments of the sand from which the Eeolian hills were composed. These
rocks vary according to the conditions to which they have been subjected,
and which often greatly alter their aspect. They become more or less
hardened or covered with a hard ringing crust when exposed to the action
of the sea or of the rains, or to the decomposing effects of the acids gen-
erated by the decomposition of vegetable matter. Sufficient time has not
elapsed since my return from the Bahamas and Bermudas to report
more in detail regarding the composition of the rocks and bottoms of
these regions. But I hope to return to this subject, and present the
results of the examination of the rocks now making by Professor Wolff.

SOME RECENT VIEWS ON THE THEORY OF THE
FORMATION OF CORAL REEFS.

It may not be out of place to pass rapidly in review some of the points
relating to the theory of the formation of coral reefs which have been
brought forward in the discussion of the subject in the new issue of Dar-

^ Trace of organic matter.

AGASSIZ: BAHAMAS. 171

win's Coral Reefs, edited by Professor Bonney ; ^ in the interesting review
by Dana of the whole subject in the American Journal of Science (1885,
Vol. XXX. pp. 89, 1G9) and in the new edition of Dana's Corals and Coral
Islands (Now York, 1890), with the Appendix on the Sandwich Islands ;^
and in tlie essay by Dr. Langenbeck ^ on the Theories of the Origin of
Coral Islands and Coral Reefs.

In this chapter I shall bring together and discuss only such questions
as have naturally been suggested by my explorations of the Bahamas,
the Florida Reefs, the Cuban elevated reefs, and the reefs of the Wind-
ward Islands. Minor points will be found referred to in otlier places of
this report, and in the articles of the Duke of Argyll and of Professor
Huxley, as well as in the more recent communications in " Nature" and
elsewhere by Murray, J. C. and G. C. Ross, Irvine, Harrison and Browne,
Bourne, Guppy, Hickson, Captain Wharton, Studer, Ortmann, Dall, Bal-
four, H. 0. Forbes, Supan, Perrier, Heilprin, Captain Moore, Sluiter,
Walther, Lendenfeld,- Kent,* and many others, as well as tlie Presidential
Addresses of Sir Archibald Geikie and of Px'ofessor James Geikie at Edin-
burgh in 1883 and 1892.

While it is undoubtedly true, as Professor Bonney states, that many
of Darwin's critics may have perused his book with overmuch haste, yet
he must admit that when explorers examined the coral reefs of districts
which apparently had nothing in common with coral reefs formed in
areas of subsidence, it was most natural that they should seek some other

1 The Structure and Distribution of Coral Eeefs, by Charles Darwin. Third
Edition. With an Appendix by Prof. T. G. Bonney. London, 1889.

2 See also A. Agassiz, Bull. Mus. Comp. Zool., Vol. XVII. No. 3, p. 121, 1889,
"The Coral Reefs of the Hawaian Islands."

3 Die Theorien iiber die Entstehung der Koralleninseln und Korallenriffe und
Hire Bedeutung fiir Genphysischen Fragen. Von Dr. R. Langenbeck. Leipzig, 1890.

* Kent, in his Monograph of the Great Barrier Reef of Australia, has added com-
paratively little to our knowledge of the theory of its formation. " He looks for the
conditions of subsidence which have made the formation of the Great Barrier Eeef
possible in the former undoubted connection of Australia with Tasmania and New
Guinea ; and if that is not satisfactory, he is quite ready to call upon a still greater
subsidence of the Australian continent as shown by its presumed connection with
New Zealand. If, as is probable, and as Mr. Kent suggests, the Great Barrier Reef
existed as a narrow fringing reef in the late Tertiary, there has elapsed more than
ample time also for its transformation into the Great Barrier Reef of to-day from
other causes than those called upon by him. The Great Barrier Reef has entirely
obliterated the Australian coast shelf itself, and it may have found upon that all the
conditions of depth for the vigorous growth, both vertically and laterally, of the
original insignificant fringing reef of the northeastern coast of Australia." — Nation,
No. 1463, July, 1893.

172 bulletin: museum of compaeative zoology.

explanation of their formation than that suggested by Darwin, and thus
call upon factors which he had relegated to a subordinate value. The
want of unanimity among the opponents of the Darwinian theory of the
formation of coral reefs has, it seems to me, nothing to do with the ques-
tion whether Darwin is right or not. It merely shows that. there are
many points of difference which are more or less apparent in the various
districts examined by recent writers on coral reefs, and that perhaps the
problem of their formation is not quite so simple as the upholders of the
subsidence theory would have us believe.

The disadvantages of a discussion of the theory of the formation of
coral reefs by one who has not faced the problem in the field are appar-
ent in the essay of Dr. R. Langenbeck. It undoubtedly is a most useful
summary, and, while the writer has endeavored to retain an impartial
attitude, his views naturally do not carry much weight, based as they
are only upon cartographic knowledge, and cunsisting in great part of
special pleading in favor of the theory of subsidence. His essay does
not seem to me to have the importance assigned to it by Huxley in his
discussion of the article on "A Conspiracy of Silence," ■^ by the Duke of
Argyll. This and similar essays can give us only interesting speculations,
of little value except when tested by observations.

Dr. Langenbeck has been unfortunate in the exposition which he gives
of the Florida and West India reefs. His description of the reefs of Cuba
and of the Bahamas differs greatly from that given by the writers from
whom he" has taken his data. But he is the first writer on the theory of
reefs who, while of the opinion that the theory of Darwin is the only one
which can explain the formation of atolls in the extensive tracts of the
Pacific where atolls are so numerous, yet acknowledges that there are
extensive districts in which the formation of reefs has not been influ-
enced by subsidence. This was the principal point at issue between the
earlier opponents of the theory of subsidence. Subsequent investiga-
tions in the Pacific and elsewhere have shown most clearly that the reef

1 I am iinabletounderstand the statement of Professor Heilprin (The Bermudas,
p. 21) regarding the authority of " A Conspiracy of Silence." It is most extraordi-
nary if, as is stated by Heilprin, the " younger school of geologists " have pinned their
faith to the statements of its author and have chosen for their leader one who has
never made any original observations on the subject of coral reefs. The great ma-
jority of recent observers of coral reefs have stated that the generally accepted
theory of their formation was not applicable to the particular district they exam-
ined. Each has endeavored to give an explanation of the facts, and they have
not " pinned their faith to any standard bearer," or to any theory, or to any leader,
or to any creed.

AGASSIZ: BAHAMAS. 173

problem is a complicated one, and that we have not heard the last word
on the subject. As far as the formation of many of the Pacific atolls is
concerned, either they have been formed by subsidence, and owe their
origin to causes different from those which in the West Indian and
other districts have formed atolls upon comparatively shallow bases, or
they have been formed by the same causes to which the latter owe their
shape independently of subsidence, or we may further assume that atolls
can be formed in regions of subsidence independently of the subsidence,
as well as in regions where no subsidence has been observed.

Darwin, and after him Langenbeck, as well as other writers on the West
Indian coral reefs, speak of the " extensive banks of sediment which have
been heaped up along the whole north coast of Cuba by sea currents." I
find it difficult to account for the origin of such a belief. That a great part
of the food supply of the animals living upon the banks has been brought
to them by sea currents is self-evident, but the mass, of sediment so con-
veyed is insignificant as compared with the accretions to the banks due
to the accunmlation and cementing of the carcasses of the Invertebrates
living upon them. There is, it is true, a long stretch of sand reefs and
islands all the way north from Key Biscayne to beyond Cape Hatteras,
but these have nothing in common with the limestone barriers forming
the basement of the Florida, Cuban, and other West Indian (Mexican
and Caribbean) coral reefs.

I am surprised that Professors Dana and Bonney should both speak
so strongly against the statement of Murray, that there are known
numerous submarine elevations, mountains rising from the general level
of the ocean bed, which nfiay serve as the foundation of a coral reef.
Surely, we know as yet too little of the hydrography of the Pacific to
sweep away the statement with the suggestion " that such a thing may
occasionally occur."

Professor Dana, who has made such a thorough examination of the
Feejee Islands, considers the conditions of the coral reefs of the islands
of the group as admirably illustrating Darwin's theory of the formation
of barrier reefs and of atolls. In spite of the general aspect of a sunken
continent made by the Feejee Archipelago, it is no proof that the reefs
owe their present condition to subsidence. It seems to me that the illus-
trations he brings up merely emphasize the great diversity of substruc-
ture of banks and islets and of large islands in the Feejee Islands
upon which fringing, barrier reefs or atolls might be found ; just as
in the Windward Islands and the Virgin Island Bank we find all pos-
sible conditions of elevation, from islets at the sea level, or sunken banks,

174 BULLETIN : MUSEUM OF COMPAuAilVE ZOOLOGY.

to islands with peaks ol over lour thousand feet, at the base of and
around which barrier and fringing reefs may be formed, or crescent and
horseshoe-shaped reefs might grow.

The somewhat insignificant coral reefs of the Eastern Pacitic, as far as
known, occur upon volcanic islands. Those of the Sandwich Islands and
the unimportant coral patches of the Galapagos grow upon a basement of
volcanic origin, while the only atoll of the Eastern Pacific — Clipperton
Rock, off the Mexican coast — holds to its trachytic nucleus much the
same relation which some of the reefs of the Caribbean hold to the vol-
canic peak they surround.^

1 The absence of extensive coral reefs off the west coast of Tropical America
is probably due to several causes, the most important of which as regards the main-
land is the large amount of mud brouglit down the shore mountain slopes during
the rainy season. Similar conditions exist along the shores of tlie Galapagos, where
corals are formed merely in patches, as along the coast of the mainland, in addition
to whatever influences detrimental to their growth may be due to volcanic action.
(See Dana, Corals and Coral Islands; Pourtales, Am. Journ. Sci., 1875, Vol. X.
p. 282; and A. Agassiz, Bull. Mus. Comp. Zool., Vol. XXIII. No. 1, 1892.)

It is interesting to add here some notes relating to Clipperton Rock, which prove
the existence of an atoll on the borders of the Panamic region, seven hundred miles
southwest of Acapulco. I am indebted to Professor George Davidson of the U. S.
Coast Survey for a sketch of Clipperton Island by Carl C. Jensen, made in 1893, as
well as for a specimen of the rocky nucleus. The latter was examined by Profes-
sor J. Eliot Wolff, who reports it to be " k volcanic rock with typical trachytic struc-
ture, of the same general class as the rock from the hospital quarry at Panama.'"

The description drawn up from the notes and chart he kindly sent me agrees in
the main with the one given by Lieutenant Griswold.and published by W. H. Pease
in the Proceedings of the California Academy of Natural Sciences (1865, p. 199).

Clipperton Rock was seen from the northeast by Sir Edward Belcher in 1839,
and described by him as a coral lagoon island with two entrances, both on the
weather side. It was, according to W. H. Pease, visited in 1861 by Lieutenant GriS-
wold, who landed upon the island, and says in his Journal : " It is girdled by a
broad barrier of coral about fifteen feet above the level of the ocean. There is no
entrance to the lag6on. The water inside is fresh and potable ; it is slightly brack-
ish. . . . The rock is a ragged pile of volcanic formation, on the south end of the
island. . . . The highest pinnacle of the rock is about 120 feet, and it covers about
two acres. It stands on the edge of the lagoon, or rather within it. . . . The lagoon
is a quiet fresh-water pond about two miles long and one broad, with a long spit
of mud running out into the middle of it, but elsewhere of a light green color which
seemed to indicate a considerable depth. [Light green does not in coral districts
indicate a considerable depth ; on the contrary, it may indicate a deptli of six to
seven fathoms at most. — A. Agassiz.] The shores are abrupt. ... I did not find
the least sign of vegetable life upon the island."

Mr. Jensen says there are no trees on the island. It is surrounded by a reef of
coral over which the sea continually breaks. Outside of this unbroken reef the
bottom deepens rapidly, and vessels must anchor in not less than fifty fathoms of

AGASSIZ: BAHAMAS.

175

In the West Indies_, beginning with the Bermudas, is the mass upon
which the so called atoll has been formed, — an island rising out of a

10° n'

CLIPPERTON ISLAND.

water. Best boat landing on the north side of the island. On the chart " Deep
water" is marked in the atoll to the westward of the trachytic rock near the south-
ern face of the island. " Great coral ridges alternating with deep water " is the
legend of the central part of the lagoon. A few insignificant islets are mapped in
the lagoon near its northwest face. The atoll is pear shaped, its axis running south-
east to northwest ; its greatest breadth is over one and a half miles, and its length
more than two.

On Mr. Jensen's chart the island is represented as a narrow ring of coral rock
surrounding the lagoon. This dry belt is nowhere more than seven hundred feet
wide, except to the southwest of the trachytic rock, wiiere tliere is a short spit ex-
tending into the lagoon, the point of which is perhaps fifteen hundred feet from the
outer edge of the island. At several points the coral rock belt is not more than two
hundred feet wide, and along the greater part of the southwest coast it varies be-
tween four and five hundred feet in width. A similar narrow stretch of perhaps
three quarters of a mile extends along the north side. Tiiis connects the two
broadei shore tracts of the eastern and northwestern parts of the island.

176 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

depth of two thousand fathoms. Has that atoll been formed by its sub-
sidence'? If so, we should have a great oceanic depression in the vicinity
of the Bermudas. We know nothing of the hydrogr&phy which does not
indicate that the mass underlying the coral reefs rises abruptly from the
ocean floor ; as to what height it originally rose, we have no data.

Examine next the submarine range of the Windward Islands, with its
varied depths, of which the islands beginning at the Grenadines and
ending with Saba are the exposed summits. Note the gradual widening
of the bank as we go north, and finally reach its widest part, upon which
are situated islands of recent limestone which have risen from a compara-
tively shallow bank, while farther south the recent limestones are limited
to the Basse Terre of Guadeloupe, or to the fringing reefs of Martinique,
St. Vincent, Dominica, and the Grenadines.

Isolated from the other islands in the line of the submerged chain, we
have Sombrero, a recent limestone island, while Bird Island and the
Barbados are the summits of two isolated banks, both of different
heights. Surely, we have in that submarine chain a sufficient variety
in the depths of the bank to account for the irregular position of the
limestone deposits now going on, and a sufficient number of points
which have reached by elevation the depth necessary and favorable for
the growth of coral reefs. Some of the localities have been elevated
above that level, as Barbuda, Antigua, Anguilla, Barbados, and Basse
Terre, while at other points — as Martinique, St. Vincent, and the Gren-
adines — the fringing and barrier reefs are still in full activity.

The Saba Bank, the bank upon which have risen Antigua and Bar-
buda, and the bank within which are found Anguilla and St. Barthol-
omew, are extensive shoals compared with the dry land of the island.
Upon the Mosquito Bank also numerous coral reefs are found, rising
from varying depths, similar to those of the Yucatan Bank.

Between the Mosquito Bank and Jamaica are a number of banks
within the 100 fathom line, some of which, like those surrounding St.
Andrews or Old Providence and the Morant Cays, have come to the
surface ; while others again, like the Serrana, Rosalind, Serranilla, and
Pedro Banks, have not reached a depth at which corals can begin to
grow, or have sunk below it. This gives in such a limited district as
the Caribbean Sea proper a sufficient number of instances of subma-
rine banks reaching all possible levels, some of which are in local-
ities where corals have developed to form fringing or barrier reefs,
others where the corals are reduced to irregular patches, and others
where, the conditions for their growth not being favorable, the banks are

AGASSIZ : BAHAMAS. 177

free from corals, their surface being more or less disturbed by the action
of the trades, thus choking off the growth of any delicate marine forms.
This leaves out of notice the great number of banks of the Bahamas,
that of the west side of Florida, of the banks extending from Cape Sable
to the Tortugas, of the banks to the south of Cienfuegos, extending from
the Isle of Pines to Cape Cruz, and, finally, the bank to the eastward of
Porto Rico upon which rise the Virgin Islands. Corals flourish upon all
these banks ; they form fringing and barrier reefs, and in a few instances
atolls, such as Alacran, the Hogsty Reef, and the atolls of the Mosquito
Bank, which I have not, however, examined myself. We cannot assume
that corals are only to grow where the base has reached the proper
depth during a period of subsidence, while they are denied that privilege
where the right depth has been reached in a period of elevation.

There is no greater distance between the elevated reef of the south
shore of Cuba and the fringing reef of parts of San Domingo and Ja-
maica than thei-e is between those of the north shore of Cuba and the
barrier reef of the Florida Keys. The time at which the reefs of differ-
ent areas in the West Indies were elevated need not necessarily have
been synchronous, nor should we expect this in a district where volcanic
agencies have been at work on so great a scale. This would go far 'to
account for the great diversity of coral reefs we meet with in the West
Indian district within comparatively short distances.

In fact, what I have seen so far in my explorations of the coral
reefs of the West Indies would show that wherever coral reefs occur,
and of whatever shape, they form only a comparatively thin growth
upon the underlying base, and are not of great thickness. In Flor-
ida they rest upon the limestones which form the basis of the great
peninsula. On the Yucatan Bank they are underlain by a marine
limestone. In Cuba they abut upon the Tertiary limestones of its
shores. Along Honduras, the Mosquito Coast, and the north shore of
South America, they grow upon extensive banks or shoals, parts of the
shore plateau of the adjoining continent, where they find the proper
depth.

Similarly, along the shores of the Greater Antilles they are found rising
from depths suitable to their growth upon the shore plateaus, as well
as upon the peaks and summits of submarine ranges which either
come to the surface or approach it. Round the Windward Islands
they thrive upon the banks which surround them. But in neither case
have they played any important part in the building up of the banks
or of the shore plateaus upon which they occur. Upon the Bahamas

VOL. XXVI— NO. 1. 12

178 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

and Bermudas the corals now growing form a thin veneer upon tlie a^olian
rock ledges, the remnants of the vanishing land formed from a coral
reef which must have flourished at the beginning of the present epoch.

But Dana himself has even more distinctly and emphatically than
any other writer stated the objections to the general application of the
theory of subsidence for the formation of bai-rier reefs. In " Corals
and Coral Islands," he says (p. 347) : " The amount of subsidence deter-
mines in some cases the distance of barrier reefs from shore ; but it by
no means accounts for the diflference in their extent in different parts of
a single group of islands. Indeed, if this cause be considered alone,
every grade of extent, from no subsidence to the largest amount, might
in many instances be proved as having occurred on a single island. Of
far greater importance, as has appeared, is the volcanic character of the
land." The remainder of the paragraph quoted above, describing the
relations of volcanoes and of zoophytes as the land-making agents of
the Pacific, seems hardly consistent with the subsidence theory.

Many of the observations made on elevated coral reefs do not distin-
guish between the elevation of a tjinn mantle on a slope to a certain
height, and the thickness of the reef itself. Furthermore, in such a dif-
ficnlt problem only the observations of trained observers can be of value.
Many of the data, and even many of those quoted by Dana, are the
records of merely casual observations of individuals with little or no
experience in tliis field.

The island of Lafu, one of the Loyalty group, the geology of which
has been described by the Rev. W. B. Clarke,^ is one of the few elevated
coral islands of which we may feel certain that the highest points, two
hundred and fifty feet above the sea, are covered by an elevated coral
reef, the elevation being connected in that case, not with volcanic action,
but with the elevation of the adjoining island of New Caledonia. Sub-
sequent investigations of the Loyalty group by Balansa^ and Cham-
beyron have shown that other islands of the group are covered by
elevated coral reefs. Ouvea attains a height of a little over fifty feet,
with an interior lagoon ; and while Lafu shows three terraces, the neigh-
boring island of Mari lias five distinct terraces, attaining a height of
one hundred meters, with a nucleus of trachyte.

From the interesting accounts given by Captain Chambeyron, in his
" Note relative a la Nouvelle Caledonie," ^ we can see how much remains

1 Proc. Geol. Soc. London, 1846, p. 61.

2 Bull, fie la Soc, de Geo^i., Vol. V. p. 521, 1873.

3 Ibid., Vol. IX. p. 566, 1875.

AGASSIZ : BAHAMAS. 179

to be learned regarding the barrier and fringing reefs and the great
lagoon of New Caledonia. We need additional information regarding
the thickness of coral reefs to complete the evidence of their great
thickness as due to subsidence. The thickness of such elevated reefs
as may without question be considered true coral reefs is not as great
as it has been supposed by various observers (see Dana, p. 274). The
more recent examinations of the thickness of the elevated coral reefs in
the Solomon Islands by Guppy, in the Barbados by Harrison and Jukes-
Browne, in Cuba by R. T. HilP and myself, all tend to prove that ele-
vated coral reefs nowhere have attained a tliickness of more than two
hundred to two hundred and fifty feet. This shows that a thickness of
two thousand feet observed in Cuba by Professor 0. W. Crosby, of twelve
hundred feet near Matanzas reported by myself, and of two thousand feet
of the white limestones of Jamaica, are not the thickness of coral reef
builders, but of marine limestones of very different character ; and that
the true reef-building corals have merely built a sheet of moderate
thickness upon these older limestones, which have been mistaken for
limestones built up by reef-building corals. The reefs of the weather side
of the island of Oahu form quite a thin sheet overlying the volcanic rocks.

Similar conclusions have been reached by the recent examinations of
dolomitic limestones, which are of great thickness (over two thousand
feet), and were until recently unhesitatingly considered as coral reef
limestones. They have now been shown by Miss Ogilvie and by Rothpletz
to be marine limestones, while the true coral reef builders formed de-
posits of very moderate thickness, a coral growth, as it were, not thicker
than one hundred and fifty feet. It is very evident that, in the exam-
ination of the coral deposits of former geological periods, great care must
be taken to distinguish between coral growths forming a comparatively
thin veneer and marine limestones of great thickness in which corals are
occasionally found.

The discussion of Lana:enbeck on the coral reefs of former e'eolog'ical
periods is subject to revision after a re-examination of the localities, in
accordance with the recent explorations of the dolomite region, and of
our more accurate knowledge of the distribution and thickness of the
West Indian reefs. The fossil reefs of the Jura were among the earliest
known. Attention was first called to them by Thurman (1859) and
Gressly (I860), but they have not been examined recently with refer-
ence to the points now in dispute regarding the formation of coral reefs.^

1 Amer Journ. of Science, Septemher, 1894.

2 See Zschokke, F. Die KorallenrifEe im Schweizeriscben Jura. Aarau, 1890.

180 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Ill uo case has Tliurman or Gressly observed ancient coral reefs of very
great thickness (one hundred meters) at all to be compared with the
thickness claimed for modern reefs by Darwin and Dana (two thousand
feet).

The evidence brought forward from the existence of recent lime-
stone beds of great thickness in Java and in the Philippines and other
localities, to which reference has been made as proof of the great thick-
ness of coral reefs, must be taken as subject to modification. One
of the most interesting of the fossil coral reefs is the atoll of Ben-
guet, in the island of Luzon, mentioned by Drasche.-^ It has walls of
five to six hundred feet above the valley, which itself is about four
thousand feet above the sea level. The surface of the limestone is filled
with red earth, often several feet in thickness. The limestones, unlike
coral reefs, are distinctly stratified. The corals and other fossils belong
to genera which still occur in the Indian Ocean of to-day, so that the
limestones of the atoll of Benguet and of ISTorthern Luzon belong to the
most recent formations of the island, and from their description differ
greatly from the coral reef limestones of to-day.

Eichthofen's ^ theory that the Schlern Dolomite was a coral reef was
adopted in the main by Mojsisovics,' and their views, although opposed
by Giimbel,^ have received general recognition until comparatively re-
cently. Miss M. M. Ogilvie, in the Geological Magazine for January and
February, 1894, perhaps first called attention to the peculiar occurrence
of the so called coral reefs in the Dolomites.

In the interesting account of coral in the Dolomites of South Tyrol
by Miss Ogilvie an entirely new view is taken of the reef-like appear-
ance assumed by the dolomitic massifs, which she traces to the move-
ment of the rocks in Tertiary times, and to the variation in the character
of contemporaneous Triassic deposits, her conclusions being that, "as
far as positive evidence goes, the coral rocks of South Tyrol in the
Wengen and Cassian period are not the majestic massifs of Dolomite,
but much less obtrusive lenticular masses of limestone," the " Cipit
limestones." ^ Miss Ogilvie has shown in the diagrams I.-IV., on Plate
II., (a) that coralline "Cipit limestone" and coral Dolomite form com-

1 Drasohe, R. von. Fragmente zu einer Geologie der Insel Luzon. Wien, 1878.

2 Geognost. Beschreib. der Umgegend von Predazzo in Siid-Tirol. 1860.

3 Die Dolomit Riffe v. Siid-Tirol u. Venetien. 1879.

4 Sitzb. d. Akad. d, Wiss. Miinchen. 1873.

5 See Plate II., February number of the Geological Magazine for 1894, and the
figure on page 53 of the same number.

AGASSIZ: BAHABJAS, 181

paratively small thicknesses of interbedded rock, and not the main body
of the mountain masses ; (6) that corals began to grow in Wengen time
on a submarine volcanic ridge on the northern edge of a great area of
subsidence, and travelled northward and inward in Cassian time ; (c) that
extensive banks of coral were found in scattered localities during the
Raibl period of shallower water and Dolomite deposit. And the final
paragraph of the paper contains the following statement : " My special
survey in a part of the district seems to me to justify, without doubt,
the position of those authorities who have contended that the immense
thicknesses of ' Schlern Dolomite ' rock were an ordinary marine deposit,
and not 'coral reefs.' "

In the descriptions of the Dolomitic region by Richthofen and by
Mojsisovics, " the thick dolomitic massifs of Schlern Sella," etc., were
regarded as dolomitic reefs, though they have originated as marine de-
posits, and not coral reefs. " The Wengen and Cassian coral reefs of
Groden, Ennenberg, and Upper Fassa have remained, with but rare ex-
ceptions, limestone. The actual thickness attained by the individual
lenticular coral reefs or the coral banks is in no case very great, seldom
more than one hundred and fifty feet, and usually much less."

Finally, Rothpletz,^ after a thorough examination of the field, comes to
the following conclusion : —

" Wir kounen also die zu Beginn dieses Capitels aufgeworfene Frage,
ob der Schlerndolomit eine Korallenriffbildung sei, jetzt mit einem
entschiedenen Nein beantworten, so bald man die Frage in dieser Allge-
meinheit noch stellen will. Die bisher besprochenen reinen Kalk- und
Dolomitmassen der oberen Trias sind allerdings vorwiegend organogenen
Ursprungs und verdanken ihre Massigkeit dem Wachstum von Thieren
und Pflanzen in wenig tiefen oder sogar seichten Meerestheilen. Aber
Korallen haben dabei durchaus keine besonders hervorragende Rolle
gespielt, wenn schon sie stellenweise wahre Rasen bildeten und wohl
auch zu kleinen echten Riff bildungen gefiihrt haben mogen. An den
meisten Orten hingegen treten sie anderen Thieren und den Pflanzen
gegeniiber so senr in den Hintergrund, dass wir sie nur als einen unter-
geordneten Bestandtheil der Fauna bezeichen konnen. Viel bedeutsa-
mer werden sie statt dessen mancherorts in den Cassianer und Raibler
Schichten. Es ist bezeichnend genug, dass der Schlerndolomit des
Schlernes selbst nur sehr wenige Spuren von Korallen aufweist, wahrend
die wenig machtigen Dolomitplatten der -Ruibler Schichten auf dem
Schlernplateau, also unmittelbar im Hangecrten des letzteren, geradezii

1 Ein Geologischer Querschnltt durch die Ost-Alpen, p. 67. Stuttgart, 1894.

182 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

aus Koralleu zusammengesetzt sind, so dass sie als wahre Korallenrasen
bezeichnet werdeu miissen. Die Dolornitisirung des Gesteines hat die-
selben keineswegs verwisclit und unkeuntlich geruacht. Wir konnen
aber auch da nicht von eigentlichen Riften, sonderu iiur vou Raseii
sprecheu, den ihre Hoheneutwickelung ist nur eine sehr geringe. . . .

"Soil nun aber fiir den Schlerndolomit wenigstens der Name 'Eiff '
beibehalten worden, wofiir neuerdings auch Suess ^ sich ausgesprochen
hat] . . . Das wird vou der Definition abhangen, die man dem Worte
geben will. Versteht man darunter steil und hoch vom Meeresboden
aufragende Massen, etwa wie die Koralleninseln des Stillen Oceans, dann
passt der Name fiir die Ost Alpen ganz und gar nicht. Begreift man
darunter aber auch die submarinen, organogenen Plateaus des Golfes
von Mexico und die Dacia-Bauk der Tiefsee, dann liesse sich der Name
' Dolomitriff,' wohl auch fernerhin auf die Alpen anwenden, doch musste
vorher eine entschiedene UmprJigung dieses Gliedes unseres Wortshatzes
vorgenommen werden."

I do not mean in any way to deny the possibility of the formation
of marine limestones by subsidence. But I must also call attention again
to their formation at great depths by accretion, and to the fact that true
coral reef builders have as yet been shown positively to attain only a
moderate thickness.

The difference in the steepness of the slope of the Florida Charts off
the Florida Reef and that of the steep submarine slope of the East-
ern Bahamas off the edging reef, I believe, can be explained on the
supposition that the Eastern Bahamas are a series of more or less
disconnected summits, with slopes no steeper than those of the vol-
canic Windward Islands to the south of them. In the one case, the
Florida reef is the belt edging a gentle slope in depths of from four
to fifteen or sixteen fathoms. In the other, the edging reef of the
Bahamas is a similar belt limited to the same depths on the brink of
a steep slope. The structure of the coral reef and its bathymetrical
range are in neither case affected by the nature of the underlying rocks,
or by the angle of their slopes beyond the bathymetrical range of the
corals.

Professor Dana himself, when speaking of the limited thickness of the
elevated coral reefs of the Solomon Islands described by Dr. Guppy, says
(Corals and Coral Reefs, p. 291) : " Such observations have great interest,
but they only prove that in coral reef seas corals will grow over any
basis of rock that may offer when the water is right in depth and other

1 Antlitz der Erde, Bd. II. p. 331, 1888.

AGASSIZ: BAHAMAS. 183

circumstances favor. . . . They are not evidence against the subsidence
theory, but simply local examples under the general principle just stated."
It seems to me that in the first part of the passage quoted Professor
Dana grants the principal factors which have been advanced in opposition
to the universal application of the Darwinian theory of coral reefs. The
objections to it from this and that quarter have little by little circum-
scribed the area within which it is applicable. It is not applicable to
the West Indian and Caribbean Sea reefs. It does not hold true for
the Bahamas or Bermudas any more than for Cuba or Florida. It
does not hold for the Sandwich Islands or the Galapagos. It has not
been accepted by Semper, by Rein, by Murray, by Bourne, by Guppy,
nor by others who have in recent years examined many of the localities
•which are considered by Dana and Darwin as typical of tb^ subsidence
theory.

Perhaps one of the best sections of coral reef rock and limestone is
that given by Julien,^ as fir back as 1867, in his description of the
geology of Sombrero, W. I. The beds examined by him vary in thick-
ness between thirty-three and sixty-six feet. From his description of
the key as an atoll, I imagine it must resemble either Hogsty Reef or
some of the smaller pseudo-atoll banks of the Bahamas, only it has been
elevated about forty feet. Some of the beds are separated by a thin
layer of limestone differing from the preceding.

The formation of hills of nearly four hundred feet in height, in a dis-
trict showing a probable subsidence of at least two hundred and fifty or
three hundred feet, shows very plainly how a limestone bed of that thick-
ness may have been formed from a coral reef flourishing at the time when
the beach which bordered the seolian islands was at least two hundred feet
higher than the present beaches on the Bahamas. It is not probable that
during the period of subsidence the reef kept up its growth with sufficient
vigor to add to the material which helped to build the seolian hills now
disappearing before the action of the sea. On the contrary, we carf only
imagine such high seolian hills as those of the Bahamas to have been
formed in a period of rest, during which the great beach of the then
existing reef constantly supplied fresh material to be changed by the
surf and the winds into sand for the heaping up of sand dunes. They
could not be formed in a district of subsidence unless the subsidence
was slower than the rate of growth of the corals, which is not the case
in the Bahamas, as the reefs of to-day, even when they come to the sur-

^ On the Geology of the Key of Sombrero, W"T . by Alexis A. Jullen, Ann. Lye.
of Nat; Hist, of New York, Vol. VIII. p. 261.

184 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

face, are not the sources from which the material for the great dunes of
the Bahama Islands is derived.

