*■*£

6*31

JChH*^

if.«S,

r*5 V

?V V

Do not circulate

HARVARD UNIVERSITY.

LIBRARY

MUSEUM OF COMPAEATIVE ZOOLOGY.

J^Wl

V^V/^W

•MfrUVwOlHA av\(\OyOG_ vJiWvJlx \o,\<\fc

*>

BULLETIN

MUSEUM OF COMPARATIVE ZOOLOGY

HARVARD COLLEGE, IN CAMBRIDGE.

VOL. XXXVIII.

(Geological Series, V.)

CAMBRIDGE, MASS., U. S. A.
1900-1903.

University Press:
John Wilson and Son, Cambridge, U.S.A.

CONTENTS.

Page
No. 1.— Notes on tlie Limestones and General Geology of the Fiji Islands,
with special reference to the Lau Group. Based upon Surveys made for
Alexander Agassiz. With a preface by T. W. Edgeworth David. By
E. C. Andrews. (40 Plates.) November, 1900 1

No. 2. — The Structural Relations of the Amygdaloidal Melaphyr in Brook-
line, Newton, and Brighton, Mass. By Henry T. Burr. (2 Plates.)
March, 1901 51

No. 3. — The Physiography of Acadia. By Reginald A. Daly. (11
Plates.) March, 1901 71

No. 4. — An Excursion to the Grand Canyon of the Colorado. By W. M.
Davis. (2 Plates.) May, 1901 105

No. 5. — The Geology of the Northeast Coast of Labrador. By Reginald
A.Daly. (13 Plates.) February, 1902 203

No. 6. — Leucite-Tinguaite from Beemerville, New Jersey. By John E.
Wolff. February, 1902 271

No. 7. — River Terraces in New England. By W. M. Davis. October,
1902 279

No. 8. — The Foraminifera and other Organisms in the Raised Reefs of
Fiji. By R. L. Sherlock. March, 1903 347

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

3 l^tf Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 1.

NOTES ON THE LIMESTONES AND GENERAL GEOLOGY OF

THE FIJI ISLANDS. WITH SPECIAL REFERENCE TO

THE LAU GROUP. BASED UPON SURVEYS

MADE FOR ALEXANDER AGASSIZ.

WITH A PREFACE BY T W. EDGEWORTH DAVID.

By E. C. Andrews.

With' Forty Plates.

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

PRINTED FOR THE MUSEUM.

November, J900.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 1.

NOTES ON THE LIMESTONES ANB GENERAL GEOLOGY OF

THE FIJI ISLANDS, WITH SPECIAL. REFERENCE TO

THE LAU GROUP. .BASED UPON SURVEYS

MADE FOR ALEXANDER AGASSIZ.

WITH A PREFACE BY T. W. EDGEWORTH DAVID.

Br E. C. Andrews.

With Forty Plates.

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

PRINTED FOR THE MUSEUM.

November, 1900.

TABLE OF CONTENTS.

PAGE

Introductory Note by Alexander Agassiz 3

Preface : Letter from Professor T. W. Edgeworth David, of the

University of Sydney, New South Wales 5

Report of E. C. Andrews . . 11

Introduction 11

General Geological Structure 13

The Larger Islands 13

Viti Levu 13

The Stratified Rocks 13

The Dolomites 13

The Lau Group and the Smaller Islands 14

The Volcanic Islands 15

The Limestone Islands 15

The Limestone and Volcanic Island j 15

Geological History and Topography .1(3

The Main Islands - 16

General Topography 16

Coastal Topography lg

Topography of the Smaller Islands 17

Mango 17

Thithia 20

Vanua Mbalavu 21

Naitamha 21

Vatu Vara .22

Yathata .22

Kambara 22

Vatu Leile 23

Wangava 24

Taviuni .24

Ovalau 24

Munia ' .25

Lakemba 25

Lines of Beach Erosion 25

Elevations of the Fiji Group 27

Number of Elevations 27

The Nature, Origin, and Age of the Elevated Limestone 30

Tuvutha 30

Vatu Vara 31

Yathata 31

vol. xxxviii. — no. 1. 1

2 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

PAGE

Mba Vatu 31

Ngillangillah . . .... .... 31

Malatta 31

Mango ... 32

Geological History of the Lau Group and its Relations to the Western

Fiji Islands 32

Metamorphism of the Limestones ..... 34

Volcanism of Fiji with Reference to Lau and the Smaller Islands ... 37

The Recent Age of the Volcanic Phenomena 38

The Volcanic Rocks 40

Of Levuka 40, 41, 42

Of Mango ... 40, 42, 43

Of Munia 41,42

OfTotoya . . . . 41

Of Taviuni 41.42

OfSingatoka 41

Of Mba Vatu 42

Of Suva 42

Of Loma Loma 42

The Limestones of Fiji 43

Of Viti Levu 43

Older Limestones of the Main Island 44

The Dolomites 44

The Shelly Limestones 44

Calcareous Rocks passing into Soapstone 44

The Lau Limestones 45

Manganese at Mango and Thithia 46

Explanation of the Plates • 47

NOV 24 1900

No. 1. — Notes on the Limestones and General Geology of the Fiji
Islands, ivith Special Reference to the Lau Group. Based
upon Surveys made for Alexander Agassiz. By E. C.

Andrews.

INTRODUCTORY NOTE

By Alexander Agassiz.

After my exploration of Fiji during the winter of 1897-98, it became
evident that many of the problems which suggested themselves while
passing from island to island could only be solved by a careful examina-
tion of well-selected localities representing the typical structures of the
islands and coral reefs of Fiji.

It seemed best that the selection of a well-equipped explorer should
be made from Australasia, and I naturally turned to Professor David of
Sydney for advice ; he most kindly interested himself in the project.
At his recommendation, Mr. E. C. Andrews, assisted by Mi*. B. Sawyer
(both of the University of Sydney), consented to undertake the explora-
tion of the most accessible and interesting of the Fiji Islands.

The collections made by Mr. Andrews have safely arrived in Cambridge ;
they will be worked up in connection with the collections which I made
myself in Fiji, and at the same time with a mass of similar material
collected during the last expedition of the U. S. F. C. S. " Albatross "
in the Tropical Pacific from August, 1899, to March, 1900.

The limestones collected by the " Albatross " represent material
brought together from the Paumotus, Xiue, the Tonga Islands, and
Guam. Of course, a considerable time must elapse before this large col-
lection can be properly examined and the results compared with those
obtained by Mr. Hill and myself from an exploration of the reefs and
the elevated limestones of the "West Indian area (Bermudas, Florida, the
Bahamas, Cuba, Jamaica, San Domingo, and the Windward Islands).

I intended to publish the report of Mr. Andrews with a running com-
mentary based upon my exploration of Fiji, but with the wider expe-
rience lately gained in the Tropical Pacific, it seemed best to issue that
report with few comments, and in the final report on the coral reefs of

4 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

the Tropical Pacific to take up the subject again and thus not duplicate
now the report on the results of the last " Albatross " expedition.

As Professor David has examined not only the collections made by
Mr. Andrews, but also his report, I herewith add the greater part of a let-
ter he was kind enough to write me relating to the collections and report
of Mr. Andrews. This report, taken in connection with the monograph
of Christmas Island J (Indian Ocean) undertaken at the instigation of
Sir John Murray by Mi*. Charles W. Andrews, illustrates how com-
plicated a question the formation of an atoll is, even when we are able to
follow its stages in the pages of the history of its elevation. And nothing
shows more distinctly than the sections of Mango, of Tuvutha, and of
Singatoka, given by Mr. E. C. Andrews, how insignificant is the part
played by corals in the economy of an atoll.2

Newport, Rhode Island,
August, 1900.

1 London, 1900. Printed by order of the Trustees of the British Museum.

2 Charts of the Fiji Group and islands of the group will be found in "The
Islands and Coral Reefs of Fiji." Bull. Mus. Comp. ZooL, 1899, Vol. XXXIII.,
Plates l-23a.

ANDREWS: LIMESTONES OF THE FIJI ISLANDS.

PREFACE.

Letter from Professor T. W. Edgeworth David of the Univer-
sity of Sydney, New South Wales, dated March 15, 1899.

Mr. E. C. Andrews, B. A., is forwarding to you by this mail for your
information his notes, maps, sections, and photographs relating to the
geology of the raised reefs of Fiji, with some additional notes on Tonga.

This information he gained by personal examination of all the islands
referred to in his report, and he and Mr. B. Sawyer, B. E. (also of Sydney
University), who accompanied him, spared no pains in attempting to
carry out your instructions as fully as possible.

With the funds placed by you at his disposal, Mr. Andrews was en-
abled to hire a cutter and a Fiji crew, and by this means he managed to
visit most of the islands of the Lau Group and also explored Vatu Leile,
part of Viti Levu, Taviuni, Totoya, etc.

His method of work after landing on an island was to explore the cliff
faces and inland terraces, measuring altitudes and distances with an
aneroid and Abney's level, by pacing and taking angles with a prismatic
compass, and systematically collecting specimens by blasting and
quarrying.

This cliff exploration was done at first by means of a strong wooden
box lowered over the cliffs by a rope. This method was discarded for
the simpler one of scaling the cliffs, after the manner of the Fiji natives,
by means of the long rope-like roots of the banyan-tree, which depended
in some cases for over one hundred feet in length over the cliffs.

After the sections of the cliffs had been obtained and systematic col-
lections of specimens made, the higher slopes and inland terraces of the
islands were explored, it frequently being necessary to cut tracks through
the dense undergrowth for a considerable distance, in order to admit of
the ground being traversed.

Mr. Andrews' short report and notes and sections speak for them-
selves, but at the same time I should like to say a few words about
his general conclusions.

He finds the following formations represented in Fiji, the older being
mentioned first : —

b BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

(1) Bluish gray hard limestones without macroscopic fossils, and dip-
ping at steep angles. This was seen only in the Singatoka area, Viti
Levu, where it underlies the bedded foraminiferal limestone.

All the specimens of this have been forwarded to you, and this rock,
as being probably the oldest hitherto observed in Fiji, should be worthy
of careful microscopic examination. It may possibly have some relation
to the Globigerina limestone of the Solomon Islands, and that recently
brought to Sydney by Mr. Danvers Power from Noumea in New Caledonia,
which is also a Globigerina limestone.

(2) Volcanic rocks, such as spherulitic rhyolites and diabasic dolerites,
which must have supplied the well-rolled pebbles in the conglomerates
of the Fiji soapstones.

(3) (a) " Bedded limestones." These are developed, as observed, by
Mr. Andrews chiefly at the Singatoka area. Their angle of dip is 15°,
and they are largely foraminiferal (forms like Amphistegina, Globigerina,
Textularia, etc., being well represented), mixed with fragments of nulli-
pores, gastropod and lamellibranch shells, polyzoal skeletons, fragments
of echinus spines, etc. As you will see from the specimens sent, this is
distinctly not a coralline limestone, but rather a foraminiferal nullipore
rock. If the planes of bedding represent originally horizontal planes
which have been subsequently tilted so that they now dip at 15°, the
thickness of these bedded limestones at the Singatoka area, as measured
by Mr. Andrews, cannot be. much less than 1500 feet.

Outcrops of similar limestones have been observed by Mr. Andrews at
the raised atoll of Mba Vatu, Vanua Mbalavu, in the Lau Group, where
they form the foundation rock upon which the raised limestone rock
rests.

(b) Calcareous fossiliferous volcanic conglomei'ates probably passing
in places into the "soapstone" formation of Fiji, the latter being a
foraminiferal submarine volcanic tuff and tufaceous foraminiferal mud-
stone.

These may be slightly newer than, or possibly contemporaneous with,
the foraminiferal " bedded limestones" of the Singatoka area.

In the Lau Group the}' are seen at Mango, where, as shown on the
left-hand side of the upper of Mr. Andrews' sections, they form the basal
rock on which the raised reef limestone rests (Plate 2).

A thickness of only about four feet is exposed at Mango. They ap-
pear to be associated in places at Mango, as observed by Mr. Andrews,
with a steatitic fine basic tufaceous rock, homotaxial perhaps with the
Suva " soapstone."

ANDREWS : LIMESTONES OF THE FIJI ISLANDS. 7

(c) Fiji "soapstone " with interstratified coral reef and coral detritus,
as well as one bed of coarse well-rolled pebbles of volcanic rocks. This
formation is at least three hundred feet thick, and its base is not visible.
It exhibits in places, as at Suva, a dip of about 10°.

As you have ample material for a detailed description of this remai'kable
tufaceous marine mudstone, I will make only the few following remarks
about it : (1) That on certain horizons in Suva at over a hundred feet
above sea-level I noticed thin interstratified bands, almost wholly formed
of angular crystals of augite and plagioclase and of lapilli of decomposed
basic lavas, from the size of a hazel-nut to that of a walnut. There
appeared to me to be every gradation in this soapstone from a true sub-
marine tuff to a slightly tufaceous foraminiferal rock, and from the latter
rock into a tufaceous detrital coralline limestone.

In one spot at Walu Bay a small coral reef is interstratified with the
soapstone, as already mentioned in your paper.1 Professor Sollas has
referred to this as indicating that some, at all events, of the soapstone was
laid down in shallow water, and is not, as Mr. G. B. Brady maintained,
of rather deep-water origin. The coarse conglomerate formed of volcanic
pebbles at the base of the coral rock at Walu Bay confirms this view. I
observed a large clam-shell (Tridacna) at the rifle butts at Walu Bay
Quarry, and was shown large teeth of Carcharodon from this quarry by
Sir Henry Barclay. Mr. Andrews has forwarded you a similar Car-
charodon tooth. The evidence of these fossil sharks' teeth alone would
prove the formation to be at least as old as Pliocene, while the occur-
rence of Tridacna shows that it is referable to the later rather than to
the earlier Tertiaries.