How far formations similar to those of the Bahamas occur in districts
where there are coral limestones of considerable thickness, and where
their original structure can no longer be detected, is an open question.
There are groups of islands in the coral reef districts of the Pacific which
from their resemblance to the Bahamas it is interesting to compare with
them. I refer specially to the Maldives, to the Louisiade Archipelago,
and to the Keeling Islands. The explanation which Darwin gives of
the disseverment of the larger Maldive atolls, that the smaller ones have
been formed during subsidence by a process of which traces can be de-
tected in the northern part of the group, does not seem natural. The
sunken coral reefs of which we have surveys show no such method of
formation, while, on the contrary, banks of which the depth is some-
what varied, forming more or less irregular knobs or folds, could give a
suitable substructure for the growth of irregularly shaped atollons such
as constitute the Maldive Archipelago and the Mahlos Mahdoo atolls.
We may compare with these atolls the circular reefs of the Mosquito and
Yucatan Banks, which have formed wherever the conditions of depth
and currents and configuration of the bottom have been favorable, in the
midst of other irregularly shaped patches of corals, many of which are of
considerable extent. The simHarity of the eastern and western ranges
of land of the Great Bahama Bank to the distribution of the atollons on
the opposite sides of the Maldive atoll group has been suggested by Pro-
fessor Dana ; but while this is marked, the structure of the islands in one
case and of the atolls in the other is, as I have shown for the Bahamas,
very different. The similarity of the Bahamas to the Louisiade Archi-
pelago suggested by Dana does not seem to me so well established. In
the latter we have high islands of older rocks, and upon the edge of the
plateau sui-rounding them barrier reefs have developed ; while in the
former we have a fringe of islands of seolian origin edged by a barrier reef

In the account which Dana gives of the breaking up of the Maldive
group into an archipelago of atolls, and of the Louisiade Archipelago,
there is nothing clearer than that subsidence has taken place so as to
bury all the high land of the archipelago, but the connection between
the subsidence and the growth of the atolls and barrier reefs of these
archipelagos is not so apparent That there are submerged banks, such
as the Great Chagos Bank, the Macclesfield and the Tizard Banks
described by Captains "Wharton and Aldrich, seems to me to show the
imnossibility of the synchronism of the rate of subsidence of a district

AGASSIZ: BAHAMAS. 185

with the rate of growth of its coral reefs. Off the west and northwest
coasts of Austi-alia, to the north of Houtman's Abrolhos there are exten-
sive coral reefs surrounding adjoining islands and skirting the shores of
the Australian continent. From their appearance on the Hydrographic
Charts they bear a certain resemblance to the reefs of Yucatan, Hon-
duras, .and the Mosquito Coast, extending as they do out on the com-
paratively shallow continental platform of that part of Australia.

One cannot fail to be struck with the general resemblance of the Keel-
ing atoll described by Darwin, and of which a chart is given by Guppy
in his article on the Cocos-Keeling Islands in the Scottish Geographical
Magazine for June, 1889, to some of the smaller of the Bahama Banks.
The tilling up of the interior lagoon from the sand driven in through the
narrow and shallow passages between the islands edging the bank, and
perhaps also from the disintegration of the islands themselves, will in
time produce, as suggested by Guppy, a bank, the interior of which will
be entirely choked by sand, while corals will continue to thrive in the
northern part of the lagoon in numerous patches, much as they do in
the Bahamas, where their growth is not affected by the movement of
the bottom, or by the sand held in suspension during heavy weather.

The sunken banks of the Bahamas, the Navidad, Silver, and Mouchoir
Banks, remind us of the Great Chagos Bank, although there is no evi-
dence to show that these banks owe their peculiar shape to the growth
of corals due to subsidence, but merely that by subsidence what was once
a flourishing coral i-eef edging a marine bank is now by subsidenoe left at
a depth at which corals can no longer thrive, while they still flourish and
grow on the edges and interior floors where the subsidence has not been
great enough to sink that part of the bank to as great a depth, or in other
localities where little or no subsidence has taken place.

The Bahamas can hardly be said to have lagoons in the sense in which
we understand an atoll lagoon. With the exception of the lagoon of
Hogsty Reef, the lagoons of the Bahamas are either huge sinks or slop-
ing faces extending from the inner side of islands on the edge of the
banks towards their open side, or they are small salt sinks like the lakes
of Ne.w Providence, the lagoon of Watling, and the many salt ponds ex-
isting on Long Island, on Inagua, on Grand Turk Island, and others.

On the Great Bahama Bank the only sinks which are enclosed by
higher banks are the extensive depression between New Providence and
Eleuthftra, with a greatest depth of six fathoms near Eleuthera, and the
two to the west of the line of cays between Highborn Cay and Harvey
Cay, with a maximum depth of four and a quarter fathoms. On the

186 bulletin: museum of comparative zoology.

Little Bahama Bank are two depressions, one to the north of Little
Abaco, the other to the south, with a maximum depth of three and
three quarters fathoms. These sinks are depressions iu the general
topography of the banks, and have no definite relation to the islands
and to their edging reefs. The slopes to the west of Cat Island and
of the open banks, such as the Crooked Island, the Caicos, and Turk's
Island Banks, are very gradual until they reach the edging coral reef in
from four to six fathoms, whence the pitch is more rapid and quite sudden
from the 20 fathom line to the 100 fathom line. There is nothing on
the Bahamas which exactly corresponds to the sounds of the Bermudas,
such as Harrington Sound, Castle Harbor, and the Great Sound, the land
which once may have fringed the greater sinks on the Bahamas having
disappeared by erosion, only flats of a more moderate depth indicating its
former existence.

As is well known, a depth of fifty to sixty fathoms has been assigned to
some of the lagoons in the Maldives, of twenty-five fathoms at Keeling
Island by Darwin, and of twenty fathoms at Matilda atoll. In the
composite atolls, like those of the Mahlos Mahdoo atoll, and other of
the Maldives, the depth in the smaller atollons is stated by Darwin to
be not more than five to seven fathoms, while the depth of the greater
lasoon — of the bank itself — is sometimes as much as two hundred and
fifty to three hundred feet. May this not be due to causes similar to
those which in the Bermudas have formed the sounds, attaining a depth
of nearly sixteen fathoms, and in the Bahamas have formed deep chan-
nels between adjacent but more distant banks and islands'? Are the
atollons perhaps, only on a larger scale, such formations as the Serpulse
reefs occurring off the south shore of the Bermudas,^ which are purely
mechanical structures, and in the formation of which the Serpulae or
corals in the case of the Maldives and Mahlos Mahdoo atoll have had
but a limited share?

The islands and islets of the two archipelagos named may be only the
summits of a bank of very irregular outline, upon which corals have
established themselves. Subsidence has taken place there to a very
considerable extent, as stated by Darwin, but the position of the corals
and the shape and distribution of the reefs have only been affected in a
limited manner by it and are not due to it any more than in such locali-
ties as the Bermudas and the Bahamas, The depth between Horsburgh
atoll and the southern end of the Mahlos Mahdoo atoll (over two hundred
fathoms) is no more surprising than the great depth of the channel be-

1 A. Agassiz, Am. Journ. of Science, June, 1894.

AGASSIZ : BAHAMAS. 187

tween the Mira por vos Bank and Crooked Island Bank, or in the passage
between the outlymg banks to the south of the Bermudas.

Granting that the deep chainiels have been formed by subsidence, it
does not follow that the present distribution and existence of the corals
on the summits of the banks is due to the same cause.

In fact, we may apply, only inverting the argument, the reasoning of
Professor Dana against a theory of elevation as explaining the origin of
barriers and atolls : " Coral reefs may, like sea beaches, be made at dif-
ferent heights on the slopes of rising land; but this is not tlie result of
elevation, which is implied ; for barrier reefs and atolls are the objects
whose origin is to be accounted for." Substitute subsidence for rising
land, and remembering that reef corals will not grow at a greater depth
than twenty fathoms, we eliminate subsidence as a factor, unless we are
prepared to accept and imagine a synchronism between the growth of
corals and subsidence in a great number of the districts in \Yhich they
flourish, — of which we have no proof.

AGASSIZ: BAHAMAS.

189

INDEX.

The foUowinj:; references to the Charts of the U. S. Hydrographic Office (H. C. No. . . .) and to
those of tlie Hydrographieal Office of the Admiralty (A. C. No. . . .), give tlie numbers of the charts
relating to tlie Bahamas, to Cuba, to the Gulf of Mexico, to the West Indies, and to the Caribbean,
which have been consulted during the preparation of this Report on the Cruise of the Steam Yacht
" Wild Duck." In the Introduction, see page !.">, reference has been made to the general charts cov-
ering the region explored by the " Wild Duck " i>ublished by the British Admiralty and the U. S.
Hydrographic Bureau, as well as by the U. S. Coast and Geodetic Survey. These charts are included
in this Index, and the localities referred to them of which there are no separate charts.

AiiACO Bahama Land, 67.
Abaco Island, A. C. 3'.J9, p. 65.
Ackliii Island, 91.
^olian Rock (Juarries, 19.

thrown up by hurricanes, 46, 58.

— '■ uuderniined, 4G.

Ai,Mssiz, L., on Salt Cay Bank, 81.

Alacran Reef, H. C. 403, 900, 1234, 1240,

A. C. 1203.
Albuqiieniue Cays, A. C. 1.511, p. 157.
Aldricli, Captain, on submerged coral banks,

184.
Alfred Sound, A. C. 2022, H. C. 422, p. 98.
Ambergris Cay, 90. ;
Andros, H. C. 26% A. C. 496, p. 50.

, Reef of, 54.

, Vegetal ion of, 55.

, White Marl off, 52.

Anegada, A. C. 130, H. C. 1002, 2003, p. 153.

Anguila, 84.

Aiiguilla, II. C. 1002, A. C. 2038, p. 152.

Bank, A. C. 1.30.

Antigua, 11. C. 1004, A. C. 918. 2000. p. 151.
Areas Cays, II. C. 403, 1234. A. C. 1829.
Arenas Cay, IT. C. 403, 1233, A. C. 1209.
Arrowsmith Bank. H. C. 966, 1380.
Ascension and Kspiritu Santo Bavs, II. C. 402,

1380. p. 101.
Athol Island, 18.
Atlantic Coast of United States, Cape Hatteras

to Key West, , ,J,^^^, U. S. Coast Survey.
. n. C. 944, Cape Canaveral to Havana,

and Bahamas to San Salvador.
Avis Island, H. C. 40, A. C. 2600, p. 154.

Bahama Bank, discoloration of, 43.
Bahama I'.anks, A. C. 399, 1490, 2009, 2077,

2075, II. C. 26"-20'i, 940.

-, Beaches of, 167.

, Bottom of, 100.

, Coralline .Vlg*, Species of, 167.

Bahama Island, A. C. 399, p. 73.

, Barrier Reef of, 73.

, Vegetation of, 74.

Bahamas, Coral Reefs of, 57, 1.36.

. Height of Dunes of. 183.

, Hydrography of, 139.

, Lagoons of, 185.

, Rocks of, 170.

, Sinks of, 185.

, Slope of, 182.

, Subsidence of. 68. 99.

. Sunken Banks of, 185.

Bah la Honda, II. C. 520\ A. C. 411.

Baitiqueri, Terraces of, 111.

Bmj.. Xuevo, H. C. .379. A. C. .391, p. 156.

Balansa on Loyalty Islands, 178.

Banes, A. C. 426. p. 121.

Baracoa, A. C. 43.5, II. C. 377^ p. 118.

, Elevated Reef of, 118.

Barbados, II. C. 1010, A. C. 950, 2485, p. 147.

, Elevated Reefs of, 177.

, , Harrison and Browne on, 179.

Biirbuda, H. C. 367, A. C. 1997, 2000, p. 152.

Barracouta Rocks, 76.

Bartlett Deep, 106.

Beecher, Captain, on Landing of Colnmbus,

88.
Beminis, The, H. C. 26% A. C. 496, p. 59.

190

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Bermudas, II. C. 27.

Berry IslaD.ls H. ('. 20% A. C. 499, pp. 26, 61.

Billy IslaniL .",2.

Bird lidck, \'i,-it (if Columbus to, 93.

Black I^uu^. 79.

Blake, Sir Ilrnry, on Landfall of Columbu<;,

88.
Blake n-itcau, l-",.
BliiSMini ( 'hainii'l, :j9.
Blue Kifid La- i, II. C. 391, 1292, A. C.

1504.
Blue Hills, 2.;.

Blue Hole i'niiit, II. C. 2G% A. C. 2075.
— to (iuineho.-. Cay, II. C. 2G>i, A. C. 393,

20(11).
Blue Holes 42.

Bouaeca. II. C ■'!!iG, A. C, 1718, p. 165.
lionne'', r. ( i.. 171.

, on Subiuai'iue Elevations, 173.

Bottom.s, Carijonate oflinie, per cent of, 109.
Bryant, Henry, V^isit to luagua, 99.
Brown Bank. 94.
Buen .\yre Island, 159.
Burrow ( "ay, 09.

, Conch hea])s of, 70.

— , Pelagic Fauna off, 71.

C.vi'.o.Nico and Livisa, Cuba, A, C. 428,

II.(,'. UiL
Caicos l!aid<, A. C, 393, p. 95.

, Mauds <if, 95.

Caleta Point, Terraces of, 110.

Campeche Bank, II. C. 966.

Canuouan, H. ( '. •">57, p. 148.

Cape ( 'atoehe, IGL

Cajie (Jracias a Dios to Gulf of Darien, H. C.

945.
C.-ipc Vei-(L 46.
Cardenas, A. C. 420, p. 128.
Cariaeou. H. ('. 357, A. C. 2872, p. 148.
Caribbean Sea and Gulf of Mexico, A. C. 31,

3(1, H. ('. 12!iO.

and West Indies, II. C. 1290.

, iiaid'is of, 145. 156.

, Cord Penis of. 145, 160.

, I'.astern Part of, H. C. 40.

, SiiliuKii'iiK" Banks of, 176.

. Wr-t Coast. II. C. 394, A. C. Sheets

XI.-XIV.. 1204. 1218, 1219, 1579.
Cattle Islau.l. '.12.

Cat Island. H. C. 26'', A. C. 393, p. 34.
C.'ixones, If'..").
Cnv Coco. 120.

CavCnnllles, H. C. 277, p. 124.
Cay Frances, 125.
Cay Lobos, 125.
Ca\- Piedr.as, 127.
Cayman Bank, 158.

Cayman Brae, H. C. 43, 947, A. C. 462,

p. 158.
Central America, East Coast of, II. C. 1120,

A. C. 763.
Chambeyron on New Caledonia, 178.
Clietnmal P.ay, H. C. 394.
Chinchorro Bank, II. C. 394, 1072, A. C. 1796,

p. 163.
Chiri(iui Lagoon, II. C. 386, A. C, 1793, p. 159.
Churchill, Uonald, Analyses of Bottoms. 108.
Clarence Harbor, H. C. 339, A. C. 2093, p. 44.
Clarion Bank, 94.

Clark, F. W., Analyses of Coral, 167.
Clarke, W. B.. on Lafu, 178.
Clifton Bluffs, 22.
Clijiperton Rock, 174.
Cochrane Anchorage, 18.
Colorado Beefs, 174.
Colnndnis Bank, H. C. 26=, A. C. 2075, pp. 37,

43.

Point. 35.

Conceiition Island, H. C. 26^ A. C. 2075,

p. 90.
Conch Cut. 49.
Coral Heads, 28.
Coral Beef, Recent Views on, 170.

, Thickness of, 177.

Co-ta Rica, II. C. 945, p. 159.

Courtown Cays, II. C. 391, 945, A. C. 1511,

p. 157.
Cow and Bidl, 23.

Cozumel Island, H. C. 402. 1380. p. 161.
Crab Island, A. C. 130, H. C. 1001.
Crooked Island Bank. A. C. 393, 2.580. p. 91.

, Slope off, 143.

Crosby, 0. W., on Yunriue, 119.

•, on Elevate<l Reefs of ('uba, 179.

Cuba, Eastern, A, C, 303, 2580, II. C, 946.
, North (,'oast of, Gibara to Ca]ie San

Antonio, H. C. 947.

, Western, A. C. 2579, H. C. 947.

, Coast of, 108.

, C(n-al Reefs of, 133, 135.

, Limestone Hills of North Shore, 120.

, Shore Platform of, 135.

Cuidado Bank, 94.

Culebra, A. C. 130, H. C. 1001.

Current Island. 27.

P.MT,, W. IE, Bahama Land Shells, 20.
, on Miocene Fossil Shells, 113.

Dana, .1. D., On Barrirr Reefs. 178.

, on I'.levation of Coral Reefs, 187.

, on the I'eejee I>-!ands. 173.

, Lonisiade .Vrchipida^o, 184.

, on the Maldives, 184,

, on Metia, •12.

. on Submarine Elevations, 173.

AGASSIZ : BAHAMAS.

191

Dana, J. D., Thickness of Elevated Coral

Reefs, 182.

, on the Tortusras, 116.

Darien, Gulf of, to (Jiilf of Triste, H. C 9G4.
Darwin, on the W. I. Ueefs, 11, 173.

, nn tile Maldives, 18-1.

Davidson, Geoi-g-e, on Cliiiperton Island, 174.

Devil's Rliiff, (;;].

D.'sirade. A. ('. SS5, II. C. 303.

Dominica, II. C. 1318, A. C. tjlt7, 1)50, p. 149-

D.iiible Headed Shot Cays, 84.

Draselie, K. von, Elevated Keefs of Luzon,

180.
Douglas Channel, A. C. 406, II. C. 2G\ 339,

p. 26.

Eastekn Bahama Islands, 85.

, Reefs of, 137.

, Sect inns across, 142.

E<linburuli Keef. II. C. 945, p. 1(!4.

Efrs Island to Eleuthera, II. C. 1241, A. C.

2098, p. 27.
Eleutlieni Island, II. C. 20'\ A. C. 2077, p. 29.
Elevated Reefs of Cuba, 179.
Escondido, Terraces of. 111.
Espiritu Santo Bay, Kil.
Explanation of Plates, 107.
Exnnia Harbors, H. C. 340, A. C. 1474.

, Slnpc olf, 144.

Esuuia Island, .\. C. 2075, p. 47.

Faiu.uw, W. G., on Bahama AIi^.t, 168.
Findlay Cay, 26'', A. C. 2(i77.
Flaniin.--.. Cay, H. C. 26=, A. C. 2075, p. 39.
<'liannel, II. C. 2G^ 1227, A. C. 405,

2i)77, 11. 2r>.
Florida, Straits of. Section across, 141.
Florida Reefs, Slope off, 140, 182.
. U. S. Coast Survey, Nos. 166, 167,

1(;8, 169.
Fdraniinifera in White Marl, 56.
Fonnii;as Bank, II. C. 373, A. C. 486, p. 155.
Fortune Island, !)2.

Lauoiin, 03.

Fd.x, (i. v., 89.
Frai;iiso Cays, 127.
French Cay, 96.

GAt.i.ioT Bank, .37.

Gla-s Window, 30.

Gli.vcr Reef. II. ('. 394, 1120, p. 163.

(iovernor llarlior, 32.

Graham Harbor. 88.

Grand Cavniau, II. C. 43, -373, 947, A. C. 462,

p. 1.-^8.
Grand Tnik Mand. 100.

Coral Heads of, 138.

Grantstowu Flat, 23.

Great Abaco Island, 65.

Great Bahama Bank, H. C. 26»-26-i, A. C.

1490, 20o;i, 2075, 2077, p. 50.

, Xortliern I'art, II.(". 044, p. 17.

Great Corn Islands, II. C. 392, A. C. 1476,

p. 164.
Great Exuma, H. C. 26^ 20% A. C. 2075.

Harbor, A. C. 1474, II. C. 340.

Great Harbor Cay Bores. 63.

(Ireat I.-aac, II. V. 2(1 ', A. C. 496, p. 60.

, I'ot Ilnles of, 61.

, Vegetation of, 01.

Great Ragged Island, 43.

Great Stirru)) Cay, A. C. 14-32, H. C. 1158.

Green Cay, II. C. 20'', A. C. 2077, p.'49.

Bores, 50.

Green Turtle Cay, A. C. 398, 78.

Grenada, II. C. 1316, A. C. 956, 2821, p. 147.

Grenadines Bank, II. C. 357, A. (_'. 2872, p.

146.
Gressly on .Jurassic Reefs, 179.
Griswuld, Lieutenant, on Clipperton Rock,

174.
Guadeloupe, H. C. 363, A. C. 885, 956.
Guajira Peninsula, 159.
Guanos Point, Terraces df, 130.
Guantanamo, A. C. 442. II. ('. 377^ p. 111.
Gulf of .Mexico, Sheets I.-VL, II. C. 31, 32,

33, 34, 35, 36, 704, 705, A. C. 392.

, Coral Reefs of, ino.

Gulf of Mexico and Caribbean. A. C. 31, 36,

H. C. 1290, 1,^ S. ('oast Survey, -,^^.
Giimbel, 180.
(juppv. Dr., on the Cocos-Keeling Islands,

185.

, on Elevated Reefs, 116.

(4un Cay, 58.

Haines Bu'kf, 62.

Haiti, West Coa-^t, II. C. 373, 948. A. C. 393.

IIaIf-:\Ioou R.'ef, ]i;5.

Harl)or Islaml, 31 .

Harrison ami lirowne. Elevated Reefs of Bar-
bados. 115.

IIarv.'y Cay, .36.

Hatchet Point. 32.

Havana, .\. C. 414, H. C. 270, p. 132.

Hawk's Bill P.ank. 40.

Hawk's Xcst -Vnchnrage. 101.

Heilnrin, Professor, on thr Duke of Argyll,
172.

Highborne Cut. II. C .'I'itl, A. C. 1717. 2077.

Hill, R. T.. on the Geology of Cuba, 100.

. on the Elevateil Reefs of Cuba, 179.

Hog Man.!. 18.

Hogsiy Keef, A. C. 393, pp. 103, 144.

. Cavs of, 103.

, Coral Heads of, 105.

192

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Hogsty Reef, Corals of, 107.

, Laifoon of, 103.

, Fauna of Lagoon, 105.

, Sections across, 10-1.

liole in the Wall, G5.

Honduras, Coast of, H. C. 394, 966, pp. 14G,

lU-2.
Honduras Gulf, H. C. 397, 1120, A. C. 1573.
Horsburgh Atoll, Depth off, 18G.
Huxley, T. iL, on A Conspiracy of Silence,

172."

Imia Hay, 111.
Inagua, A. C. 393.

, Great, 97.

, Little, 99.

Jamaica, A. C. 446, 459, 486, H. C. 347, 373,

p. l.'Ju.
Jaragua, A. C. 433, H. C. 518\

, Terraces of, 113.

Jensen, Carl C., ou Clipperton Island, 174.

Joulter Cays Bank, 38.

Julien, Ale.xis A., on Sondjrero, 183.

Kebxinc; Island, Depths of Lagoon, 18G.
Kent, W. S-, on the Great Barrier Reef of

Australia, 171.
Kingston, A. C. 255, 450, H. C 348.

Lake Cunningham, 21.

Lake Kiilarney, 22.

Langenbeck, Dr. R., 171.

. on Floriihi and West India Reefs,

172.
Lanzanillii Channt.'l, 127.
Lark Channel, 39.

Light House Reef, H. C. 394, 1120, p. 1G3.
Limon, H. C. 390, 1293, A. C. 2144.
Linyard Cay, 81.
List of Figures in the Text, 195.
Little Aliaeo, 77.
Little Bahama Bank, A. C. 399, H. C. 944,

p. G4.

, Cays of, 77.

, Reefs of, 138.

Little Cavnian, II. C. 43, 373, 947, A. C 462,

p. ns.

Little Corn Islands, H. C. 392, A. C. 1476,

p. 1G4.
Little Sale Cay. 76.
Little San Salvador, 34.
Long Cav, 25, 97.
Long MJind, II. C. 26°, A. C. 2075, p. 44.

, Caves of, 45.

, Veiretation of, 45.

, yKoiian Cliffs of, 40.

Los Koques, 159.

Mackie Bank, 38.

Madagascar Reef, H. C. 1235.

Maldives, Depth of Lagoon, 186.

Manati Port, A. C. 418.

Margarita Island, H. C. 374, A. C. 230.

Mariei, H. C. 520^ A. C. 413.

Marie Galante, H. C. 363, A. C. 95G, 885.

Mariguana, A. C. 393, p. 94.

Martmiipie, H. C. 1009, A. C. 37, 9-56, p. 150.

Mata, A. C. 435, H. C. 377".

Matanzas, A. C. 415, H. C. 270. p. 129.

Mathew Town Road, Inagua, S. W. Pt. Reef,

A. C. 2025, p. 98.
Mays Cape, 1 10.
Memory Rock, 75.
Middle~Ground, 137.
]\Iira por vos Bank, A. C. 408, p. 93.
Misteriosa Bank, 11. C. 9GG, p. 158.
Mojsisovics on Dolomitic Reefs, 180.
Molasses Reef, 97.
Monito Cay, 128.

Montserrat", H. C. 1011, A. C 254, p. 151.
Morant Point, A. C 2.J5.
Morant Cays, A. C. 255, H. C. 347, 373, p,

155.
Mores Island, 70.
Morgan's Bluff, 54.
Morrison Cav, H. C. 945, p. 164.
Moss Hill, 23.
Mosquito Coast, II. C. 394, 945.

Bank, 163.

Mouchoir Bank, A.C. 393. p. 102.
Mugeres Harbor, II. C. 402, 1379.
Mulatas Archipelago, 159.

Naran.jo, Limestone Hills of, 122.

Nassau Harbor, H. C. 949, A. C 1452, p. 18.

Navassa Island, H. C. 379, A. C 4G1. 486,

p. 155.
Navidad Bank, A. C. 393, p. 102.
Negro Head to Turneffe Cays, H. C. 399.
Nelson, Capt. R. J., on the Bahamas, 3.
Nevis H. C. 1011. A. C. 487, p. 151.
New Providence, H. C. 26^, 335, 1377, A. C

1489, 2077, pp. 18. 23.
Nipe, A. C. 427, p. 120.
North Cav, 25.
Northrop, Dr. John L., on Cays of Nassau

Harbor, 35.

■, on the Bahamas, 7.

Nuevitas. A. C 416, H. C. 520». p. 123.
Nurse Channel, II. C. 340, A. C. 1494,
p. 41.

Ocean Holes, 41.

Ogilvie, Miss, on Dolomitic Reefs, 179. 180.
Old Bahama Channel, H. C 26% A. C. 2009,
p. 140.

AGASSIZ : BAHAMAS.

193

Old Providence, H. C. 395, 945, 1372, A. C.

1334, p. 157.
Opisbo Shoals, H. C. 403, 1239, A. C. 1830.
Orange Cay, 55.
Orchila Island, 159.

Padre, A. C. 419, H. C. 520^ p. 123.

Panama, 159.

Pan de Matanzas, 129.

Paredon Grande, 125.

Pearl Cays, H. C. 392, A. C. 1503, p. 164.

Pease, W. H., on Clipperton Rock, 174.

Pedro Bank, 11. C. 373, A. C. 450, 48G, p. 156.

Pimlico Cav, 27.

Plana Cays", A. C. .393, p. 94.

Port Cortez, H. C. 398.

Port Nelson, H. C. 1113, p. 90.

Porto Rico, A, C. 130, 2600, H C. 40, 1001,

p. 154.
Potter Cay, 35.
Powell Point, 33.

Queen's Channel, 39.

Queen's Stairvvay, 19.

Quita Suefio Bank, H. C. 394, 945, p. 157.

Racoon Cut, H. C 341, A. C. 1470.
Ragged Island, H. C. 26", 341, p. 41.
Ragged Island Harbor, H.C. 339, A. C 1472.
Redonda, H. C. 1011, A. C. 2600, p. 161.
Richards, T. W., Analyses of Bottoms, 108.
Richthofen, F. von, on Dolomitic Reefs, 180.
Ripple Marks un bottom, 49.
Roatan Island, H. C. 394, p. 165.
Rock Harbor, 72.
Roncador Bank and Reef, H. C. 373, 395,

945, 1-374, A. C. 1478. p. 157.
Rosalind Bank, H. C. 373, .394. p. 156.
Rose Island, II. C. 339, pp. 18, 25.
Rothpletz on Udloniitic Reel's, 179, 181.
Roval Island, 27.

Rum Cay, H. C. 26", A. C. 2075, p. 90.
Russell Island, 27

Saba, A. C. 130, 487, H. C. 1002, p. 151.
Saboney, Terraces of, 112.

, Limestone Hills, 117.

Sagua la Grande, A. C. 2-384, H. C. 1311.

p. 127.
Saintes, H. C. -362, 363, A. C. 885, 956.
Salina Fl.at, 45.
Salt Cay Bank. H. C. 944, 947, A. C. 6-59,

1217, p. 81.
Samana Island, A. C. 393, p. 94.
Samphire Cays, 26.
Sand Bores, 37.
San Domingo, Corals of, 154.
San Salvador to San Domingo, A. C. 1266.

VOL. XXVI. — NO. I. IJ

Santa Cruz, A. C. 1-30, 485, H. C. 1002, p.
153.

Gap, 131.

Santiago de Cuba, H. C. 1003, A. C. 443, p.

112.
Schooner Cay Flats, 33.
.Seal Cays, 96.

Sea Water, solvent action of, 48.
Serrana Bank, H. C. 373, 394, 945, 1374,

A. C. 486, 1478, p. 157.
Serranilla Bank, H. C. 379, 394, A. C 1498,

p. 157.
Seven Hills, 24.
Ship Channel, H. C. 26", 1227, A. C. 1509,

2077.
Silver Bank, A. C. 393, p. 102.
Sisal Keefs, H. C. 403, 1235, A. C. 1206.
Six Hill Cays, 96.
Soboruco, 108.
Sombrero, A. C. 130, 484, 2600. H. C. 371",

1002; U. S. Coast Survey, 1859, p. 152.
South America, North Shore, H. C. 40, 964,

A. C. West India Sheets, VIII.-XI., 394,

395, 390, 1579, p. 158.
St. Andrews, II. C. 391, 945, A. C. 1511,

p. 157.
St. Bartholomew, H. C. 1002, A. C. 2038,

p. 151.
St. Eustatius, H. C. 1011, A. C. 487, 2600

p. 1.50.
St. .lohn, A. C. 130, H. C. 1002.
St. Kitts, H. C. 1011, A. C. 487, 2600, p. 151.
St. Lucia, II. C. 1261, A. C. 956, 1273.
St. JIartin, H. C. 1002, A. C. 130, 20-38, p. 151.
St. Thomas, A. C. 1-30, 2183, II. C. 1001, 1002,

p. 154.
St. Vincent, H. C. 1279, A. C. 791, 956, p.

149.
Suess on the Straits of Florida, 13.
Swan Islands, H. C. 966.

Tanner Net, Haul with the, 8.
Tarpon I'nint, 33.

Thomas, H. L., .Journal of Columbus, 89
Thunder Channel, 39.
Thurman on .lurassic Reefs, 179.
Tillinghast, Handkerchief Shoal, 14.
Tobago Island, II. C. 354, A. C. 505.
Tobago Anchorage (Tji-enadines), A. C. 2872.
Tongue of the Ocean (South part), II. C. 26',

A. C. 2077, p. 39.
Tortola, H. C. 1002, A. C. 130.
Triangles, H. C. 403, 12-39.
Turk's Islands, H. C. 1000. A. C .393, 1441,

p. 100.
Turk's Islands Passage, Pelagic Life of, 97.
Turneffe Cavs, H. C. 394, 399 (to. Negro

Head), 1120, p. 16-3.

194

bulletin: museum of compaeative zoology.

Utill.v Island, H. C. 396, A. C. 1532, p.
1G4.

Vera Ckuz, Anchorage of Anton Lizardo,

H. C. 907, A. C. 523.
Vegetation of New Providence, 21, 24.
Virgin Gorda, H. C. 5(59, 1002, 2008, A. C. 130.
Virgin Island Bank, H. C. 2008, A. C. 106^

106S 130, p. 153.
Virgin Passage, 11. C. 965.
VivuriUa Island. 105.

Wa-sherwoman Cut, H. C. 26^
Watling Island, A. C. 393, p. 80.

, L.-indfall of Columbus, 88.

, Reef Harbor of, 89.

Wax Cay Cut, H. C. 339, A. C. 1495, 2077.
Wharton, Capt. W. J., on Submerged Coral

Banks, 184.

, Sketches of Blue Holes, 42.

Wide Opening, H. C. 1220, A. C. 2100, 2077,

p. 51.
Wild Duck, Track of, 3.

William Island, 52.

West India Islands and Caribbean, A. C. 761-

703.
West Indies, Sheet No. 2, H. C. 946, A. C.

393.

, Elevated Reefs of, 177.

, Gulf of Mexico, and Caribbean, H. C.

1290.

-, Submarine Ranges of, 176.

, Submarine .Scenerv of, 165.

Whale Cav Channel, A. C. 398, H. C 998,

p. 80.
Windward Islands, II. C. 40, A. C. 956, 2000,

p. 145.
Wolff, J. E., on Clipperton Island Trachyte,

174.
Woollendean Cays, 71.
, Marl of, 71.

Yucatan Bank (Northern End), A. C. 1205,

II. C. 900, 1234. 1235, 1380.
Yumuri, A. C. 435, H. C. 377«.
Yunque, 119.

AGASSIZ: BAHAMAS. 195

LIST OF FIGURES IN THE TEXT.

PAGE

90

Cow and Bull (Eli-utlu-ra)

Section across New Providenee 23

The Glass Window 30

Highest Hills of Cat Island 34

yEolian Hills, Clai-ence Harbor 44

jEolian Hills and Clitrs, southern part of Long Island 46

yEolian Cliti's south of Cape Santa ;\Iaria 46

Water-worn Islet, Conch Cut 49

West Shore of Andros, Wiil(> Opening 51

Young Mangroves, Wide Ojieniug 53

iEolian Piocks thrown up above high-water mark, Gun Cay 58

Gun Cay " . . . 59

Devil's Blair 63

Hole in the Wall 65

Mores Island 70

High Rock, Bahama Island 73

Little Sale Cay Rocks 76

Salt Cay Bank 82

yEolian Clitfs, South Anguila 84

Landing Place of Columbus, according to Captain Becher, R. N 88

Landfall of Columbus, according to Sir Henry Blake 88

The Plana Cays " 94

Samana or Atwood Cay 95

Six Hill Cay 96

Long Cay 97

Hogsty Reef 106

Terraces at Caleta Point 110

Gran Piedra and Terraces near Saboney 117

Elevated Reef (Sobornco), Entrance of Baracoa 118

Cuchillas del Pinal 120

Terraces and Entrance to Nipe 120

Outlet of Port P.anes 121

Limestone Hills back of Xaranjo 122

Patches of Sobornco, Cay Frances 126

Pan de Matanzas , 129

Terraces, Guanos Point, Matanzas 130

Santa Cruz Gap 131

Clipperton Island 175

AGASSIZ; BAHAMAS. 197

EXPLANATION OF THE PLATES.

PLATE I.

Cliart of the Bahamas, prepared by the Hydrographic Bureau, Commander C. D.

Sigsbee, U. S. N., Ilydrograplier.

PLATE II.

Fig 1. Ilogsty Keef, from a sketch sent by Captain W. J. Wharton, R. N ,
Hydrographer to the Admiralty. A few soundings made by the " Wild
Duck " have been added outside tlie 100 fatiiom line.

Fig 2. Longitudinal section across Hogsty Reef, from the 120 fathom point on the
east to the 140 fathom line on the west.

Fig. 3. Transverse section from the 120 fathom line on the north to the 206 fathom
point on the south.

Fig. 4. The same as Figure 2, on a larger scale, to show tlie slope of the lagoon from
the eastern ledge to tlie main entrance of the lagoon.

Fig. 5 Section across the lagoon near the eastern ledge.

Fig. 6. Section across the lagoon to tlie west of South Cay somewhat to the east of
the entrance.

Fig. 7. Section across the lagoon at its widest part, near the middle.

PLATE III.

Charts of the vicinity of three prominent Blue Holes, from sketches kindly sent by
Captain Wharton, R. N., Hydrographer to the Admiralty.

Fig. 1. Thirty-eight and seventeen fathom holes. These holes are south of the

eastern extremity of Thunder Channel, Tongue of the Ocean.
Fig, 2. Tliirteen fathom liole.
Fig. 3. Twenty-four fathom hole.
The holes m Figures 2 and 3 are immediately north of Blue Hole Point.

PLATE IV

Fig 1 Section across the Straits of P'lorida, from Jupiter Inlet to Memory Rock,
across the Little Bahama Bank to off Pensacola Cay.

Fig. 2. Section from Hillsboro Inlet across the Straits of Florida and the North-
west Providence Channel, to Mores Island and across the bank to ofi
Great Abaco.

198 BULLETIN: MUSEUM OF COMI'AEATIVE ZOOLOGY.

Fig. 3. Section from Fowey Rocks, across tlie Straits of FloriJa to Gun Cay, tlie

Great Baliania Bank, the Nortlieast Providence Channel, to Koyal Island

and off Kleutliera.
Fig 4. Section from Orange Cay across the Great Bahama Bank to Andros Island,

the Tongue of the Ocean, Nassau, to off Eleuthera.
Fig. 5 Section across Salt Cay Bank, the Santaren Channel, the Hurricane Flats

to Andros, the Tongue of tiie Ocean to Green Cay, to Harvey Cay across

Exuma Siumd to off Cat Island.
Fig 6. Section across Salt Cay Bank, the Santaren Ciianncl, the Hurricane Flats,

the Tongue of the Ocean to Great Exuma, Exuma Sound, Conception

Island, and off Watling's Island.
Fig. 7. Section from Cape Gracias a Dies across the Mosquito Bank to Kosalind

Bank, Pedro Bank, and to Portland Biglit, Jamaica.

PLATE V.

Fig. 1 Section from Buenavista Bay (Cuba) across the Old Bahama Cliannel.
Fig. 2. Section from the Coast of Cuba acrtjss Romano and Paredon Cays and the

Old Bahama Channel to Guinchos Cays.
Fig. 3. Section from the Coast of Cuba across Romano Cay and the Old Bahama

Channel to a pt)int somewhat west of Lobos Cay.
Fig. 4. Section from the Coast of Cuba to Romano Cay across the Old Bahama

Channel to Diamond Point.
Fig. 5. Section from the Coast of Cuba to Cay Sabinal across the Old Bahama

Channel to Blue Hole Point.
Fig. 6. Section from Pan de Sama to off Cay San Domingo and to Ragged Island.
Fig. 7. From Point Azules (CapeMaysi) to Inagua.
Fig. 8. From Cape Isabella (Haiti) to Turk's Islands.
Fig. 9. E'rom Cape Cabron (Haiti) to Navidad Bank.
Fig. 10. E>oni Castle Island (Acklin Island) to Hogsty Reef to Inagua.
Fig. 11. From Cay Verde to South Cay (Mira por vos Bank) to Castle Island

(Crooked Island Bank).
Fig. 12. From Cay Verde (Long Island) to Bird Rock (Crookt-d Island Rank).
F'ig. 13. From Cay Sal Bank across the Nicholas Channel to Baliia de Cadiz (C^uba).
Fig, 14. Section from Coffin's Patches across tlie Straits of Florida to Elbow Cay

(Cay Sal Bank).
Fig. 15. Section across Cay Sal Bank from Elbow Cay to Anguila.
Fig. IG. Section across the Straits ol Florida from K<jy West (Sand Key) to Havana.

PLATE VL

Fig L Section off Alfred Sound (Inagua Island) to 222 fathoms.

E'ig. 2. Section off the southeast spit of Fortune Island (Crooked Island Bank) to

300 fathoms.
E'ig. 3. Section off Exuma Harbor to 266 fathoms, west side of Exuma Sound.
Fig. 4. Section off No Name Cay to 500 fatlioms (south of Green Turtle Cay,

Little Bahama Bank).
Fig 5. Section running northeast of Lagartos across the Yucatan Bank.

AGASSIZ : BAHAMAS. 199

PLATE VII.

Fig. 1. Section along the Virgin Island Bank from the eastern extremity of Porto

Kico to off Anegada.
Fig. 2. North and south Section across the Virgin Bank through the Island of St.

John.
Fig. 3. Section from the west side of Saba Bank to Saba and across the Anguilla-

St. Bartholomew Bank.
Fig. 4. Section across Montserrat and Antigua.
Fig. 5. Section across Martinique.
Fig. 6. Section across Grenada.

Fig. 7. Section across the Grenadines Bank to off Barbados.
Fig. 8. Section from Aves Island across to Dominica.
Fig. 9. Section from Aves Kidge across St. Lucia to the Barbados Ridge.

PLATE VIIL

Contour Chart of tlie Caribbean Sea, prepared by the Ilydrographic Bureau in 1885,
Commander J. R. Bartlett, U. S. N., Ilydrographer for the " Three Cruises
of the Blake."

PLATE IX.

The Smaller Banks from Navidad to Watling Island, to illustrate the transition from
such islands as Great Inagua, Watling Island, Rum Cay, Mariguana to
Crooked Island and t!aicos lianks, to such banks as Turk's Islands and
Mira por vos Banks, and finally to the three easternmost sunken Banks
of ]Mouchoir, Silver, and Navidad, From Ilydrographic Chart No. 910,
Admiralty Chart No. 393.

Fig. 1. Crooked Island Bank.

Fig. 2. Caic(« Bank.

Fig. 3. Great and Little Inagua Islands.

Fig. 4. Mariguana Island.

Fig. 5. Turk's Islands Bank.

Fig G. The northern part of the same bank, to show the mass of beads left to the
soutli of Grand Turk Island by the disintegration of the former Grand
Turk land. Reduced from Ilydrographic Chart No. 1000.

Fig. 7. Mouchoir Bank.

Fig. 8. Silver liank.

Fig. 9. Mira por vos Bank.

Fig 10. Navidad Bank,

Fig. 11. Conception Island and Rum Cny.

Fig. 12. Watling Island.

PLATE X.

From Hydrographic Charts Nos. 26*, 26°, 944, and 946.
Fig 1. Little Bahama Bank. See Admiralty Chart No. 390, for a larger scale.
Fig. 2. The northern extremity of the eastern pnrt of Great Bahama B;ink. including

New Providence to Eleuthera, and the Middle Ground to the southeast

of New Providence.

200 bulletin: museum of comparative zoology.

Fig. 3. A part of the &ame Bank from New Providence as far as the Douglas Chan-
nel, on a larger scale than Fig. 2.

Fig. 4. The west face of Crooked Island Bank, from the southern spit of Fortune
Island to Bird Hock, showing the disintegration of the lowlands of the
tanks, and the formation of lagoons and the line of soundings taken by
the " Wild Duck " to 300 fathoms. Reduced from Admiralty Chart
No. 393.

Fig. 5. The southwestern edge of Exuma Sound, showing the formation of the outer
chains of cays from Great Exuma Island on the narrow shore plateau
bounded by the 100 fathom line, and the disintegration of the inner face
of Great and Little Exuma Islands to form the low cays, flats, shallows,
and sand bores on the adjoining parts of the bank. The dotted line indi-
cates the line of soundings taken by the "AVild Duck " to 206 fathoms.
Reduced from Ilydrographic Chart No. 20°.

Fig. 6. The southern extremity of Long Island, showing the transition of tlie west-
ern face of the central part of the island into the bank flats, and the
breaking up of the west face of the southern promontory of Long Island
by the SaUna Flats. Keduced from Hydrographic Chart No. 26°.

PLATE XI.

Fig. 1. The Cays and Bores to the eastward of Exuma Island. Soundings in feet.

Ilydrographic Chart No. 340.
Fig. 2. Clarence Harbor and the adjoining Cays and Bores. Soundings in feet.

Ilydrographic Chart No. 2093.
Fig. 3. The northeastern extremity of Andros with the Joulter Cays and Bores.

Soundings in fathoms. Hydrographic Chart No. 1496.
Fig. 4. Washerwoman's Cut Channel^southof Andros), leading from the Tongue of

the Ocean to Hurricane Hole. Soundings in fathoms. Hydrographic

Chart No. 2077.
Fig. 5. The Ragged Islands Cays on the southeastern extremity of the Great

Baliama Bank. Soundings in feet. Hydrographic Charts Nos. 341, 1399.
Fig. 6. Russell and Royal Islands, on the northeastern extremity of the Great

Bahama Bank. Soundings in fatlioms. Hydrographic Chart No. 2098.
Fig. 7. Green Turtle Cay, with the adjoining cays on the eastern face of Little

Bahama Bank. Soundings in feet. Hydrographic Chart No. 398.

PLATE XIL

Fig. 1. The Channels between the Sand Bores leading from the southern extremity
of the Tongue of the Ocean on to the Bank. Hydrographic Chart
No. 2009.

Fig. 2. The Beminis, Gun Cay, and the adjoining flats to the eastward. Hydro-
graphic Chart No. 1490.

Fig. 3. The three Biglits dividing the Island of Andros to the south of Wide Opening.
Hydrographic Cliart No. 1406.

Fig 4. Great Stirrup Cay and the Sand Bores to the west of Great Harbor, on the
northeastern point of the western shank of Great Bahama Bank. Hydro-
graphic Chart No. 1432.
The soundings in Figures 1-3 are in fathoms, those of Figure 4 in feet.

AGASSIZ : BAHAMAS. 201

PLATE XIII.

Fig. 1. North Shore of Cuba from Port Padre to Cay Guajaba, showing the harbors
of Nuevitas to Padre, and the mode of separation of the shore cays from
tlie main island.

Fig. 2. The Cays of the North Shore of Cuba, from Cay Guajaba to Cay Coco.

Fig. 3. The Cays of tlie North Sliore of Cuba from Buenavista Bay to Lanzanillo
Cay, and the Boca de Caaete.

Fig. 4. The Western Cays of the North Shore of Cuba, from Boca de Canete, Sagua
la Grande to Cardenas Bay.

Fig. 5. The Cays of the Western Extremity of Cuba and the Colorado Pcefs. from
Inez de Soto Cay as far as Cape San Antonio. All figures from Hydro-
graphic Chart No. 947.

PLATE XIV.

Fig. 1. The Cays of the Coast off Sagua la Grande.

Fig. 2. Cay Confites, one of tlie outer cays on the edge of the plateau of the north

shore of Cuba. Admiralty Chart No. 2384.
Fig. 3. The Harbor of Baracoa, one of the most easterly indentations of the north

shore of Cuba. Admiralty Chart No. 435.
Fig. 4. Alfred Sound, the reef harbor at the northwestern e.xtremity of Great

Inagua. Admiralty Chart No. 2022.
Fig. 5. Port Guantanamo, a fiask-shaped harbor on the south shore of Cuba between

Cape Maysi and Cape Cruz. Hydrographic Chart No. 377''.
Fig. 6. Port Banes, one of the flask-shaped harbors of the north shore of Cuba,

Admiralty Chart No. 420.
Fig. 7. Port Padre, a characteristic flask-shaped harbor of the north shore of Cuba.

showing the separation of a part of the shore land by the narrow passage

connecting Padre and Malagueta forming a cay still closely joined to

the main island. Admiralty Chart No. 419.

PLATE XV.

^olian Pocks at the foot of the Queen's Staircase, Nassau.

PLATE XVL

Distant View of Nassau across the Grantstown Plain

PLATE XVII.

Eroded Mangrove Flat off a beach on the south side of New Providence.

PLATE XVin.

North Shore of New Providence near the Tea House, Nassau.

PLATE XIX.

Inland Vegetation. Pine Barren with Palmettos near Lake Cunningham, New
Providence.

PLATE XX.

Vegetation along the North Shore Road, New Providence, nenr Nassau.

202 bulletin: museum of compakative zoology.

plate xxi.

Green Turtle Cay, witli tlie adjoining Cays to tlie nortliward, Little Bahama Bank.

PLATE XXII.

Beach on Green Ca}', east side of the Tongue of the Ocean.

PLATE XXIIL

Great Lagoon, Watling Island.

PLATE XXIV.

Part of North Shore of Great Isaac. On the northwest corner of the west shank of
the Grrnt Bahama Bank.

PLATE XXV.

General View of a part of the South Shore of Great Isaac.

PLATE XXVI.

A Nearer View of a part of the South Sliore of Great Isnac. iEolian rocks greatly
honeycombed.

PLATE XXVII.

A Stretch of the Western Coast of Llcuthera, between the Glass Window and
Governor's Harbor.

PLATE XXVIII.
A Part of the West Side of Eleuthera, immediately south of the Glass Window.

PLATE XXIX.

The Cliffs of Eastern Face of the same part of Eleuthera, looking north towards
Harbor Island, from a point south of the Glass Window.

PLATE XXX.

Honeycombed ^Eolian Pocks on the West Fnce of Eleuthera near the Glass
Window.

PLATE XXXI.

Elbow Cay, the Doulile Headed Slmt Cays, and Salt Cay Bank.

PLATE XXXII.
Booby Island on the eastern edge of Douglas Channel, Great Bahama Bank.

PLATE XXXIII.
Morgan's BlufT, near the northern extremity of the eastern face of Andros Island.

PLATE XXXIV.

Stretch along the western shore of Cat Island.

AGASSIZ: BAHAMAS. 203

PLATE XXXV.

Flamingo Cay, on the eastern edge of the southeastern part of the Great Baliama
Bank.

PLATE XXXVL

The Cays of Conch Cut, on the western side of Exuma Sound, Great Bahama Bank.

PLATE XXXVIL

A Part of the Western Shore of Crooked Island north of the French Wells (Crooked
Island Bank).

PLATE XXXVIIL

A part of the West Shore cf Fortune Island, south of the settlement (Crooked
Island Bank).

PLATE XXXIX.

Memory Rock, on the west side of the Little Bahama Bank.

PLATE XL.

Great Abaco Light and the Hole in the Wall, southern extremity of the Little
Bahama Bank.

PLATE XLI.

The " Yunque," west Of Baracoa, Cuba, seen from the harbor.

PLATE XLII.

Yumiiri Cafion, entrance to the Yumuri Valley, seen from the east shore of the
harbor of MataiiZMS.

PLATE XLIIL
A part of the Yumuri Valley.

PLATE XLIV.

Elevated Reef (First Terrace) running along the north shore of Cuba between
Matanzas and Havana, near the Canasi River. Cliffs composed of older
limestones.

PLATE XLV.

Entrance to Havana Harbor. Soboruco Reef in foreground.

PLATE XLVI.
Patch of Soboruco near Morro Castle, Havana.

PLATE XLVn.

Morro Castle, entrance of Santiago de Cuba. Patch of Soboruco at right-hand
corner.

AGASSJ7

T UK It A n A.>!i A S

ii^

w=«SS"

/

^

A .

"N

r ^

/ ..^ \.

^' '_- .--7 >*,^—" '

r -' .i^;;;i^^5 \ir^' ^

|fj,y

-, V

«^.

.0' o-„,,/^

- ■ ■ - ^

;;.^ ^/

i".,*i

.iibASSiz -Bahamas

fi-g leyU « On* >rUe,

s^l,„a. .OntXU,

j /

/ '

/

J

H Fie.3.

\

*\

"'^*'*»-.

^

X

\

\

ia\

\

\

\

\

■ii\

\ — 'i

StaLa^l

<)/.<• .»i:(.

j i ! 1/^ 1 1

■ ! i / \' ■ ; !

i i 1 / \; '

i ■ / \ ) ' I

■■! ! / \i :

-

\ i / 'V! i ; 1 : ! 1 i 1/ VJ

\ 1

Fi,6 \

j i

) j

"10!

♦ 1

Sealt¥U '^iftfe

}• Fi ^ S

R^.2.

•a'\ *«' W*9t tx>ny

Fi^.l.

yiO

gsty

"£•£■/

jKw. c^, ^o^'

^uU, C^ *,'<J,** ' '

J}eptha ui fathoms.

B3tet:M3MU«.

AGASS/Z BANAMAS.

PlML

From a Surve^t by LieuT. G.B. Lawrance \844-.

4o' 35'

7/1 ' , r- ,— , 1 1 1 1 1 1 ■ 1 1 i — 1 ■ '1 ■ ■' ' 1 1 1— , . T^'

^ -i- "* ^ 4t ^ 4i

Ai 44 4>. '^ ^^

S
5 ^ 5

41 4J 4t 4i

4s 4i 4i

45

s

4'. 4j <t 44

*^ ^ 4, -^ ^

44 4S
4-, K

4* ^4

44 « *» ^:'3

4i ^ ^.

44 4, '^ 4. -^^

''i 4-V

^ ^ -l 4J

41 - ^

^-^^KoU 4t ^-^^ ^-^

•f.-. 4i 44 *5 ^

^5 4i

^i 44

'^ '^ ^* 44 44

** '^ 44

44 -K

4-J ^ •.. ..• 4i
4S ^ 4j *i \j 44 <K

^. ^ ^ 44 - " ^' .,

44 ^. 44 4i *» -^i

^ ^ . '■'- * - H ., ., '• ., ., -

44 '..'f,,. 38 fins hole vo*:'d.amete,. ..2|; ■** ■^ 4j
* ' in eri potcV,. ■'^ ■■•••■

44 >'.*•' *• *-^ ^ '^ ^

*\ 44 ••.-"•I 4.J ^

/^ ^ ^ ^ *s ^

&

'■3s- ■'^ . . 45 44

83

5'il
76

40W««t Long:. 76

5-
35-

From a Survey by Comm" Edward Barnett 1844.

lO"

l(>-iO-

25'

—

6 *

5J

10'

-

—

,5* 6

6

kz-'

6 Si

H

6 Hue'

k

6
6i

,13i '♦HoU
— ' ci-1

" SI

6 6

a.wd g

SJ

N

a'

7

St

54

4

76'30'

Z9'

SounMn^s in. Fojthomj .

3
From a Survey by Ojmm? Edward Barnett 1844.

J

N.
22

O' 23' 38'

5t St ^^ '■•■
54 ^i 5t

N.
22

If
7i

54 5i
St

^^ 54 1 5t

, 5i if
5i ,, 54

/5'
/-*■

5i
6

54

Cs>i\d.'Hd.

sx

so' \Nest U<ng. 23' ^'

Sotxn^in^S tla ^AtnomS ■

SouJiAin^ tn. FaJOtjoms^

i>3ASsiz -Bahamas

PLATE ?;.

Iforixonfn/ Sr/r/r, J. 25 " 60 nautical miUs
ntfrcaf Scafe. I'^JOOO fhlhorns
Deptha in faihoms.

2 -Bahamas

1 i

•a .s

St. Nicholas k rrrrtA

1? Channel S "JBA

HorUonlnl .Srnle. 1.2o " 60 nautimlmiUa
Vnttcal Sntte. I'- 1000 faUioni «

Depths u, /alh«m<, ^^^_ 1^ ,__^_^

,..^ SIS -Bahamas.

Alfred Sountl

Fig-.l.

oiif ^^Uf = .¥'

Sniirxfin^.s i it feet

5 g a Afvel

Riming" >'•
Eof l^g'arlo.

^'uf ;it<iii li i\ Ilk

Gre at F.xunia
Harbour

F15.3.

Srn I.<!X-el

Fk/s. 2.:i.5: Soii/ir/iii '/s I II I'lilluims

Sea I^evet.

llorizoiih'l Srnle. J.^n " 'SO nmi/irnl niit,-:
Vftlirnl Scale, j"- JOOO lulhoilia
Depths ill fnjifoitis.

Ajassiz-Bahakas,

PLATE >ni.

7 lo s 9 13 15 IS 17

Horizonhtl Scale. 1.^6 - 60 nauUcal mil^
]e,lirnl Srale, 1' 1000 fathoms
Ueplhs ui /athoms.

BA:1A«AS

.^ortugtt

«- .,,_..-ol'-''

:^

b-*.«l

yo

oofi" V: '', ' ■

" ' . - ■ .1 t ,^j — y,

^-*<* \ ■*' -^

^

/

f \ ^

^"-VU ^ '

■^,C >^i

^'1?

CONTOUR MAP

OK THE

CARIBBEAN SEA

1885

I*tvparvtf /roin tUtfn fiirtushefl hy f^*"
U.SHydro^raphir Office,
pK based on, the deep-sea srtundXnefS of the

I'SCSStrBlaJre and. the U.SJ'.C.Str.Alhetfross.

//!l

-iitflotivinjlit '^ ^

-■«••

■J h^/ /

u:

^^

• ' 'r^..

-^^^V

JAMA \

D M 1 N C t

c>^'

\

\ '. / \.

I'' M(.\"C.Uj^

j i .' / / \\

6, / / ,'.0; / '. I * «onOttd«.r^Jt

^ , \ft';Birai

TANNER D ceP

..k

S^-^?^:^^^^^'^

Tbcta^

^MartBJ-ita T

'jifi^i>e4

\ Trinidad

\ /

\\

>^-"

.■■■■f

Af.VAIM>K>

/

Fig I ,j ' T

S,U,..n 1-. Jf,*^^,,.,,,, |„. r,

)/

-Ai

\^

...

c, •(■ I \ A r, r A

i s I. .\ X o

A I i: II s

B .1 A- l<

■i

.» Mi.liUi Inn rt*iK* jB

, i{,]ji..,s~. . ^'•■•«5,';„'

F,.2 |«v.,s„„«sw

,\,„li..,jvi> l.nm-'

.\fttrr ijrj i /.■ n .\ y /i

Fig 6

(!)•

^>^4.

.\fAl*n: i: AN \ ISLANIl

s«i.. i--

.««*»» vaU"/

s.,:,.>C^-

T ,kl I^^V

/ /. \' /:■ I:

ti .V .V /I

i^ij..

SinriliKin-

DEPTH IN FATHOMS

3 /ril/inm.i

10
100

F.g.lZ

i/3»

Conc.-pUou ra

, Fig-iO

Agassiz-Bahavas,

CLARE IVOK
If-VHBOXJR Ci5^

coy N^

#^S£^^^

'^-O k. Fi

, 'm^^^ AVasKerwomans

man^Tiet ify ttbe^it here.

f « *^ i'fe/

l^uQi Cav

^o^"4S

^^i /.J". J. 7 soundings in iiet
otkers in fixihams

,^

'S? <q;

1 N Wv

^^■^■"

/ '-^■^'

../■■■■■

/ '-■

\( N

I

f.«<f

A;

•Trfevl

■--.V

/ 7

it'

. E ■

'5 - i?>r^-a',;-

'■:-i-^.^^'.

wm'//'///'//''/y/////"''////////////////f

<

en

8

<

I — I

o

>

X

<

-J

Oh

<

■J.
•S.
<

Z

UJ

>

h
z
<

5

X

w

<

%

1

n -mm

I

'I

c

m

N3

'ir.

in

<

f

>

O

a:

z

u.'
<

I — i

UJ

>

o
or
o

<

■A

U

z

UJ

Q

>

o

UJ

z

UJ
X

h

q:

,n

z

UJ

u
w

Q

>

O
CC

a.

w
2

Z

g
<

h
O

UJ

>

Q
2

<
_l
2

w
o

%

Q

>

o

Oh

o

► — '

<

H
W
O

w
>

w
o:
c

00

X

<

(T.

ir.
■f.

■3/".

>

<

u

UJ

_1
h

h

2

UJ
UJ

cc
o

<

X

X

tl

Q

<

en

CO

O

z

h

2

6
o
<

-J

cq

u

<

<

■f)

<
w

O

l-U

o

LU

o

O

c-J

<I1

iC

<
<

<

q:
o

o

q:

o
I

h
D
O

■s.
r.

<
<

<
UJ

cc

o

o

u

O
J-

CO

X

f ,

b
o

en

X

f^l

m

<

UJ

X
h

UJ

_j

UJ

•X.

O

h
in
<C

O
u

h
CO

UJ

br.

vi;

>

X

to

<

U

h
D
UJ

u

cc

bJ

i:

h:
o
z

o

CJ

h
CO

X

"i^

<

<:
w

H
W
uj

o
o

cr

Q

o

o

w
o

X

X
X
X

<

a,

I

en

<

o

CQ
UJ

cr.

<

<

m
o
o

ri

t/2

^^fV^'

'-0

q:

Q
<

Ll.

Ij
m

c/;

z

<

a:

'i f

•• ti

V »

<

J

■r.

r-
X

e

IT.

Q

<
CO

<
O

UJ
CC
O

d:

CO

h

CO.

X.

<
u

O

o

0-

r.

h

D

L)

O
D

Q
Z
<

Q
UJ

^r
O

o

q:

u

uJ
o

CO

u

>

X

X
X

bd
r"
<

CL,

in

IT.
(T.

<

Q
<

00

z

D
H
QC
O

ui
o

CO

h

X

<

X

O

o

o

X

I

X

o

o
o
<

<

h
<:

w

o:
o

iif «''t

X

4i

UJ
D
O

z

D
>

5j:

<

_1

X
<

<

O

<
o

D

S

D

x:

<
Oh

c

a;

;/:

<

>

iSISfe.;;?S?3

X

^Ai

<

CQ
D
U

Lu

Q
UJ

a:
o

X
a:
z

UJ
Lj-'

q:

Q
UJ

<
>

u
-J

UJ

K
O
DQ

<

X

<
2
<
>

<
X

o

H

UJ
O

z

<

h
w

<

z
<
>

<
X

UJ

h
CO
<

o

O

or
q:
o

o

U

D

cr
O

CQ

O
CO

u,
O

3:

u

<

PQ
D
O

W
P

O
O
<

h

<

m

O
H

W
O

o

(X

a:
o

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXVI. No. 2.

A VISIT TO THE BERMUDAS IN MARCH, 1894.

By Alexander Agassiz.

With Thirty Plates.

CAMBRIDGE, MASS., U.S.A.:

PRINTED FOR THE MUSEUM.

April, 18?5-.

TABLE OF CONTENTS.

General Description 209

^OLiAN Hills and Dunks. —Plates IX. to XII., XV. to XX., and

XXVIII 221

Fossils 229

The Sounds and Lagoons. — Plates II., IV., VI., VII., XIV., and

XXVII. . . : 2.31

Distribution of the Corals 235

Ledge Flats and Patches. —Plates IL, XV., and XVIL to XXVL 237

The Serpuline Reefs. — Plates XXL to XXVI 253

Pot-Holes. — Plates XXIX. and XXX 269

North Rock. — Plate VIII 270

Proto-Bermuda. — Plate II 273

Explanation of the Plates 279

No. 2. — A Visit to the Bermudas in March, 1894. By
Alexander Agassiz.

GENERAL DESCRIPTION.

Before completing my article ou the Bahamas I was anxious to visit
the Bermudas. During my visit I chartered a sea-going tug, and was
thus able in a comparatively short time to cover every interesting spot
on the shores of the islands and on the inner and outer ledge flats.
During the spring of 1894 I spent about a month in their examination/
and find that the story of their present condition is practically that of
the Bahamas, with the exception that at the Bermudas we have an
epitome as it were of the physical changes undej-gone by the Bahamas.
One cannot fail to be struck with the insignificance of tlie corals as com-
pared with those of Florida, of the Bahamas, and of the Windward Isl-
ands. It is true that on the ledge patches inside of the so called " ledge
flats " the Gorgonians and Millepores are very flourishing, but the devel-
opment of the true reef builders, of the massive corals, is insignificant ;
while the absence of Madrepores is remarkable, and changes the whole
aspect of the coral growth.

I shall have little to add to the description of the Bermudas as given
by Nelson, Rein, Thomson, Rice, and Heilprin, but I am inclined to take a
different view of the part which the corals now growing there have played
in the formation of the reef ledge flats. The corals have not added any
material part to the reefs ; they form only a thin veneer over the disin-
tegrated ledges of feolian rock which constitute the so called reef off"
the south shore of Bermuda and the ledge flats of the reef ring near
the outer edge of the Bermuda Bank. yEolian rock ledges underlie the
coral growth not only on the patches off" the south shore and on the
ledge flats of the outer reef, but they also underlie the so called patches
and heads forming the flats which extend on both sides of the main
channel and divide the lagoons or interior waters of the bank into irreg-

1 Notes from the Bermudas. By Alexander Asrassiz. From a letter to Professor
James D. Dana, dated Bermuda, March 12, 1894. American Journal of Science,
3d ser.. Vol. XLVII No. 282, June, 1894.

VOL. XXYI. — NO. 2. 14

210 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

ular sounds, like Murray Anchorage. The passage of the shore seolian
rock ledges into the coral patches can easily be traced, off both the
north and the south shores, as will be seen later on.

I have to thank his Excellency, the Governor of the Bermudas, Gen-
eral Lyon, for permission to make soundings and dredgings among these
islands with the view of studying the coral formation. To the Hon.
Archibald Alison, Colonial Secretary, I am indebted for assistance in
many ways during my visit to the Bermudas, especially in obtaining in-
formation from the government officers, and for a fine specimen of float-
ing pumice stranded upon the south shore ; and to Captain Can-, R. N.,
in charge of the Bermuda dockyard, for information regarding Ireland
Island and the flats. I have to thank General Eussell Hastings and
the American Consul, the Hon. Marshall Hanger, for their interest in
my explorations, and to Mr. John C. Watlington 1 am indebted for
statistics regarding the temperature of the sea water at different sea-
sons of the year.

The slope of the mountain of which the Bermudas are the summit
varies considerably, judging from the three sections given on Plate II.
The slope off North Rock (Plate II. Fig. 3) is steeper than the slope
off Castle Harbor (Plate II. Fig. 4). Off North Rock the distance from
the 100 fathom line to a depth of nearly 1,400 fathoms is about six
miles, while off Castle Harbor the 100 fathom line is nearly eiglit miles
from a depth of less than 1,250 fathoms. Ott' the Argus Bank the 100
fathom line is about ten miles from a 1,370 fathom sounding (Plate II.
Figs. 1, 2). These sections show the slope of the island to be steeper
off the north face than on the south side of the island.