4. Raised limestones. These are formed, according to Mr. Andrews'
observations, only to a very limited extent of coral in situ, such as would
have formed a true coral reef. True coral reef limestone was usually
found forming a capping about one hundred feet in thickness at the top
of the raised limestone. Such is the case ki several places at Mango, and
also at Tuvutha, Thithia, and Kambara, etc.

Much of this limestone has been elevated from 800 feet to 1050 feet
(Vatu Vara) above the sea. The base is usually not seen along the
seaward faces of the islands, except in the case of Mba Vatu and
Tuvutha and Mango. At Mango there is a good section a short dis-
tance inland, showing the limestone resting on a foundation of marine
calcareous rock with abundant fragments of decomposed basic or ande-

1 American Jour. Sci., Vol. V., February, 1898. " The Islands and Coral Reefs
of the Fiji Group." By Alexander Agassiz.

8 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

sitic volcanic rocks, the highest level of the junction plane being about
400 feet above the sea. As no trace of this calcareous volcanic con-
glomerate can be seen in the sea-cliff less than half a mile distant, and
about 400 feet high, it follows that the coral rock must thin out con-
siderably against the inland foundation of the island, while it thickens
rapidly seawards, as shown on Mr. Andrews' section.

This, I think, points either to stable equilibrium of the earth's crust
in that neighborhood for a sufficiently long period of time to have
enabled the reef to grow out seawards on a rock formed partly of its
own talus, partly of foraminifera and nullipores, or it points to upward
growth of the reef during a slow subsidence. Mr. Andrews inclines to
the former view. Had either elevation or subsidence, as the case may
have been, taken place fairly rapidly and continuously, the result would
probably have been the formation of merely a veneer of coral rock lying
on the foundation rock.

It seems to be very difficult to obtain a reliable natural section show-
ing the internal structure of this raised limestone, for two reasons : —

(i.) As the elevation progresses, the tendency is for each newer ring
of coral and talus formed at successively lower levels to hide from view
the base of the previously formed ring. See the description of Vatu
Leile by Mr. Andrews. This masking of the older limestone formation
by newer formations and thick soils is more conspicuous on the leeward
side of the islands than on, the windward, for while narrow terraces of
coralline limestone, mostly now covered by deep soil, were added to the
western side during the elevation of the atoll, not only was coral growth
checked, but the sea even made some inroads into the earlier limestone
rock, as proved by the wave-worn beach lines up to over fifty feet above
sea-level.

(ii.) Chemical solution is constantly forming stalagmitic or tufaceous
crusts over the cliff faces, sides of ravines, and even sides of caves in
the limestone, and this material, of course, either hides the original
rock from view, or by infiltration changes its original structure con-
siderably. The limestones are much dolomitized in places, so that
almost all original structure of the organisms composing them is
obliterated.

5. Andesitic rocks, later than the raised limestones.

(a) Andesite and coral agglomerates. These are well developed at
Mango. The inclusion of blocks of angular raised limestone, up to four
or five feet in diameter, in this agglomerate proves that it is newer than
the raised limestone, as Mr. Andrews is convinced that these included

ANDREWS : LIMESTONES OF THE FIJI ISLANDS. 9

blocks belong to the raised limestone, and not to the older "bedded
limestones," which latter are largely foraminiferal.

(b) Massive flows of andesite like those of Mount Rupert in Mango.
These cap the coral agglomerates (a) ; they have a very fresh appear-
ance, are in no places capped or encrusted by the raised limestone, and
invest the raised limestone and enclose inliers of it. For these reasons
Mr. Andrews considers these andesites newer than the preceding agglom-
erate and than the raised limestone.

These lavas rise in dome-shaped masses to a height of 700 feet above
the sea, and attain a thickness of about 300 feet.

6. Basalts. These form small outcrops quite inconspicuous as com-
pared with the andesites.

They are olivine basalts, having a very fresh appearance, and Mr.
Andrews considers them to be of later origin than the andesites, and
consequently the newest of the volcanic rocks of Fiji.

7. The pi'esent coral reefs. As regards the origin of the raised lime-
stones of the Lau Group, it would, of course, be premature to advance
any definite conclusions pending the examination of the collections made
by yourselves and by Mr. Andrews.

At the same time, a brief statement might be made here as to Mr.
Andrews' present opinion about them and my opinion as to their relation to
the limestone penetrated to a depth of 1114 feet in the bore at Funafuti
Atoll, Ellice Group. Such thin bands of the limestone rock from Lau
as Mr. Andrews classes as true raised coral reef seem to me very closely
allied to the bulk of the rock penetrated at Funafuti. Mr. Andrews con-
curs in this opinion. In the Lau Group, however, such true raised coral
reefs proper, in Mr. Andrews' opinion, form only a very small proportion
of the whole thickness of raised limestone. At Funafuti, however, they
constitute by far the greater proportion of the whole thickness of rock
proved in the bore. Whereas, then, at Funafuti, in my opinion, there is
little calcareous rock that is not true reef, at Lau, in Mr. Andrews'
opinion, there is little visible calcareous rock that is true reef. Mr.
Andrews thinks that probably the raised limestones of Lau acquired
their great thickness of over 800 feet by outgrowths of coral reef on a
bank of coral talus and foraminiferal nullipore limestone during a long
period of crustal quiescence followed by intermittent upheavals with
several intervening periods of quiescence which led to the formation
of the terraces.

Most of the raised limestone reefs of Lau are distinctly raised atolls,
notably Tuvutha, of which the old lagoon floor is about 350 to 400 feet

10 bulletin: museum of comparative zoology.

now above sea-level and about 200 feet below the general level of its
well-defined rim.

Terraced structure was certainly very plainly visible in a photograph
by Mr. Andrews of Yathata.

In the case of Yathata, it is probable that whatever may be the age of
its oldest and highest limestone area, whether it be late or middle Ter-
tiarv, it has representatives of limestones in its lower terraces of various
intermediate ages between Tertiary and Recent, as these terraces are
probably not terraces of erosion, but terraces of growth during pauses in
the elevation.

That the latest elevatory movements have taken place in recent
geological time, is proved by the freshness and good state of preserva-
tion of the wave-worn grooves in the sea cliffs at Vatu Leile and else-
where up to a height of 50 feet above sea-level, as detailed by Mr.
Andrews in his notes forwarded to you. Mr. Andrews' observations
emphasize, I think, the need upon which you have so often insisted
for making a separate study of each group of coral reefs and considering
them in relation to the local geographical, biological, and geological
conditions.

The whole question of the exact mode of origin of the massive raised
limestones of Fiji appears to me to be one which can only be satis-
factorily answered by careful examination of thin slices prepared for the
microscope of your own collections, supplemented by those forwarded
by Mr. Andrews.

Sydney, New South Wales,
March 15, 1899.

ANDREWS: LIMESTONES OF THE FIJI ISLANDS. 11

REPORT OF E. C. ANDREWS.

Introduction.

I left Sydney, New South Wales, on Wednesday, June 1, 1898, arriv-
ing in Suva on the 10th of June. Some few days elapsed before I could
get away from Suva for the Singatoka River, owing to the absence of
suitable cutters. During that time I examined the area around Walu
Bay, and searched the shore west of the islands of Lambikoof Vao for
calcareous rocks while tools and requisites for blasting purposes were
prepared.

A Tavua cutter was chartered through the kindness of Captains
Calder and Woolley of Suva, and a rapid glance taken at the Singatoka
River limestone, including also a visit to the famous dolomitic cliffs, thirty
miles up the river. Returning from Singatoka to Suva, a Kandavu
cutter was chartered for work in the Lau Group. While the cutter
made the direct passage to Mango, I took passage in the S. S. " Maori "
and a Taviuni cutter to the same island, calling and making rough notes
en route on the islands of Ovalau and Taviuni. Arriving at Mango, we
had to wait for the cutter, which occupied some fourteen days in the
passage to windward.

Three weeks were taken up altogether in exploring Mango and ex-
amining its cliffs. From Mango our course was laid to Munia and to
Loma Loma, and from thence we coasted along the eastern shore of
Vanua Mbalavu with the intention of visiting Mba Vatu and Ngillangillah.
Two weeks were spent in examination of the cliffs at this locality.

The section of North Ngillangillah being completed, we revisited
Loma Loma and Munia and scaled the principal heights. Thence a
dead beat was made to Tuvutha and the ascent of that island effected,
and then we continued on our way to Lakemba. The local cave was
explored. We then steered for Kambara, and spent part of two days
examining the same, and our course was laid to Totoya and hence to
Vatu Leile via the Solo Light on Kandavu, where the raised coral reefs
were examined. From Vatu Leile we made for Thuvu, and another visit
was paid to the limestone cliffs, and a brief examination made of the
recently upraised reef skirting the Thuvu-Singatoka sea-beach.

12 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

In this excursion I was accompanied and assisted by Mr. B. Sawyer,
B.E., of Sydney University.

Afterwards I made short excursions to Mango, Vanua Mbalavu, Nai-
tamba, Yathata, Yatu Yara, and Thithia to correct my first impressions
of this limestone group.

I sailed from Suva on October 28, 1898, with the intention of obtain-
ing samples of the Tongan limestone for comparison with the Fijian
raised reefs.

I reached Sydney from Tonga via Auckland on the 17th of December,
1898. I wish to acknowledge here my sincere thanks to Prof. T. W. E.
David of Sydney University for his valuable help and criticism ; to
Mr. B. Sawyer, B. E., of Sydney University, my associate in the Fiji ex-
pedition ; to Captain D. Calder of the A. U. S. N. Co. for invaluable
help in all matters of equipment ; especially also to Dr. B. Corney for
valuable hints and data; to Captain Woolley of Suva; to his Excel-
lency, Sir G. M. O'Brien, Governor of Fiji ; and to the Hon. W. L.
Allardyce for permission to blast ; to the Hon. J. M. Barron, owner of
Mango, for his great hospitality ; to Mr. J. N. Lennox, H. H. Steinmitz,
Prince Batu Lala, the Hon. J. Berry, Rev. J„ Burns of Lakemba, and
many others of the planters and settlers of Lau and the Singatoka River
for their hospitality and help.

The present expedition was undertaken primarily, as detailed in your
letter of instruction to me, with the idea of securing as many varieties
of limestone as possible from the Fiji area ; of selecting typical cliff's for
sectional purposes, of collecting as many fossils as possible, of study-
ing the topography and general geology of the Lau Group, and ascer-
taining if possible the basal limestone or volcanic rock supporting the
more modern deposits, such as reefs. The extraordinary influences of
erosion, the total lack of watercourse exposures except in caverns with
walls concealed by long deposition from the roofs, the obscuring of cliff
exposures by redeposition of calcareous matter in stalactitic form, the
enormous amount of secondary calcite arising from solution of the corals,
the dense vegetation, and the general absence of highways, all conspire
to hinder good progress in geological researches in the Fiji Islands.

The means of transport in these, as in most of the South Sea Islands,
is unsatisfactory. Many islands, of which Naitamba, Vatu Vara, and
Yathata are types, are approachable only in tiny craft, and that also only
in fine weather.

ANDREWS : LIMESTONES OF THE FIJI ISLANDS. 13

General Geological Structure-
The Larger Islands.

Viti Levu.1 — In this, the largest island in the Fiji Group, nay researches
were confined to an examination of the Suva district, of a ten-mile strip
stretching from JSTandronga Singatokawards, and a short excursion ex-
tending over some thirty miles up the large. Singatoka River. Viti Levu
and Vanua Levu stand out distinctly from all other members of the
group, and bear the imprint of a hoary antiquity compared with the
volcanic islands of Taviuui, Ngau, and Kandavu, as well as with
the limestone islands constituting the Lau division of the Fiji area. Pro-
ceeding along the banks of the Singatoka River, interesting outcrops of
limestone of high dip were observed. Volcanic agglomerates of porphy-
rite and masses of andesite rock, apparently identical with similar rocks
in the younger islands, have burst through the more ancient sedimentary
rocks. The volcanic rocks exhibit various stages of decomposition. The
hills roll, cone after cone, behind each other to a height of 3000 feet, and
the main road follows the curves of the lower hill shoulders, thus reveal-
ing numerous dykes, and thin beds of strata, dipping at high angle, which
are crossed again and again in a few hours' walk. In places these strata
contain a bright green cherty rock, interbedded with softer brownish
layers.

The Stratified Rocks. — Argillaceous rocks and soft crumbling beds
occur with high dip. They are brown in color and at times effervesce
with acid.

The Dolomites. — These appear to underlie the strata just enumerated.
They are principally centred in an immense block overlooking the river.
They consist of two hills 1000 and 1500 feet respectively above the
river, and reaching, in the case of the smaller one, into the water at the
base. They are practically perpendicular, having a slope of 85° in places
aud 90° in others, and are uniformly hard and homogeneous in character
from base to summit. They are tunnelled with caverns. To the eye or
ordinary lens they are non-fossiliferous. Acid scarcely causes any effer-
vescence. According to report this rock is found along a line roughly
northwest and southeast, but I saw no trace of its development else-
where.

Of the evidently later limestones that more or less fringe the shore of
the island near the mouth of the Singatoka River, one notices, com-
1 See A. Agassiz, Bull. Mus. Comp. Zool., 1899, Vol. XXXIII. p. 110.