On the southwest face off Long Bar 1,250 fathoms is found at a dis-
tance of five miles from the 100 fathom line, and the 1,000 fathom line is
only two miles from the 100 fathom line. Northeast from the East
Ledge Flats 1,2G0 fathoms is found at a distance of five and a half miles
from the 100 fathom line. South-southeast from the Southwest Breaker
960 fathoms is found two and a half miles from the 100 fathom line.

The distance of the 100 fathom line from the sea edge of the ledge
flats varies but little, thtmgh it is true that off the south shore it is
nearer as a whole, and it comes to within one mile and a quarter of the
shore off Castle Harbor, and on the north side it is as much as three
and a half miles from the Nortli Ledge Flats, and east off Mills Breaker
it is nearly five miles (Plate 1.).

The Bermudas (Plate I.) form a hook-F<haped line of islands, the main
island running northeast from Gibbs Hill to Castle Harbor, which is

AGASSIZ: BERMUDAS. 211

bounded on the north by a chain of islands of which St. David is the
most prominent, and on the east by a chain of smaller islands, the con-
tinuation of the spit east of Tuckerstown, which is the easternmost point
of the main island. Outside of Castle Harbor on the north extends
a chain of islands which bound St. George Harbor, the largest of
which is St. George Island. Endless islets and rocks flank the south
shore from Tuckerstown to Church Bay, where tlie main island makes
a sweep to the southwest as far as Wreck Hill. It is separated by a
narrow chamiel from Somerset Island, to the northward.

From Somerset Xarrows a chain of small islands extends to the north-
east, on the last of which is built the dockyard. The north shore of the
main island is flanked by numerous islands, which form Hamilton
Harbor, Port Royal Bay, Great Sound, and the connecting waters to
the westward of Spanish Point (Plate III.). At tlie eastern end of tlie
main island a narrow channel opens into Little or Harrington Sound,
which covers the greater part of that end of the island. The position
of the former Bermudian land is indicated by isolated rocks, like North
Rock (Plate VIII.), the Pilchard Dicks, the Southwest Breaker, the
South Reef, the Mills and Northeast Breakers, and others rising above
the general level 6i the broad belt of ledge flats which completely sur-
round the summit of the Bermuda mountain. These flats leave only
here and there a narrow passage into the interior sounds, bounded by
the many belts of flats crossing the inner waters. The principal passages
are the Narrows, or Ship Channel, Hog Fish Cut, Chub Cut, North
Rock, and Mills Breaker channels, in addition to a few insignificant boat
passages. The Narrows is really the only channel navigable for heavy
draught vessels.

The Lagoon, lying to the northward and northwestward of the
islands, is bounded by the curve of the outer ledge flats. They are sub-
merged at low water, except at a few points such as the North Rock,
Mills Breaker, Southwest Breaker, and others marked on the , chart
(Plate I.). The depth of the Lagoon is in general from seven to nine
fatlioms, though a few of the deeper points are twelve or thirteen
fathoms. Between the outer ledge flats and the islands are found the
many secondary flats, called Elies Flat, Cowground Flat, Brackish Pond
Flats, Green Flats, Bailey Bay Flats, Three Hill Shoals, and the like,
which consist of endless patches of ]\Iillepores and Gorgonians, reaching
many of them to within a few inches of the surface. The Gorgonians
and Algae which cover the patches have grown upon the remnants of
ledges of the proto-Bermudian land that attest to its former existence

212 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

at so many points, not only of the outer reef ledges, but on the flats inter-
vening between them and the Lagoon (Plate XVIII.). Between these
patches separated by deep water, and plainly seen by its discoloration, it
is comparatively easy to find one's way when the water on the banks
is not rendered millvy by winds stirring up the bottom.

Lagoons similar to those between the reef and the islands are Great
Sound, Port Royal Bay, Hamilton Harbor, Harrington or Little Sound,
St. George Harbor, and Castle Harbor. These soxmds, as the inner
lagoons are called, differ from similar sinks in the outer lagoon ])y being
bounded in part or wholly by land, while the sounds in the outer lagoon
are more or less indistinctly limited by flats formed of "coral head"
patches (Plate I.).

Harrington Sound is connected with the outer lagoon by a narrow
channel, the Flats Inlet; it is, in fact, in the condition of many of the
smaller bays on the south shore, where the sea has only corapai-atively
lately encroached upon the interior sinks. I refer to such bays as
Sinky, Hungry (Plate XV.), Devon, and the like. In the case of Little
Sound, the opening at Flats Inlet broke through the ridge separating
the outer lagoon from an extensive sink, the smaller elevations of which
have entirely disappeared, with the exception of a few islands and shoal
patches. Castle Harbor and St. George Harbor were probably similar
sinks, the outlines of which are still indicated by the line of islands
protecting in part the southern face of Castle Harbor, while tlie out-
line of the narrow sink forming St. George Harbor is shown by the
islands which separate it from the Narrows, and by St. David and
Long Bird Islands, which are the remnants of the ridge dividing St.
George from Castle Harbor. But both these ridges are now broken
through, so that Castle Harbor connects with St. George Harbor and on
the north with the lagoon forming Murray Anchorage.

The feolian hills extending eastward from the northern entrance to St.
George Harbor are most characteristic, and the saddles separating them
are of varied elevations, and show how readily the sea could, after a
very limited subsidence, find its way into sounds like Castle and St.
George Harbors at a time when they probably appeared much like the
nearly closed Harrington Sound (Plate V.).

As Professor Heilprin has stated, it is merely a question of time when
Harrington Sound will be more of an open bay than of a land-locked
lagoon, presenting in time the appearance of Castle Harbor, and finally
perhaps that of Great Soimd.

An examination of the chart (Plate I.) reveals faintly the position of

AGASSIZ: BERMUDAS. 213

similar lagoons, the limiting land of which has been entirely washed away.
Such are the three small lagoons situated in the flats to the south of East
Ledge and to the west of Mills Breaker, showing channels leading into
them from the bank similar to the Narrows which lead from the edse
of tlie bank to the lagoon forming Murray Anchorage. A number of
such open lagoons facing south are found extending into the inner
■?dge of the reef flat, to the eastward and westward, directly south of
North Rock, as far west as the Eastern Blue Cut. On the outer face
of the reef south of Western Blue Cut, at the Chub Cut, the Chub
Heads, and between High Point and the Chaddock Bar, a number of
such open lagoons can be traced. Long Bar is a flat ledge which
once may have formed the barrier of a narrow lagoon open at its two
extremities.

The so called reef off tlie south shore I look upon as a series of ledges,
the remnants of the cliffs of the shore when it formerly extended to the
present position (Plates XXL, XXIII.) of the reef One cannot fail to
read tlie mode of formation of this I'eef on seeing the work of destruc-
tion which has been and is still going on all along that coast (Plate XIX.).
The shore of the island is gradually being eaten away at all the low-
points leading either into sinks like those of Sinky or of Hungry Bay, or
into more elongated sinks like those which will be formed when the ponds
lying close to tlie shore to the westward of Tuckerstown are invaded by
the sea (Plate XIV.). The next process is the formation of a line of
islands, such as still protect Castle Harbor and St. George Harbor on
tlieir sea face (Plate XXL), oi', as is well seen at Sinky Bay, where the
line of rocks to the north and south of the opening is still connected
with the shore line by a beach or neck, but which will soon disappear
and change that bay and the one to the north of it into an extension of
Whale Bay. On the sea face of that part of the coast extends a long line
of isolated rocks, islets, rocky patches, and sunken ledges, which plainly
tell of their former connection in a continuous ridge. Parts of these
ledges are worn to the water's edge, forming flat ledges or mushroom-
shaped rocks overgrown with Algse and Serpulte, and likewise the Ser-
pulfe atolls (Plates XXIV.-XXVI.) and boilers of the shore line, similar
to those which form the outer reef, and are separated by a belt of water
varying from one and a lialf to four fathoms close to the lee side of the
reef. At some points of the shore it is difticult to separate the line
of the outer reef, and of its ledges extending towards the shore, from
the ledges wliich form the Serpulse reefs of the shore line itself
(Plate XXVL).

214 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

The amount of material which is kept in constant movement by the
action of the sea within the outer line of ledges is very great, and is con-
stantly increased by the additional material derived from the breaking
up of the outer ledges by seas of unusual violence. The outer as
well as the inner ledges become disconnected ; passages between them
are opened with four fathoms of water, or even more. The outer
slope of the ledges is greater ttian the inner slope, the depth of water
increasing in some places rapidly from the ledges which are awash at
low water to six or seven fathoms, or even up to twelve off Castle
Harbor. Fi'om that depth, judging from the soundings on the hydro-
graphic charts, the slope is quite gradual, twelve fathoms usually being
found at a distance of nearly a mile from the outer ledges ; off the outer
side of the western reefs the slope is somewhat steeper, and off the north-
eastern face of the reefs from the Western Blue Cut to North Rock the
slope is slightly flatter.

Two small disconnected banks exist to the southwest of the Ber-
mudas with a least deptli, the one of twenty-one, the other of twenty-
four fatlioms (Plate 11.). A depth of over 1,000 fathoms has been
sounded between the main Bank and the Challenger Bank, while one
of 580 fathoms has been oi)tained in the channel between it and the
Argus Bank. The bottom in these is stated to be coral sand, and these
banks may, like the Bermuda Bank itself, be the summits of volcanic
peaks which have risen from a greatest depth of more than 2,000 fathoms
at a distance of about nine miles from the 100 fathom line.

The Bermudas present quite a difterent physiognomy from that of the
Bahamas. The hills are more diversified in shape, many of tliem quite
conical, surrounding lowlands or sinks of considerable extent and of great
variety of shape. Along the northern shore of St. George Island, ex-
cept towards the east end, the hills are undermined as they reach the
water line, forming low cliffs in marked contrast to those of the south
shore, or on the waters of Harrington Sound and St. George Harljor and
the northern coast of Castle Harbor, where in certain places the water has
undermined the cliffs, and from them have dropped off huge masses leav-
ing vertical sides of considerable height, from fifty to seventy feet. On
the north shore only a few short stretches of the coast are occupied by
sand beaches. Shelly Beach is the only one of considerable length. On
the south shore we find sand dunes and many long stretches of broad
sand beaches (Plate XL) wliich sup])ly the material for the dunes that
are still in some places forcing their way inland over the ancient dunes,
as at Tuckerstown, Elbow Beach (Plate XII.), and Whale Beach.

AGASSIZ : BERMUDAS. 215

Professor Heilprin is of the opinion that perhaps the natural arches at
Tuckerstown were formed at a time when the relations of land and water
were different from what they are now; though it is difficult to ascertain
precisely what his views are. Here, as at the Bahamas, many sinks occur
in which, owing to the porosity of the rock, but little water can remain;
if the sinks are at some height above the sea level, this undoubtedly per-
meates the interior of the island, and whenever the sinks reach below
the level of the sea they are filled with brackish water. There is very
little difference in the appearance of the rocks composing the islands.
We have by no means the variety in tlie appearance of the seolian rocks
met with in different islands of the Bahamas.

Solution has undoubtedly played some part in producing many of the
fantastic spires of seolian rocks one meets in the Bahamas and Bermudas.
The undermining of the shore cliffs, the mushroom shape of many of
the isolated rocks and sliore ledges, both above and below low-water
mark, is in part due to the solvent action of sea water. This is
especially well seen in the Baliamas on every island or islet at many of
the channels leading from tlie sea face on to the bank.^ On the Ber-
mudas this is perhaps best seen in the formation of the pinnacles in the
comparatively quiet waters of Mullet Bay and of Castle Harbor, and in
the undermining of all the shore cliffs and the cavernous and honey-
combed condition of the older ledge patches between the islands and the
reef flats and those of the ledge flats themselves. Close to the Causeway
at the northwest part of Castle Harbor there is quite a patch of low pin-
nacles from two to two and a half feet in height, which seem to show an
active solvent action by the sea.

But the solvent action of the salt water cannot be compared in effi-
ciency with the destructive mechanical action of the sea ; the latter has
to a great extent been arrested by the covering coat of Gorgonians, Mille-
pores, Algi3e, and Corallines, as well as of the more massive corals found
thriving upon the heads, patches, ledges, and ledge flats of the inner and
outer waters of tlie Bermudas. But these heads, ledges, etc. do not, as has
been stated by former observers, owe their existence and their gradual
increase to the corals, as they consist of teolian rock with only a protect-
ing veneer of corals over their surface, constituting a coral gi-owth, and
not a coral reef.

In the region of the Everglades of Florida,^ the process of solution of

1 See A. Agassiz, The Bahamas, Bull. Mus. Comp. Zool., Vol. XXVI. No. 1,
1891, p. 49.

2 The Topography of Florida, by N. S. Shaler, Bull. Mus. Comp. Zoiil, Vol. XVI.

216 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY,

the limestone rocks has been an important factor in shaping the general
plane of the region. No one can find their way any distance into the
Everglades without being struck with the deeply corroded and honey-
combed aspect of the rocks, and the numerous sink-holes, due to the
effect of the rain water saturated with acids derived from the decaying
vegetable matter.

Heilprin has well shown that the lagoons and sounds of the Ber-
mudas are not kept open through solution, and do not owe their origin
or increase to that cause. ^ But I think he has underestimated the effect
of solution on the cliffs and ledges where exposed to the action of the sea.
The solution effected by the percolation of fresh water is clearly seen in
the sinks,^ pot-holes, and caverns opening out on all the cliff exposures
and in many of the cuts in the honeycombed surfaces of the feolian
rocks wherever laid bare. ~ There is hardlv a vertical wall or cut for a road
which does not show some trace of the solvent action of water percolat-
ing through the seolian beds and covering patches of the edges of the
strata with a stalagmitic coating, so as to obliterate their dividing lines
or form small stalactites from bed to bed.

A comparison of the base of the sea faces of the cliff ledges with the
sides of the mushroom-shaped rocks below low-water mark will clearly
indicate the different kind of work accomplished by solution by sea
water acting upon the more or less submerged vertical faces, and that
exhibited by the action of fresh water some yards above the high-water
mark. The effect of the solvent action of the sea water is readily traced
above the high-water mark as fiir as the waves or spray can reach, and
the encroachment of the sea water upon the area partly honeycombed by
fresh water is most instructive (Plates XXVIIL, XXX.). Above high-
water mark the area exposed to this action is very considerable, and by
the solvent effects of sea water upon the limestone area exposed be-
No. 7, 1890, p. 145. See also Murray and Irvine on Coral Reefs and other Car-
bonate of Lime Formations in Modern Seas, Proc. R. Geog. Soc. Edinb., 1889-90,
Vol. XVII. p. 79.

1 Bermudas, p. 44.

2 Professor Dolley accounts for tlie formation of banana holes by the action of
decaying vegetation collected in the holes, kept moist by the action of rains under-
going fermentative changes by the product of which the soft calcareous rock is
dissolved and leaches away. This process undoubtedly acts as suggested by Pro-
fessor Dolley, but only to a limited extent, as the most active honeycombing takes
place on the surface of barren islands, where the vegetation has long ago disap-
peared, but where the surface is exposed to the combined action of rain and of salt
water spray, as at Glass Window, Great Isaac, and other places in the Bahamas.
See A. Agassiz, Bull. Mus. Comp. Zcxil., Vol. XXVI. No. 1, 1894, pp. 30. 60.

AGASSIZ: BERMUDAS. 217

tween high and low water mark a very considerable amount of lime must
be removed, probably as much as the sea water will hold in a saturated
solution. Additional lime may also be taken in solution, while in-
numerable minute particles are held in suspension during rough weather.
At the time of our visit, the wind having blown for two days quite per-
sistently from the northwest, the whole bank was fairly milky white in
a belt extending from five hundred to seven hundred feet from the north
shore. The isolated patches on the edges of the greater ledge Hats were
also indicated l^y tl\e white waters surrounding them, while the deeper
parts of the outer sounds were indistinctly indicated by a more bluisli
tint of the water.

The so called patches and reefs appear like diminutive vertical cliffs,
or parts of cliffs, which once were broken off from tlie larger shore line
cliffs then existing. Of course their faces and surface have been to a
certain extent modified by the growth <jf corals upon them, but tliat
only to a very limited degree. The coral patches are built up over a
substratum of seolian rocks.

The islands as seen fi'om the north present in general the features of
the seolian hills of the Baliamas, although they have a more varied out-
line (Plate IV.). There are more conical hills, and there are not as
many of the distinctive lines of ranges of reolian hills trending in one
direction. Many of the islands in Hamilton Harbor are quite bare, their
surface indurated by the action of rain and more or less honeycombed,
as is the case in the majority of the Bahamas. The absence of vegeta-
tion is also marked over a great part of the western extremity of Spanish
Point. There is a very fine a^olian cliff on the bay at the foot of Admi-
ralty House (Plate XVI.).

On the north shore the trees and shrubs of the main island do not run
close to the water's edge. From the somewhat scantier groves of ju-
niper crowning the summits of the successive ?eolian hills run broad
grassy undulating slopes, terminating in the comparatively low vertical
cliffs which characterize the north shore of the principal island. All the
way from Spanish Point to Harrington Sound there is an unbroken line
of these low vertical cliffs (Plate XXVITL), with the exception of a
few insignificant sandy beaches breaking in occasionally, the only promi-
nent exception being Shelly Beach. The islands of St. George and St.
David are comparatively bare (Plate V.), and the general aspect of the
country at the eastern extremity of Bermuda reminds one of the aspect
of the long lines of barren islands so common in the Bahamas. On the
southern and western slopes of St. David junipers grow more abun-

218 BULLETIN : MUSEUM OF COMPAKATIVE ZOOLOGY.

dantly, and the narrow band separating Castle Harbor from Harrington
Sound is comparatively well wooded (Plate YI.). Though there has been
over that district a greater denudation, judging from the general appear-
ance of the outcrops of isolated patches of weathered seolian rocks, which
form here and there low vertical cliffs.

The hills to the eastward of the flats extending on the north shore to
St. George Island and the islands protecting the entrance of the harbor
are comparatively bare of vegetation. The few stunted cedars and other
bushes growing on the north side of that part of the Bermudas ai'e mainly
limited to the southern slopes. In fact, all the eastern part of the Ber-
mudas, including the territory around Harrington Sound, shows far
more than the central part of the principal island the effect of denudation
which has taken place both there and in the western district of the
group. The aeolian rocks crop out in all directions, greatly weathered,
and are near the surface changed to hard ringing limestone.

Toward the summits of the hills and in the saddles passing into the
interior of the main island the cedars are more abundant, and in the
lower and better watered valleys palmettos grow in small groves, form-
ing the same contrast with the Bermuda cedars which they do in
similar localities with the pines in the Bahamas. In all directions the
wild verbena is to be found, either as the only growth in the more bar-
ren districts, or encroaching to a great extent on the open spaces of the
wooded parts of the islands. In the valleys and lowlands of the islands,
the soil is sometimes of considerable thickness. This is especially the
case in the series of longitudinal sinks parallel with the south shore,
which extend from Hungry Bay through the greater part of Devon.

At the western as at the eastern extremity of the islands we find a
similar diminution in the vegetation. The southwest end of Somerset
Island is bare, and the surface greatly worn. On the rest of the island
there are fewer junipers than on the main island, and they are smaller
and more scattered, as they are on the east end of the main island
all the way from Wreck Point to High Point.

The difference in the aspect of the Bermudas and Bahamas is per-
haps due to the prevalence of the trade winds in the latter, where they
have had the tendency to build sand dunes bearing in one direction, their
sea slope being abrupt, while the bank slope is more gentle, thus forming
long lines of dunes lapping and running in one direction. This leaves
long, narrow valleys between the ranges of dunes, and, as is well seen at
Nassau and Andros, produces a greater monotony in the outlines of the
islands as compared with the varied landscape of Bermuda. These valleys

AGASSIZ : BERMUDAS. 219

often become sinks (Plate XIV.) or elongated pools, and when broken
into by the sea soon leave a line of cays parallel to the main shore.
The winds are more variable at Bermuda, the islands being several
degrees north of the limits of the trades, while the Bahamas are on
their northern edge.

The general aspect of the Bermuda vegetation is characterized by the
presence of the Bermuda juniper, which has assumed in these islands
the prominence which the pine has taken on the Little Bahama Bank,
on Audros, and on ^"^ew Providence. Comparatively few of the plants so
characteristic of the shores of the most barren of the Bahamas are met
with near the shore line.

The Alga; and Corallines which I collected at the Bermudas were
kindly examined for me by Professor Farlow. Off the sand beaches of
the south shore on th6 bottom of the interior sounds the calcareous Algae
consisted mainly of species of Peuicillus, of Bostricha, of Udotea, and
of Halimeda, identical with West Indian types. On the rims of the
serpuline atolls (Plates XXIV.-XXVI.) were collected species of Turbi-
naria, Galaxaura, Blodgettia, Dasyclades, Codiuni, Laurencia, Dyctyota,
Eucheuma, Sargassum, Zonai-ia, Caulerpa, and Janca, all of which also
occur on the outer ledge of flats and inner patches.

Professor Moseley collected a number of marine plants at the Ber-
mudas which have been described in the Journal of the Linnaean Society,
and his collections of the flowering plants formed the basis of the ex-
tended Report on the Botany of the Bermudas by Hemsley,^ where a
full account of the earlier sketches of the flora of these islands will be
found. Professor Ciiarles S. Dolley has also given an account of the
Botany of the Bahamas.^

The principal accounts we have of the geology of the Bermudas are
those of Captain Nelson.^ The geology in Jones's " Naturalist in the
Bernnidas" (1859) is taken mainly from Nelson's Report. Rein gave a
most interesting sketch of the geology of the islands in the Bericht. ii. d.
Senckenbergische Naturf. Gesellschaft for 1870, page 140; he was fol-
lowed by Sir Wyville Thomson,* who spent a short time in the Bermudas,
and next in order came the visits of Rice,''^ of Fewkes,'' and of Heilprin.''
Darwin is of the opinion that the Bermudas "have a close general re-

1 Voyasre of H. M. S. " Clia]len!,^pr," Botany, Part I., W. II. Ilemsley, 1884.

2 Proc. Phila. Acad. Nat. Sciences, 1889, p. l.m

3 Trans. Geol. Soc. of London, V. 103, 18-37.

* The Atlantic, I. 289, 1877. 6 Proc. P,ost. Soc. Nnt. Hist., 1888.

5 Bull. U. S. Nat. Museum, No. 2-5, 1884. ^ Tlie Bermuda Islands, 1889.

220 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

semblance to an atoll," although they differ from one in several respects
wliich he enumerates. Dana and Sir Wyville Thomson regard the Ber-
mudas as part of an atoll. Thomson thought the islands were formed " ])y
the raising of the weather edge of the reef above the level of the sea." ^
Professor Rice also considered it an atoll, but he was careful to distin-
tinguish between the present outlines and those which belonged to the
original atoll.

Professor Rice suggests the following heads as his explanation of the
geological history of the Bermudas : —

" 1. A subsidence in which the original nucleus of the island disap-
peared beneath the sea, the characteristic atoll form was produced, and
the now elevated beach rock was deposited.

" 2. An elevation in which the great lagoon and the various minor
lagoons were converted into dry land, and the vast accumulations of
wind-blown sand were formed which now constitute the most striking
peculiarity of the islands.

" 3. A subsidence in which the soft drift rock around the shores suf-
fered extensive marine erosion and the shore platform and cliffs already
described were formed."

As regards 1. This is the natural explanation which would be given
by the Darwinian theory of the formation of coral reefs to account both
for the disappearance of the original nucleus and for tlie formation of an
atoll. To those who do not accept the theory the disappearance of the
nucleus is of course explained by subsidence also, but by the subsidence
which followed the formation of the dunes from an extensive ring-shaped
coral sand beach of which the material was derived from a reef growing
upon the upper plateau of the Bermudian mountain, from a depth of less
than twenty fathoms.

2. Whenever the accumulations from tlie reef were sufficient' to
build up a beach reaching the surface, all the conditions necessary for
the formation of sand dunes existed, and we need not call upon either a
subsidence or an elevation to account for tlie existing condition of the
Bermudas.

3. The subsidence which I imagine to have taken place after the
building of the dunes from tlie broad beach surrounding the original
nucleus — now sunk or disintegrated — is, it seems to me, quite suf-
ficient to explain the existing condition of the dunes and cliffs of the
Bermudas, if the interpretation I have given of the base rock is the
correct one.

1 The Atlantic, I. 302.

AGASSIZ: BEKMUDAS. 221

I shall have occasion to refer to the views of Heilprin in the course of
the followHjg pages. Regarding the statement he makes on the nature
of the Dolonntic reefs of the Tyrol, I would refer the reader to my Re-
port on the Bahamas, page 170, wliei-e 1 have given an abstract of the
views of the latest researciies on the nature of the dolomitic reefs, views
whicli are diametrically opposed to those advocated by him.

Heilprin says : " If it ever existed (the atoll condition), it has been
completely masked by overgrowth ; . . . the facts, such as they are,
show with sufficient clearness that the present islands and reefs have
little or nothing in common, beyond occupying position, with a pre-
existent ring." ^ Yet it is on observations gathered in a district thus
characterized by him tliat Heilprin bases his assent to tlie Darwinian
theory of coral reefs, and he dissents from those wht) hold opposite views
with a vehemence which might be excused in one having an extended
acquaintance with coral reefs.

^OLIAN HILLS AND DUNES.

Captain Nelson was the first to call attention to the ccolian charac-
ter of the I'ocks of the Bahamas and Rernmdas. This chai'acter saute
aux i/eax in every dii'ection. In the Bahamas the vertical clilfs of the
weatlier side of tlie islands show this to perfection, and here and there a
quarr}^ or a cut leaves no doubt that the substructure as well as the
superstructure of the island is all of the same character. On the Ber-
mudas one comes upon quarries of all sizes at all points, close to the sea
level or near the highest summits, and at all ])Ossil)le intermediate ele-
vations. Tlie rock everywhei'e ])i'esonts the same structure. There are
also endless rock cuts for the passage of roads (Plate IX.), giving excel-
lent exposures of the a'oliau strata twisted and turned in every possil)le
irregular manner according to the direction of the then prevailing winds,
or we may come across a patcli exposed in a cliff or in a deep cut where
the strata run parallel for quite a distance. As in the Bahamas, the
surfice of these Eeolian rocks has liecome indiu'ated by the percolation of
fresh water tln-onuh its mass, and has formed here and there the thin
riim'ing coating so comiiiou all over tliose islands, where the surface is
not so well protected by v(>getation as it is in tlie Beinnudas. Thi-ough-
out the islands avc come upon evidence of the extensive denudation
and erosion which have affected the a-olian rocks of the islands and

1 Bernuula Islands, p. 40.

222 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

worn them into the varied forms they have assumed, either along the
more exposed shores or in the sheltered bays and inlets and sounds.
Some of the seolian pinnacles off' Castle Harbor have assumed the most
fantastic shapes, due to the combined action of the weather and of the
solvent and wearing action of the sea and rain.

Sir Wyville Thomson -^ has also given an excellent account of the
general characteristics of the teolian formation. Heilprin has called
attention to the comparatively insigilificant part which corals play in
the supply of the material which has gone to form the a^olian hills of
the Bermudas, and which, as in the Bahamas, is made up of many other
organisms. Among them Nullipores, Corallines, broken shells, and
Millepores take a most important place. In some localities, where the
seolian rocks have not become well indurated, it is not infrequent to
have secondary dunes formed from the sand derived from the breaking
down of one of the softer cliffs, the dunes covering to a certain extent
the older eeolian hills, much as the seolian sand of the south coast climbs
over the faces of the older hills.

The fine co'ral sand, which is so often spoken of as washed up on the
shores by the sea,^ is not, strictly speaking, coral sand, but is primarily
composed of fine sand derived from decomposed a^olian rock. This ma-
terial is derived from the disintegration of the shore cliff" ledges, and
from that pounded off by the sea from the outer reef ledges, together
with the broken shells of the mollusks living upon the flats and the
small amount of material supplied by the breaking up of the massive
corals and Gorgonians forming the coral growth upon the ledges, the
ledges themselves consisting of seolian rock covered by Algte, Corallines,
Serpula?, and Millepores. On the south shore this fine sand is blown
far inland, forming dunes which cover extensive tracts;^ at Middleton
Bay beach they run up over the surfices of the older solidified dune's,
and reach to a height of over one hundred feet from high-water mark,
encroaching upon the vegetation near the lee summit of the saddle
through which they are lilown. A row of small dunes has formed on
the edge of the beach south of Whale Bay ; a larger dune has also been
piled up inland within the line of the beach dunes, extending over an
older but smaller solidified dune (ipoliau hill); just as the beach sands
at Elbow Beach (Plates XL, XIL) have run to a height of more than one
hundred feet, although here the sand dunes do not extend as far inland.

1 Voyage of H. M. S. " Cliallenger/' Tlio Atlantic, I. 310.

2 ll)ir1., I. r,07.

3 See tlie excellent accounts of the chines by Tliomson, Tbiil., .312.

AGASSIZ: BERMUDAS. 223

It is probable that the other non-calcareous rocks and minerals which
have occasionally been found may have been brought here by floating
trunks and roots of trees, as is tiie case in many of the oceanic islands.
We should, however, not forget the possibility of their being the frag-
ments of the volcanic summit around which the proto-Bermudiau reef
was first formed, a summit which has completely disappeared, either ow-
ing to subsidence or to disintegration, or to both combined. There exists
in the collection in the Government Building a piece of fine-grained
seolian rock of a reddish tint from the north shore near Warwick Road,
in which is embedded an angular fragment of basalt, or some eruptive
rock.

For a coral island the elevation of tlie Bermudas is very considerable.
The highest points are Sears Hill, 260 feet, Gibbs Hill, 240 feet, and
Piospect Hill, 222 feet ; a number of points reach an elevation of
nearly 150 feet. On the Bahamas, with the exception of the highest
points of Cat Island, which are said to reach 400 feet, the greater num-
ber of tlie seolian hills do not rise to more tlian from GO to 100 feet,
very many of the islands attaining a height of not more than from 20
to 40 feet, and only a few summits reaching over 200 feet. But it
should be remembered that the heights named are not due to the ele-
vation of coral reef rock, but to the height attained by the asolian hills
which constitute the dry land of the two groups.

The Bermudas and Bahamas ^ offer an example of the thickness that
a recent limestone deposit may attain during a period of rest. Assuming
for the Bernuidas a probable subsidence of 70 feet and a greatest ele-
vation of 260 feet, we get an asolian coral limestone of 330 feet in thick-
ness, the material of which has all come from a reef which itself was
^probably not thicker than 120 feet, or a total thickness of 450 feet.
When we remember how readily these coral limestones are changed into
hard ringing rocks, we introduce a new element into the discussion of
the mode of formation of huge masses of limestone, especially in the
region of the trade winds.

The beach rock and the so called base rock which have been observed
at the Bermudas belong, I believe, to tvvo different types. The former,
the beach rock, consisting of coral or other sand, is deposited in strata
dipping to the sea at a slight inclination, and is characteristic of all
coral reef districts where sand is accumulated along a shelving line of
coast. This frequently becomes hardened and changed to a ringing
limestone, and is composed usually of rather coarse particles, but not

1 A. Agassiz, Bull. Mus. Comp. Zoiil., Vol. XXVI. No. 1, p. 183.

22-4 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

necessarily so. The base rock, considered by some of the writers on the
Bermudas to underlie the ajolian hills, I look upon as the modified part
of the lower portion of the teolian strata chano:ed into a hard riui'inf
limestone in wliicli all traces of stratificAtion have often disappeared
(Plates XVL-XVIIL).

Heilprin argues tiiat the beach rock has been elevated and is still
found at an elevation of 12 to 16 feet; that it "antedates the last
subsidence, ... is at least as ancient as the lagoons and sounds, and
probably much more ancient. Indeed, there is nothing that could lead
one to suppose that it is not the original rock which was formed when
the island first came to the sm-face. Although now exposed on the sea
border, it is really an interior rock, as is proved by the broad band
of land which must have been i-emoved from the seaward side of the
existing cliffs." i But this neither indicates elevation, nor that it is an
original beach rock, since at the western extremity of the Bermudas, at
Ireland Island, it is underlaid by true a;'olian beds fifty feet below low-
water mark. It does not seem to mo tliat beach rock is found at any
greater elevation than that at which it could have been thrown up (and
subsequently cemented) during a hurricane or violent gale.

The shore platform of which Professor Rice speaks appears to con-
sist only of modified seolian strata, changed into hard ringing rock by
the action of the sea, and of a shore platform eroded to ledges, as he
himself describes them. He well says, when speaking of the relation
between drift and beach rock on the south shore, "If we conceive the
seaward face of the dune to \)e restored, it would certainly in some
localities extend beyond the narrow shore platform into the area now
covered by the sea."

Can we not find a simpler explanation of the formation of the Ber-
mudas than the one suggested by Rice 1 Instead of a subsidence
during which the nucleus disappears, followed by an elevation during
■which the seolian hills were formed, and then by a subsidence during
which the present soft drift rock was eroded, as is suggested by Rice,
we need only a single subsidence to explain all the phenomena, if, as
I have suggested, base rock is only modified seolian rock, and beach
rock has been forming continuously, and the seolian hills were formed
at the time when the atoll was one gigantic annular beach constantly
receiving fresh material from the outlying reef. This primordial reef
has disappeared, and its remnants exist perhaps at depths of from twenty
fathoms or more near the edge of the bank.

^ Bermudas, p. 43.

AGASSIZ : BERMUDAS. 225

Darwin's suggestion that the fringing reef on the south side of the
Bermudas is evidence of recent elevation, does not, in view of the fact
that the reef is made np of aeolian ledges, need any discussion. Nor is
the reason given by Dana regarding the cause of the great difference
in the amount of dry land on the north and south side of the atoll a
satisfactory one, if the ledges are a^olian ledges, which were the hrst to
disappear after subsidence began. 1 am inclined to look npon the
present state of things as due to the former existence of lower teolian
hills on the northern edge of the islands ; ^ but his views would apply
for proto-Bermudian times.

My oljservations lead me to look npon the beach rock of the Bermndas
as consisting mainly of the larger and heavier seolian materials, which
either have not been carried so far or blown to so great a height as the
li'-hter wolian sand. The effect of the intermittent submersion of the
seolian rocks exposed at low-wate • mark seems to be to cement the par-
ticles on the exposed liui;.s of thj knife-edged strata by a process very
similar to that going on ni all the dee[) road cuts on the islands. By it
all traces of stratitication arc grad lally lost, and an upper crust running
over the exposed surface is foni ed irrespective of the a^olian layers.
Thus a belt of comparatively hard rock is formed, covered with a crust
ringiug to the hannner, which at first sight appears to be imconformable
■with the teolian strata. A closer examination invariably reveals at no
great distance traces of the continnation of the aiolian stratification,
which contimie plainly visil)]e to high-water mark, to points below it,
and at intermediate heights. "Where the sea bi-eaks violently against a
vertical cliff, this cementing effect, accompanied by the disappearance of
the evidence of stratification, c;in be traced in some cases well above
high-water mark, whei-e it giadually jiasses into the region honeycombed
and pitted by the action of the rains. Snch parts of the rocks cannot
be distinguished from the base rock, and they have all its characteristics
except that the cementation is not rpiite so complete (Plates XV. -XVII.
and X.VVIII.).

Here and there alnng tlie beaches beach rock is forming, as in some
parts of (ireat Turtle B;iy, of the shore south of "Whale Bay, in Whale
Bay itself, and lietween short pi-ojectinu' headlands where the (h'hrix from
the outer and inner ledizcs accunndates in greater quantity. This beach

^ It «eems somewhat Iinznnlou'; to Mtt(>nijit, ns T\ice lins done, to correlate tlie
movcmontA of elev;iti')n ami subsidence of what proliah'y is a volcanic cone — of
which lie lias, as he thinks, foiiinl evidence — with iliose of the iVmerican conti-
nent. (Rnll. Nat. Mus., No. 25, p 18 )

VOT, XXVT. XO. 2. If,

226 bulletin: museum of cOxMparative zoology.

rock is generally readily recognized as such. It is formed in fairly thick
layers, from two to six inches, and always dips toward the sea at a very
moderate angle, and has nothing in common with the eeolian strata
against which it abuts. Parts of it may be ground up again by a storm
should the calm between heavy surfs not continue sufficiently long for
it thoroughly to consolidate.

All along the south shore one can find patches of beach rock dipping,
as observed by Professor Pace, to the sea at a slight angle, — the modern
beach rock of to-day, formed from the remodelling of the material thrown
up from the outer ledges. This beach rock is formed similarly to that
of the Florida Eeef, where it plays so important a part in its economj^,
while at tlie Bermudas it is of comparative insignificance. It is often
difficult to separate the beach rock from the base rock, but if, as I believe,
what is called "base rock" is only modified seolian rock, the latter has
Dot the importance attributed to it by Professors Rice and Heilprin.
On White Clifl:' Bay there are some a3olian cliffs dipping at a sharp
angle into the sea, showing remarkably well tlie transformation of the
thin aiolian layers into massive compact beds of base rock, in which the
dip of the strata can scarcely be detected, obliterated as it has been by
the cementing and solvent action of the sea water acting upon tliem.

At Hungry Bay, Middleton Bay, and many points on the south shore,
and on the north shoi'e at Ireland Island, and on the north shore of St.
George, there are numerous localities where it is possible to observe the
transition of the inclined or horizontal feolian strata above high-water
mark to the solid ringing limestone characteristic of tlie "base rock."
On the shore of Godet Deep to the west of Heron Bay, at the foot of
Gibbs Hill, the " base rock," can be seen passing gradually from the a'olian
beds into the solidified ringing limestone characteristic of the intratidal
limits. The action of tlie sea cements the strata together, so that all
trace of their a-olian structure is lost. In many cases, however, we can
trace the continuation of the reolian stratification indistinctly, so that I
am inclined to consider what is termed "base rock" as due merely to
such cementing action of tlie sea; the more so, as similar {)henomcna
are clearly observable all along the Cuban coast on the shoi-e edge of
the elevated reefs between low and liigh water mark, where tliere is no
question of an underlying base rock.

Professors Rice and Heilprin both speak of the "base rock" as dis-
tinguishing the old beacli formiition. ;ui'l as indicating the position of
the former sea i)order. It seems to nie tliat this basal rock is eeolian
rock which has become excessively indurated by the action of the sea

AGASSIZ : BERMUDAS. 227

water upon the lower part of the eeohan strata, and from being friable
and crumbling, as they are above the reach of the sea, have been
changed into a solid compact limestone, which rings under the ham-
mer and can be chipped off in sharp-edged flakes. Tiiis is simdar to the
hard ringing beach rock now forming, and does not, it seems to me,
indicate the position of the former sea border. Almost anywhere on the
south shore one finds the base of aeolian cliffs consisting of strata dipping
inward, changed as high up as the sea can reach, into this hard com-
pact ringing limestone. A similar " base rock " fringes all the Bahama
Islands ; inland at Nassau, as at tlie Bermudas, a few steps from the
shore inside of the " base rock," the reolian structure is clearly defined
in quarries and wells extending below the water line, but the sea, acting
merely by percolation, has not changed their thin edges and cemented
them as it has on the sea face of the shores where the strata are fully
exposed to the action of the sea, and are in addition exposed for a
longer or shorter time to the atmospliere during low-water periods or
during the intervals between consecutive breakers.

Rice says of the locally called base rock, " that it does not uni-
formly underlie the softer rocks, nor is tliere any evidence that it is
older than they." ^ A part of the confusion between base and beach
rock seems to me to have arisen from considering the ledges of seolian
rock as reef rock, and from the fact that there are a few localities on
the south shore where beach rock is actually forming from aeolian rock
sand, derived from ledges in deeper water, mixed with broken shells and
fragments of corals and Millepores, all of which particles are cemented
by the deposition of lime held in solution in the water percolating through
its masses.

Rice further says, " That there can be no absolute distinction between
beach rock and drift rock will be manifest from the consideration that
the two formations are in their origin strictly continuous." Yes, but
their origin is not the same; the beach rock of to-day is formed in great
part of the aiolian rock of former days. I would go one step farther in
believing that the base rock is by no means usually beach rock, but that
beach rock is a very local phenomenon, and is younger than the feolian
rock, and belongs to the present epoch, and has been forming at different
levels, as it is forming to-day in favorable localities, from the time the
islands began to subside, as well as before that time. I am at a loss to
know what Rice and Nelson can mean by reef rock, unless it be the
thin crust of coral growth upon the ledges. I am inclined to adopt

1 Bull. Nat. Mus., No. 25, p. 9.

228 BULLETIN : MUSEUM OF COMPAUATIVE ZOOLOGY.

Thomson's view/ that the Bermuda limestone is entirely aeolian, and
tiiat the base rock does not underlie the softer seolian rock," but to
modify it by the above statement regarding the formation of base rock
and of beach rock as now forming mainly from the remodelling of older
material.

Ptice accepts the Agers Island strata as beach i-ock, as well as the stra-
tum at the south end of Ireland, a statement from which I must most
emphatically differ. These beds contain marine shells in seolian strata.
The base rock of the islands of Hamilton Harbor appears, as far as I have
observed it, to be due to the induration of the lower strata of the seolian
rock exposed to the action of the sea. There are but few islands in
Hamilton Harbor of which the seolian strata do not dip at a considerable
an<ile towards the low-water mark line into the sea. It is true that in
some cases there are also a^olian rock strata parallel with the sea, every-
where lying neai'ly, if not quite, horizontal, but these strata present the
character of leolian rocks modified by the action of the sea.

On Grace Island, Hawkins Island, and off the Quarantine in Hamil-
ton Harbor, we traced most distinctly, at low-water mark, seolian strata
dipping at a high angle into the sea, and yet at many points they have
passed into what is called base rock by the complete obliteration of the
knife-edged strata from the abrading, the cementing, and the sdveut
action of the sea. The third island south of the Quarantine, Post Island,
and Darrel Island all present the same phenomena of modification of the
seolian rock into base rock. In the interior of a cavern leading into
Quarantine Island we could trace the dip of tlie seolian rocks into the
sea, and the same was the case at Hanson's Island. On tlie southwest
side of Post Island the fine seolian lamination coulii still readily be
detected below high-water mark through the " base rock " coating.

Nowhere in the P>ormudas have I found corals above high-water mark
or higher, the presence of which could not be accounted for by the action
of high winds or waves during hurricanes, and surely the presence of
caves above high-water mark is not in a limestone district an indication
of elevation. If the explanation I liave given of the formation of base
rock is correct, its existence at a height of a few feet above high-water
mark is not a proof of elevation.

1 Thomson, Atlantic, T. .'507.

- There wus no "base rock" found while cutting through the aeolian strata
during the excavation of the Ireland Island Dry Dock.

AGASSIZ: BERMUDAS. 229

FOSSILS.

On a small island to the south of Agers Island I found quite a number
of species of marine shells identical with those now living embedded in
nearly horizontal ivoliau strata a few feet above liigh-water mark. Also a
bank of Chama evidently thrown u]) or blown up during a luu'ricane,
much as we find Strombus on some parts of the Bahamas thrown up in
great banks high above high-water mark. That marine shells should
thus be thrown up or blown up to such considerable heiglits in what
may, in proto-Bermudian time, have been a protected sound, as well
as is Hamilton Harbor, is not extraordinary. We need only recall
tlie great violence of the hurricanes which sweep past and over the Ber-
mudas, durino; which vessels have drau'ged their anchors in the sheltered
inner harbor of Hamilton, wliere the wind and sea have a comparatively
limited range.

Below that, but in seolian strata, these fossils extend to low-water
mark, apparently embedded in the " base rock." These lower strata
have at first sight all the appearance of beach rock ; they consist
mainly of particles larger than a?olian sand, which probably have not
been l)l()wn a great distance upward from their base. But those strata,
consisting of larger brecciated fragments, have, like other a^olian beds,
been changed into the hard ringing limestone so characteristic of
nearly all the exposures below high-water mark.

During the very low tides which prevailed for the last days of my
visit at the Bermudas, I was able to trace the existence of ajolian beds
imderlying the fossiliferous beds with the base rock lying between them.
The fossils are embedded in aeolian rock, and in certain spaces, whicli
liave become cemented so as to destroy the laminations, they appear
to be embedded in the base rock. The existence of tiiese fossiliferous
beds above high -water mark in the islands of Hamilton Harbor has led
Rice to assume a period of sliglit elevation ns having occurred in the
Bermudas, and further to maintain tliat much of the interior of the
islands is underlain by beach rock, a statement with which Professor
Heilprin agrees in the main. The size of the material and the broken
shells tlirown up at Slielly Beach show how high up material similar
to it may, even under ordinary circumstances, have been thrown up and
become embodied into aeolian beds without its being any indication of
a period of elevation.

230 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

The fossiliferous strata of St. George Island/ mentioned by Professor
Eice, seem to me to belong to the same category of rocks which crop out
in Hamilton Harbor. They are seolian rocks containing many species
of Lucina, Chama, and the like; and as all the marine shells near
Agers Island must have found their way into these ceoliau strata under
the action of the winds or the sea, tlie parts of the strata below high-
water mark have here also been partly changed into the hard ringing
limestone of the base rock. Professor Heilprin also found marine shells
in seolian rocks.^ In a hill to the eastward of Stone Hill I have found a
few recent land shells in very friable seolian rock.

An interesting collection of rocks and sub-fossils from the seolian rock
quarries and other localities is preserved in the Government Building.
It contains among other specimens a small collection of casts of Telli-
nas and Lucinas obtained from a submarine cut in Tomlin's Narrows,
sixteen feet below low-water mark. Tliis would indicate the existence
of a bay or sound, as in our day, at the time when the level of the sea
was higher, before the land had by subsidence obtained its present level,
and it is no indication of elevation any more than the presence of the
living shells of to-day as fossils would indicate such a change. It also
contains the shell of a tropic bird egg found in seolian rock quarries in
the Middle Road in Devonshire, the bones of a snipe's obtained from
seolian rock near St. George, and marine shells from a bed twenty feet
above the sea and one hundred feet away from it, occurring in seolian
rock on the main road from the north side of Gibbs Lighthouse
Hill. There is also in the Government collection a large Turbo, which
was found in a cutting in feolian rock at the east end of Hamilton ;
this Turbo is said to be extinct, but is found sub-fossil in the highest
seolian hills.

1 I would consider tlie peculiar concjlomerate of Stocks Point as only the Iiiglier
limit of a local beacli rock, which may have been thrown up in a locality specially
exposed to gales or hurricanes, and limited in extent. It contains fragments of
the underlying drift rock, and resting upon it, as T^ice observes, is the ordinary drift
rock. But on both sides of the beach rock we find jeolinn drift rock reaching to the
sea, which would indicate either a fault or that the conglomerate was older than
tlie a?olian hills of the Bermudas, neither of which suppositions Is in accordance
with other facts observed in the vicinity.

- "At several points, more particularly along the north shore, I found marine
shells (Lucina, Tellina, etc.) embedded in unquestionable drift rock, and indeed it
could hardly have been expected that such association should not occur. . . . The
same is also true in a measure of the occurrence of land snails. . . . One of the
commonest shells of the lower drift rock is the large Tiirho {Liroiin) pica, a shell
which appears to be very abundant about the' coast." Heilprin, Bermudas, p. 35.

AGASSIZ: BERMUDAS. 231

THE SOUNDS AND LAGOONS.

Plates II., IV., VI., VII., XIV., ami XXVII.

The sounds are sinks and depressions filled with sea water, as was first
suggested by Rein, and none of tliem are secondary atolls. They owe
tlieir origin either to the breaking through of low saddles dividing sinks
from outer lagoons, or to subsidence, allowing the water of adjacent
lagoons or the sea to flow in over separating ridges, or to both these
causes.

Professor Heilprin gives an excellent description of the rapid waste
which the islands are undergoing, and of the formation of the sounds, on
pages 36 and 37 of his Bermudas.

The improbability of the sinking of the roofs of large cavernous areas to
form the sounds, as has been suggested by Rein ^ and Fewkes,^ does not
militate against local disruptions on a limited scale, of which, as Heilprin
states, there is abundant evidence.^

The lagoons of the south shore between Tuckerstown and Newton Bay
are brackish pools separated by low hills from the sea (Plate XIV.). In
many places it would require comparatively sliglit inroads of the sea, or
but little subsidence, to change them into diminutive harbors or sounds,
similar, but of course on a smaller scale, to Castle Harbor or Harrington
Sound. The shore platforms of Harrington Sound and Castle Hai-ljor
are similar to the ledges which extend off the clitfs from the outer shores
of the islands (Plates VI., XXVK.).

Harrington Sound seems to have been formed in exactly the same
way as the smaller liarbor indentations of the coast. Its shores present
all the phenomena of disruption by waves exiiibited by the outer shores,
although in a less degree. The action of the sea is of course much less
powerful, yet is sufficient to have undercut the cliti's, and in some places,
as on the nortli sliore of the sound, they are fully as high as many of tlie
more striking cliffs formed on the sea faces outside by the splitting off
of large slices of the pcolian hills.

We find in Harrington Sound islands, islets, and many honej'combed
ledges (Plate XXVII.), pinnacles, and mushroom-shaped rocks, due to

1 Kein, Rericht., 1870.

2 Fewkes, J. W., Troc. Bost. Soc. Nat. Hist., 1887, p. 518.

3 Bermudas, p. 45.

232 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

erosion or to the solvent action of the sea, difiering in no way from those
made along the nijrth and south shores of the islands. Along its shores
there are numertus ledges running out froui or parallel with them, ex-
tending between the small rocky promontories, which if cut oft" would
form a series of patches close to sliore similar to those which extend from
the north shore inside of the banks towards the outer ledge flats and
upon them. Gorgonians, corals, and other growths, have settled upon
the ledges since the time when they and the saddles have sunk or liave
been eroded to their present level, giving the sea access into the interior
of the various sounds so characteristic of the Bermudas. This is admi-
rably shown by the cutting of tlie sea into St. George Harbor, so as to
give access to it both from the outer and inner waters of the bank.
These passages are narrow, so that corals do not get a sufficient suj)p]y
of fresh water, and hence are far less couunon than on the shelves of Cas-
tle Harbor, which is freely connected with the sea on its southern expos-
ure. In Harrington Sound the connection with the lagoon is still less
open, a narrow cut on the north being the only opening through which
the inner waters of the bank gain admission to it. Castle Harbor in the
same way is connected freely with the sea on the soutJj, and but slightly
with the inner waters of tlie bank, through the same opening which con-
nects tliem with St. George Harbor.

A narrow cut separates Somerset Island, wliich forms the western
boundary of Great Sound, from the main island. Tliis sweeps round to
the eastward and forms the southern flank of Port IJoyal I!ay, which is
separated from Great Sound by the line of islands extending outward
from Tucker to Darrel Island (Plate II.).

The de])th of water in tlie sounds is veiy considerable, not only in the
sounds themselves, but also in the inner waters of tlie reef, which liave
been called lagoons, but are hardly such in the sense in which that term
is understood. It would greatly conduce to accuracy to call the inner
basins of deep water — surrounded on one side by tlie outer ledges of
the reef, and on the other either by tlie connecting patches of ledges or
by the islands in part — sounds also, for such tliey undoubtedly are, and
were sounds similar to those now existing and known as Great St)und,
Port Royal Bay, Hamilton Bay, St. George Harbor, Castle Harbor, and
Harrington Sound.

In Great Sound we find from ten to eleven fathoms. In Harrington
Sound as much as twelve fathoms is found in several spots. In the outer
sounds we do not as a general rule find so luroat depths. What we may
call the Brackish Pond and Bailej- Flats Sound has a gencial depth of

AGASSIZ : BERMUDAS, 233

only from five to six fathoms. The great sound known as Murray An-
chorage, to the northeast of Eailey Bay Flats, is somewhat deeper, and
varies from seven to nine fathoms. To the north and northwest of
Murray Anchorage the water is still deeper, varying from seven to ten
fathom^;, with a deeper bight to the northeast of Three Hill Shoals, where
the depth is twelve fathoms close to a spur of the East Ledge Flats.
These depths all run close to the five fathom line, which may be called
the inner edge of the outer flats or ledges extending fi'om East Ledge to
the Ledge Flats north of Blue Cut. On these the depths vary from
one and a half to four fathoms, with occasional deep holes, with a
white sandy bottom, or islets separated from the edges of the Ledge
Flats or inlets running in from the sounds, and patches surrounded by
from seven to ten fathoms. To the westward of Three Hill Shoal the
clear bottom averages from seven to eight fathoms. In the sound be-
tween Brackish Pond Flats and Elies Flats the depth varies from six
to ten fathoms, with occasionally a five fathom sounding between the
numerous isolated patches of ledges inside of the live fathom line. The
extensive sound to the west of Wreck Hill, extending to the Western
Ledire Flats and southwest of Elies Flats, varies from seven to eleven
fathoms (Plate II.).

In St. George Harbor the bottom is hard in five fathoms. The dredge
brought up many specimens of Toxopneustes and Echinometra. We
found only a few patches of Gorgonians and of massive corals in the
harbor itself, while in Castle Harbor, which has a fi'eer communica-
tion with the sea, the patches of corals on the ledges are quite nu-
merous, having much tlie same cliaracteristics as tijose of the Ledge
Flats.

In Harrington Sound the growth of (Jorgonians and massive corals is
also less prominent than in Castle Harbor, and the development of Gor-
gonians, A.lgte, corals, and corallines in these sounds, as well as in the
sounds at the western end of the islands (Hamilton Harbor, Port^Eoyal
Biy, and Great Sound), seems clearly to indicate that as fast as they
l)ecame comiected with the outer sounds and in proportion to the acces-
sibility of the sea, corals have gradually found their way into these
sounds, and have also developed iu proportion to it, being less abundant
in sounds indifferently connected with tlie open sea. All the conditions
of the coral growth indicate a comparatively recent inroad of the sea,
first into Castle Harbor, next into St. (reorge Harbor, and finally into
Harrington Sound. The corals have found their way into the sounds
much as the corals forming the veneer of the outer reefs have found

234 bulletin: museum of comparative zoology.

their way from the West Indies in the truck of the Gulf Stream, or
perhaps have been derived from the corals forming the proto-Bermudian
reefs, which in their turn were introduced from the West Indies through
the same agency.

It seems probable that the cedars dredged up in the excavations for
the channel in St. George Harbor were floated into the basin from the
adjoining hillsides ; but in the case of the red earth coming from the
excavation in Ireland Island, the site of the present dockyard was prob-
ably a banana hole, which during the subsidence sank to its present
level, say hfty feet or so below low-water mark.

The conditions of growth of the corals in the sounds of the Bermudas
do not seem to me to have any bearing on the growth of corals in the
lagoon of an atoll. The lagoon of an atoll swept by the currents, with
its rim pounded upon by the surf, and the Bermudian Sound, with its
comparatively quiet expanse of water formed imder such diff'erent condi-
tions, do not seem to have many features in common.

There are near Harrington Sound, between it and Castle Harbor,
three small sounds in the process of formation, which present all the
characteristics of the larger sounds, only on a most diminutive scale in
proportion to the range of the sea they enclose. The two most interest-
ing are one to the east of Harrington House, and one called Webb's
Pond, on the road to St. George, after passing the Flats. The latter is
an irre^ ilarly pear-shaped, miniature sound, about 200 by 180 feet, and
perliaps 200 feet from the north shore. At its southern extremity there
are low crumbling cliff's. The depth is said to be fourteen feet. Both
of them aie merely sinks close to the sea, but only connected with it
under ground, and perhaps filled by pei'colation of the sea through the
seolian rock. The tide rises and falls in both.

Spittle Pond, on the south shore, is a brackish sink surrounded by
grassy shores, which barely reaches high water-mark. Between Tuckers-
town and Newton Bay there are also a couple of brackish ponds, the
shores of which are protected by mangroves.

While the sounds undoubtedly indicate subsidence, they are not
lagoons surrounded by corals, such as we find in atolls, and should not
be compared to them. They are sinks or low tracts, which have become
connected with the outer waters into which corals have found their way.
Such sinks we find ready to be changed to sounds or pseudo lagoons at
many points of the Bermudas ; as, for instance, along the South Shore
road from about Walker Bay nearly to Hungry Bay, there are a series of
low valleys about at the sea level, and separated from the sea by a ridge

AGASSIZ : BERMUDAS. 235

of low seolian hills. A further slight subsideuce would change them
into shallow harbors by the rushing in of the water over the lowest of
the dividing saddles. A similar low tract extends to the south of Gov-
ernment House between it and Hamilton, with an outlet into Boss Bay.

DISTRIBUTION OF THE CORALS.

The Bermudas are the most northerly limit ^ where reef-building
corals are known to occur, unless we can call coral districts areas where
Astrangia, Primnoa, and other northern types, like Caryophyllia and
Lophohelia, are to be found.

The absence of Madrepores, to which Thomson called attention, is
very striking. There are thus wanting the very elements to supply the
bulk of the material broken off and thrown up by the sea to accumulate
as beaches or islets.

It is interesting to note that the littoral marine fauna of the Bermudas
is the same as the shallow water fauna of the West Indies, and that its
existence here is one of the finest examples of the effects of great oceanic
currents in shaping the geographical distribution of animals the em-
bryos of which are pelagic a sufficient length of time to be transported
to this their northern limit from the Bahamas and other parts of the
West Indies. During my stay at the Bermudas, every day when the
wind blew from the southwest or west the common West Indian
Physalia appeared in great numbers.

Besides marine animals, floating masses of wood coming from the West
Indies are frequently stranded on the shores of the Bermudas, these
sometimes carrying fragments of rooks. A large ellipsoidal mass of
floating pumice, measuring eleven inches in length, was picked up off the
south shore by the Hon. Archibald Alison. A similar float, thrown up
on the south shore, is preserved in the Museum of the Government
Building. This piece is filled with red earth.

1 The minimum temperature of the surface of the sea occurs in January and
February, when it varies between 59° and 63°. In March it varies from 62° to 66°.
In April its maximum has already risen to 71° ; in May the maximum is 76°,
minimum 70° ; in July the minimum is 70°, maximum 82^° ; in August the maximum
is 85°, minimum 82.^° ; in September the maximum is 8.3°, and the minimum 75°.
The temperature tlien falls rapidly from 69° and 74° in October, down to 61° and
65° in December. The minima are remarkably low temperatures for a coral reef
district. The above data were kindly furnished me by Mr. John C. Watlington, of
Hamilton.

236 bulletin: museum of comparative zoology.

Wallace lays great stress upon the pumice thrown up from the sea as
being a possible explanation of the source of i-ed earth. Although I
examined the beaches of the south shore many times, I never succeeded
in finding a single piece of puraice. Red earth is abundant, both at the
Bahamas and Bermudas, in localities to which drift pumice could not
have access.

I was greatly struck with the apparent want of adaptation to their
surroundings of the coloring of many of the Bermudian marine inverte-
brates. The dark violet Diadema and Echinometra are very common on
the faces of the steep rocky patches, as well as in the sandy hollows of
the surface of the bank. In the one case they are part of the brilliant
patchwork forming the coloring of the reef surface; in the other, they
stand out most prominently against the whitish Nullipores. Diadema
in Florida, as well as in the Bahamas, is often found in colonies entirely
filling the bottom of some sandy depression in the midst of a white
field of surrounding coralline bottom, the patches of brilliantly col-
ored corals and Gorgonians themselves standing out as a whole in
striking contrast to the whitish bands of coralline or eeolian sand sep-
arating them.

Professor Heilprin has greatly added to our knowledge of the fauna of
the islands,^ which was previously mainly derived from the sketch of their
Natural History by Jones, ^ and has also given a list of the species of
corals belonging to the islands.^

The low tides prevailing during the last days of my stay at the Ber-
mudas enabled me to note the luxuriant growth of I\Iillepores and Gor-
gonians on the surface of the many patches which were nearly awash
during these days. The flats extending to the north of Ireland Island,
and the flats to tlie soutlieast of the North Rock Ledge, were specially
noteworthy for tlieir abundant coral growths.

On passing through Mangrove Bay to reach Hogfish Cut from Great
Sound, we dredged Oculinas from the deepest part of the bay ; the bottom
in four fathoms is very fine sticky silt, almost marl. Corals in Hamil-
ton Harbor, Great Sound, and Port Royal Bay are limited to a very
scattered growth along the shores of the islands of these sounds below
low-water mark. In the deeper parts of the sounds Oculina? grow to
considerable size. Zoanthidse are abundant on the higher knolls of the
outer reef, and also on the serpuline atoll-like structures. Porites is
quite common on the outer reef. This is in marked contrast to the

1 Bermudas, p. 97. ^ Bermudas, p. 98.

2 The Naturalist in tlie Rermudas.

AGASSIZ: BERMUDAS. 2^7

abode of Porites in Florida, where it is usually most abuudant on flats
more or less sheltered.

The corals at the Bermudas as well as at the Bahamas do not gener-
ally reach the surface. They form a more or less connected belt of coral
growth in from five to six fatlioms on the inner edge of the flats to the
outer limits, the beginning of the broken ground, the co^-als extending to
eight or ten fathoms on the southern sea faces of the flats.

LEDGE FLATS AND PATCHES.

Plates II., XV., and XVII. to XXVI.

The ledge flats, patches, or coral heads, are names given to different
parts of the reef, wliich has universally been considered to owe its exist-
ence to the growth of corals, and much of the confusion existing regard-
ing the structure of both the Bahamas and Bermudas is due to the
fact that corals have been assigned a part in the building up of these
islands which they have never performed.

The flats consist, not of coral heads, though they are often so called
both here and at the Bahamas, but of ledges of ^olian rock rising from
a depth of five to six fathoms or more (Plate XVIII.). These ledges,
with their nearly vertical sides and their slopes deeply honeycombed,
drop rather abruptly into the coralline bank bottom, which forms more
or less extensive irregularly shaped patches separating the ledges. The
surHice and sides of the ledges are veneered by corals,^ Gorgonians,
and Millepores ; the sharper edges of the ledges are covered by incrust-
ing masses of Millepores, and calcareous and other Algse grow in great
profusion between the corals and Gorgonians. The Millepores, Gorgo-
nians, NuUipores, and calcareous and other Algee, are by far the most
abundant growth on the inner patches. On these we find only compara-
tively few of the larger Mseandrinas and Astreas. The massive corals
increase greatly in number as we approach the outer edge of the reef,
and the finest and most numerous specimens appear to grow on the
outer sea face in from five to seven fathoms of water. Beyond that, or
even at lesser depths, in five to six fathoms, the broken ground begins.
This consists mainlv, as far as 1 could ascertain from the observations
of others and gather from my own notes and dredgings, of Gorgonians,

1 Thomson thinks that the patches have been built up by the corals. Atlantic,
I. .304.

238 bulletin: museum of comparative zoology.

masses of Algse and of coralline Algse, and Nullipores, with compara-
tively few massive corals. As far as I am aware, this broken ground
does not extend on the northern sea face of the ledges as far out as off
the southern ledges, where the fishermen report its existence to a depth
of from sixteen to seventeen fatlioms.

The ledge patclies rise in steps from the coralline bottom depths,
much as they fall in successive ledges off the shore clifi's. As we ap-
proach the bank edge of the flats, the ledges become smaller, the depth
of water increases, and the sand spaces between the patches increase,
often forming long tongues extending into the main body of the ledges
of the flats. Many of the ledges near the edge come quite close to
the surface, and a great number are awash at low -water, although the
depth of water between them is greater than at the point we might call
the crest of the ledges. The nearer we come to the breakers, the greater
becomes the wear of the sea slopes of the ledges, so that in many
places their slope is qtiite abrupt, from two or three to five fathoms, and
a somewhat gentler slope extends from that point seaward to form the
broken ground.

The high seolian cliffs of the south shore probably extended to the
outlying reef, which is itself only a series of ledges running parallel to
the coast, the crests of which are bare at low water. On these and the
inner irregular flat ledges which dot the bottom over greater or smaller
areas between the outer ledges and the shore grow corals and Gorgo-
nians, — a comparatively thin veneer, which supplies, when dead or
beaten off by the surf, a part of the material which goes to form the
sandy beaches of the south shore, — though by far the greater mass of
the material is derived from the disintegration of the ledges themselves.
So that the submarine remnants of the ancient seolian hills supply the
material which to-day creeps over their faces and finds its way inland,
much as they in their own time must have crept over the lowland exist-
ing within the limits of the proto-Bermudian coral reef.

From the observations I have thus far made, it seems to me as if
the corals now growing at the Bermudas had, as at the Bahamas,
played a very unimportant part in building up the mass of the reefs.
It is true that some of the flats are largely formed of coralline coral
and ajolian sand, derived in part from the coral patches which line
their faces. But as yet no islands or islets have been formed by their
disintegration, showing that the coral growth is not rapid ; and al-
though In some of the patches along the inner edges of the flats and
on some of the connecting patches the corals have attained sufficient

AGASSIZ ; BERMUDAS. 239

thickness to conceal the original ledges, and perhaps in many cases to
build them up from a couple of fathoms or so to within the limits of
low-water marlc. Tlieir growth can in no way be compared to the mas-
sive coral reef structures we find in the West Indies and Florida.