14 bulletin: museum of comparative zoology.

mencing with the older, a series of rocks which, excepting the oldest,
conform to the same dip and strike. They comprise a vast group of
Limestones and so-called " Fiji soapstones " that form the nearer coastal
scarps, reaching in places to a height of 330 feet above the sea. The old-
est of these Singatoka rocks is a compact bedded blue and apparently
non-fossiliferous limestone having a high dip (50°) (Plate 3, Fig. 1).
To this succeeds a series of soft rocks having an average dip of 15°
(Plates 7, 33). The junction of this highly inclined blue lime-
stone with the softer overlying sti'ata could not be distinctly seen
(Plate 32). The next group of strata in ascending order is undoubt-
edly large and comprises a sandy species of limestone. This gives
place to a shelly and foraminiferal zone ; another series contains fossils
of Pecten and other lamellibranch types, with here and there a stra}-,
weathered, indeterminable coral (Plate 3, Figs. 1, 2). Still another
series comprises a dense, homogeneous, macroscopically non-fossiliferous
zone, and this includes a six-foot belt of coral reef, containing Porites
or Monti pora. This, in turn, overlays a two-foot belt of red clay, and
the top of the reef consists of a thin belt of finely laminated lime-
stone. To this succeeds a great development in layers of a brown,
friable rock that will scarcely bear the weight of the hammer (Plate 3,
Figs. 1, 2). A gap in the section occurs at this spot, due to drift
sands of later origin hiding the underlying rock : but two or three miles
toward Thuvu similar layers of fossiliferous limestone are again picked
up, and in these occurs a belt of hard red limestone.

At Thuvu the limestone has disappeared and its place is taken by
hills 200 to 300 feet high of inclined " soapstone " (Plate 4).

The series of friable and dense limestones just considered lies back
about two miles from the sea, the Singatoka having deposited its enor-
mous flat seawards in front of it so that the cliffs abut directly on
it. On the seaward of the lower limestone formations and of the old
Singatoka flat a modern elevation has exposed a reef wall in places as
much as twenty to twenty-five feet above high-water mark (Plate 5).
This reef face is vertical.

"The Lau Group" and the Smaller Islands.1

These may be divided for purposes of classification, as has been done
before, into : The volcanic islands ; The limestone islands ; The volcanic
and limestone islands.

1 See A. Agassiz, /. c, p. 17.

ANDREWS : LIMESTONES OF THE FIJI ISLANDS. 15

The Volcanic Islands. — These include islands like Kandavu, Taviuni,
Ngau, Nairai, Mbengha, Totoya, and Moala. Some of the islands, like
Muuia and Ovalau, are formed mainly of andesitic agglomerates (Plate
6) laid down rapidly in thick strata. In these beds occur layers of vol-
canic ash, finely comminuted, and filled with beautiful augite crystals.
Generally this agglomerate consists of angular blocks so firmly welded
in places as to render the junction of individual blocks almost indistin-
guishable. On Vanua Mbalavu layers of volcanic conglomerate occur
with rounded pieces, and this as much as 500 feet above the sea. Again,
as at Taviuni, on a large scale, and more or less on almost every island in
the group, streams of lava occur, the so-called " vata loa" of the natives.
On islands like Taviuni, Mango, Thithia, and Vanua Mbalavu, this
vesicular lava appears so distinct as to seem but the product of yester-
day (Plate 8). The basalt appears to have followed the andesite out-
bursts. Basaltic and andesitic rocks are seen everywhere on Taviuni
and Mango. Taviuni contains numerous craters, one being nearly 3000
feet above the sea level, with its rim fully 4000 feet above the sea.

The Limestone Islands.1 — These are very few in number and are small.
Vatu Vara, Waugava, Wailangilala, Katavanga, and Namuka are the
only ones I know to be wholly limestone. Fulanga, Karoni, and Aiwa
I believe may be (in the Lau division) added to this list. The Yasawas
are limestone, 800 feet high. With the exception of Fulanga, each
of the above named is extremely small. They consist of compact lime-
stones, very flinty in composition, and emitting showers of spai'ks when
struck with a large hammer.

The Limestone and Volcanic Islands.2 — These constitute a numerous
and important group. Besides the two main islands, it comprises Mango,
Vanua Mbalavu, Tuvutha, Naiau, Lakemba, Kambara, Thithia, and
Naitamba. These islands were, originally and generally considered,
lands possessing and presenting as many as six or seven terraces or indi-
cations of elevation and subsequent reef extension. Subsequently came
the upheaval which produced the volcanic agglomerates and lava masses.
These burst through and destroyed a great part of the cliff slopes, and
more or less filled the old inland lagoon areas. Such has been the case
particularly with Mango, Thithia, Lakemba, and Vanua Mbalavu. At the
latter island the force was so immense as to leave only the north and
south extremities intact, carrying the middle reef formations as high as
500 and 700 feet above high-water mark, leaving but a miserable meta-

1 See A. Agassiz, /. c, p. 43.

2 See A. Agassiz, /. c, p. 88.

16 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

morphosed remnant to represent the pre-existent land now showing
merely as chalcedonic blocks littering the razorback as it extends S-like
on the surface of the sea.

Basal rocks or basal limestones are of unfrequent occurrence. At
Mba Vatu — the northern division of Vanua Mbalavu — huge stratified
limestones occur as high as 250 to 300 feet above the sea. To the north
of these dipping limestones the land is occupied by " terraced " forma-
tions, and to the south it abuts on a great andesitic massif. Fourteen
miles south the coralline limestones lie unconformably on a bedded red
laterite.

At Tuvutha and Mango a very decomposed conglomerate or agglom-
erate of andesitic origin underlies a bright red limestone in the case of
Mango and white limestone in the case of Tuvutha. The cementing
material is very calcareous. These rocks bear the stamp of great age,
aud have evidently a great development beneath the coral islands.

The Geological History and Topography.
The Main Islands.

The General Topography. — The topography of southwestern Viti
Levu is characterized by numerous almost perfect cones, with smaller
cones on their slopes.

In the Namosi, or central district of Viti Levu, sharp knife-like sum-
mits and aiguilles rise to 3000 and 4000 feet above the sea. These yield,
on disintegration, fertile soils, especially in the valleys.

The great dolomites of the Singatoka River constitute a decided topo-
graphical feature. They rise almost perpendicularly with exceedingly
large and sharp projections in the upper half, the result of long-con-
tinued erosion. From the base the strata rise straight as masonry for
400 or 500 feet ; above that they become more or less broken. They
appear to be exposed in part by the removal of softer overlying strata.

The Singatoka River itself has changed its course, and viewed from
the huge tilted dolomites dominating the river, its old bed, now about
50 feet above the modern stream, constitutes a marked feature in the
scenery.

Coastal Topography. — The coastal topography of the larger islands
consists in the main of cliffs of elevated and shelly limestones of ancient
date, scattered amongst huge rolling slopes of volcanic hills extending
seaward. . In these elevated shelly strata reefs of undoubted coral origin

ANDREWS: LIMESTONES OF THE FIJI ISLANDS. 17

occur. These have eappings and bases of so-called " soapstone." This
alternating structure is still more marked along the Tamavua River,
near Suva, where many successive layers of soapstone and of limestone
are visible in one exposure (Plate 39). Terraces and beaches are not
distinct topographical features of the larger islands, as they are in the
members of the Lau Group.

With respect to the reefs surrounding the islands of the Fiji Group,
we may notice that on the southwestern portion of Viti Levu the reefs
are fringing.1 Strange to say, the Singatoka River, large as it is (Plate
3a) and bringing down a volume of fresh water comparable with that of
the Rewa River, causes no sensible diminution of the fringing reef.

There is one small passage only, a very few yards in width, through
which the escaping water at low tide pours out at the rate of six to nine
miles per hour. Even this passage has a reef base nearly awash at
low tide. For a distance of seventy miles towards Suva the fringing
reef continues. But beyond this point, in a direction extending towards
Ovalau, the reefs gradually recede for miles from the shore.

Vanua Levu is surrounded by an even greater mass of reef.2

Topography of the Smaller Islands.

The islands of the Lau Group and such islands as Taviuni and Kandavu
present the appearance of huge sloping lava flows as at Taviuni, as higher
pitched slopes at Ngau, Nairai, and Vanua Mbalavu ; or, again, as solid
limestone masses rising from the sea. These mural fronts are roughly
subcircular in shape, are some 400 or 500 feet high, and may exhibit
vertical cliffs looking out upon large skirting sand flats, or, again, they
may exhibit a slope suggestive of an inverted cup or truncated sugar-
loaf. This shape is of general occurrence, its uniformity all over the
group suggesting it to be the original slope of the reef before elevation.

Terrace structures and upraised lines of ancient sea erosion cut into
the cliff's are also a frequent feature of Lau scenery. A detailed exami-
nation of Mango's topography and more or less complete notes on other
islands visited is furnished here as a supplement to these general
remarks.

Mango (Plates 1, 2, 9-17). — This island is subcircular in shape,
and as seen from the sea appears like a huge mass that has been
upheaved about 500 feet above the level of the sea. The summit
seems to be a vast flat broken in two places by great dome-structures

1 A. Agassiz, I. c, pp. 110, 118.

2 A. Agassiz, I. c, p. 121.
vol. xxxviii. — no. 1. 2

18 bulletin: museum of comparative zoology.

rising above the general level, and almost equal in area to the base,
owing to the wall-like front the island presents to the sea. The island
is small, being eleven or twelve miles in circuit and three and a half
miles in diameter. A strip of fertile flat about 100 yards in width runs
round the greater part of its circumference and relieves the ruggedness
of what would otherwise be an iron-bound coast. In one place this
feature is found where the huge cliffs plunge abruptly into the water.
Mango possesses a great number of vertical cliffs, some of them as high
as 400 feet (Plate 9), and where the limestone ring or girdle of the
island persists, this cliff structure is varied only with the " sugar-loaf"
formation.

On a closer examination traces of a " terrace " up to 200 feet above
high-water mark, will be found on the northern and southern aspects of
the island. Splendid patches of raised reef, with Pontes and Astraean
corals in abundance in situ, lie scattered over the sand fiats and up the
lower slopes of the cliffs. These are subsequent to the first upheaval.

The practised eye soon detects a want of uniformity in the limestone
ring (Plates 12, 15, 16, 17). The wall structure is replaced by lava
flows, and at the centre of flow the coralline rock is absent. Glancing
aloiiu the line of cliffs from some point where they still preserve then-
original shape, the escarpment is seen to dwindle away towards the
centre of flow of the disrupting lava, passing from great cliffs to in-
significant wedge-shape, still preserving, however, the general level of
the ring as it vanishes beneath the lava. Continuing the line of sight,
the limestone is picked up again at the same level a little further ou,
now, however, merely a few limestoue islands rising above the lava
sheet. The volcanic mass pours through the gap made in the old lime-
stone ring and spreads out, fan-shaped, away from the cliffs, carrying the
island structure farther out to sea than formerly by some 400 yards
(Plate 11). This is the case on both the north and south of the island.
On the northeast a vast natural break occurs in the old ring. This gap
represents a former channel of communication between the outer sea and
the central lagoon.

A coil of rope laid on the floor, sand poured gradually over any two
opposite sides of the rope-ring till the cone-shaped sand masses rise con-
siderably above the sub-level of the ring and hide the rope at that spot,
a V-shaped gap cut in the rope at a spot bisecting either of the rope
semicircles left, — this is the contour of Mango, with the rope substituted
for the limestone girdle, the sand-cones for the andesite masses, and the
V gap for the old passage from the sea to the central lagoon.

ANDREWS : LIMESTONES OF THE FIJI ISLANDS. 19

On the northeast of the island there is a small lagoon (Plate 12)
having a longer diameter of 1000 yards (Plate 1) and a shorter one of
400 yards. A channel ten yards in width connects it with the sea
through a narrow limestone belt, 50 to 100 feet high. There is a
trace of an old reef on its margin.

A fringing reef of modern date with enclosed lagoons conforms in a
general way to the shape of the land. These lagoons are deep holes,
true lakes at low tide, some of them are fully nine fathoms deep. The
floors are usually sandy, with scattered corals, while the steep walls of
the lagoon are often formed of growing reef. The reef itself is much
littered with huge blocks of coral debris, known locally as "nigger-
heads " and " horse-heads." Other larger blocks occur, some of them
40 feet in height. They result from cliff ruptures, and have undercut
bases (Plates 13, 14). In the deepest lagoon several small islands
occur, partly volcanic, and partly limestone agglomerates lying in a
cementing of andesite tuff.

Along the shores of the island "beach rock" is well developed, and
at two or three spots along the shore streams of fresh water crop
up between low and high water marks from beneath this beach-rock
formation.

A rapid glance at the centre of the island reveals a limestone ring
broken up by great volcanic outbursts of andesitic lava, dipping in most
cases towards a great central depression (Plates 15-17). The geometrical
shape of this hollow is lost, owing to the intrusion of a vast andesite
mass near its centre ; from the sides and from the slopes of the lava
streams forming the great amphitheatrical enclosure, vast alluvial masses
have spread over the ancient limestone floor of the hollow. The old
base or floor crops up now and again as a broken and weathered lime-
stone, or as huge isolated limestone masses projecting above the alluvium
or the lava slopes. At levels of 250 and 50 feet respectively above
high-water mark, remnants of old reefs are encountered along the inland
cliff exposures. The numerous caves of Mango are objects of in-
terest. One of these occurs in a large bluff 600 or 700 yards
inland. It is little more than a vertical hole 100 feet deep, with several
minor lateral ramifications. Other large caverns gape on the coastal
bluffs. They open out on the seaward faces of the cliffs in almost inac-
cessible positions. One or two penetrate the cliffs for distances exceed-
ing 60 yards. In all these large subterranean passages the examination
of the calcareous masses through which they have been formed is ob-
scured from various causes. In one or two places, however, the lime-

120 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

stone could be examined. Its general appearance was that of calcareous
patches found on the modern reefs, containing no whole corals, but full
of small fragments of reef organisms set in a fine, hard, compact white
matrix.