Oflf tlie noi'th, as well as off the soutli shore, the patches nearest the
land are merely ledges consisting of larger or smaller iiieces which have
become separated from the shore cliffs by the action of the sea, or else
they are more extensive patches marking the position of small islands,
islets, or rocks which were once more or less closely connected with the
main island, and wliich now run as ledges parallel to the shore line.
These ledges, if close to the shore, are barely covered by Algae and a few
barnacles, or Mytilus, or isolated corals, or such animals and plants as
we find on the immediate shore line. Farther from the shore they be-
come overgrown with a greater profusion of Algae and Nullipores. As
we proceed from the north shore to the ship channel, we gradually come
upon ledges on which are found corals and Gorgonians, Algae and Nulli-
])ores occuiring as on the ledge flats, but in less profusion. It is on
tlie outer ledge flats, which have probably been under water longest,
that we find the most abundant growth of corals. While I do not deny
that some of the ledges have been increased in height, and slightly in
width, by the corals covering them, yet the corals have played but an
insignificant part in building up the ledges tliemselves. The ledge flats
are tlie remnants of the proto-Bermudian eeolian land worn down by
the action of the sea to a certain level, and upon these seolian ledges
forming the underlying foundation of all the patches, the coral reefs —
viz. corals, Corallines, Gorgonians, and Algae — have grown, but only as
a comparatively thin veneer upon the pre-existing aeolian ledges.

Tlie surface of many of the ledges outside of Hungry Bay on the
south shore, exposed at low water, is covered with coral growth, espe-
cially the ledges on the inner face of the outer patches of the south
shore reef. The ledges exposed at extreme low water- are in-egnlarly
shaped, rising from two to four fathoms of water ; the}' are greatly under-
cut and abraded, and show signs of the solvent action of the sea. The
vertical and sloping faces of the ledges near shore are covered and pro-
tected from wear by a thick growth of Algte and Corallines, similar to
the growth which protects the upper face of the ledges. But on the upper
surface there is in addition an abundant growth of Serpula?. Between
the outer ledges and the shore, more or less protected by the isolated
outer patches extending to the reef, a sort of lagoon is formed. In this
lagoon are found numerous ledges ; then, closer to the shore, overgrown

240 BULLETIN : MUSEUM OF GOMPAKATIVE ZOOLOGY.

with Algee, Coi'allines, and Serpulse, there is deeper water, with corals and
Gorgonians. Many of the ledges within the lagoon consist of the minia-
ture serpuline atolls and reefs described in another section of tliis Report.

On the inside of the outer serpuline reef ledge, corals and Gorgonians
flourish, according to the depth and the position of the ledges intervening
between the outer reef and the shore. Off Sinky Bay the bottom outside
of the outer reef ledges is hard. Off Castle Harbor, as far as the channel
leading into St. George Harbor (Plate XXL), we can readily trace the
gradual formation of islands and islets originally constituting the con-
tinuous barrier to a sound formed by the breaking through of the lower
saddles of the ridges dividing it from the sea. The outer row of these
islets and islands difiers from the inner one in having comparatively
wide ledges, projecting round the base of a central pinnacle more or less
undercut. As the central pinnacles are cut away, they leave only a nar-
row ridge on the broad platform, the ridge itself also disappears, and on
the outer line of ledges have grown Algae, Serpulse, and other organisms,
which prevent in some cases the further wearing away of the whole
ledge, protecting its most exposed parts. Tlie sea breaking upon the
upper surface of the ledge soon forms the more or less regular serpuline
atolls and "boilers" of the south shore wliicli will be described later on.
They are found on all the breakers on the outer side of the reef ledge
flats, like Mills Breaker, the North Rock, and others; serpuline reefs
extend off the headland on the west of Church Bay. The Southwest
Breaker is the westernmost of the line of serpuline reefs skirting the
south shores. It has three " boilers " on it, a long one and two smaller
ones, with a small serpuline atoll to the soutliwest of the main ledge.

There are on the eastern part of Castle Harbor itself a number of
ledges coming to within a foot or two of the low-water mark. They
are covered with corals and Gorgonians, Algae and Corallines. The Gor-
gonians are not very flourishing, but the Nullipores and Alga' grow in
abundance between the massive corals. The patches are separated by a
flne sandy bottom. On the outside of Castle Harbor there are many
coral ]iatclies, boilers, and ledges extending outward of the outer ledge,
to a. depth of from seven to eight fathoms, with an occasional ledge
rising from ten fathoms (Plate XXL). But, as a rule, outside of the
outer reef ledges we come upon the " broken ground."

The line of reefs to the south of the islimd extends uiibi-oken from
St. David Head to off High Point. Tlie ledges are all anilian shore
cliff's which have become separated from the island by the action of the
sea, then beaten away, abrad 'd and eaten into by the surf, and, accord-

AGASSIZ : BERMUDAS. 241

ino- to their position, the depths in which they are found have heen
transformed into the pecuhar agglomeration of reefs and ledges off the
south shore. . They form an incipient ledge flat, as it were, of which
the outer line is still very prominent, and which the outer breakers
have not as yet undermined and eaten away, so as to leave, as they
do round the ledge flats, only a few isolated rocks cropping to the sur-
face. Some of these ledges are a hundred and fifty feet in length, and
even more, with a breadth exposed at low water varying from two or
three feet to thirty or forty feet ; others are only small pinnacles a few
feet in diameter. All, however, present nearly vertical faces, and rise
abruptly from two and a half to four fathoms. They are all more
or less undercut, eaten away, of irregular mushroom shape, and the
breaking up and disintegration of the exposed pinnacles after they
have been so undermined as to break from their base supply a large
amount of the material thrown up on the beaches.

A section along the slope of the sea beach of the south side of the
island shows first a shore line of flats, ledges, and pinnacles, then a sec-
ond or a third row of mushroom-shaped undercut rocks, some reaching
to above low- water mark, others barely awash, or a few feet below. A
fe'w of the ledges may still be surmounted by a^olian rock pinnacles,
while the submerged surface of other flats is eitlier protected by Algse,
Corallines, or Serpuhe, and according to their depth they are changing
or have been changed into serpuline reefs. A few of the ledges in
deeper water inside of tlie outer line of ledges are covered with corals
and Gorgonians. The outer row of ledges forming the reef do not
difter from the rows of rocky ledges inside of the reefs, or from those
close to the shore. There are on the outer lines, however, no ledges
surmounted by pinnacles, most of them having been changed into boil-
ers, or into long ledges with winding or S-shaped vertical walls, the sur-
face of which is protected by Alg^, Serpnlye, and other growths. Outside
of the outer row of ledges we come upon the broken ground bottom,
which consists of flat ledges extending from five to fifteen or more
fathoms. Upon these in the shallower parts flourish the massive corals
and Gorgonians, while over the deeper parts extend mainly the Gorgonians
and Algfe, as well as Corallines. Such broken ground bottom occurs off
Chaddock Bar, off Long and Little Bar, off the Chub Heads, and all the
way from the Southwest Breaker outside of the south shore reef to off
Castle Harbor and off" St. David Head. Similar broken ground occurs
wherever on the Admiralty Chart it is marked r, — off the Mills Breaker
Channel, outside of the North Rock Channel, the Eastern and Western
VOL. XXVI. — xo. 2. 16

242 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Blue Cuts, and Chub Cut, as wen as between Long Bar and the "West
End I.edge Flats. There are outside of the reefs many areas. of rock}'-
bottom, marked r on the chart, the remnants probably of extensive
seolian ledges.

Nowhere do we find more fantastic shapes in the pinnacles remain-
ing on some of the ledges than those which are seen to the south of
Nonesuch Island, and extend to the eastward toward St. David Head.
The islands pass into pinnacles, into ledges, and finally into boilers, in
regular succession, and in proportion to the exposed condition of their
position. Similarly eroded pinnacles are also seen on a smaller scale,
but of fully as fantastic shapes, in St. George Harbor, in Mullet Ba}-,
and on the south side of the causeway on the western side of Castle
Harbor.

The patches outside of the reef off the south shore can be clearly seen
extending a short distance to sea, separated by irregular white patches of
sand. The inner ledges, forming the patches between the outer reef and
the shore, are most capriciously distributed. Outside of the reef off the
south shore the corals do not seem to thrive, and the broken ground is
comparatively barren, though we find an occasional patch where Gor-
gonians, Algse, and massive corals are more abundant. The coral
growth is more that of the broken ground than of the reef flat ledges
or of the connecting patches.

Heilprin has noted the great importance which the Millepores take in
the composition of the bank sand bottom. On the south shore, where
Serpulse are so abundant, the fragments of their pinkish shining tubes
can readily be distinguished in the coarser fragments of the sand thrown
up on the many beaches along the shore. While in the Bahamas I was
struck by the importance of the Millepores in the economy of the reefs.
They seem to be far more abundant there than upon the Florida reefs,
where the Madrepores take an extraordinary development, while they
are absent in the Bermudas.

The south reef extends at a distance from the coast of about one
thousand to fifteen hundred feet throughout its length from the en-
trance to St. George Harbor to the eastern side of Hogfish Cut. It
has nothing to do with a barrier reef as such. It is a barrier ledge
of seolian rocks derived from the old shore line, and not a barrier
reef formed by corals, as Heilprin would lead us to suppose. The
description which he gives of the work of destruction going on upon
the banner reef which skirts the southern coast is somewhat unfor-
tunate, as the material of which he speaks as "blocks of coral and

AGASSIZ : BERMUDAS. 243

of coralline . . . detached and broken," is derived from the rehandling
of tlie ledges of a3olian rock of the former sea-shore. The corals now
growing play an inhnitesimal part in the forming of the sand dunes
which "stand on the eminences which to-day are tlie Bermudas." His
description would apply to the original reefs from which the Bermudian
hills were formed, but is scarcely applicable to the work doing in our
day.

Heilprin, after quoting Dana's description of the reef of an atoll,
finds it largely applicable to the condition of the Bermudas, an opinion
to which exception must be taken. From what has been said it will
be seen that the Bermudian coral reefs have little if anything in com-
mon with the coral reefs of an atoll. Certainly no more erroneous state-
ment could be made than that " the more seemingly favored patches are
the creations of the surf themselves." Tlie Bermudian reef corals are,
like the Bahama reefs, submerged, rarely come to the surface, and have
not supplied any considei'able part of the material which has gone to
build up an extent of land either in the Bahamas or the Bermudas. In
the Baiiamas the corals flourish most jjrofusely in depths of from five
to twelve fathoms; at tlie Bermudas six to seven is their limit, and
those on the sea face of the ledges do not seem any more abundant
than those on the edges of the flats. I was not able in the several
sections I made across the sea faces of the southern reef to find
the unbounded profusion of coral growth which Heiliirin observed.
In fact no one has better shown than he that the coral reefs which
now encircle the Bermudas have had no share whatever in their
formation, and I fiiil to see how the fact that subsidence has given
to these islands their outline of to-day has any bearing upon the theory
of the formation of coral atolls by subsidence. Any land surface ex-
posed to the action of the inroads of the sea owing to its subsidence
would have been eroded to some extent according to the nature of the
rocks composing it. The subsequent formation of a thin veneer of coral
reefs upon its sunken ledges would not liave any bearing on the theory
of the formation of thick masses of limestone by subsidence. It may be
interesting, in this connection, to refer to Heilprin's statements " that
tlie present form of tlie Bermuda Islands bears no relation to the rino; of
an atoll," and that " the existence of an atoll is not demonstrable."^

I fully concur in what he says regarili.ng the subsidence which
followed the elevation of the islands to their greatest height. Heilprin
was impressed by the absence of loose boulders of rock (coraU).

1 Bermudas, p. 4(3.

244 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY,

This is natural, for we find on the ledges of the Bermudas mainly seo-
lian rock masses readily crumbling to pieces, a thin coral belt, and but
little solid material to be sliaped into boulders by the sea, and similar
to that of reefs studded with massive corals and Madrepores, which are
usually crowded with boring Annelids, Sponges, and jMoUusks, and which
thrive in such localities, but find nothing to feed upon iu the seolian
rocks forming the base of the ledges of the Bermudas, or in the jcolian
sand flats, the bottom of which is constantly kept iu movement. One
finds only occasionally on the beaches' of the south shore very limited
deposits of flattened pebbles composed of corals, fragments of seoliau
ledges, and shells of Nassa.

We can readily follow off the north shore of St. George the transition
of the eeolian sliore ledges into mushroom ledges, or other patches gradu-
ally becoming coated with coral growths as they come nearer the main
channel into deeper water towards Mun-ay Anchorage. We find here also
a few serpuline atolls and fragments of vertical walls protected by Algte
or other growths. Sargassum, Algte, and Corallines are especially abun-
dant on the inside ledges.

The ledges I have examined immediately north of the main channel,
the southernmost patches of the connecting k-dges, all present a very sim-
ilar structure. They are deeply eroded on the sides and surface. Some-
times one side drops nearly vertically from a depth of two to three feet
at low water to six or seven fathoms. The top is more or less flat, re-
sembling the ledges near shore, and differing from them only in being cov-
ered by a thick growth of Algfe and Corallines, which pi-otects their sharp
edges and ridges from the eftects of the sea. Tlie other faces are more
or less sloping, dropping in steps much as the shore cliffs do, and they
are more or less undermined and honeycombed. One can sometimes trace
what may perhaps have been the low-water shelf of the ledge before
the subsidence had reached its present height, when it was a part of
the old shore line cliff", or one of the outlying rocks or islets.

On the Devil's Flats there are large patches whicli have in some cases
been covered with bank sand, leaving the feolian rocks exposed oidy on
the outer edges, where they are covered with the usual coral growth.
In Port Royal Bay, iu Great Sound, and in Hamilton Harbor there are
many rocky patches rising above tlie sandy bottom on which Oculina^ are
growing. Along the north shore tlie rocks are genei-ally thomughly hon-
eycombed innnediately above high-water mark ; between that and low-
water mark they show signs of abrasion and of the solvent action of the
sea. The low shore cliffs usually extend from low-water mark outwards

AGASSIZ : BERMUDAS. 245

in shelves of very varying width, terminfiting either abruptly, or passing
into deeper water, eitlier l)y one or more steps, or by a gradual slope.
The slielf immediately above low-water mark is usually protected against
abrasion by species of Algje, or small barnacles, or patches of Serpulae
or Mvtilus, or a thin coating of NuUipores, while below or at low-water
mark Sargassum and coralline Algse begin to grow.

It is easy to trace out on the chart the former connection of the flats
with the present land surfaces in all directions. The evidence obtained
from an examination of a number of ledge patches between the north
shore and the south side of the main ship channel is, most conclusive that
all these patches are only teolian ledges, parts of the clitls which once
were connected with that shore and have become separated from it by
causes similar to those now acting upon its cliffs. Upon these ledges
have gradually grown Algje, Corallines, a few Porites, Gorgonians, and
Millepores. When the patches are close to the shore Algre predominate.

The shore cliffs extending into the sea usually have vertical or steep
faces, and one can readily follow their indented and honeycombed out-
lines to a depth of three to four fathoms or more, where the base of
the cliff passes abruptly into the coarse bank bottom. An examination
of the patches to the north of the main ship channel shows ledges with
tlie same structural features, excejit that their surface is more thickly
coated with corals, Corallines, and Alga^, as well as Millepores ; we find
also a few indistinct serpuline atolls on these patches, but their number
cannot be compared with those of the south shore. Many of the ledges
are only protected by small barnacles and Algae. Off Bailey's Beach
there is a row of isolated cliffs and ledges forming an outer barrier to
the bay, the remnants of the hills which once separated what now forms
Bailey's Bay from the sea. The interior patches of ledges are more iso-
lated ; they stand out vertically, or nearly so, in from five to six fathoms
of water, while those in deeper water nearer the outer belt of ledges may
be more or less choked and covered up by the masses of coralline and
*olian sand constantly accumulating and forming there.

Some of the ledges which are not too far removed from the shore
line, like those off" Bailey's Beach or some parts of the north shore off
Spanish Point, are most instructive as showing ledges which still are
capi)ed by seoliau pinnacles, of which the aeolian stratificatif)n is most
distinct. In the submerged base of these pinnacles the stratification
has completely disappeared, and between high and low water mark the
seolian rock has been changed into a hard compact ringing limestone,
more or less worn and honeycombed by the solvent, as well as the

246 BULLETIN : MFSEUM OF COMPARATIVE ZOOLOGY.

mechanical, action of the sea. This irregular honeycombed and cavernous
sui-face extends to the base of the ledge, where it passes into the coarse
bank bottom. The base of the ledge may spread somewhat, or it may
have been greatly denuded above low- water mark, so as to form a wide
base for the seolian pinnacle surmounting it. The interesting feature,
however, is to trace the gradual increase of coral and GJorgonian as well
as Nullipore and Coralline growth upon these ledges below low-water
mark, as we examine tliem both in deeper water and at a greater distance
from the shore. So that when we reach a certain distance from shore
where ledges surmounted by ajolian pinnacles are rare, and where we
find only ledges reaching up to low-water mark, we soon pass into the
coral patches, where the coral growth has become so vigorous that it
appears at first glance to have been itself the builder of the patches,
having so completely buried under its coating the seolian ledge which
constitutes its foundation. Unless one has traced the gradual develop-
ment of these coral patches from teolian ledges through all their transi-
tions, such an intei-pretation would be most natural.

The rocks and ledges off Craw Point out to the ship channel, and the
rocks and ledges off the north end of Shelly Beach (the Stags), leading
to the outer patches as far as the south side of the main ship channel,
all tell the same story. We have everywhere the gradual change of an
seolian cliff" which has become detached from the shore passing into a
ledge, and, according to the distance fi-om the shore and depth of water
becoming a ledge coated with Millepores, Algae, Corallines, and coral
growth, known as " coral heads." The more massive corals and forests
of Gorgonians thrive better on the patches near the flats, or on the ledge
fiats themselves. There is a fine lot of patches to the westward of
Mangrove Bay ; they are seolian ledges close to Ireland Island, which
gradually pass into coral and Gorgonian patches as one goes to the
westward.

To the west of Mangrove as well as to the west of Daniel Island the
patches are in comparatively shallow water, and are surrounded by
great stretches of sand, the ledges being more widely separated and
cropping out in greater number close to the outer edges of the sand flats.
Gorgonians and Millepores flourish mainly on the inner flats, while
corals grow, but not in abundance, on the outer ledges. These sand
flats with pretty steep slopes seem to be due to the disintegration of
great numbers of ledges which must have yielded more readily than
ledges elsewhere on the bank to the destructive agency of the sea. An
elamination of the ledges of the great sound bounded by the Daniel

AGASSIZ: BERMUDAS. 247

Island Flat shows them to be similar to those of corresponding position
which we examined oft' the north and south shores of the main island,
and there is nothing to show that thej, any more than the ledges just
mentioned, owe any considerable part of their increase to coral growth.
We found here many patches of limited extent, with nearly vertical or
steep faces, greatly honeycombed, and worn and covered with Algte,
Corallines, and coral growth, some of them rising from seven fathoms up
to near low-water mark.

The corals on tlie ledges of Brackish Pond Flats increase in profusion
on the patches as they increase in distance from the main channel. But
the appearance of the animal and vegetable growth on the ledge is prac-
tically the same on all the ledges of the Bermudas ; it is a question qf
quantity mainly. The greatest profusion of corals and Gorgonians, as
far as I have observed them, has been found on the ledges of the flats
of the northern, northwestern, and northeastern parts of the banks.
The ledges and patches to the west of Iceland and Somerset Islands are
connected with the patches to the westward, and form a continuous line
of flats as far as the Western Reef Flats. They constitute a series of
proto-Bermudian clifl" ledges which have been worn away from the shore
cliffs, or from the edges of former lagoons and sounds, and have been
overgrown by a tliin veneer of corals, Millepores, and Gorgonians.

The west shore of Somerset has been greatly encroached upon by the
sea ; its northern extremity has been divided into a number of islands
terminating with Ireland Island and the islets flanking it. One of
its extensions forms the western line of rocks and islets of Mangrove
Bay. It was formerly connected with the spit running from High Point
to Wreck Hill, but the sea has eaten its way through, and the islets run-
ning north from Wreck Hill are the western barrier of Wreck Bay. All
along the shores numerous mushroom-shaped rocks are seen, either iso-
lated or still connected by a basal ledge, especially in the ledge running
south of Daniel Island. From Wreck Hill Bay to Ireland Island innu-
merable ledge patches are found, with from two to six fathoms between
them. These patches extend in a wide flat to the westward, forming
what is known as Elies Flat until they join the western ledge flats to
the eastward of Chub Cut, where there is a narrow and somewhat intri-
cate channel with four to five fathoms of water leading from the inner
waters to the outer bank. This channel separates the Western Ledge
Flats and the flats to the eastward of the Blue Cuts. The bottom of the
channel is covered with massive corals, Gorgonians, and Algae. An
examination of patches which reach out from the shore, and have been

248 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

disconnected from it but comparatively recently, gives us the key to
the formation of coral heads and of the ledge flats. Many of these
patches are not as yet covered with coral growths of any kind, and their
origin can still be plainl}^ traced. From these we pass to more distant
patches, in somewhat deeper water, on which Millepores, Gorgonians,
Algae, and corallines have begun to obtain a foothold, but of which the
ledge structure is still apparent. Some of the ledges and patches on the
north side of Spanish Point show* admirably the passage from the jeolian
rock cliffs, which have fallen into the sea covered only with a thin coat-
ing of Algae, to ledges with a more abundant growth of Corallines and
Algae, and finally to patches with corals and Gorgonians at a greater
distance from the shore.

The mouth of Wreck Bay is protected by a number of islands running
across the opening of the harbor, the remnants of the land which once
connected Somerset Island with the main island. The rocks and islets
to the westward of Mangrove Bay are the continuation to the south of a
series of ledges which connected it once with Ireland Island. Traces of
the proto-Bermudian land are found in the many patches of ledges to
the westward of Ireland Island which extend towards Green Flat and
thence to the west of Mangrove Bay, reaching out to Cow Ground Flat.
The patches between Wreck Bay and Daniel Island reach out to Elies
Flat and connect with the Chub Cut Flat. They are the northern
boundary of an extensive sound bounded on the south and west by the
flats reaching to the w^estward of Hogfish Cut, and sweeping northerly
east of the Chub Heads to join the Western Ledge Flats. Similarly,
south of Chub Cut, Elies Flat is the sunken boundary of a smaller sound
bounded on the north by the western extension of the Cow Ground Flat.

On the western ledges we find large patches of sand intervening be-
tween the ledges on which corals grow ; sometimes these sand patches
form long sand bars with coral-bearing ledges only on the windward
edges, the lower ledges having been triturated into sand which is more
or less shifting according to the direction of the wind.

The formation of the sand flats from the disintegration of the seolian
rock ledges shows how little material the corals have sup))lied to form
the flats ; they often come up close to low-water mark, and yet no coral
sand islets have been formed anywhere on the ledge flats, either near the
outer reefs or on the interior flats. These sand patches gradually pass
into the ledges forming the outer flats, where the coral growth is most
abundant and gradually diminishes on the sloping ledges of the sea face
to a depth probably of twelve fathoms. The finest corals and Gorgonians

AGASSIZ : BKKMUDAS. 249

appear to have their limit at a much less depth, in from five to seven
fathoms. Beyond this deptli the broken ground sets in, which the tish-
ermen state that they can trace to seventeen or even twenty fathoms.

The ledges on the sides of the causeway connecting St. George with
the hills to the east of Harrijigton Sound, together with the flats which
connect them, indicate tlie former existence of a chain of hills which have
been disintegrated. Tliese flats form the platform of tlie east and north
sides of Castle Harbor, and they give us an explanation of the sand
flats to the westward of Ireland Island. The causeway flats are literally
packed with coralline Algse.

To the northwest of Western Blue Cut there is a stretch of coral
covered ledges, which, like the ledges to the north and east of a line run-
ning: west of Ireland Island, are somewhat isolated, and have remained
disconnected from other ledge flats. They have not, like the Devil's
Elat, and those to the westward and southwestward of them, been
pounded and ground up to form coarse sand ledge flats with ste;ep
slopes, from the surface of which scattered aeolian ledges b.arely projt^.ct
high enough to allow a scanty growth of Millepores and Gorgonians.

On the inner side of the reef massive corals do not as a rule seem to
exist beyond four to six fathoms, the point at which the great expanses
of coralline bottom begin, and which extend nearly unbroken to the great-
est depths of the inner waters on the banks.

The greatest width of the belt in which corals grow from the inner
edge of these flats or patches to the outer six or seven fathom limit is
about three miles at the eastern extremity. This is nearly the width of
the land and water belt included between tlie island of St. George and the
entrance to Castle Harbor. The belt between Ireland Island and Gibbs
Hill is however considerably wider than any of the ledge flats. At the
western end it is not more than a mile, the ring of ledge flats being wid-
est west of Mills Breakers, and diminishing towards North Rock. The
ledge flats are much narrower along the whole western and southwestern
face of the Bernnidas, The gradual shelving slope of the ledges which
have been abraded on the sea face of the flats is well seen between the
entrance to Hogfish Cut and the Western Blue Cuts. The Little Bar
and Chaddock Bar form two wide spits with a gradual slope from two
to seven fathoms, covered by Gorgonians, Corals, Millepores, and Algae.
Long Bar is a similar ledge, separated however from the Western Ledge
Flats by a channel of from six to seven fathoms, the bottom of which
is covered with corals and the attendant Algfe and Sargassum. These
bars are full of just such ledges as have been described, only they have

250 BULLETLN- : MUSEUM OF COMPARATIVE ZOOLOGY.

"been the first to feel the effect of the inroads of the sea upon the proto-
35ermudiau hind, and have been abraded to a greater depth. "We find
ou their slopes as runeh as four to eight fathotos as a genei-al depth ; the
channels between the ledges are in most cases overgi-own with coral and
Algse. Oataide of the six or seven fathoms at which they still flourish,
Goi^onians and Alg'* extend ou the broken ground down to a depth of
seventeen to tweutv fathoms, with here and there an irregular sandv
patch between the l&iges. In the channels between Long Bar and Little
Bar" and the Ledge Flats, which have a general depth of sLs to eight f:\th-
oms, corals and Gorgouians grow in patches which are separated by coarse
bank sand. The same slope similarly overgrown extends to the Chub
Cut, and from the Western Blue Cut it becomes narrower towards the
Southern Ledge Flats.

On crossing the Ledge Flats opposite the southern end of Long Bar, one
meets the same ledire^, but more worn and covered bv a largrer number
of massive corals and Gorgonians. In the channel between the South-
west Ledge Flats and Long Bar, which is itself made up of patches similar
to those of the outer edge of the Southwest Ledge Flats, the heads and
patches do not come so near the surface, they form patches of massive
corals, Gorgonians, and Corallines, or Algae separated by areas of clear,
coarse l>ank sand bottom. Such is the character of the outer rim of the
reef, wherever we examined it, to the westward of Hogfish Cut, beyond
Chaddock Bar, outside of Chub Cut, to the west of the Blue Cuts, out-
side of North Rock Flats, and to the south of !Mills Breaker and off the
outer reef of the south shore. The ledge patches and coral heads increase
rapidlv in height and number as we approach the outer edge of the flats
and Gorgonians, Corallines, and massive corals become more abundant
also in the spaces between the patches. The bulk of the corals and
Gorgonians do not seem to grow beyond ten to twelve fathoms ; beyond
that depth Gorgonians, Algse, and Corallines preponderate, and cover
the bottom.

Off Hisrh Point extend the Bream and Kitchen Flat Ledges. Thev
are like all the other ledges of aeolian rock, with more or less vertical
honevcombed sides. On Chaddock Ledge there is a depth of two to five
fathoms. It is, like Long and Little Bars, made up of ledges in some-
what deeoer water. It is continuous with the flat ledges to the west of
Hog Fish Cut, and not separated from them by a channel. On Chaddock
Bar there is a fine growth of Gorgonians, of Corals, of Corallines, and of
Algae, which stop in from six to seven fathoms, where we pass into the
broken ground described above. The bottom of the channel of Hogfish

AGASSIZ : BKUMUDAS. 251

Cut in eight fathoms, as it passes out on to the bank becomes hard, and
is covered with Thalassia.

On the inner edge of the reef to the north of Three Hill Shoal,
starting from Mills JJreaker Passage, we could observe in the close
network of ledge patches no differences between them and those of
other localities. The massive corals are perhaps tiner and more nu-
merous than elsewhere on the reefs. The Ma,'andrinas and Astrseans
are more abundant, as well as the Gorgonians, Algse, and Corallines.
The submerged faces of the reef ledges, as examined through the water
glass, show no difference from those of similar ledges, such as we see
risin<^ from six to seven fathoms of water to a depth of two or three
fathoms at low water, on which corals, Gorgonians, Algae, and Coral-
lines have not as yet obtained a foothold. In the deeper parts of
the interior sounds, in from ten or more fathoms (sixteen at the out-
side), the bottom sand is much coarser than we find it in the shallower
patches somewhat protected by the reef ledges, or in the reef bights in
which patches of sand run in a considerable distance between the ledges.
The reef ledges close to the edge, with nearly vertical or very steep sides,
in from ten to eleven fathoms, are often separated by deep passages
covered with sand, though occasional patches of Gorgonians and Algae
or Corallines grow over this bottom, and form connecting bottom strips
between the ledges. A considerable amount of dead material accumu-
lates at tlie foot of the reef patches and ledge flats, and, according to its
position, is being slowly ground into the characteristic bank sand bottom
composed of fragments of Millepofes, Corallines, Algae, Gorgonians, and
Nullipores.

The " breakers " known as special rocks on the outer edge of the reef
flats, such as Soutliwest Breaker, the Mills Breaker, North and Northeast
Breakers, and many others, of which the North Rock is the most promi-
nent, are the remnants of islands and islets or of ranges of a?olian hills
which once rose upon the outer reef flats, and surroimded the now
snuken sounds, the lagoons and waters of the inner part of the Bermudas
to the northward of the islands. They have by most observers been con-
sidered as owing their origin entirely to the growth of the corals we find
thriving upon the sm-face of the ledges which compose these pat.che.s.

There are also three or four breakers bare at low water between the
North Eock and the Pilchard Dicks. The Southwest Breaker is the
westernmost of a series of ledges parallel to the south shore extending
to the eastward as far as the entrance to St. George Harbor, the geolian
character of which can readilv be observed. The inner ledges extending

252 bulletin: museum of comparative zoology.

from tlie south shore reef towards the main island are the remnants of
the platforms of rocks once rising above higli-water mark, or forming
perhaps small islands, rocks, and islets across the bays of the proto-Ber-
nuidian land. We find to-day such islands and rocks separating Castle
Harbor from the sea, those across the mouth of St. George, or islands
beloniiini;- to an outer line of ledges which may be entirely disconnected
from land promontories, or form, as they do across Whale Bay, Sinky
Bay, and parts of other bays, an outer barrier protecting the south shore
somewhat from the beating of the surf till they have crumbled and in
turn been reduced to ledges bare only at low-water mark. The true
character of many of the ledges forming the flats or the connecting
patches is hidden b^' the coral growth. But both on the north and on
the south shore we can follow the passage of the seolian rock ledges
as they recede from the shore, from nearly bare ledges still connected
\vith tlie sill ire cliffs to the coral patches. The ledge at Briggs Flat is
mainly covered with Gorgonians and Millepores. We find there but few
heads of massive corals; the}' are small Maeandrinas and Astiaeans, to-
gether with an abundant growth of Sai'gassum, Algje, and Sponges. The
Sponges are more abundant on the connecting ledges, if I may so call
the patches extending from the north side of the main channel towards
the flats, than tlicy are upon tlie outer ledge flats.

As far as wo can judge from sucli an examination as can be made in
crossing the reef flats from tiie inner waters to tiie ()])en sea, in the sec-
tions across tlie reef at Hogfish Cut, across the Western Ledge Flats at
Little Bar and opposite the west end of Long Bar off the Chub Heads,
across Chub Cut, across the Blue Cuts, across at the Northwestern Ledge
Flats, across at the North Rock, Northeast Flats, Mills Breaker Passage,
and the main channel, all the " coral heads " or patches seem to be
growing on the tops of pinnacles of seolian rocks, or of flat ledges, or of
mushroom-shaped tables, or of large irregularly shaped ledges rising
sometimes gradually in irregular shelving strata, or in nearly perpen-
dicular steps, fVom six or seven fathoms of water to near the surface.