Thithia (Plates 18-21). — Thithia and Mango might be studied in
detail with almost equal advantage. Whatever statement holds good
in the description of Mango is equally applicable to Thithia. Thithia is
a trifle larger, being fourteen miles in circuit and four miles in diameter.
Both have the same vertical cliffs of limestone, the same forest-clad lime-
stone slopes (Plates 18, 20, 21), and the same uniformly inclined lava
rock bursting to the sea. Mango is not so completely overwhelmed by
volcanic material as the sister island, twenty miles away, and shows the
limestone of the interior more nearly approximating to its former state
than is the case with Thithia. Mango exhibits terrace formations and
unmistakable signs of a very recent uplift, while at Thithia only traces
of terraces can be detected. Formerly Thithia possessed an outer ring
of raised coralline limestone origin, with a central hollow (Plate 19).
That hollow has been so encroached upon by lava outbursts that there
is at present but a hint of its former extent. Thithia might now be
described as a circular island, indented slightly at one end, and sloping
steeply to the water, and with four or five bold black cliffs and forest-
clad rocks to break the continuity of the volcanic slopes. These lime-
stone rocks at times project but slightly through the down-like slopes
of lava that have almost buried them. The development of rugged cliffs
and isolated calcareous blocks comprising the broken country, hereto-
fore mentioned, represents the survival of the old limestone perimeter
(Plate 19).

To the north are evidences of three platforms, respectively 400, 250,
and 50 feet above tidal influence, the lower one being particularly well
marked, with the middle one rising above it precipitously.

Nevertheless from this chaos we can restore the former condition of
Thithia. It finds its parallel in Mango. Untouched in the volcanic
uplift, the gap through which the ancient sea ebbed and flowed to the
inner lagoon still exists (Plates 20, 21). Its 300 yards' wide entrance
is protected on both sides by bold rocky scarps 400 feet high, and the
whole line of gap may be traversed for nearly three quarters of a mile
inland (Plate 20), with no part more than four feet above high-water
mark. At its head lie great boulders similar in appearance to those
disposed along the modern shore lines of Lau. As with the modern
shore rocks, so here at the terminus of this old sea-way (Plate 21), their

ANDREWS: LIMESTONES OF THE FIJI ISLANDS. 21

bases have been indented by the earlier sea at a former period. The whole
locality has the aspect of a mangrove swamp left dry by the receding
tide, and on a first visit to the place it is hard to disabuse oneself of the
idea that here is a modern tide-affected area. Mr. J. N. Lennox, part
owner of Thithia, assures me that there is no knowledge of the sea enter-
ing that gap.

The undercutting of cliff bases by sea-erosion is generally confined to
limestone areas. At Thithia we have an instance where a volcanic agglom-
erate has undergone a similar carving by the action of waves.1

As with Mango, huge caverns occur, showing a great development of
stalagmite and secondary calcareous incrustations covering the walls.
The internal structure (15 yards in) reveals a rock similar in composi-
tion to the material of the general cliff face, containing very few corals,
and is in hand specimens non-fossiliferous, as is the rule in the raised
limestones of Fiji.

Vanita Mbalavu (Plate 22). — This is literally "the long land." It
is a distorted S shape. The belly and middle curves are composed of
andesitic lava, andesite agglomerates, andesite basalts, and " talasinga,"
or burnt earth of the natives. This "talasinga" is a recognizable
feature in almost any Fijian landscape. It is a soil arising from volcanic
rock decomposition, and is of such a bright red that at times it is used
as a pigment. The terminal points of the S consist of high bluffs of
elevated limestone, reaching in the north, at Ngillangillah and Mba
Vatu, a height of 500, and at the south a height of over 400 feet, at
Malatta and Susui islands. The length of the main island is 15 miles,
and disposed along the curving backbone of the island are the points of
eruption which have so altered the former topography. At the south of
the island the volcanic rock has tilted the limestone rocks.

Vanua Mbalavu, with its associated islands and islets, lies in a barrier
reef of pronouuced type (Plate 22). A small patch of the lagoon
separates the reef from the mainland to the west, while to the east it
retreats for many miles from the land.

Naitamba (Plates 23, 24) lies 20 miles north of Vanua Mbalavu. It
resembles Mango and Thithia in most particulars, but differs as to the
extent of the volcanism. The northern extension of the old limestone
ring uow lies quite inland, owing to andesite extension at the base of the
cliffs. To the south the cliffs have been fractured and lifted to a height
of 600 feet above sea-level.

1 See A. Agassiz, I. c, Plates 62, 63.

22

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Vatu Vara (Plates 25, 26). — The mass of the island rises as a vast
truncated pyramid, presenting along the summit an almost perpendic-
ular cliff 200 feet high. It possesses a sub-horizontal summit of some
forty acres. On closer examination, the top is seen to contain innumer-

VATU VARA SEEN FROM MANGO.

able pits and depressions, varying from G to 30 feet in depth. Traces of
five uplifts are visible on its ascent. Three of these are " terrace"
formations. Two are in form of beach-erosion lines.

Tathata (Plate 27). — Seen from the southeast, it looks like a hat.
From the east it appears to consist of six steps arranged symmetrically

TATHATA SEEN FROM NGILLANGILLAH.

with respect to a central dome. On the western side of the island this
symmetry is broken by a volcanic mass that has welled up to the 500 feet
level, scorching and whitening the limestone. A small island called
Kaimbu is included in the same lagoon. They are 500 yards apart, and

YATIIATA AND KAIMBU FROM MANGO.

the water between is about 6 feet deep. On Kaimbu a line of beach
erosion exists a few feet above high-water mark. The dominating cap,
840 feet in height, is " sugar-loaf."

Kambara. — It is much as Mango may have been, one solitary ande-
site mass of 30° slope, rising 470 feet above the sea to mark its era of

ANDREWS: LIMESTONES OF THE FIJI ISLANDS. 23

active volcanism. This dome of lava x burst through the old cliff ring
and carried the broken fragments up, subsequently, on the top of a
rising lava.

Vatic Leile (Plates 28-31). — This interesting island differs materi-
ally in many ways from the other Fiji islands. Topographically, it
resembles the Vavau group, while lithologically it shows affinities with
the Lau Islands.2 The eastern coast is flat (Plate 29), rising, like the
northern shore of Tonga, but a couple of feet out of the reach of the high
tides. As at Vavau, a deep deposit of soil covers the rocky base, and
prevents geologizing to any extent. Through this flat long tongues of
basalt have been protruded, reaching into the lagoon. On the western
side are magnificent examples of former elevations. On a cliff face 100
feet in height (Plates 30, 31), are no less than four well-preserved lines
of beach erosion. These, with the modern one, constitute a marked fea-
ture in the landscape. They are, in cross section, like a quadrant of an
ellipse with the longer semi-axis held horizontally. The most pronounced
is that immediately above the modern line of beach erosion, and persisting
for so many miles of such a height and flatness of floor as to have earned
from the natives the name of " The Great Walk." The floor is 6 feet
above high-water mark. It is 6 or 7 feet from floor to roof, and is eaten
back into the cliff some 10 feet. The second is 14 feet, and the third
and fourth are respectively 35 and 45 feet above the same datum line
(high-water mark). It may be remarked, in passing, that a line of beach
erosion corresponds to a terrace on the opposite side of the island. But
it must not be forgotten that these evidences of erosion do not stretch
as geometrical lines throughout the length of the western cliffs. Only
in one favored spot can the four be seen at once. On the main
cliff the third and fourth lines are very faint, while the second is
invisible.

Corresponding to " The Great Walk " (Plate 31) is an upraised reef
spreading out from the cliffs, which at this spot have retreated slightly
inland. The reef is bounded to the rear by a cliff 100 feet high and by
the upper part of " The Great Walk." This flat is about 6 feet above
the tides. So recent is the uplift that hundreds of loose corals lie
over the platform. The platform itself consists of reef-debris rock, as
hard and dense as any Lau limestone, while here and there scattered
coral are stuck in it like stray pins in a cushion.

By walking behind the sea-cliff, another set of rocks about 50 yards

1 See A. Agassiz, /. c, Plate 78.

2 A. Agassiz, /. c, Plate 100.

24 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

inland is observed, and here at the same level, 6 feet above high-water
mark, occurs the raised reef with scattered corals and the line of beach
erosion. The traces of a raised sea-beach from 7 to 11 feet above the
present tidal limit are still fresh at their feet. These inland cliffs rise
some 30 or 40 feet above the flat and are encrusted with coral reef forma-
tions. At 16 feet above high-water mark a small patch of disintegrated
raised " beacli rock " may be seen.

At Thuvu, as at Vatu Leile, the dead corals litter the reef and project
above its surface. It seems to have a general level of 6 to 8 feet above
the tide, with a higher development to the rear, rising from 16 to 20
feet. Traces of a more elevated beach (25 feet) also occur (Plates 30,
31). The reef has grown on the older inclined limestones.

Wangava. — It has rugged slopes rising steeply (50° to 60°) from
base to summit, to a general level of 300 to 350 feet above the sea.
Half-way up the slopes traces of a " terrace " may be seen. The ceutre
of the island is depressed, containing a lake which rises and falls with
the tide, as I am informed by Hon. W. L. Allardyce. (A similar case
occurs at Nomuka, in the Tongan Group.)

Taviuni (Plate 8). — It is of blunted wedge-shape, and its axis,
determined mostly by volcanic cones and craters, lies approximately
north-northeast and south-southwest. As seen from the south, it is
cone shaped. An examination of the western slopes reveals flows of
basaltic lava partially hiding masses of older lava flows of steeper
slope. To the east it falls precipitously, and the later basalt does not
appear to have influenced the topographical features of this division of
the island.

An inland lake occurs at the 2800 feet level. It is either contained
in an extinct crater or has been formed by the closing in of dome-shaped
masses on one another. A curious feature in this old lake is the re-
peated occurrence of long narrow belts of water running in all directions
through the bog. Their edges are sharply denned by the transparency
of the water when contrasted with the surrounding bog.

Ovalau (Plate 6). — One of the most rugged and picturesque
islands of the group. It is contained wholly within the Viti Levu
barrier. It is 2100 feet in height, and is composed of huge beds of
evenly dipping strata carved into deep valleys, aiguilles, cliffs, and dome-
shaped masses. These strata are andesite agglomerates, in places con-
sisting of augite-andesite blocks loosely aggregated, in others of angular
blocks firmly welded together, aud again beds of andesite ash occur full
of augite crystals.

ANDKEWS: LIMESTONES OF THE FIJI ISLANDS. 25

Munia. — It is described most accurately by calling it " a miniature
Ovalau." The same rugged cliffs of andesite agglomerate occur on a
smaller scale, the total height here being 1000 feet.

Lakemba. — The mass of the island is composed of andesite hills,
crowding thickly one upon the other and rising at the highest point to
720 feet above high-water mark. The former state of the island is
represented by a mere fringe of coralline limestone on the southwestern
extension of the volcanic mass. The whole configuration of the older
limestone has been altered, unless, indeed, Lakemba was at one time an
extensive limestone island. A little bigger volcanic display, and not a
vestige of the older limestone would have remained, except perhaps as
blocks in andesite ash beds, as is the case with some of the Mango
limestones.

The great attraction of Lakemba is its cave, piercing the fragment of
limestone that skirt the volcanic slopes. The white residents describe
it as an immense tunnel, easy of examination, and running for about two
miles through the hill. Measurements by myself reduce this great
length to 500 yards. Even thus the statement is misleading, for the
cave disappears in the hill, describes an arc of a circle, and by so doing
reappears round the corner of the same bluff that it started at. From
entrance to exit is about 200 yards measured round the outside of the
cliff, and the exit is 100 feet above the entrance. The cave is more
easily explored than any I have visited, the floor being fairly smooth,
wide, and of even slope. Generally the roof is about 50 feet above the
floor. The walls, where clean, reveal the structure of a raised reef,
composed of corals and fragments of mollusca.

Lines of Beach Erosion. — These may be divided into two groups, the
ancient and the modern lines of tidal wear.

In the islands of the Lau Group, and confined almost exclusively to
the limestone formations, the effects of recent beach erosion are very
pronounced. r "When a group of islands, of which the hundred isles of
Ngillaugillah (Plates 22, 35) and Mba Vatu may be cited as illustra-
tions, have their windward face overlooking a fairly deep lagoon or are
placed directly opposite an opening like the Ngillangillah Passage,
we find small islands quite undermined by wave agency, while others
project in mushroom form above the surface of the lagoon (Plate 35).
On the windward side of the Ngillangillah Group the erosive action
has extended as much as 8 to 10 feet vertically and into the cliff face
from 14 to 15 feet.1 To the leeward this cutting into the cliff rarely

1 See A. Agassiz, /. c, Plates 73, 74, 92, 93.

26 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

exceeded -i feet vertically or laterally. Other islands furnish equally
good examples of this modern erosion, but only in cases where the
cliffs faced a wide expanse of lagoon or sea, with no intermediate reef.1
South of Thithia a remarkable mass of volcanic agglomerate occurs
that exhibits similar weathering, inducing thereby a toadstool shape in
the block.2 Fine erosion effects are visible on the great outlines of the
Mango cliffs, which have been carved into the most bizarre possible
forms. Equally curious also are some of the under-cut cliff-bases at
Fulanga,8 where islands filling the lagoon of the raised reef exhibit every
stage in this undermining of the cliffs. Besides the islands just enu-
merated, Vatu Leile,4 Naitamba, and Kambara show equally striking
results due to erosion. In some cases the redeposition of carbonate of
lime has taken place to such an extent as to almost obliterate the traces
of the erosion-line (Plates 28. 30, 31). Much iron as ferric oxide is
mixed with the redeposited lime material, and imparts a rich-brown or
red color to the mass.

At Vatu Leile this redeposition of calcareous material charged with
ferric oxide has gone on very largely in the line of erosion known as
" The Great Walk," parts of which it has completely obliterated, while
other parts are half filled with large masses of the deposit. At the
southern extremity of Vanua Mbalavu, another line of beach erosion
occurs tilted at 15° to the horizon. Here again the cliff has almost
regained its original shape by subsequent refilling of the wave-worn
groove.

Most of the islands, as Thithia, Lakemba, Taviuni, Vanua Mbalavu,
Vatu Vara, Kambara, and Vatu Leile, have more or less well developed
flats extending along the bases of the hills, whether volcanic or lime-
stone. The Mango flats vary from 100 to 250 yards in width, and
throughout this extent are fairly level. They are most persistent along
the windward or southern side. Two flats were examined, one on the
south and one on the northeast of the island. In both cases they lay at
the bases of limestone cliffs 400 and 500 feet in height. The seaward
edge of the flats is perhaps higher than the middle portion. To make
sure of the origin of this raised portion, two holes were sunk till the
harder basal rock was reached. In both cases the old beach rock was
found, and this at a distance of nearly 200 yards back from the present
similar formation.

1 See A. Agassiz, I. c, Plate 94 (Ongea).

2 See A. Agassiz, I. c, Plate 62.

8 See A. Agassiz, /. c, Plates 82, 83, 84.
4 See A. Agassiz, /. c, Plates 9G, 100. 102.

ANDREWS: LIMESTONES OF THE FIJI ISLANDS. 27

It would thus appear that some of the flats are wind-blown formations,
and not the results of recent upheaval, as in the case of Lakemba, Thithia,
and Vatu Leile. [The reef-flats are due to submarine erosion. — A.
Agassiz.]

Elevations of the Fiji Group. — From the consideration of the raised
coralline limestone alone, it is clear that a greater elevation may be
claimed for the northern division of the Fiji Group than for the southern
one.1 The Yasawas (800 feet), Thikombia-i-ra (630 feet), Tuvutha (800
feet), Vatu Vara (1050 feet), and Yathata (840 feet), are lofty heights
compared with the elevated islands stretching to the south of terraced
Tuvutha. The elevations decrease in altitude as we leave Tuvutha and
retreat from the equator. At 18° south latitude, we have Naiau with
580 feet, at 19° south latitude Kambara and Wangava from 300 to 350
feet, and still farther south Fulanga with 260 feet of uplift. Thus it
seems that the great uplift or series of uplifts reached its maximum on
or about the 17th parallel of south latitude. Whether this uplift was
due to one elevating influence, as suggested by J. Stanley Gardiner,2
or due to intermittent periods of uplifts and repose, is now to be
considered.

X umber of Elevations. — As mentioned before, Mr. Gardiner considered
the probability of one uplift for Lau. In this he was influenced by the
apparent lack of visible "terrace" or even incipient "terrace forma-
tions," and from the repeated occurrence of huge vertical cliffs noted by
him in his researches conducted in the Lau Group. But the evidence
for repeated uplifts in place of a rapid and single rise seems incontest-
able when read in the light of the following facts.

MBA VATC FROM MANGO.

Appended is a summary of the numbers of indications of upheaval in
individual islands, and from a careful study of the first 50 feet of Vatu
Leile, Vatu Vara, and Yathata, it seems probable that many former
traces of elevation have been completely obliterated, due to the subse-

1 A. Agassiz, /. c, p. 131.

2 Proc. Canib. Phil. Soc, 1898, Vol. IX. Part VIII.

28 bulletin: museum of comparative zoology.

queut degradation of the limestone,1 and the actual upward movements
outnumber those which can be observed.

Ngillangillali
Mango,
Ovalau,
Vatu Vara,
Naiau,

,3(4?).
2.
1.

4(5?).
2.

Mba Vatu,

Thithia,

Yathata,

Tuvutha,

Lakemba,

5.

3.

7.

4. Plate 3, Fig. 3.

2.

Wangava,

2.

Kambara,

2.

Vatu Leile,

5. Plate 30.

Of all the elevations the trace of the last upheaval is specially well
preserved in several of the islands. Speaking very generally, it occurs
approximately at the same height above high-water mark in the group.

At Ovalau an old reef is exposed in a small creek at the back of the
main hotel. This watercourse has cut through a loosely aggregated
mass of andesite blocks and pebbles, several feet in thickness, and now
runs along a coral floor exposed by its own erosive and denuding action.
This raised coral floor is almost indistinguishable from the modern reef-
flat. The highest tides do not quite cover it.

At Vatu Leile the evidence of a recent uplift is undeniable. It can
be seen on the eastern edge of the island in the form of a flat removed
but a few feet from the reach of the tides (Plate 29). On the west it
takes the form of an enormous groove extending for miles, flat underfoot
and arched overhead (Plates 28, 31). At Yathata we have the same evi-
dence as at Vatu Leile, and at Lakemba we have an extensive flat stretch-
ing along the south of the island and reaching from the present reef inland
as far as the foot of the volcanic slopes (Plate 34).

If we consider the vertical cliffs of Lau as the original steep submarine
slopes rather than due to sub-aerial and beach erosion subsequent to their
upheaval, then it seems reasonable to accept a much smaller loss to the
islands by erosion than a first glance would appear to justify. Some of
the islands are known to consist of limestone masses topped by sub-
horizontal caps. The terraces, again, of the older cliffs appear from
certain points as perfect level plains (Thithia, Tuvutha, Yathata), and
the truncated cone structures of some island masses suggests the contour
of the original reef.

If we admit the persistence in the general level of such raised plat-
forms as Tuvutha, Vatu Vara, and Yathata, and of the original slope of
the raised reefs, then the amount of loss by erosion seems confined to
such pits and cracks as are seen on the summit of Vatu Vara, which

1 A. Agassiz, I. c, pp. 58, 78.

ANDREWS: LIMESTONES OF THE FIJI ISLANDS. 29

although in places 30 and 40 feet deep, have not altered materially the
original contour of the summit.1

Observations taken in the Windward Isles of the Group show a
remarkable similarity in this slope. Many of the angles of inclination
to the horiton of the limestone bluffs averaged 50° to 55°, with a final
100 feet of steeper slope. This top part might go even higher than 75°.
The truncated " sugar-loafs " are a topographical feature perfectly unmis-
takable (Plate 27).

On examining these slopes closely, they are found to consist of black-
ened and weathered needle-like formations. The minute ridge-structure
mentioned before is developed to such a degree as to render climbing up
the steep faces dangerous, owing to their being so worn away as to yield
to the slightest touch in places.2 Struck with a hammer, these eroded
blocks, of perhaps five hundredweight, and attached to the rock mass by
a column only one or two inches in diameter, ring like bells, and a very
slight blow is often quite sufficient to send them crashing down the hill
to the debris heap below. Notwithstanding this weathering, the re-
mainder still marvellously preserves the original contour of the reef as
it rose from the waters.3

At the base of these escarpments great talus heaps accumulate. These
were formed, apparently, in many cases, under submarine conditions.
The base assumes a less highly pitched angle of slope than the upper
development. Furthermore, denuded of its debris covering, the basal
rock, projecting here and there in rocky ribs, possesses a less slope than
the upper part.

Again, the angle of inclination may exceed 80°. Then the cliff varies
altogether. It is no longer black, but more commonly assumes a yellow-
ish-white or brown color, and the face is encrusted with secondary cal-
careous material. This veiling of the cliff face hides otherwise easily
determinable characteristics of the limestone. Some few of these per-
pendicular exposures are most likely due to rupture subsequent to
upheaval. But this is exceptional. At the north of Mango are three
cliffs (Plate 9) close together, 300 feet above the talus, and so conspic-
uous as to constitute landmarks far out to sea. At the base of one of

1 For a statement of the opposite view see A. Agassiz. /. c, p. 74, and Am. Jour.
Sci., Jan., 1900, p. 42 ; ib. Feb., p. 110; ib. March, p. 196; ib. May, p. 372.

2 See A. Agassiz, I. c, Plates 97, 98.

3 I think Mr. Andrews is mistaken. The immense amount of denudation which
has taken place in the Fiji Islands and other elevated reef islands in the Pacific
I have examined shows that the original reef slope is not preserved to any extent-
It has usually been greatly obliterated by denudation and erosion. — A. Agassiz.

30 bulletin: museum of comparative zoology.

these lies an immense block 100 feet long and some 30 feet thick. This
is a recent fall, while all around and below it are thousands of blocks
with isolated specimens as much as 100 tons weight. At Thithia there
is an exposure of limestone preserving a beautiful slope of from 50° to
60° for a couple of miles. In one place, however, this continuity is
broken. A huge block of some 20,000 tons (150 X 50 X 60 feet) has
slipped down slightly, and should it fall over would expose a cliff over
100 feet in height. The precipices occupying the upper 300 feet of
Vatu Vara may have been formed similarly.

Perhaps, however, the majority of cliffs represent originally steep
submarine slopes. Again, on the western side of Vatu Leile the cliffs,
100 feet in height, persist for miles and are quite vertical (Plates
28, 30, 31). Yet on these cliffs are to be found no less than four
well-defined lines of old beach-erosion. Therefore these cliffs must
have been submarine. There has, however, been a certain period of
quiescence since the last uplift, for a modern fringing reef has formed
outside of the cliffs. The next uplift would give a terrace structure
to Vatu Leile. Similar evidence is given from Yathata and Vatu
Vara (Plates 25, 27). In many cases there are no outliers indicating
an erosion so long continued as to eat back the " sugar-loaves " into
enormous 400 feet cliffs and to have got rid of the vast accumula-
tions of debris that would result from such a dismantlement of the
gigantic vertical exposures. For it must not be forgotten that many
high cliffs gather comparatively little limestone debris at their feet.
In some instances the cliffs abut directly onto a sand flat with no
intervening talus to lessen the angle of ascent.

The Nature, Origin, and Age of the Elevated Limestones.

Tuvuthd (Plate 3, Figs. 3, 4) furnishes positive evidence of four
elevations as terraces, occurring every 200 feet. Yathata is equally
characterized by its terraced appearance. Vatu Vara presents steep
acclivities and very poorly developed terraces. Vatu Leile empha-
sizes the short and frequent intermittent uplift, while Mango, Kam-
bara, Mba Vatu point to long-protracted periods of stable equilibrium.

Of all the islands in the Lau Group, none perhaps are more remark-
able than Tuvutha,, Vatu Vara, and Yathata. They are closely allied
in structure, as regards height and traces of elevations. Tuvutha is,
perhaps, the most remarkable of all. Its summit is represented by a
curious peak (Plate 3, Figs. 3, 4), 800 feet above the sea, and rising 150

ANDREWS: LIMESTONES OF THE FIJI ISLANDS. 31

feet above the rest of the island. This mass consists of a compact
coralline limestone, fairly horizontal, though eroded into pits and cracks.
This cap is of several acres' extent only. The sides are steep. On the
sea-side it presents a steep declivity to the water. It rests on the rim
of the central hollow some two miles in diameter. This, the second
" platform," is merely a thin coralline rim dipping sharply towards an in-
land crateriform cavity, some 250 feet below the level of the "platform."
The floor consists of weathered limestone, and is densely wooded. At
one end a gap occurs through which the material of the once continuous
platform appears to have been passed out in solution. Below this again
is another platform or terrace, also well marked (Plate 3, Figs. 3, 4).

Vatu Vara has been described before. It is mentioned here again
merely to emphasize its height (Plate 25), its terraces and sea-erosion
at the 800 feet, 600 to 700 feet, 350 feet, 25 and 15 feet levels.

Yathata is similar to Vatu Vara, especially in its sub-level cap (Plate
27) of several acres only and its well-defined steps or terraces here.
Another interesting point at Yathata is the comparative thinness of the
coral rock.

Mba Vatu and Ngillangillah appear from the deck of a boat to consist
of huge coralline masses either as coral reef or reef debris (Plate 35).
This would give them a thickness of coral growth amounting to at
least 500 feet. By searching, however, an exposure was discovered of a
beautifully bedded limestone underlying the coralline rock. These strata
consisted of compact brown and yellow stone, apparently non-fossiliferous,
and reaching up through the mass to a height of nearly 300 feet. This
may occur as a base for the whole mass of Mba Vatu, but the coral
growths have seized upon it as it approached the surface, and terraces
have spread out from it, thus hiding it from view. The cliff of bedded
limestone stands out boldly amidst its coralline surroundings, and appears
to owe its exposure to-day to some past cliff rupture whereby the
encrusting coral growths were separated from the stratified rock. Three
miles away, and on the other side of the island, similar beds occur, but
now they are but 50 feet in height and have very little coral growth
attached to them. So that on this block 500 feet in height there is
at most not more than 200 feet of coral growth.

Malatta. — To the south of Vanua Mbalavu, and 14 miles from Mba
Vatu, a small patch of limestone, about 100 feet in height, has escaped
the general disruption of the limestone by the volcanism. Here an
old line of beach erosion is found, tilted at 15° to the horizon, and
almost obscured by abundant redeposition of calcareous matter charged

32 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

with ferric oxide. A small hole exists near the present sea-level in
the limestone, and by crawling in one finds a small cave about 8 feet
long and 3 feet wide. The sides are composed of red volcanic clay
most distinctly bedded. Above this is found a reddish, hard lime-
stone, and a very short distance away the coral rock itself occurs. Hot
springs also exist throughout this limestone area.