Passing through Chub Cut to the outside of the reef, we find in four
to five fithoms large Mteandrinas, Astra^ans, and fine Gorgoiiians, together
with the usual accompaniment of Millcporcs, Sargassum, Corallines, and
other Algae. As we pass into deeper water tlie massive corals become
smaller ; in seven f itlioms they are quite small and not inimerous, and the
whole lutttom becomes thickly covered witli (;oi-gouians, Corallines, and
Sargassum.

An examination of the charts of the Bermudas will show many places

AGASSIZ: BERMUDAS. 253

outside of the reef to a depth of twenty fathoms which are marked
rocky {>•). These spots are most probably the outcrops of £BoHan ledges
of the proto-Bermudiau hill lands projecting slightly above the sandy
bank bottom and forming a part of the broken ground. Beyond that
depth (twenty fathoms) the lead brings up what is called coral bottom,
made up in great part of a^olian sand and of fragments of Corallines,
Algjie, and the like. Tliere are also patches of this rocky bottom inside
of the reef ledges, as for instance close to the Western Blue Cut, where
all the hauls of the dredge only brought up small quantities of the bank
sand bottom.

The bottom over the Bermuda Bank is quite uniform in character.
The greater part of it is covered with seolian sand of different degrees
of coarseness, and more or less mixed with fragments of coralline Algse
and of Millepores or Gorgonians.

In other localities the surface of the old seolian rocky ledges is exposed,
and is comparatively bare of seolian sand, as in some of the sounds, and
the bottom may be called rocky. On this Oculinse grow in profusion in
the deeper waters of the sounds, or the more massive corals where the
sea has free access to the sounds.

To the westward of Wreck Hill there is a small extent of bottom in
seven fathoms of water covered with very fine mud, much like the white
marl off Andros. A similar patch of marl occurs to the eastward of
Ireland Island.

THE SERPULINE REEFS.

Plates XXI. to XXVI.

The serpuline reefs described by previous observers are perhaps the
most interesting structures of the Bermudas. Thev are most numerous
off tlie south shore, constituting miniature atolls and barrier and
fringing reefs apparently formed by the upward growth of SerpuL-p.
While Serpulse undoubtedly cover a great part of the surface of the
structures, yet Algse, Corallines, barnacles, mussels, and other inverte-
brates, are found to be fully as abundant as the Serpulee, which in many
cases play only a secondary part in the organic covering. In fact, it
would be as ci^rrect in some localities to call them Algse or Coralline
atolls. Neither the Serpuhe nor the A]ga\ nor any other organisms,
have to any considerable extent built up the vertical walls of the differ-
ent kinds of diminutive reefs so charactenstic of the south shore. The

254 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Serpulse, Algae, Corallines, and other growths have only protected the
surface of the mushroom-shaped seolian rock ledges which form these
structures from the action of the breakers. They have not built up the
raised rims of the atolls, or the crescent-shaped or the horseshoe-shaped
reefs, or the vertical walls forming the irregular convolutions and curves
of the broader ledges.

Before my visit to the Bermudas I accepted the explanation given by
older writers of the mode of formation of the atolls, as due to the
accelerated growth of Serpulte on their outer rim. I was therefore
greatly surprised, on hammering at some of these structures, to find
that the vertical walls were not built up, as is generally believed, of ser-
puline limestone, but were composed of seolian rock, and to discover that
in many cases the elevated rim was protected by the hard ringing crust
so characteristic of limestones exposed to the action of the sea, and fur-
ther to find that the coating of Serpulie, of Algse, of Corallines, and of
Nullipores was quite superficial.

Some of the serpuline atolls are circular and quite regular in outline,
others crescent-shaped, while others are apparently formed by the ac-
crescence of two or three atolls. Some of the circular atolls are sym-
metrical, with a central depression, at the bottom of which more or less
sand has gathered. The rim of these atolls may project from a few
inches to one and a half feet, or even more, from the nearly vertical base ;
its surface is completely covered by a thick growth of different species
of Algse, Zoanthida', Corallines, and Serpulfe. The rim varies greatly in
width ; in some cases it is not more than eight to ten inches, in others
from one to five feet, and in some cases there is only a small circular
pot-hole or a very circumscribed area left bare of growth in the centre.
The rim is often greatly developed on the weather side, forming a
crescent, tapering gradually to a thin wall on the opposite side. The
crescent is often open for a great part of the circumference, the weak wall
of rock forming its lee edge having been carried away by the breakers.

On the outer reef the ledges which are awash are similarly constructed.
It is true there are few of the regular atoll shape, by far the greater
number being long ledges of compoinid atolls made up of diminutive
crescent-shaped reefs. Upon these leilges low vertical walls have been
cut out varying from six to eighteen inches in height, following all sorts
of curves, rising like a succession of S-shaped loops of circular or cres-
cent shape, or re-entering curves, running in all possible ways, and which
at first sight would appear to be all due to the growth of vegetable and
animal life which covers the top and sides of the walls.

AGASSIZ: BliRMUDAS. 255

The Algae, Serpulae, Corallines, Mytikis, and the whole growth which
goes to form the serpuline atolls, form but a thin coating upon the
ledges of seolian material upon wliich they happen to have grown.
Underlying this animal and vegetable coating we find the peolian rock,
which on some parts of the ledge may still be protected by the hard
rinsini' crust so characteristic of Bermudian and Bahamian limestone.
The inner parts of the pool or atoll within the raised walls is com-
posed of softer material, or of material whicli has not been protected by
animal and vegetable growth from the destructive agency of the sea.
The serpuline atolls are seolian i-ock ledges which once were a part of
the south shore cliffs at the time when the shore line was farther
to the southvvard and had not yet begun to yield to the inroads of
the sea.

The protecting growth of the atoll has little to do with the formation
of the wall forming the rim of the atoll ; in some cases it has undoubt-
edly grown up perhaps twelve to eighteen inches above the wall itself,
but the deep lagoons and steep vertical walls of the serpuline atolls so
characteristic of the southern side of the islands have been formed, I
believe, by the mechanical agency of the breakers. Tiiese diminutive
atolls are large pot-lioles excavated l)y the surf and sand, and the varied
forms of circular or of crescent-shaped reefs, of barrier reefs, and all the
possible modifications one finds on the south shore of the Bermudas, are
primarily due to the mechanical action of the sea. All these structures,
from a circular or elliptical atoll to a barrier or fringing reef, with all
their possible modifications, are due to the action of the surf and the
sea in wearing away the surface of the mushroom-shaped rock, which
is either softer than the surrounding parts or is not protected by the
covering coat of Algae, Corallines, or Serpulae.

One can off the south shore trace the whole process from the time
when the large fragments of shore aeolian rock fall by undermining into
the sea, until they are changed by the action of the surf into mushroom-
shaped ledges surmounted by pinnacles, and next into the stage when
the pinnacle has in turn been undermined and dropped alongside of the
ledge to become the holding ground of coral and other growths. The
surface of the flat ledge which formed the base of the pinnacle is now
freely acted upon by the breakers. According to the nature of the
upper crust, and to the extent of protection given to it by the covering
coat of animal and vegetable life, the sea acts upon it, and we have hol-
lowed out diminutive circular atolls, crescent or horseshoe-shaped struc-
tures, as well as the curved, straight, or convoluted or looped vertical

256 BULLETIN : MUSEUM OF CUxVlPARATIVE ZOOLOGY.

walls of broader ledges which stand up from the bottom and seem to
have been built up by the organisms covering the surface.

The serpuline atolls are of all shapes, depending primarily upon that
of the slab from which they happen to be formed. We may imagine
one of the shore slabs or ledges more or less overgrown with Algae
and Serpulse exposed to the action of the incessant breakers of the
south shore. The sea face of tlie ledge either shjpes rapidly or is more
or less vertical, sometimes undercut or worn to a mushroom-shaped
table. According to the hardness of the protected edge of the ledge or
of its surface, it becomes more or less broken through by incipient pot-
holes, which expose the softer a;olian rock to the action of the sea.
With each tide the wearing action increases, until a circular pool is
formed, in which sand is constantly triturating and grinding away the
softer surfaces. Tlius a miniature inner lagoon becomes excavated, not
more than a few feet in depth, and surrounded by a more or less regular
rim ; the depth of some of the shallower lagoons varies from twelve to
fifteen inches. On a ledge in which a pot-hole has been formed the sea
thus washes at first into a shallow dish, or into a series of dishes which
are soon run together, and thus a straight or curved or S-shaped verti-
cal wall may be excavated on the edge of a ledge of seolian rock, the
inside depth in one case being eight feet.

The serpuline atolls take their greatest development towards the
western part of the south shore. Otf Great Turtle Bay we find the
same extraordinary development of the serpuline atolls and reefs which
we traced farther to the eastward, off" Hungry Bay and off Elbow Bay.
There is hardly a sunken ledge on or along or off the south shore of
which the surface is not protected in some way by Algae and Serpula?,
and covered with structures which are directly the result of the action
of the sea upon the friable aeolian rock of which the ledges are com-
posed. It is indeed a remarkable sight to see, as far as the eye can
reach in either direction, this narrow belt of ledges which have been so
sfrangely modified by the action of the sea and the protecting agency
of the animal and vegetable growth upon its surface.

The presence on the south shore of so many striking circular atolls
and horseshoe-shaped, crescent, or curved rings, or partial rings, and
S-shaped walls, withdraws the attention from the far greater number
of mushroom-shaped blocks and ledges which no longer reach the sur-
face, owing to the wearing of the aeolian rock of which they are com-
posed. The atoll-shaped ledges have attracted more notice, not on
account of their greater number, but maiulv from the interest centring

AGASSIZ : BERMUDAS. 257

in such structures. For alougside the atolls, either rising to the surface,
or near to it, or always covered at low water, there are other ledges, of
endless differing shapes, which do not attract the eye, but which play as
important a i)art in the economy of tiie ledges ofl" the south shore as the
atoll-shaped structures themselves, and which give us the clue to their
formation.

Everywhere on the shores of the Bermudas where active degradation
of the coast is going on we meet with a number of ledges, or pinna-
cles, or islets, or mushroom-shaped rocks, which have been fiishioned
by the sea into a nearly circular or elliptical form. Stmietimes a num-
ber of these isolated rocks may stand in a row above high-water mark,
the stems almost eaten away from the ledge upon which they stand.
When the top tumbles over, tlie support, or a part of it, may remain well
above low-water mark. It is upon these ledges of all sizes, ft'om a foot
or so ill diameter to long elliptical or irregularly shaped masses of fifty to
seventy feet in length, or even more, tliat the sea begins to act, and to
shape the serpuline atolls of tlie st)uth coast, though they are not con-
fined to it, as I siiall show hci'eafter. Standing on some parts of Elbow
Beach, one may follow tlie irregular muslu'oom-shaped rock ledges stand-
ing between high and low water mark to those at and beyond that point
into deeper water. We may note the changes which gradually take place
as the protective growth upon these irregular ledges, at first bare, trans-
forms them into the atolls, or crescent-shaped or S-shaped structures
forming the reef off the south shore. One can watch at low tide and
see the breakers combing in over the rim of the little atull scouring the
lagoon, and the supm'fluous water ihiwing over its sides. The sea breaks
over the edge, carries off" such loose fragments as may have been started
by the preceding rollers, and scours the inside with the sand it may have
brouiiht in, in addition to what it finds inside.

Some of the crescent-shaped serpuline reefs are formed on ledges bare
at low water extending out from shore. They form low vertical walls of
from twelve to twenty-four inclies in height, running in a series of irreg-
ular curves, a kind of festoon as it were, protecting tlie inner lagoon or
lagoons of all sizes and sba])es which have been gouged out by the waves.
It is not uncommon on the south sliorc to find fine sand deposited on
the flat ledges near low-water mark, and kept in place by the growth
of a thin sward of Algae; this, together with the thin crust formed over
its surface, hardens the mass, keeps it in place, and enables it to resist
the moderate action of tlie breakers.

(loinii- ^vestward from Great Turtle Bay to Warwick Bay we find the
vor, xxvi. — \o 2. 17

258 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

same condition of things, — an outer ledge of boilers together with irregu-
lar inner rows of ledges running close to the line of low-water mark, and
gradually passing into the mushroom-shaped ledges which still form a
part of the shore cliffs.

The district extending from Sinky Bay west and east is specially in-
structive, as showing the method of irruption of the sea through low
shore clift's to furm small boat bays, and the gradual passage of these
shore clift's to lines of rocks and islets running parallel to the coast, and
cutting out such bays as Whale Bay, Bailey Bay, Warwick Bay, Great
Turtle Bay, etc. We may next follow the passage of these cliffs to
submerged ledges, and their transformation into the boilers off" the
south shore and the outer line of boilers forming the so called reef
off the south shore. The most striking of the serpuline reefs are the
fringing and barrier reefs, and their outlying atolls, off one of the points
at the east end of Whale Bay, together with the lines of atolls and
variouslj' shaped serpuline reefs extending to the eastward. Nowhere
perhajis on tlie south shore do we see so clearly the transition of the
isolated mushroom rock ledges surmounted by eeolian pinnacles into the
ledges which are to become serpuline reefs, as in the district between
Great Turtle Bay and the bay at the foot of Gibbs Lighthouse.

From the descriptions given above of such a variety of reefs formed
by the serpuline ledges, and of the action of the sea ui)on them, we
may obtain on a small scale an illustration of the mechanical theory
of the formation of some coral reefs. This may be specially applicable
to the formation of compound atolls, as has already been suggested.
We find off the south shore, in the same area and suliject to identical
conditions, patches which assume the shape of atolls of fringing or bar-
rier reefs all within a stone's tlirow of one another. But in this case
the structure of the foundation gives us the explanation of their fori"na-
tion, for the shape of these diminutive reefs is primarily determined by
tliat of tlie ledge, and not by the growth of the Serpula?. The different
shapes of tiiese diminutive reefs can be traced to the manner in which
the sea has acted upon their feolian sul)structure. It may have only
honeycombed the surface of a large cliff fragment, and left it as it fell,
merely covei-ing its diminutive spires and hollows witli a thin layer of
Serpuhr, INIytilus, Alga% and Corallines. It may have washed off from the
shore cliffs slabs of a?olian rofk in such a manner that as they lie on tlie
beach the strata are horizontal, and, the edges having become cemented
by the action of the sea, the ilivision lines become obliterated, and over
their surface has grown an animal and vegetable covering. It is not

AGASSIZ : BERMUDAS. 259

an uncommon thing to see on some of the beaches large shibs of base
rock, upon the hard rhiging surfiice of which grow AlgfB ; these collect par-
ticles of sand, and thus form a coating from a quarter to three quarters of
an inch thick, upon which larger Algie then flourish. In the intermediate
spaces grow tlie Serpulse, Mytilus, and Corallines, whicli soon conceal the
surface of the ledge by their protecting coat. Should this slab alone or
with adjoining slabs form an extensive ledge far enough out from the
beach to be exposed to the action of the breakers, its nearly vertical
sides would form a rampart over which the sea combs and pounds down
over the edge of the slab, striking beyond the outer edge well toward its
interior, according to the size of the breaker. If there exist at the
points reached by the breakers any weak spot in the protecting crust, or
any incipient fracture, or any ditference in the hardness of the upper
layer, the sea soon makes an inmad upon it. It grinds out the softer
interior parts, which are carried off, and thus forms the beginning of a flat
shallow saucer-shaped cup on tlie inner part of the ledge. The outer rim,
on the contrary, protected either by a hard crust or by a growth of Algae
and of Serpulne, remains intact, and gradually rises higher and higher,
partly from the additional growth of the Serpulae and other calcareous or-
ganisms, but mainly by the grinding away of the interior of the ledge to
form a basin, which little by little becomes deeper. The organic growth on
the outer rim is more vigorous than in the basin itself, either on its sides
or on the bottom, where the sea breaks and is at work grinding away the
protecting growth. The Algae, Serpulse, and other growths become less
abundant in proportion to their distance from the outer weather rim,
until, towards the centre of the atoll, the inner ring or cup or slope is cov-
ered with sand. We thus have diminutive atolls, barrier reefs, or crescent-
shaped or horseshoe-shaped structures formed out of tlie jeolian rocky
ledges. Should the sea face of the slab be harder than the faces of the sides
or of the leeward side, some of those strata are soon broken through,
and gaps made in the rim, forming a crescent-shaped wall or arc witli
its greatest height seaward, the wall gradually falling to leeward to the
level of the ledge, and the raised edge sloping towards the horns of the
crescent on either side. Tlie lee face of the crescent-shaped atoll is in such
a case thoroughly scoured by the outward rush of the sea, which, carrying
with it a certain quantity of the sand that fills the depression, runs
off the lower lee side with considerable velocity. The depth of the ntoUs
varies from a few inches to six feet or more. The lee faces of both horse-
shoe-shaped and crescent-shaped atolls are frequently so rapidly removed
as to leave nothing but a vertical wall of from one to two feet on the
weather face of the ledge.

260 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

The diminutive reefs formed off the south shore of Bermuda are of
all possible shapes, — atolls with regular rims of the same width on
all sides, atolls with the sea face rim wider than that to the leeward,
and parts of rims of circular atolls of horseshoe or crescent shape, or
only of parts of arcs of greater or less extent. We also find belts of
small atolls on ledges of considerable width, and atolls of an elongate
type ; others are dumbbell-shaped, foi-med evidently by the breaking
through of division walls of circular atolls. We also meet with chains of
atolls, each one forming a link as it were, or irregular parallel chains,
which, when the separating walls are broken through, give the elements
for all the possible figures assumed by the ledges off the south shore.
When such reefs are formed on the shore ledges, we have all the possible
types of fringing and barrier reefs, or combinations of these, forming
diminutive reefs with low vertical walls apparently most irregularly
jjlaced, often as if their existence in their varying shapes and positions
could be due only to the upward growth of the Serpnla^ and Algte. But
I believe that the vertical growth of Serpulte and Alga3 is not of itself
sufiicient to account for the existence of the vertical walls, and that they
are due only in small part to the upward growth of organic material,
and in a great measure to the action of the breakers upon the aeolian
rock ledges, probably in the manner I liave just described.

The barrier reef off the small spits to the east of Whale Bay at tlie
Targets may he described as a vertical wall surrounding three sides of a
rectangle, the dia^ronal of which is somewhat over fiftv feet ; the sea fnce
corners are well rounded and the side walls formed of short arcs. The
distance fi-om the shore edge of the ledge to the outer wall is about
twenty-two feet. The greatest width of the rim is at the two outer
angles, where it varies from five to six feet; the inside edge of the sides
of the rim gradually passes into the shore ledge, being in a general
way parallel to the trend of the sides of tlie barrier i-eef. The raised
part of the rim varies from twelve to sixteen inches in height. The
outer vertical wall has a height on the west of eight feet, and on the
east of about six feet, the slope of the ledge being more or less con-
centric round the deepest part as a centi'e, and sloping sideways towards
the shore edge. The outer rim at two points is gouged out into two
smaller elongated pits. The rim is everywhere well protected b}' Algae,
Corallines, and Serpulre, the Algfe growing on the more or less level
platform of the rim, and on the outer and inner faces of the vertical
walls of the barrier reef. The ledge is greatl}' undercut, and its outer
faces present all the irregularities of wear by water so characteristic of

AGASSIZ: BERMUDAS. 261

shore ledges. The inner slope is covered by a thin growth of Corallines
and smaller Algae, whiVh do not seem to thrive as well as where
they receive the direct force of the breakers. Close to the westward
of this diminutive barrier reef are ledges which may be regarded as
typical of the changes which have taken place in a bare aiolian ledge
just dropped from the sliore clilfs until it becomes a typical boiler of
the south shore. One of these is sepai'ated from the western edge
of the barrier reef by not more than six feet, and has a depth of water
of more than ten feet in the passage between them. To one side of this
dumbbell-sliaped atoll, which is nearly thirteen feet on its longest axis,
is another atoll on a mushroom-shaped ledge seven feet in diametei-, with
a regular rim and a pot-hole of three feet in depth. This is separated
from the adjoining mushroom-shaped ledge by a gap of four feet, with a
greatest depth of ten feet, and eight feet on the shore edge ; it is sep-
arated from the shore and a large ledge to the south by a deep passage
of nine feet in width. The length of the larger ledge, iiregularly shaped,
is over forty feet, and its width varies from twenty to thirty-five feet.
Its outline is formed bj' curved walls, the edges of irregularly shaped
elongated or circular dumbbell-shaped pits gouged out from a broad
platform of tsolian rock. On the lee part of the ledge an irregular
rectangular pot-hole has been formed, live feet in depth on one side
and four on the other, sloping upward toward the broad outer rim.
Comparatively slight variations would change the surf^Tce of the ledge
into an atoll with a narrow rim following the outlines of the ledge ; or
it may becf>me divided into two irregularly shaped pits by the coales-
cence of the few smaller pits now upon the platform, and a single curved
wall, the remnant of the face of one of the circular pits, would form a
division wall ; or the sea may break through to a greater extent than it
has done already and leave on this ledge only disconnected fragments of
wall of varying shape, in wluch it would be difficult to recognize the
walls once limiting circular or dumbbell-shaped pits. Outside of the
larger ledge is a pear-shaped atoll with a broad sea face or rim sloping-
inward, and a circular pot-hole about three feet in depth, the rim of
which is narrower to the leeward.

All these ledges are deeply undercut and abraded, and are mushroom-
shaped ; their faces are vertical or nearly so, and all show traces of the
action of the sea upon the pillar forming the base. The ledges T have
examined close to the shore, half-way to the reef, or on the outer line of
ledges, all present modifications of the ledges described. Their ultimate
shape depends upon many local factors, and they show but a small nuin-

262 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

ber of the possible intricate figures that are found on the aeoUan ledges
off the south shore. The Algae, Serpulce, and other organic growths
which thrive upon the edges of the ledges and spread down upon the
vertical faces, still further protect tlie sides of the mushroom tables
from being washed or eaten away by the forward and backward rush of
the sea.

Off the same locality an elliptical atoll forty by thirty feet, standing in
a depth of ten feet at low water, formed the spreading top of a mushroom-
shaped ledge from the vertical sides of which all traces of the seolian char-
acter of the substructure had disappeared. Where not overgrown with
Algaj and Corallines, it showed the peculiar gouging out and honev
combing character of all shore rocks, either when exposed at low water
or where extending below it. The inner part of tlie weather rim was not
quite parallel with the outer outline ; it projected in the centre and was
somewhat scalloped in outline, with a few small deep pits. The rim
varied in width from five and a half feet on the weather side to one
foot in its narrowest part on the leeward edge. I could not detect that
the weather rim was perceptibly higher than the leeward rim, though
this is not unfrequently the case.^ The pot-hole was six feet in depth
nearer the western edge, gradually decreasing in depth to three feet
on the opposite side. A crescent-shaped lagoon eighteen feet in diame-
ter, greatly undercut, especially on the weather side, as all these
ledges are, had a rim five feet wide at its widest part, gradually taper-
ino- to four or five inches at the two extremities of the crescent. Both
this and the elliptical atoll just described had a narrow shelving plat-
form on the inside of the weather rim. The lee edge of this crescent-
shaped atoll was worn away five feet lower than the weather rim, A
circular atoll twenty-five feet in diameter was surrounded by an irregu-
larly elliptical rim twice as wide on the weather as on the lee side, form-
ing a scalloped pot-hole with a greatest denth of three feet. Along the

1 Heilprin speaks of the Serpulae as occurring " in dense bundles," and where
the surf beats liardest " tlie Serpula growth was most largely developed, and to such
an extent as to form a raised rim or barrier to the more protected inner side."
But, as he himself says, (and I am not quite clear whether he attributes the for-
mation of the atoll to tlie Serpulae,) " the breaking in on all sides of the surf has
created a number of more or less irregularly oval islets with depressed centres, or,
more properly, with elevated borders." In the one case he says, " The depression
is merely a negative one, being such by reason of a somewhat more rapid growth
developed only from the water line, or within the surf," but he feels satisfied, how-
ever, that the two structures (these serpuline atolls and the coral atolls), while seem-
ingly alike, have practically little or nothing in common.

AGASSIZ: BERMUDAS. 263

south shore I observed a great number of ledges with nearly flat tops,
from which the surmounting ajolian pinnacles had been worn away ; some
of them were just awash at low-water mark, otljers reached half-tide mark,
and a great number were sunken ledges. On these last, if forming the
inner part of the outer line of ledges, ai-e found Gorgonians and corals,
giving them to a certain extent the appearance of a coral reef.

Off White Cliff Bay there is an excellent specimen of an irregular
crescent-shaped barrier reef formed upon a ledge barely in contact with
the beach at low water. The wall on the inside is formed of short irregu-
lar steps, and the inner area of the ledge is thickly overgrown with cal-
careous Algse, more or less covered with oeolian sand washed from the
friable parts of the ledge.

The narrow pedestals which are the bases of some of the pinnacles of
seolian rock are often the remains of extensive flat ledges on which
the different organic growths characteristic of boilers have obtained a
foothold. The parts of tlie tops and sides of these ledges which have
not as yet been covered by such a growth, or only partially so, plainly
show that they ditter in uo way except in size from the smaller mush-
i-oom-shaped feolian rock ledges, and are formed by the same agencies.
The larger piimacles of pcolian rock like those still standing off" Whale
Bay will become, when they fall, large flat ledges upon which the sea
wlien breaking digs out irregular pot-holes of all shapes and sizes.
The remnant of the l)ase of the pimiacle becomes, when worn away by
the sea, either a shelf or a flat corrugated ledge, and the more or less
vertical sides below low-water mark are worn away by the wash of the
sea into mushroom-shaped ledges.

At the east end of Elbow Beach several patches of linneycombed shore
rock ledges have been left stranded in the midst of the lieach sand sur-
rounding them. These ledges, if exposed to the action of the sea, would
soon be worn flat, and according to the angle of their stratification would
be dug out into atolls or other irregularh^ shaped structures protected
by Algfe and Serpulte. The same serpuline growths and similar convo-
luted walls occur on the North Rock and tiie adjoiniu"; ledLres on the
north. Similar structures, forming more oi- less distinct reefs, also occur
ff" tlie north sliore and elsewhere on tiie edge of the reef flats, but they
are not as numerous nor so well defined as a rule. They consist, how-
ever, always of the same mushi'(.)om-shaped ledges, sometimes still sur-
mounted by tlieir piimacles of ceolian rock, and of others abraded to the
level of low-water mark, or even well below it. The shore ledges on the
south coast are acted upon by short, sharp breakers, formed in froi>» 'me

o

264

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

and a half to two fathoms of water, beating upon the ledges much as
a small waterfall drops upon the rocks at its base. On the north shore
the breakers and sea do not act with the same regularity as on the
south shore. There, owing to the existence of an outer line of reefs,
the conditions are more uniform than on the north shore, where the
effect of the winds upon a comparatively broad stretch of sea are far
more variable.

On the north shore the surface of the ledges between high and low
water mark is protected mainly by a small species of barnacle, clusters
of small Mytilus and incrusting Nullipores, and a few species of small
Alga?. At or below low-water mark Sargassum begins. Serpula? are
not as conunon as on the ledges off the south shore. On the north shore
serpuline atolls are most numerous on the ledges off Spanish Point in
an easterly direction for a distance of four or five miles.

The mode of formation of the peculiar and intricate windings of the
vertical walls which crop up to the surface on the summits of the ledges,
and which take on such complicated curves can readily be explained
from the manner in which the ledges themselves yield to the action of
the surf from the wash of the sea, and also from the angle at which the
seolian strata lie when attacked by tlie waves.

The following diagrams will further explain the mode of formation of

the various serpuline reef struc-
tures whicli have been described.
AB (Fig. 1) is a piece of shore
cliff' which has become isolated
from the sliore ; tlie seolian lam-
ination is clearly seen above
high-water mark. Below high-
water mark it is honeycombed
and eaten away, leaving the
rcnlian pinnacle supported only
by a slender stem i-ising from
an extensive base more or less
covered with Alga'. Serpuhc,
and other gmwtlis. The surface
of the ledge, as well as tlie base
of the mass extending below
low-water mark, is more or less eaten away, and when the feolian pin-
nacle (A) has fallen off a mushroom-shaped mass is left, the upper sur-
face of which mav be above or below low-watcr mark. All trace of

/

Fig. 1.

AGASSIZ: BERMUDAS.

265

HW.

^i= — j L.W. -nl

Figs. 2,3.

aeolian stratification has been obliterated by the cementing and solvent

action of the sea water. If the base, the mushroom-shaped ledge, is

stratified horizontally (Fig. 2)

the result of the wash of the

breakers upon any part of tlie

top left unprotected will he to

dig out circular or elliptical

atolls (Fig. 3) like A or B. In

one case, A, the atoll will have

a rim of nearly the same width,

while in the other case, B, if

tlie softer parts of the top are

on the lee side, the atoll will

have a wider rim on the

weather side from B to A, or

a pot-hole may also be formed

between A and B. The pot-holes of these circular atolls are usually

from three to four feet in depth. But in some cases I have measured

them between five and

si.\ feet, and eveil more. '^' '^'

In others they are only
a few inches deep. I
have not observed any
growth of Serpulge of
greater thickness than
from twelve to eighteen
inches.

Should the aeolian
strata dip towards the
lee side (Fig. 4) a horse-
shoe-shaped atoll is
formed, as indicated by
the dotted line AB.
The I'im is widest at A,
Fig. 5, graduall}^ becom-
ing narrower and lower
tow^ards the lee side as
it nears B, the whole or
the greater part of the
ledge having been carried away by the pounding of the surf, leaving

t,:w -!!%

266

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

a liigh narrow wall with a deep opening at B between its extremities.
Should the surf break through the sides at E or F, or both, we should
have curved vertical walls left, apparently built up by Serpulse, in reality
walls of aeolian rock which may be dug out as I have suggested, either
in the case of Figure 2 or of Figure 4.

When the pinnacle finally drops off, it will in its turn be attacked by
the sea, and go either to form a smooth ledge, to be covered with AlgtB
and Serpulae according to tlie depth in which it lies, or may in its turn
be attacked in a similar way to the ledge, and changed to an atoll or a
crescent-shaped serpuline reef according to the dip of the strata. Before
the breaking off" of the pinnacle, Algse and Serpulse have already begun
to grow upon the flat part of the ledge, and protect it to a great extent
from the action of the sea. When the sea no longer washes round the
pinnacle, but breaks on the ledge at low water, and finds a part which is
not protected by Algag or otherwise, it begins to erode it, the sand formed
acting like a churn, and thus little by little forming a deep hole in the
centre of the mushroom rock.

)00

Fig. 6.

In the case of a long and wide ledge, there are formed upon it a num-
ber of secondary pits and atolls, or pot-holes, as indicated by the heavy
lines of Figure 6. Let the walls of these break through and connect
adjoining pot-holes and we obtain a vertical wall, of irregular outline,
such as is indicated by the dotted line of the figure, which is a diagram
of one of the ledges of the outer reef off' the south shore. The wash of
the sea may break through the continuous wall, leaving only discon-
nected parts standing, or we may have the outer walls on the edge of
the ledge left, forming a long trough.

A flat ledge projecting from the base of a shore cliff", if eaten into in
the same way by the surf (Fig. 7), may be worn into a circular reef
witli vertical walls, of which the top is protected by Algfe and Corallines
or SerpuljE, with a pot-hole near its outer wall, in this case eight feet deep.
We might call this a diminutive barrier reef. If the walls are parts of
diminutive barrier reefs, the shore cliff" behind them may disappear, and

AGASSIZ : BEllMUDAS.

267

thus leave a wall standing apparently isolated from the level of the
led-j-e. It is not surprising that a cursory examination of tliese walls and
atolls should have induced the earlier observers to attribute the growth
to Serpulse.

Kice accepts the theory that the serpiiline reefs are due alone to the
upward growth of Serpulaj. He has also observed the circular ridges of
coral on the outer ledges (Millepores) similar to the serpuhne reefs,
except that they are less elevated, tlieir upward growth being limited
by the inability of the corals to survive an exposure above the water.
These circular reefs
are either serpulineor
edged by Millepores,
and are called boilers ;
oif the south shore
they are generally ser-
puline, with only here
and there a Millepore
boiler on the inner
patches. Along the
east shore of Harring-
ton Sound there are
a number of such Mil-
lepore atolls and barrier reefs which are merely rims of pot-holes pro-
tected by a growth of Millepores. A number of these Millepore reefs
extend eastward on the northern side of the main channel off Spanish
Point. Along the shores of Harrington Sound we meet occasionally one
of the vertical walls or irregularly shaped pot-holes of a projecting ledge,
the top of which is protected by a growth of Millepores, much as the Ser-
pulpe and Algae protect the walls of the pot-holes off the south shore. On
many of the patche? to the north of the main channel, in a line from
Ireland Island to St. George, we find a similar growth of Millepores pro-
tecting the exposed ridges of aenlian rocks. Thomson also considers the
serpuline reefs as due to the agency of Serpulse.