Mango. — On the summit of a typically shaped andesite dome, and
carried to its present position by the eruption of this same lava mass,
a massive bedded limestone occurs (Plate 2, Fig. 1). It overlays a
greatly decomposed volcanic conglomerate, gray to greenish in color. A
number of gasteropods are scattered throughout the mass of the con-
glomerate. The stratified beds immediately above the decayed vol-
canic material are a dense, compact, reddish limestone, with fragments
of fossils in places. It is very similar to the basal limestone at Mba Vatu
and Malatta.

The bulk of this volcanic conglomerate cannot lie at a depth much
below sea-level, if, as seems probable, its thickness is comparable with
that of the outlier shown on the left hand of the section of Mango given
on Plate 2, Fig. 1, which has been pushed up into its present position
by an outburst of andesite.1 We can easily avoid the necessity for ex-
plaining away huge deposits of coral reef origin, since the overlying reef
material occupied but a fairly thin crust.

Tuvuthd is as instructive as Mango. An old volcanic conglomerate
almost identical in appearance with that seen on Mango occurs on the
sea beach. A newer andesite lava has broken through and poured over
it and also towards the limestone cliffs, leaving a gap between the two.

Geological History of the Lau Group and its relations to
the Western Fiji Islands.

It is now possible to suggest the genera] lines along which the present
features of Lau were developed. Calcareous deposits, which afterwards
formed the " bedded " limestones, were laid down on the floor of the
ocean, the submarine plain which later was to form the foundation of
the Lau Group. The existence of a basal volcanic conglomerate below
the reef rock at Mango and Tuvutha shows that there succeeded a period

1 This does not imply that compact basal limestones project above the floor of
the central hollow, but that the undisturbed and the disrupted rock are similar in
their upper portions.

ANDREWS : LIMESTONES OF THE FIJI ISLANDS. 33

of volcanism, probably in Tertiary time, during which masses of volcanic
material were heaped up along a fairly well-marked north and south line
of weakness. The highest of these masses of volcanic rock, such as
Yathata, Tuvutha, and Vatu Vara, came within the reef-building zone
of corals, and reefs commenced to form upon them. Alternating epochs
of upheaval and stable equilibrium followed, during which the reefs grew
outwards over calcareous banks as well as the bedded limestone. This
is demonstrated by the long lines of erosion cut into the clitfs of Lau
and the sub-horizontal terraces above them. This state of unrest must
have continued, for a long period of time, to have produced the vast
masses of limestone like those of Mango and Tuvutha.

Another volcanic phase occurred in recent time. The andesite out-
bursts, which disrupted the old reef rock and bedded limestones, formed
the coarse agglomerates with large included lumps of coral, and capped
the latter with domes of lava. Many of the limestone islands became
centres of eruption. At first numerous explosion craters were formed,
and blocks of andesite and reef limestone were hurled from the vents
until large islands were piled up by the ejectamenta. The former island
of Vanua Mbalavu was almost completely wrecked as a result of the
volcanic explosions.

Another instructive feature in these outbursts is the fact that the
bases of the great limestone cliffs became in many places the centres of
disturbance. If the eruption was extensive, then the cliff vanished and
a heap of calcareous agglomerate was all that was left to attest to its
former existence. If the disturbance was smaller, then the volcanic
matter welled up from the cliff base without wholesale fracturing of the
limestone.

Subsequent to, or pei'haps immediately after the first paroxysmal out-
bursts, the andesites welled up in dome-shaped form and buried the coral
agglomerates, produced by the earlier explosions, as well as the surround-
ing limestone. A final phase in the eruptions is marked by the later
eruptions of olivine basalts. These found vents at the bases of the older
andesite domes, and at Mango are represented by small tongue-shaped
flows and minor masses.

From these fragmentary traces of basal rocks and from these indubi-
table signs of elevation, it is possible to understand the mode of formation
of the Lau limestones and their subsequent elevation. Along the sub-
marine plateau and running a little west and east of a meridional line,
volcanic masses were erupted along the axis of regional elevation ; and
on these as a base great accumulations of volcanic ash and conglomerates

VOL. XXXVIII. — NO. 1. 3

34 bulletin: museum of comparative zoology.

were deposited, mixed with accretions derived from the tests of animals
provided with calcareous shells ; or again great stratified limestone de-
posits were found which are comparable to the compact limestone strata
of the Singatoka area (Plate 3, Figs. 1 and 2), and which, like them,
may be non-fossiliferous, while others again may contain scattered corals
and shells ; and that these bedded rocks came within the reef-building
zone, either by uplifts like those traced to-day on the Vatu Leile and
Vatu Vara cliffs ; or by their own increase in thickness during a period
of stable equilibrium ; or again by the joint results of elevation and
accretion. Much of the limestone mass shown in the sections of Mango
and Tuvutha,1 (Plates 2 and 3, Figs. 3 and 4) is doubtless (in its
lower part) composed of compact bedded limestone.

After the first period of elevation succeeded periods of rest, during
which such masses as the second "terrace" of Yathata and Tuvutha
were formed. While Yathata, Tuvutha, and other islands show broad
flat land at this stage, Vatu Vara shows but an incipient " terrace "
structure. This might be accounted for if Vatu Vara was building on a
very steep slope. In this case the corals could advance but slowly on
their own talus, while if the other islands had broad fiat patches of land
to build on, or craters, they could easily assume their present great exten-
sion at this level.

From the next point of upheaval (the 400 feet level) to the present
sea-level, the evidence points to smaller upheavals, with intervening
periods of protracted rest. One of these periods of rest was of much
longer duration than the others, and marks the 250 to 300 feet level.

The Metamorphism of the Limestones.

The limestones near Suva are generally interstratified with the " Fiji
soapstone," and this is especially noticeable in the wedge-shaped mass at
Walu Bay. This superimposing of the " soapstones" upon the calcareous
rocks has suggested 2 the possibility of the induration of the limestones
of Lau by exposure and water soakage, and that the immunity of the
Suva limestones from this metamorphism is due to the impervious char-
acter of the " soapstone " cap. Even a cursory examination of the Singa-
toka River area, where limestone occurs in all varieties of hardness and
compactness, shows this not to be the case. There we have an escarpment

1 The volcanic peak drawn as the base of the summit of Tuvutha, Plate 3,
Figs. 3, 4, is merely provisional, as there is no direct evidence of its being such.

- J. Stanley Gardiner. The Coral Reefs of Funafuti, Rotuma, and Fiji. I 'roc.
of the Camb. Thilos. Societv. 1898.

ANDREWS: LIMESTONES OF THE FIJI ISLANDS. 35

(Plate 7) more than a mile in length and exceeding a height of 300 feet,
the rock is stratified, and dips at a uniform angle of 15°, so that as the
surface of the ground is nearly level, each individual bed crops out at
the surface. One series of beds comprise layers of loosely compacted
calcareous sand ; others again are as hard as marble. The rocks are
very similar to those at Suva. At the Singatoka, however, there is no
trace of soapstone nearer than Thuvu, several miles away. There is no
non-porous capping to the Singatoka bedded limestones.

The induration, therefore, of the raised reef limestones appears to have
been quite independent of the presence of an impervious overlying rock
such as "soapstone."

In the Lau area, the metamorphism of soft coral into hard dense rock
is emphasized as strongly at the sea-level as on the island summits. The
recently elevated Vatu Leile reef-flat, the latter some 6 feet above high-
water mark, although strewn with hard corals, is hard and flinty as any
limestone observable anywhere. This statement must not be considered
to have universal application. This hardening extends only to the
greater part of the reef-flat, — that part composed only of reef-debris, —
and not the real coral masses representing a former belt or patch of
growing reef. The raised coral masses found in the lower cliffs of Mango,
Ngillangillah, Kambara, and Vatu Leile, although possessing an exceed-
ingly hard shell ringing and emitting sparks under the hammer, were
nevertheless soft and cellular within. So soft is the portion beneath this
hard exterior that individual corals may be crushed beneath the fingers.
At the same level, however, rock not formed of a coral mass is generally
solid throughout.

This can be seen in a less degree on the " Liku " or cliff coast of the
Tongan Group. There whole cliffs of crumbling material, similar to those
observed at Singatoka (Plate 38), are exposed to the rollers of the Pa-
cific. All along the coast of Vavau great masses of rock are perpetually
falling and revealing a whole harvest of corals and shells more or less
indurated or changed to translucent calcite, and lying in a soft white
matrix. But where the cliff has preserved its integrity and still shows
the original slope of the reef seawards (Plate 36), there a hard crust of
limestone protects the soft interior. This is also observable on the fallen
masses, the resulting debris of the soft coral cliff.

On the raised limestones of the Thuvu-Singatoka area a reef has
grown, which has since been elevated. While the greater part of this
reef mass is dense and hard as the Lau stone, the older basal rocks are
still soft as those of the Walu Bay area.

36 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

The rock of the second, third, and higher elevations has generally
undergone alteration (induration) to a depth of which the limit cannot
be ascertained by blasting, and this induration ha3 been more extensive
than that which has affected the lower and newer limestones, and has
spread through the coral reefs proper as well as through their associated
coral debris. At the 300 feet level on the sides of Ngillangillah, an ele-
vated reef mass 50 feet thick and consisting of corals, such as Porites and
Montipora, shows that the pores of the corals have been almost wholly
obliterated through solution and secoudary addition of lime. At the
higher levels on Kambara, the coral structures have changed to calcite
in many places, and lie in a crumbling calcareous base. But in most
cases all coral structure has vanished, the spaces formerly occupied by
the corals being represented by dome-shaped cavities (Plate 37).

At Vatu Vara a white dolomitic rock is filling the pores of the corals,
and they are gradually disappearing. This dolomitization is seen at
different levels in the Lau islands. Thus at Mango, 240 feet above
high-water mark, numerous patches of white rock are found which
hardly effervesce with acid. The dolomitization observable on the
slopes of Vatu Vara is most pronounced, and its mode of occurrence
suggests that it took place from the cause above mentioned during
long pauses in the elevating movement ; but the apparent absence of
such dolomite patches in the present lines of beach erosion is opposed
to the supposition of the coincidence of dolomite belts with old beach
lines, as the same dolomite may have resulted from the absorption by
the limestone of magnesium salts from sea-spray.

Another factor in rock alteration is the development of calcite along
the higher levels. This may take the form of banded, colored calcite,
filling large gaps, which is so common a feature in the older rocks of
Mango, or it may take the form of transparent rhombohedra, about one
inch in diameter. On the old, undisturbed raised reef flats, numerous
patches of this calc-spar formation may be noticed. They are very com-
mon on the sub-level summit of Mba Vatu. They occur at Mango less
frequently, while in Naitamba we find the finest calcite crystals ; they
frequently exhibit a radial disposition, and occur in numerous patches,
and are almost invariably connected with old reef flats or lagoon areas.

From the foregoing considerations, it seems reasonable to infer that
the agents which produced the hardened shell of the lower limestone
cliffs of Lau and those which formed the secondary ferruginous calcareous
rock of the beach-erosion levels and compacted the " beach rock," are
similar if not identical ; and these, together with the more pronounced

ANDREWS: LIMESTONES OF THE FIJI ISLANDS. 37

dolomitization at the bases of the old terraces, appear to make up the
sum of the more conspicuous metamorphic phenomena at work in the
limestones of Lau.

The Volcanism of Fiji with reference to Lau and the
Smaller Islands-

The erupted material of Viti Levu appears to consist of andesites,
porphyrites, and andesite tuffs. At Suva great beds of volcanic boul-
ders occur beneath the i*eef limestone, intercalated in the " soapstone "
at Walu Bay. Some of the rounded blocks are upwards of 15 inches in
diameter, and are surrounded, in individual cases, by a thick layer of
soft-brown earthy crust, the result of the rock decomposition. They
consist of porphyrite, quartz, and felspar porphyry, andesite dolerites,
and andesites. As the base of the quartz and felspar porphyry is
minutely sphenelitic, the rock probably represents an old rhyolite flow.

In the smaller islands the prevailing rock-types are andesites, occur-
ring as lava and agglomerate and basalts, the latter occupying a much
smaller area than the former. The influence of the volcanism on island-
making will be appreciated when it is stated that Taviuni, Kandavu,
Ovalau, Ngau, Nairai, Totoya, Moala, and Matuku are wholly volcanic,
while islands which to a casual observer seem wholly limestone, such as
Mango, Thithia, Vanua Mbalavu, and Naitamba, are two thirds volcanic,
only the remaining third being limestone. Vanua Mbalavu affords
splendid examples of andesite agglomerates and conglomerates covered
with andesite lava, and exhibits numerous small outlying islands formed
of audesite tuff's covered with lava and pierced with dykes of the same
material.

Ovalau, Munia, and Vanua Mbalavu are, in great part, composed of
huge, uniformly dipping strata of andesite blocks (Plate 6). These
blocks are angular, and present stages of coarseness varying from the
finest grained rock with a few scattered pyroxene and plagioclase crys-
tals to blocks consisting of huge idiomorphic crystals of augite and
felspar crowded in a fine base.

Insular masses, such as Ovalau and Munia, composed of andesite
agglomerates, present very high angles of slope to the sea. In fact, the
coastal scarps can hardly be described as slopes. They are very steep
ascents broken by precipices 400 or 500 feet, sheer in places. The soft
ash beds throughout the mass allow weathering into gorges and cliffs to
take place very rapidly. The slopes of andesite flows of Lau (Plates

'SH bulletin: museum of compakative zoology.

23, 27) and islands lying in the neighborhood of the Koro Sea present
very steep angles.1 This angle in places exceeds 30° or even 35°.
The slope is generally littered with huge spheroids of andesite. The
basalt forms a much more gentle slope, as may be seen by a traverse
of Taviuni. Usually the basalt is in very inconsiderable quantity and
appears to be of younger date than the andesite.

The Recent Age of the Volcanic Phenomena.

There appear to have been two distinct flows, one of andesite, the
other of basalt. Both appear to be very recent, even subsequent to the
last upheaval at Lau. Wherever volcanic rock is found in the smaller
islands of Fiji, whether that island be wholly volcanic or part volcanic
and part limestone, the eruptive rock is disposed of in gradually sloping
hills, with heaps of lava blocks scattered indiscriminately, or takes the
form of streams of volcanic stones littering the beaches.

At Mango, Thithia (Plates 15: 16, 19), Vanua Mbalavu, and Yathata,
cliffs of limestone exist forming inliers in flows of andesite lava. It
would seem that a great stream of andesite descended from higher levels,
parted near what is the inlier, and then swept over the cliff on either
side, and became confluent again at its base, where the lava was by de-
grees piled so high as to be nearly on a level with the top of the cliff,
and so produced a steep embankment sloping in one direction back
towards the base of the cliff inliei-, and in the other direction towards
the general trend of the lava flow. The contour of the lava bank as it
lies under the cliff suggests that of a viscous mass flowing round a large
obstacle and grown stiff before it had completely enveloped the
obstruction.

In Mango huge sections of strata may be seen exposed in gullies that
have cut through the lower parts of the hill slopes (Plate 2, Figs.
2, 3). These strata have a decided dip, and consist of andesite tuffs
with large angular blocks of limestone scattered through the rock mass.
These blocks are as much as a ton in weight, and consist of coralline
rock indistinguishable from the present reef mass. A similar phenomenon
occurs in the 9-fathom lagoon to the south of Mango, where the tuff beds
are similar to those before described (vide map of Mango, Plate l). The
volcanic agglomerate is therefore clearly newer than the raised reef lime-
stone, and is in turn capped in places by a flow of andesite lava.

The shape of the island of Vanua Mbalavu has been determined by a

i See A. Agassiz, I. c, Plates 34, 46, 48, 50, 57, 58.

ANDREWS: LIMESTONES OF THE FIJI ISLANDS. 39

line of volcanic vents, while its north and south extremities consist of
lofty limestone escarpments. Silicihed corals lie on the summits of the
razorbacks and higher slopes of Vanua Mbalavu, and are evidently
examples of silicihcation by contact metamorphism through percola-
tion into reef coral of superheated water derived from the volcanic
rocks.

At Mango and Thithia similar silicihcation may be noticed.

On Mango, a spring draining the north side of a volcanic mass 600
feet in height suddenly disappears in an enormous mass of limestone.
This point of disappearance is at the 50 feet level above high-water
mark, but almost immediately above this another andesite mass rises
seawards to a height of 150 feet above the sea in such a position that it
is interposed between the point of disappearance of the stream and the
sea-shore. The distance thence to the coast is 500 yards ; the stream,
which is of considerable volume, reappears on the beach 600 yards away.
It is far mere probable that the water worked its way through a lime-
stone substratum than that it percolated through the more or less imper-
vious andesite. If this is so, the volcanic rock must here be above and
newer than the limestone.

At the south of Vanua Mbalavu a hot spring bubbles up through the
limestone near the tidal zone. The temperature, though below boiling-
point, is high enough to scald the skin. At times the water in the ad-
jacent lagoon becomes heated over a small area. About 30 yards distant
from the first spring, another hot spring occurs in the midst of a lime-
stone fissure.

Both these springs are in close proximity to the junction line between
what, in my opinion, is one of the latest intrusions of andesite and the
old reef rock.

At Naitamba, Tuvutha, and Kambara 1 the old lagoon floors are still
intact.2 Mango and Thithia were doubtless once possessed of similar
lagoon hollows prior to the great covering of volcanic alluvium which
now hides the old limestone floor. Consequently, if this reasoning is
right, these lagoons antedate the chief volcanic outbursts.

Calcareous deposits obscure everything along the coasts of Lau with
the exception of the volcanic rock, whether in dome-shape, dyke-form,
or as tuffs. The enormous rolling hills of andesitic lava that have spread
out from the cliffs have no trace of limestone or calcareous material at-
tached to their sides, except at Kambara, where whitened, burnt blocks

1 See A. Agassiz, /. c, p. 74.

2 At Kambara certainly not. — A Agassi*.

40 bulletin: museum of comparative zoology.

of limestone litter its rounded volcanic hill ; these, however, represent
unmistakably disrupted cliff masses, and are the broken fragments
resulting from the volcanic outbursts.

The Volcanic Rocks.

A good collection of volcanic rocks of the more common types from
the Fiji area was secured, and a brief summary of their macroscopic
and microscopic characteristics is given below.

As regards the general macroscopic features of the Fiji andesites, they
may be described as compact, reddish, or greenish rocks weathering into
spheroidal blocks, and showing dark green augite and plagioclase pheno-
crysts in a fine-grained base. The development of the large augite
prisms is very pronounced.

In section, two generations of plagioclase are distinguishable : first, a
crop of large idiomorphs with corroded edges, and showing well-developed
concentric zoning. These also show albite twinning, and their general
outline is tabular, although often they occur in nests or bunches, one
crystal being partly or wholly enclosed by others. Next, a crop of small
lath-shaped sections. The " felt-like aggregate of feldspar microlites " so
characteristic of andesites is also well marked. The augites constitute
an important factor in the mass. They are usually represented as
monoclinic prisms, yielding octagonal sections, and showing concentric
zoning, and twinning parallel to the orthopinacoid in a marked degree.
Rhombic pyroxene polarizing in high colors and almost indistinguishable
from olivine in thick sections is also common. The augite is often
replaced by viridite. The following are some detailed descriptions of
some of the thin sections I prepared of the Fiji volcanic rocks : —

No. 1. Levuka. Specific gravity, 2.9. A compact augite-andesite.
The augites are in the form of phenocrysts and yield fine octagonal sec-
tions. Concentric zonal structure and twinning parallel to the ortho-
pinacoid is common. Two generations of plagioclase, the older and
larger showing kaolinization and concentric zoning, the newer crop con-
sisting of lath-shaped sections. The augite has been replaced by viridite
in places, and much magnetite is scattered throughout a felt-like base.
This is a fragment from a lava agglomerate.

2. Mango. Specific gravit}T, 2.57. Taken from a lava flow. An
extremely dense andesite with idiomorphs representing the first crop of
plagioclase. The "andesite" ground mass is well developed, and large
particles of magnetite occur in a great development of magnetite specks.

ANDREWS : LIMESTONES OF THE FIJI ISLANDS. 41

3. Munia. Specific gravity, 2.70. Taken from an andesite agglomer-
ate. Vesicular to cellular lava. Phenocrysts of plagioclase twinned on
albite type. Shows the " felt-like base." The felspar is much decom-
posed and the magnetite is well developed.

4. Totoya. Specific gravity, 2.7. Sample from lava flow. This is
an andesite, much decomposed. Rock generally green, owing to decom-
position of augitic ground-mass into viridite. The felspar of the first
generation shows advanced kaolinization, and great abundance of lath-
shaped crystals of felspar.

5. Munia. Specific gravity, 2.55. Specimen from lava agglomerate.
A cellular andesite lava. Iu section, shows a few pj-roxenes polarizing
in high colors, also plagioclase idiomorphs in a fine " andesite " base.
Zonal structure is found in the felspars and spherulitic structures are
common.

6. Taviuni. Specific gravity, 2.66. Vesicular lava. Described more
fully later. A fine-grained black rock. Yields in section phenocrysts
of plagioclase set in a base of closely packed lath-shaped felspars and
the " felt-like " material common in andesites. Magnetite is common,
and also olivine or pyroxene.

7. Singatoka River. Specific gravity, 2.71. Porphyrite in agglomer-
ate form. Large greenish plagioclase crystals in brown base contain-
ing second growth of minute felspar crystals.

8. Levuka. Specific gravity, 2.8. Reddish andesite with porphyritic
development of green augite showing perfect idiomorphic contours.
Felspars of first generation much decomposed and corroded in the
andesite magma. This felspar occurs in nests. Zonal structure well
observed in felspars. Spherulites present as well as much magnetite.
Abundance of pyroxene decomposition products.

9. Totoya. Specific gravity, 2.7. A greenish augite-andesite. Similar
to Xo. 4, but not so decomposed. Shows well-developed plagioclase
crystals in fine base with scattered magnetite as fairly large grains or
well-distributed specks. Second crop of lath-shaped felspar crystals.
Augites and pyroxene phenocrysts, the former showing zonal structure
at times. Fair examples of spherulites.

10. Mba Vatu. Specific gravity, 2.85. Selected from tongue of
compact reddish andesite lava, with glistening felspar crystals. Shows
in section well-defined plagioclase crystals and phenocrysts of augite
with the usual accompaniment of lath-shaped felspars. This rock
exhibits the " felt-like base structure " with great development of
magnetite specks. Twinned augites occur.

42 BULLETIN: MUSEUM OE COMPARATIVE ZOOLOGY.

11. Levuka. Specific gravity, 2.8. Very compact andesite agglom-
erate, almost identical with (8).

Mr. W. E. Wooluough, B. Sc. of Sydney University, has kindly
examined and described No. 12.

Xo. 12. Suva. Specific gravity, 2.84. Olivine dolerite.

Plagioclase by far the most abundant constituent of the rock, making
it almost andesitic. They take the form of large tabular crystals with
abundant concentric-zoned inclusions, mainly magnetite and a little
apatite. The zoning is very perfect. Augite is present in large yellow
idiomorphic crystals showing twinning and traces of ophitic structure.

The olivine present is a highly ferriferous variety, some of the crystals
being blood-red. Most of the olivine is in the form of large grains, some
show distinct outlines. Some of it is almost undecomposed, while some
is almost opaque with secondary serpentine and iron-oxides.

The magnetite recurs in scattered grains, and is distributed abundantly
throughout the ground mass, which is crypto-crystalline to glassy in
texture. In one cavity there is a mass of small spherulites of a zeolite.
These give under crossed nicols a dark cross at about 30° with the prin-
cipal planes of the nicols.

13. Mango. Specific gravity, 2.75. Selected from compact andesite
flow. Shows the usual structure and appearance of plagioclase. The
pyroxenes are altered ; some show beautiful 8-sided crystals. Much
magnetite present.

14. Taviuni. Specific gravity, 2.70. Selected from vesicular andesite
Under microscope shows two crops of felspar in glassy base. Great
development of magnetite.

15. Taviuni. Specific gravity, 2.72. Vesicular olivine basalt. The
olivines are quite fresh. Crop of lath-shaped felspars.

16. Mania. Specific gravity, 2.70. Taken from an andesite agglom-
erate. Dark-colored rock showing usual two crops of plagioclase in fine
base. Pyroxenes are altered.

The following three rock slides were described by Mr. W. E. Wool-
nough, B. Sc, of Sydney University.

17. Loma Loma. Specific gravity, 2.92. This appears to be a coarse
dolerite with phenocrysts of striated felspar, and apparently greenish
pseudomorphs after olivine.

18. Sicva. Specific gravity, 2.57. A rhyolite.

Groundmass is crypto-crystalline, containing abundant grains of
limonite, also containing very abundant small spherulites which are
quite free from limonite.

ANDREWS: LIMESTONES OF THE FIJI ISLANDS. 43

The phenocrysta are quartz orthoclase, plagioclase, and a little decom-
posed biotite. They are all more or less idiomorphic aud are more or
less corroded by the magma. They all contain inclusions which are
gaseous, glassy, or lithoid. Apparently there are no liquid inclusions.
The felspars are all clear and glassy and are distinguished from the
quartz and from each other by the nature of the twinning. The in-
clusions are more or less symmetrically arranged.

There are three or four decomposed basal sections of biotite. These
contain grains of magnetite.

19. Mango. Specific gravity, 2.75. Olivine anamesite.

The plagioclase is of two generations, — the older in tabular, cleaved,
and corroded crystals, showing zoning in places, and the newer crop is
in the form of acicular microlites felted together. The olivine is very
abundant and in places shows large idiomorphic grains. Only slightly
serpentized. The augite is scattered through the rock in almost color-
less granular aggregates. Magnetite occurs as irregular grains in great
abundance. Apatite is also present in small stumpy prisms, especially
as inclusions in the older felspars. The groundmass of the rock is
crypto-crystalline and colorless.

Professor David of Sydney University also describes a section from
Mango.

20. Mango. Olivine basalt.

This olivine basalt taken from the beach at Mango has in thin section
a very fresh appearance. It is formed of irregular granules of olivine
fairly free from decomposition, though slightly serpentinized at the edges ;
clear phenocrysta of plagioclase and smaller felted crystals with small
augites and numerous magnetites in a crypto-crystalline base.

The magnetites are all bounded by sharp edges, and are remarkably
free from traces of decomposition.

The whole rock has a decidedly fresh appearance.

The Limestones of Fiji-
Viti Levu.

The Older Limestones of the Main Island. — This is a compact limestone
of bluish-yellow color, and possessing a dip of 50°. It appears to be
connected with the dolomites of the upper river. It will be remembered
that this rock underlies the shelly limestone of the Thuvu-Singatoka
area.