The existence of such a variety of reef forms under the same conditions
naturally suggests whether the explanation by mechanical causes, such
as finds here its application on a small scale for serpuline reefs, may
not be a natural explanation also of the formation of the complicated
systems of atolls of the Carolines, Marquesas, and Maldives, and whether
small simple atolls like the typical ones of the Pacific, and many with
complicated outlines, may not in many cases be due to the scouring ac-

FiG. 7.

268 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

tion of the sea working on a larger scale upon substructures easily yield-
ing to its effects. When examining'the Hogsty Atoll of the Bahamas I
could not fail to be struck by the overpowering mechanical force inces-
santly at work. Huge breakers were constantly pouring an immense
volume of water over the windward sea face of the atoll, filling its cup to
overflowing, and it could find no outlet except over those parts of the
sides which were lower than the windward face, or through which, near
the lee ends, passages of considerable depth had been eroded, or, finally,
through the still deeper channel between the lee extremities of the reefs
forming the entrance to the atoll. That is, given a bank of suitable depth
upon which corals can flourish, and upon a belt of a certain width, they
will form a protective coating to the underlying rocks, just as the serpu-
line growths protect the rims of their diminutive atolls. Corals will
naturally, from the centrifugal action of the sea, grow on the outer faces,
and most abundantly in the direction in which they find least resist-
ance in the way of detritus and other accumulations. The sea breaking
over them will excavate a lagoon, and break through the sides or lee
face to allow the water to flow out through the points of least resistance,
and through the entrances to the lagoon. All these causes are impor-
tant factors in any theory of coral reefs, and show the complexity of the
problem, and the impossibility of framing a single hypothesis to explain
the formation of coral reefs in all parts of the world.

Before seeing the serpuline atolls of the Bermudas it had occurred to
me that the configuration of the Hogsty Reef and the formation of its
atoll might be due to mechanical causes. We may imagine a bank ol
the proper depth, whether formed during subsidence or elevation is im-
material, on which corals begin to grow and form a barrier to the surf.
The breaking of the surf over this living and protecting barrier digs out
the least resisting portions of the surface of the bank, and the material
thus dug out finds its way out on the opposite side. Little by little a
lee channel is thus formed by the scouring of the mass of water poured
over the reef into the incipient lagoon, and a lagoon may be formed on a
large scale in the manner described for the formation of the serpuline
lagoons of the Bermudas. We may imagine the Hogsty Reef at one time
to have been a bank formed by a series of small, low a3olian hills, which
have been worn away and have disappeared from the same causes which
acted on a larger scale at the Bahamas. The Hogsty Bank was thus
brought by subsidence and erosion to its present level, or nearly so,
after the growth of a barrier of reefs on the remnants of the teolian hill
ledges, then began the action of the surf in eating away the central part

AGASSIZ: BERMUDAS. 269

of the bank, scouring the lagoon to the depth and dimensions it has
now attained. There is nothing to show that the depth of the lagoon of
Hogsty "Eeef is due to subsidence, or that the rate of growth of the reef
was synchronous with it, and thus formed the outer rim of the lagoon.
On the contrary, judging by analogy and by the conditions existing in
the Bahamas and Bermudas, we are led to infer that the lagoon of Hogsty
has been formed by a mechanical process, that it is due to the action of
the surf acting as an immense force pump driving the water over the
weather face of the reef^ out through the lee opening of the lagoon, or
the openings of the sides of the ring, much as the diminutive serpuline
atolls and crescent-shaped reefs or barrier reefs have been formed off the
south shore of the Bermudas.

POT-HOLES.

Plates XXIX., XXX.

Roots and stems, after being decomposed, may form branching cavities
which if filled with stalagmitic matter would give rise to columnar struc-
tures. Such formations can be traced in the more recent dunes of the
Bahamas and Bermudas, and of the Sandwich Islands. There is a type
of pot-hole which imitates these structures ^ and has been confounded
with the branched bodies, but which I do not believe to be organic struc-
tures at all.' They are the so called palmetto stumps, of which Rice has
given an excellent description,* and which I imagine to be mechanical
structures of a similar origin with the serpuline reefs (Plate XXX.).

To the north of the Devil's Hole, on the road skirting the east shore of
Harrington Sound, we find a flat ledge of teolian rocks which is literally

1 That water falling from a small height does thus excavate deep holes at the
foot of falls is well known. Any country ditch dammed by a sluiceway will show
this effect. It can be seen at the foot of every waterfall, and it occurs on the
largest scale in the Mississippi, where the scour of the river below New Orleans
has since the building of the jetties excavated a depth of between sixty and
seventy feet in some cases, — a depth about the same as that of many lagoons of
coral atolls.

2 I cannot agree with Thomson (Atlantic, I. 3:10) in his explanation of the
mode of formation of the pseudo palm stems, who considers them to have been
formed on the bottom of caves by the dropping of stalagmite, and thus forming a
single or double or dumbbell-shaped stem.

" Are not some of the tubes to which Professor Dolley ascribes a vegetable origin
merely small pot-holes such as I have figured on Plate XXIX. ?
* Bulletin of the National Museum, No. 25, p. 27.

270 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

riddled with pot-holes, many of which ha?e become connected to form
dumbbell-shaped cavities or other irregular forms (Plate XXIX.). They
are in every respect similar to the serpuliue atoll pot-holes off the south
shore, but are much smaller, and become changed many of them into the
so called palmetto bases.

An irregularly shaped pot-hole with a raised rim is one of the promi-
nent features of one of the ledges on the north side of the causeway
leading to St. George. In this case the greater hardness of the rim has
formed the protective coat for the low vertical wall of the pot-hole
(atoll). The transformation is readily followed in the pot-holes which
are a little removed from the immediate action of the diminutive, but
short and sharp, breakers of Harrington Sound. The inner cavity of
the pot-holes becomes lined with a harder coating, being acted upon
much as is the surface of the aeolian rock. When the walls of adjacent
pot-holes are eaten away or worn away by the action of the rains and
of the sea, there may be formed an outer coating from thg adjoinmg pot-
hole, and thus an irregular cylinder will be left standing above the sur-
rounding area. If the upper part of this is in turn disintegrated, we
may have left a deep cup, or merely a ring of the base, or a side of the
cylinder, or merely the harder inner coat of the bottom, — all of which
stages resemble more or less the base of a palmetto (Plate XXX.).
These pot-holes are often so close together (Plate XXIX.) that it is dif-
ficult to imagine a grove of palmettos the stems of which could be
packed in the area occupied by the pot-holes. Captain Carr, R. N.,
called my attention to a locality on the east shore of Ireland Island
where these pot-holes are numerous, and where one can trace all the
transition stages just described from the pot-boles of the shore of Har-
rin-yton Sound.

NORTH ROCK.

Plate VIU.

The North Rock is undoubtedly the most interesting monument left
of the former extension of the Bermudian land. Owing to the difficulty
of landing, it has not been visited frequently, but we were twice success-
ful in landing at the North Rock Ledge.

Excellent descriptions of the North Rock pinnacles and ledges have
been given by Rice,^ and also by Heilprin in his volume on the Bermudas.

1 Bulletin of the National Museum,

AGASSIZ: BEKMUDAS. 271

The flat ledge surrounding Xorth Rock presei;its no features diflfereut
from those of similar ledges off the south or north shore. Its surface
rises -here and there in low ridges, which are the remnants of the last
pinnacles to be eroded ; near the edges, and wherever the action of the
breakers reaches, it has been dug into so as to form pools, pits, and pot-
holes of various depths and shapes, and has been honeycombed in all
directions, according to the quality and hardness of the rock and the
extent of the protection afforded the rock surfaces by the growth of Algae,
Corallines, Serpulae, and other organisms.

The North Rock Ledge is deeply undercut, and, like many of the larger
islets off the main island, its sides are more or less vertical, or steep
slopes deeply honeycombed and cavernous, and overgrown with Algae,
Corallines, Gorgon ians, Millepores, and massive corals, much as any
similar ledge or cliff or patch. The greater part of its upper surface
is protected by the hard ringing aeolian rock characteristic of the ex-
posed iutertidal spaces. The pinnacles which remain rising above the
general level consist of aeolian rock, and the lower base rock seemed
to me to differ in no way from similar aeolian rock as modified by the
action of the sea in other localities. It is possible that the fossil Cyprasas
stated by Rein to have been found at the base of the pinnacles may
be only such shells as have been collected from the -serpuline rock
in which they had become embedded. In some spots on the ledge it
attains a thickness of from twelve to fifteen inches, and is full of boring
Mollusks and of shells which have found a foothold in the cavities of the
honeycombed rock. The serpuline rock itself often becomes quite hard,
and might easily be mistaken for true hard ringing seolian rock, but is
readily distinguished from it by the presence of the many sharp pinkish
fragments of the tubes of Serpulae.

The many Chamas and other Mollusks living on the edge of the North
Rock Ledge at low-water mark would, if thrown up and embedded in
ajolian or serpuline rocks, present all the characters of the so called fos-
sils found in the a;olian rocks of the islands of Hamilton Harbor. At
Agers Island and on the shores of the other islands of Hamilton Harbor
we find at very low tides many specimens of Chama which would have to
be carried but little way to become aeolian fossils. Neither Heilprin nor
myself found any fossils on North Rock. Heilprin considers the lower
portion of the North Rock pinnacles to be unquestionably pits and pot-
holes of the inner surface of the beach rock.

The outer edge of the northeastern part of the ledge is protected by
SerpuliB, Algae, and Corallines, forming low vertical walls as well as the

272 BULLETIN: MUSEUM OF COMPAEATIVE ZOOLOGY.

division walls of the irregular pot-holes of the ledge, and in places they
overhang to a very considerable extent. The coating of Algge and of
Serpulae on some of the walls of the interior pools is quite thin, as upon
breaking off the edges we came upon the ringing aeolian rock. On the
northeast edge of the ledge a deeper pot-hole has formed a regular bar-
rier reef, the edges of which are covered by Serpulae, Algae, and Corallines.
On the interior of the ledge their growth is less vigorous. Outside of
North Rock there are a few ledges, both to the southwest and to the
northeast. Those towards the inner side of the reef are the outer reef
patches, on which grow Millepores, Gorgouians, and the usual growth of
corals, Algae, and Corallines, which have completely hidden the nature of
the ledges. The ledges which form the continuation of North Rock to
the east and west, on the contrary, still plainly show that their structure
in no way differs from that of the North Rock Ledge. Outside of North
Rock " broken ground " extends to ten or twelve fathoms, but no patches
of corals could be seen bevond five or six fathoms. We mitjht call the
outer ledges of the reef near North Rock a flat of coral heads and of
Gorgonians. On the North Rock Ledge fiat, Zoanthus, Millepores, Algae,
and Corallines are most flourishing.

Winding our way towards St. George Island from the North Rock,
we picked our passage between the many mushroom-shaped ledges.
Many of them came to the surface or nearly so, and on two of the ledge
patches rather nearer the inner line of the ledge flats than to the North
Rock I broke off a piece of hard ringing aeolian rock in every respect
similar to that of the North Rock Ledge flats.

We could not have a better example of the true nature of the reef
ledge flats and patches than is exhibited by the North Rock Ledge and the
adjoining patches. Those of the ledges nearest to the North Rock show
the hard ringing aeolian rock which marks the North Rock Ledge ; some
of the patches aire separated from it by water five to ten fathoms deep.

We can easily imagine the whole of the ledge flats of the vicinity
to have been made up of ax)lian rock ledges and pinnacles, very much
like those off the north shore, on which corals, Gorgonians, and the like
have little by little become attached, and have finally grown over not
only the sides, but the upper surface after it became eaten away well
below low-water mark.

There may have been on the northern reef flat ledges lower lines of
hills than those now existing on the Bermudas, or low hills like those
separating St. George and Castle Harbor, or along the west end of St.
George Island and on the north side of Harrington Sound.

AGASSIZ: BERMUDAS. 273

PROTO-BERMUDA.

Plate II.

Rein, Thomson, Rice, Fewkes, and Heilprin, all agree to the former
greater extension of the Bermndian land, and Thomson,^ speaking of
the North Rock, and of Pulpit Rock off Ireland Island, says there can
scarcely be any doubt that the dry land of Bermuda at one time occu-
pied a space considerably larger than it does at present.

We may readily reconstruct the proto-Bermudian land from the exist-
ing charts (Plate II.). Beginning with the line and clusters of islands
running from Ireland Island to St. David Head, these must in earlier
times have been somewhat wider. The main island must have extended
south beyond the line of the reef, and dry land must have completely
barred the access of the sea to the sinks which on the east constituted
Harrington Sound and Castle and St. George Harbors, St. George Isl-
and itself probably forming the western edge of the Ship Channel val-
ley. On the west the main island reached to Hogfish Cut valley, and
Somerset and Ireland Islands were probably connected with a range of
seolian hills running from Chub Cut across Elies Flat. On the east,
Ireland Island was connected with Spanish Point by a ridge which
isolated Great Sound, Port Royal Bay, and Hamilton Harbor sinks from
the outer lagoons. Great Sound and Hamilton Harbor both probably
beuig disconnected sinks, and both isolated by low saddles from Great
Sound sink. The ledge flats to the west of Hogfish Cut and to the north
of Chaddock, Little, and Long Bars, which pass to the east of Chub
Heads as far as Chub Cut, formed, in connection with the Elies Flat, hills
of which the ledges are the remnants, the barrier separating a great
sound larger than any now existing from the adjoining proto-Bermu-
dian sounds. Of these we can trace four others of great size. One
bounded on the southwest by the hills of Elies Flat, on the northwest by
the hills of the ledge flat extending north from Chub Cut, on the north
by the line of flats running east in the direction of Three Hill Shoals till
they strike the eastern face of the sound formed by the Brackish Pond
Flats, the eastern boundary reaching towards Spanish Point and sep-
arated from the shoals north of it by the Ship Channel valley. The
second sound is enclosed by the Brackish Pond Flats on the west, by
the Bailey Bay Flats on the north and east, and by the main island on the

1 Thomson, The Atlantic. I. 318.

VOL. XXVI. — NO. 2. 18

274 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

south. The third is the Murray Anchorage sound, limited on the west
by the Bailey Bay Flats, on the north and east by the Three Hill
Shoals and the flats west of Mills Breaker. The third sound opened by a
narrow deep valley (the Ship Channel) towards the sea, and communicated
by a wide passage with the fourth sound, bounded by the ledge flats of the
northwestern part of the Bermudas, — flats which extend unbroken from
the north of Western Blue Cut to the Eastern Ledge Flats, — and on the
southern edge by the line of the Three Hill Shoals and by the western
extension of the Bailey Bay Flats, This fourth sound is in reality a
double sound, as the western part is separated from the eastern by a
narrow line of seoliau heads, indicating probably the position of a cross
line of dunes connecting the Ledge Flats and Bailey Bay Flats. Its
northern edge was deeply indented, as is indicated by the many tongues
of deep water cutting into the width of the ledge flats (Plate II.). Two
smaller and indistinctly connected sounds are similarly indicated to
the west of East Ledge Flats and to the west of Mills Breaker. On the
eastern ledge flats there are also a number of deep pockets, already re-
ferred to, as well as many deep bights running into the ledge flats,
indicating the position of valleys running more or less at right angles
to the trend of the ledge flats.

The Southwest Breaker and Chaddock, Little, and Long Bars are the
base of lines of seolian hills, once running parallel with the edge of the
western and southern Ledge Flats hills.

This proto-Bermudian land must have resembled the Bermudian land-
scape of to-day, and has been reduced to its present condition by the
same causes which we see at work to-day on the islands of the group, and
which have acted more vigorously either on the faces most exposed to
the prevailing winds, or upon aeolian hills of a lower altitude than those
of the main island.^ The proto-Bermudian sounds vary in depth from
six to twelve or thirteen fathoms, the deepest being the sound to the
north of Three Hill Shoal.

I also agree with Rice and Heilprin that the amount of subsidence
must have been " sufficient to account for the depth of water which
marks the lagoon and inner sounds," and that " before this subsidence
took place probably the entire area now covered by the Bermudian

1 " The prevalence of powerful winds on the south side would tend to elevate this
%ide of the island, wliile the opposite side, not feeling this influence in any marked
iegree, would remain comparatively low and flat. In a period of subsidence the
low side would naturally be the first to succumb to the waters, an'd would undergo
submergence long before the elevated slopes. And this is precisely what appears
to have taken place in the Bermudas." — Heilprin, Bermudas, p. 42.

AGASSIZ: BERMUDAS. 275

archipeJago, and much more, were dry land." But I do not think that
"it was at this time, doubtless, that the great sand dunes were ele-
vated." On the contrary, I do not imagine the dunes to have been
elevated, but to have perhaps been blown to their greatest height in the
manner suggested from a broad coral sand beach.

Heilprin adopts the suggestion thrown out by Rice, that these seolian
accumulations could only have been formed at a time when large areas
of reef, and not a simple atoll ring, were exposed above the water level,^
while it is perfectly true, as Heilprin says, that all the sand formed at
the present day is derived from the destruction of the existing land
masses, and not as a product of the disintegration derived from the
growing reef, — a statement which by the way hardly agrees with the
graphic description of the destruction of the coral reef off the south
shore given by him a page or two before.

It does not seem to me necessary that there should have existed very
wide areas of coral reefs for the formation of the Bermudian seolian hills.
A reef of a width of 1,200 to 1,800 feet, such as are known to exist,
seems to me ample to supply the material necessary for the formation
of the seolian hills, especially if, as may hjive been the case judging by
the soundings off the reef, there also existed a comparatively wide shal-
low bank outside of it. Taking the subsidence of the Bermudas as
twelve fathoms, this belt could have been a belt of twelve to twenty
fathoms, a shallow bank, the wear and wash from which would alone
supply a large part of the material needed for the formation of the
Bermudian dunes, and carry the topography now prevailing pretty well
over the whole of the bank inside the ten or twelve fathom line. In
addition, we should of course also have the supply derived from the reef
itself. The amount of material which can be supplied from a compara-
tively small reef and its adjacent bank I saw well illustrated at the
Sandwich Islands. A number of dunes were constantly travelling inland
from the beach at Spreckelsville, and had covered to a considerable ele-
vation a great part of the isthmus connecting the two islands which con-
stitute Maui. Very high dunes, from 120 to 180 feet, of nearly a mile
in width, occur on the shores of the Baltic, the material of which is
derived from a comparatively narrow beach range.

1 fully agree with tliose who before me have examined the Bermudas,
and who consider that subsidence has brought about the existing condi-
tion of the islands and sounds. But that is a very different thing from
assigning to the corals now growing the formation of the islands owino' to

^ The Bermurla Islands, p. 46.

276 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

this subsidence. That tlie proto-Bermudian land was of elliptical shape,
and owed its existence to the action of winds sweeping over an extensive
coral beach, from which was gathered the material which now form the
solidified seohan hills of the Bermudas, no one can question. But there is
no evidence to show that the original annular coral reef was formed during
subsidence. That reef has disappeared, and nothuig is left of it except
the remnants of the seolian ledges, extending to sixteen or seventeen
fathoms outside of the reef ledge fiats, ledges which owed their existence
to the material derived from it, — the former seolian hills of the proto-
Bermudian land. Remnants of such ledges and former seolian hills are
the rocks forming the outer ledge flats, the breakers all along the south
shoi'e, the Mills Breaker, the !N"orth Rocks, the Chub Heads, the South-
west Breaker, and others.

The evidence of the extent of the subsidence which has taken place at
the Bermudas is very clear. It is based upon the depth of the sounds
and of some parts of the lagoon, the existence of aeolian rock at a
depth of fifty feet below low-water mark, and the excavation of red
earth from a depth of forty-eight feet below low-water mark on Ireland
Island,^ and at eighteen feet below low-water mark in the entrance to
St. George Harbor. Numerous caves and caverns occur in the Bermu-
das, which have been fully described by previous writers. Many of them,
although extending far below low-water mark, could only have been
formed when the islands were at a greater elevation than at the present
day. In the caves, as well as smaller ponds or embryo sounds close to
the shores, the porosity of the seolian rock, as well as its cavernous and
honeycombed structure, is indicated by their connection with the sea.
The water rises and falls in the caverns with the tides, a.id the beat of
the diminutive waves against the subterraneous shores of the caves closely
follows that of the sea outside.

1 Sir Wyville Thomson has given a section of the rocks exposed during the ex-
cavation for the basin of the dry dock at Ireland Island (Atlaiitic, I. 319.) Huge
stalactites and stalactites covered with Serpnlas extending below low-water mark
in some of the caverns clearly indicate the effects of the subsidence. Trunks of the
Bermuda cedar have been found in the red earth at a depth of forty-eight feet in
the excavation of the dry dock, and h;ive also been dredged from the bottom of
Hamillon Harbor in a depth of five fathoms. Tliis is of no great value as evidence,
since the stumps may have fallen in from the surrounding hillsides, have floated off,
and become water-logged. According to General Lefroy, one of the great bogs of
the main island extends to a depth of forty or fifty feet below the sea level, and
indicates the depth to wliich one of the sinks of the proto-Rernuidinn land renched
before the subsidence took place which resulted in the present configuration of
the islands.

AGASSIZ: BEKMUDAS. 277

At the Bermudas we find nothing corresponding to the ocean holes of
the Bahamas, unless it be that such cavernous sinks and indentations as
occur all along the shores, and which can be detected on the edges of the
reefs and ledges, or in the circular areas where the depth is sometimes
greater by two to three fathoms than over the adjoining area, may cor-
respond to ocean holes, but only on a smaller scale of depth.

Heilprin thinks " that the height of land in the archipelago was formed
during a period of elevation." It seems to me more natural to suppose
that the Bermudas were formed during a period of rest, when the level
of the reef was stationary, and they were flanked by a broad sand beach
with flats perhaps bare at low water. These would supply an abun-
dant material for the formation of such dunes as we now find, and may
imagine to have existed on the northern edge of the Bermudas and ou
the flats which determine the shape of the proto-Bermudian lagoons.

While I fully agree to the all-important part which subsidence has
played in shaping the Bermudas as they now exist, I cannot trace any
connection between these facts and the proposition that " the existence
of an atoll in the present position of the Bermudas is not demonstrable."
We certainly have a group of peolian hills formed from an annular
ring of coral reefs which flourished when the land was at least seventy
feet higher than at present. But the facts we observe on the islands
to-day do not shed one ray of light on the question of the Darwinian
theory of the formation of atolls. The position of the reef to which the
Bermudas owe their origin can only be surmised, — and probably very
correctly, — but we cannot state that it was formed during a period of
subsidence, and have no data regarding this point. Subsidence has
given to these islands their present configuration. Bat it is begging
the question to state that the formation of the proto-Bermudian coral
reef, about which Heilprin himself is careful to say we know nothing, if
it does "not prove the correctness of the Darwinian theory of the forma-
tion of coral islandg, measurably sustains it." ^

We may also agree with him in the conclusion that the present form
of the Bermuda Islands bears no relation to the ring of an atoll, ^ that

1 " The question as to what form of coral structure the Bermudas actually are
— what constitutes their fundament, and how they were built to their existing
level — still remains unanswered, and possibly we may never be able to answer." —
Bernmdas, p. 47.

" Heilprin says: "In the case of the Bermuda Islands, which limit the field of
my own investigations in this direction, I am confident that, whatever may have
been the original construction of the region, the present lagoon features have been
brought about through subsidence ; and tliis conclusion was reached before me by

278 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

the existence of an atoll in the present position of the Bermudas is not
demonstrable, and that the lagoons and sounds were formed, as Rioe
first showed, during a period of subsidence. i But granting all this,
what is the connection of an island group which according to him has no
relation to the ring of an atoll, of an island for which the existence of an
atoll cannot be demonstrated, with the Darwinian theory of the forma-
tion of coral reefs 1 How can the conclusions an-ived at by Heilpriu be
reconciled with the following statements made by him : —

" It will be seen tliat these results, so far as they go, are in absolute
harmony with the views which Mr. Darwin entertained regarding the
structure of these islands. They do not prove the correctness of the
Darwinian hypothesis of the formation of coral islands, but they measur-
ably sustain it ; on the contrary, they are largely opposed to the require-
ments of the substitute theory which has been recently proposed. Ele-
vation and subsidence are both shown to have marked the region in its
development, and these conditions are more in consonance with the
Darwinian hypothesis than with any other." ^

Investigators have been carried away by the simplicity of the theory
of subsidence propounded by Darwin, and it is only of late years since a
mass of observations have been made which could not be explained by
the prevailing theory that we have at last realized how complicated the
problem is. Heilprin, as well as others before him, has truly said, "We
may not yet have fathomed the true method of the formation of coral
islands." But I must differ from him in tofo when he says, " but such
evidence as I was able to obtain at the Bermudas failed to convince me
of the erroneousness of the time-honored theory of subsidence." *

The exploration of the Bahamas and of the Bermudas has brought
into prominence a condition of things relating to the formation of coral
reefs, the bearing of which had not been realized before. It is perhaps
one of the most significant examples of how little we as yet know of the
history of the formation of the coral reefs.

Professor Rice, who seems to have been amply satisfied with the subsidence
theory." — Bermudas, p. 75.

1 Heilprin has well stated the conditions of the disintegration of tiie land when he
says : " The difficulty in the problem entirely disappears if we admit subsidence, and,
as has already been seen, tiie positive evidences of subsidence arc ample. On no
other theory, it appears to me, can tlie waste of the cliffs on the south shore be
explained. The direct evidences of subsidence, moreover, do not come from a
single point in the archipelago ; they are found from Ireland Island and Hamilton
Sound, through tlie main island, to St. George."
• 2 Bermudas, p. 46. ^ Bermudas, p. 21.

AGASSIZ: BERMUDAS. 279

EXPLANATION OF THE PLATES.

PLATE I.

Bermuda Islands. Reduced from U. S. Hydrographic Chart, No. 27.

PLATE IL

Fig. 1. The Challenger and Argus Banks to the sojithwest of Bermuda.

Fig. 2. Section from Somerset Island across the Challenger and Argus Banks.

Depth in fathoms, horizontal scale, 1 inch = 10 miles.
Fig. 3. Section north-northeast from North Rock into 1,370 fathoms.
Fig. 4. Section southeast from Castle Harbor into 1,240 fathoms.

PLATE III.
The Bermuda Islands from Gibbs Hill.

PLATE IV.
Mullet Bay, St. George Island. Characteristic ^olian Hills in the background.

PLATE V.
Bare Hills on the North Shore of the Western Entrance to St. George Harbor.

PLATE VI.
Shore of Harrington Sound.

PLATE VIL

Webb's Pond.

PLATE VIIL
North Rock.

PLATE IX.
Deep Cut through ^olian Rock, Warwick.

280 bulletin: museum of comparative zoology.

PLATE X.
Sink, Warwick, South Shore.

PLATE XI.
Sand Dunes, Elbow Beach.

PLATE XII.
Sand Dunes back of Elbow Beach.

PLATE XIII.
Devil's Hole. Characteristic hardened ^olian Rock.

PLATE XIV.
Lagobn near Tuckerstown.

PLATE XV.
Entrance to Hungry Bay.

PLATE XVI.
JEolian Cliffs near Admiralty House, North Shore.

PLATE XVII.
Pulpit Rock off Ireland Island.

PLATE XVIII.
Half-sunken Ledge, North Sliore, off Boat Harbor.

PLATE XIX.

^olian Cliffs, South Shore, near Middleton Beach.

PLATE XX.
^olian cuffs. South Shore, Elbow Bay.

PLATE XXI.
Ledges and Islets off Castle Harbor.

PLATE XXII.
.Solian Cliffs with SerpuUne Atolls, South Shore.

AGASbIZ: BERMUDAS. 281

PLATE XXIII.
^olian Rock Pinnacles and Serpuline Atolls, South Shore.

PLATE XXIV.

Serpuline Atolls, South Shore.

PLATE XXV.
Serpuline Atolls, South Shore.

PLATE XXVL
Fringing Serpuline Reefs, South Shore.

PLATE XXVIL

Pitted and Honey-combed ^olian Rocks, Harrington Sound.

PLATE XXVIIL
Ledge showing Passage of ^olian Rock to Base Rock, Boat Harbor, North Shore.

PLATE XXIX.
Pot-holes, Harrington Sound.

PLATE XXX.
Pseudo-Palmetto Stumps, North Shore.

iSASsiz -Bermudas

BERMUDA ISLiVXDS.

* m

'Li-^.^r ^i^.--': -i.-y-., ^v , oY;/. o ^

■S- •>. •!•/%'''.

.2X ^

AoASSiz -Bermudas

Fig.:.

Ch«llengerBanh.:

Argus Bank -

MorixorUal .^ecUe 1=10 Miles.

2i I^£ jnm Sattti RetX
SeaXevel i OrvMOt ■

Dep&ls "in Fathoms

BMeiMl,lii>i Bstion

<

tn

J
J

CO

DO

m

o

o

cr

C//

<

Q
D

a:

UJ

CD

Ul

X

[-

>

<

CD

H

o

o

LJ
O

in

■'\

O

u

U

z

<
cc
h

2

U

Ll

q:
O
I
cc

re
h

q:
o

2

>

s:

r

tr.

CC
C
X
in

C

z

O
•J?

z

p

o
z

q:

CC

<

X

\

;■■>

<

■j:

1-

P
Z

o

CO

m
m
w

r. I

^l

I,' !'■

/ »t ■

'•f'.

If ' ''

fv^^a 'i//. ^1

^H'«.!«^§:'Jr''^

r/.' II .*

^ I til* ,vV'ut'' Ji'.**t?JilM fit

o
o

K

I
h

O

z

Agassiz "Bermudas.

Plate IX.

i

' ^■

IP

ARTOTYPE, E BltRSTAOT,

DEEP CUT. WARWICK.

X

<

<

z

X
u
<

w
m

o

m

w

CO
UJ
Z

D
Q

Q
Z

<

CO

H
<

►J
Ph

■u

Tn
m

-««S*fJ I

O
<

UJ

o

CQ

_1

UJ

u.

o

u
<

CG
CO
Z

Q

o

X

en

1—1
>

>

<

o

h
en

cc

UJ

u

D
H

Z

o
c
c
<

>
<

en

>

cc
o
z

D

X

o
h

UJ

u
z
<

!_

z

UJ

>

UJ

o

X
CO

X

h

o
z

CO

u.
u

i- IH

^ IP

^

"4n^HM9|^iKsHL9H|^^

■iF^^P^

r-*

z

o

<

J

a

s

a;

cq

Q

<

CO

Q
<

w

o
o
p:

h

D

cc
o

X
ui

X
H
CC
O

UJ

o

Q
UJ

<
o

X
X

in
3

V

-i1

o

<
w

CD

z

o
h
w
►J

Q
Q

<

:z

!j
o

z

<

o

Agassi z "Bermudas

Plate XX.

'€

5^.

1 ,^:

CLIFFS. SOUTH SHORE.

<
0-

■fl

'Jaiiilp

rmi/i

1

1.1 j^^Kfj^

^ fJHMIM

of

■^^^1^1

O

l^^l^^l

DQ

m^i^^i

d:

^^f^^^i

<

^mp

:r

UJ

J

h

c/:;

<

u

u.

Ll,

o

t/)

H

Ld

4 *■;

-J

;

CO

l#fi

H*

rt
T3

u

V

bD

CO

O
h
<

-J

D
CL

cr
u

CO

Q
2
<

CO

Lx.

U-

_1
CJ

UJ

o

X

CO

X!
X!

H
<

P-.

<1

O
h

t f

:3

Q
<

W
J
O
<

Z

I — I

Ph

>

l-H

>^

X

w

H
<

■r.

ifl

UJ
K
O

CO

X

o

CO

CO

J
O
<

Z

D

CL

cn

UJ

C/}

>

w

<

X3

o

X
m

H

O

in

U.
O

CO

C
h
<

w

a.
w

(/2

tr,

oi

>

X
X

w

<
Ph

u
q:
o

H

o

CO

(/)"

U-
UJ
UJ

cc

UJ

z

Du
CX
UJ

CO

O
Z

o
z

u.

>

<

1-1

3h

Q

z

o

Z

O
h

o
z

s

<

X

in

u
O
q:

z

<

o

Q

m
s

o

o

UJ

z

o

>

w
<

t .

U]

X

h
cr

Q

z

CO

a:

z

<

^

cu

H

Q
Z

o

CO

Z

o

H
O
Z

<

CO

W

O

X

h

o

w

K
O

X
(J)

X

(X
O

:z

a.

CO

O

h

0.

Harvard MCZ Librar

2044 066 302 399

Hii^

kfn

t tr &tWd*LIL}^LIrl«l.»l.n^l>I.»KkCft.I.».1