44 bulletin: museum of comparative zoology.

The Dolomites. — A rock of an extremely compact nature and homoge-
neous in texture. Generally it is white or yellowish brown in color. It
is apparently non-fossiliferous. It weathers to that bluish color on the
escarpment faces so admirably seen in many Silurian limestones. The
lillipiitian mountain-range structure is well exemplified on weathered
surfaces. — Specific gravity, 2.71.

The Shelly Limestone of the Main Island. — In patches it is white in
color, but generally preserves a tint varying from light yellow to rich
brown or red. In the softer sandy beds there is no fracture, as the rock
crumbles beneath the finger, but in the sparely fossiliferous beds the
fracture is flat. This freestone exhibits a wonderful evenness of texture,
consisting of small calcareous sand particles compacted into solid rock.
In other beds, again, although area for area there is a general uniformity
of appearance, still in any small individual block there are vast deviations
from any common standard of homogeneity that may be set up. Con-
cretions occur which, when struck with the hammer, emit spai'ks, and
are exceedingly difficult to break, while round them may be clustered
roulettes of foramens lying in loosely cemented sand. These petri-
factions have conchoidal fractures, and are exceedingly compact and
minute in structure. A cliff face may present the curious effect of a
soft exposure, dotted over with hard sub-cylindrical protuberances stand-
ing out five or six inches from the vertical face. All the beds are
extremely porous. A small patch of coral reef has been sandwiched
in between a bed of greasy clay and a fine-grained limestone. From
a short distance the cliffs with their uniform and persistent lines of
stratification, their warm brown and yellow tints, and their "rock-
shelter" development, have striking similarity to some of the Triassic
arenaceous sandstones of the. Sydney district, but the age of the Fiji
rocks is of course very different, being probably Tertiary.

What has just been said concerning the Thuvu-Singatoka area is
true also for the calcareous beds near Suva. Here also the rock,
considered in belts, is fairly homogeneous.

Calcareous Rock <i„<l Coral Rock passing into Soapslone. — On the west
side of Suva harbor beds occur that pass gradually from soapstone to a
sandy limestone rock. There is no sharp line of demarcation. The one
shades insensibly into the other. Exteriorly it often pi-esents a rough
and hard appearance, but breaks down under the hammer into a sandy
clay rock. At various times large serrated Carcharadon teeth, carapaces
and plastra of turtles have been unearthed from the Suva stone. The
rock is rarely cavernous and contains few fossils.

ANDEEWS: LIMESTONES OF THE FIJI ISLANDS. 45

The Lau Limestones.

(a) A soft, porous stone, full of corals in situ, shells, and a little cal-
careous cement. The corals are no more compacted than one would
expect to find in the reef area. The stone is usually white or yellow in
color. A thin crust of hard material has gathered over the soft inte-
rior. This crust, as well as the reef or coral talus itself, is extremely
cavernous. This represents the raised coral material of the last
upheaval.

(6) A dense white rock, ringing under the hammer, generally exhibit-
ing little or no trace of well-defined fossils. This is the raised limestone
found near sea-level ; it is contemporaneous with (a), and represents the
part of the "raised reef" platforms that is composed of the firmly
cemented reef debris.

(c) A compact stone ringing under the hammer containing corals.
The color is white or yellow. It is pitted and cavernous, and weathers
into long needles. This forms the roughest and most broken country in
Fiji. It is the older coral mass of the high levels.

(d) A rock similar in hardness, but presenting a marked difference in
weathering, due to the difference of erosive influences on a homoge-
neous rock and one varying considerably in minute structui'e. The
coral reef is honeycombed, while the rock under consideration shows
perfect examples of the minute-ridge structures. It is, as a rule, non-
fossiliferous and homogeneous in structure. Such rock is found at the
higher levels and represents the compact debris of pre-existing reefs.
This is well seen at Ngillangillah.

Of other formations we may note a stratified limestone, white, yellow,
and brown in color, exceedingly compact and free from honeycombing or
cavernous weathering. Very fine-grained rock, and apparently barren
of fossils. This is the basal limestone underlying the reefs to the north
of Vanua Mbalavu.

A brown to red ferruginous mass, found encrusting the coastal cliffs
at sea-levels. In the huge gaps cut out by the sea in the limestone cliffs
these ferruginous masses simulate stalactitic and stalagmitic growths.
Redeposition of calcareous matter charged with higher oxide of iron
than was present in the rock prior to solution, or from the magnetite
mentioned before, explains this stalagmitic growth. In some places this
redeposition has advanced so far that the former sea-ledges are partly or
wholly obscured.

46 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Manganese at Mango and Thithia.

At Mango a great deposit of manganese occurs at a low level (about
100 feet above high-water mark). The earth is decidedly black for a
<;reat distance from the surface, being as much as 30 feet thick in
places. Large pieces of manganese lie about the ground as scattered
blocks.

At Thithia smaller pieces occur. One I picked up was a sphere 1 inch
in diameter.

At Naitamba the blocks are of irregular shape, as at Mango.

ANDREWS : LIMESTONES OF THE FIJI ISLANDS. 47

EXPLANATION OF THE PLATES.

PLATE 1.
Geological map of the island of Mango.

PLATE 2.

Sections across Mango :

1. From Mt. George to Prince Edward's Island.

2. From Mt. Fairhall to Mt. Rupert.

3. Southeast from Mt. Rupert to limestone cliffs.

4. Section across Mango from northeast to southeast.

PLATE 3.

Sections across Singatoka, Tuvutha, and Mba Vatu :

1. Section across Singatoka River limestones.

2. Section across Singatoka River and sand dunes.

3. Outline of Tuvutha as seen from Mango.

4. Section of Tuvutha at right angles to Figure 3.

5. Section through Mba Vatu and Koro Mbasenga.

The position of the older volcanic rocks in Figs. 3, 4, of the sections of
Tuvutha has of course not been observed, and is merely diagrammatic.

PLATE 3*.
Singatoka River, 5 miles from mouth.

PLATE 4.
Horizontal soapstone strata, Thuvu.

PLATE 5.
Raised reef at Thuvu, 11 to 16 feet above high-water mark.

PLATE 6.
Stratified andesite agglomerate near Levuka, Ovalau.

PLATE 7.
Exposure of bedded limestone nearly a mile long, Singatoka Cliff.

PLATE 8.
Volcanic rocks, " The Naivaka," Taviuni.

48 BULLETIN: museum of comparative zoology.

PLATE 9.
Cliff of nearly 300 feet vertical exposure, northern part of Mango.

PLATE 10.
Slope of later andesite, limestone cliffs in the rear, Mango.

PLATE 11.

Old gap through which interior basin of Mango was once connected with the sea
25 feet above high-water mark, shows also dense vegetation growing on the
bare limestones.

PLATE 12.

Miniature lagoon on Mango, eroded probably during a period of rest at the 200-foot
level ; the old gap (Plate 11) is visible above the pier.

PLATE 13.

Undercut and weathered limestone blocks, Mango. Island of Yathata in the
distance.

PLATE 14.

Weathered block of limestone, Mango.

PLATE 15.

Central part of Mango from Mt. Rupert, shows the old gap leading from the inte-
rior basin to the sea and the recent andesite outbursts nearly overwhelming
the limestone ring.

PLATE 16.
Southwest part of the interior basin of Mango ; view taken from Mt. Ryder.
Lettering on diagram facing Plate 16 :

a. Mt. Rupert, a recent andesite dome.

b. Block of limestone pushed up by andesite c.

c. Andesite that has carried up b.

I. Patches of limestone left or not quite covered by lava or alluvium.
/«. Limestone agglomerate in volcanic tuff.

PLATE 17.

Central basin of Mango, taken from Mt. Ryder. This view is the continuation of
Plate 16. Joins the right of the plate.
Lettering on diagram facing Plate 17 :

a. Andesite hills GOO feet high, beyond this the limestone ring, almost buried

in its inland development.

b. Block of limestone partly buried by the flow from Mt. Ryder.

c. Limestone mass capped by coral reef 240-300 feet above high-water mark.

d. Volcanic slopes and alluvium downs ; the right-hand member of d shows

fan-shaped spread of lava as it welled round the cliff c.

ANDREWS : LIMESTONES OF THE FIJI ISLANDS. 49

PLATE 18.

Limestone cliffs at Thithia. On the left is shown the typical limestone cliff slope ;
in the central part is shown a stage in vertical cliff formation.

PLATE 19.

General andesite outburst on Thithia. The little tufts of vegetation on the left
mark the site of limestone.

PLATE 20.

Old lagoon outlet on Thithia.

PLATE 21.
Gap in raised limestone, Thithia.

PLATE 22.

Part of Vanua Mbalavu Lagoon, seen from the 600-foot level behind Loma Loma.
Yama Yama is the andesite cone in the middle distance with Munia over it
and Susui to the right.

PLATE 23.

Andesite slope, Naitamba.

PLATE 24.

Naitamba, looking over the central hollow towards the old lagoon outlet and
the broken cliff of the old limestone ring.

PLATE 25.
Vatu Vara (1100 feet high) from the beach.

PLATE 26.
Vatu Vara. Two lines of marine erosion 15 and 30 feet above high-water mark.

PLATE 27.

Sub-horizontal summit of Yathata, showing the sharp line of demarcation be-
tween the volcanic and limestone formation indicated by the marked difference
in the vegetation. The foreground is a steep andesite slope which has poured
over into the limestone in the middle ground. The lava flow is covered with
grass, the limestone with a thick forest.

PLATE 28.

Cliff on the western coast of Vatu Leile, showing the vertical cliff, from 100 to 110
feet in height, the long lines of former sea erosion, the " Great Walk," six feet
above high-water mark, and the present shelving beach.

50 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

PLATE 29.
Flat coast of east side of Vatu Leile.

PLATE 30.
Cliff on west side of Vatu Leile, showing five lines of marine erosion.

PLATE 31.

Vatu Leile cliff of west coast, showing the breaking away of the modern line of
marine erosion and that called the "Great Walk," six feet above high-water
mark.

PLATE 32.

Junction of raised reef and stratified limestones, Thuvu-Singatoka beach.

PLATE 33.

Stratified limestones, Thuvu-Singatoka beach.

PLATE 34.

Andesite butts, Lakemba.

PLATE 35.
Ngillangillah (520 feet) and the Bay of Islands.

PLATE 36.

Liku, or sea-coast of Vavau (Tonga Islands), showing on the left the disintegration
of a soft white cliff and on the right the original slope (? A. Ag.) of the cliff.

PLATE 37.

Portion of cliff, 60 feet above high-water mark, showing near the bottom of photo-
graph a large Astrasan, partly removed by solution, and near the top of the
picture, cavities left after the removal by solution of coral heads, Mango.

PLATE 38.
Soft cliff face, with hard protuberance, Singatoka River.

PLATE 39.
Upper part of Walu Bay quarry, Suva, showing indistinct lines of stratification.

LOUR INDEX

MANGO

'

m

Sections across MANGO.

,,. „f twmtonr

■'■- .

1 I

■ ■ ■

n/W /«yr lunrMsntr Utii

Section across MANGO from N E to 5

■ ■

I :

! I .,.■,.,,

■

Section across SlNGATOKE RIVER LIMESTONES

mi&jm Ira

■■ i
s Safbtmmbling strata.

fll fl'U bt/OllijiHif

■

Cliffs 90 to !O0' hi£ih

Section across SlNGATOK z R IVER LIMESTONES and SAND DUNES Section bent at A and B Distance 7 miles.

. ■■■/ OH/nrj totAek

tone f 1 Hard may/a// Umt*tone,al?n older blue bmfttiw | 1 Atulfsltr Q ] Sand (■>■>

Outline of TUVUTHA as seen from MANGO

■
■

... .■■.....■■

■

Section oFTuVUTHA taken at ri?ht angles to abov

J*rf/*E1S

Section through MBA VATU and KORO MBASENGA

, Mtvtitrti rrtf flat (ISO '<""".'
i Limestone ag,/lei,lrfuif -i. •■-.

63
%

a

<

mm

Andrews Fiji.

Plate 4 .

.

iLIOTVPE PRINTING CO.. BOSTON.

HORIZONTAL SOAPSTONE STRATA, THUVU

fc.

>v ^1 5

D
<

<
>

O

U
h

<

Id
O

o
o

<

w
h

Q

z

&7£

-

<

c

<

<

O

h
<
o
2

o

a
z

o
h

Andrews Fiji.

Platf. 8.

<DREWS PHOTO.

*S. HELIOTYPE PRINTING CO.. BOSTON

VOLCANIC ROCKS. TAVIUNI.

Andrews Fiji.

Plate i o.

ANDREWS PHOTO.

THE HELIOTYPE PRINTING CO. BOSTON.

ANDESITE SLOPE, MANGO

o
a
z
<

<
a

Q

O

r

■s

2

O

o

c
<

3
h
<

is
u

1-

—

k .A- *! ' - Ji

■.i * • ;l ; '• V

5 Mm . ■

w

'X.

o

o

X
W

Z
o

h

P

K
W

:-
<

w

m

ex

F<

o
<

>

>

-z
c

O
O
Z
<

Iz,

O

h

IT
<
0,

cn

a,

Fh

5!

^-

-TO±il

o

<

c

W

>
J3

i=3

>

PL,

c
<

<

DP

a:
h
2
W

u

Andrews Fiji.

Plate 18.

31

ANDREWS PHOTO.

THE HELIOTYPE PRINTING CO., BOSTC

LIMESTONES CLIFFS, THITHIA.

<

T3

C
<

33

h

33
r

b
w

S-

D

GQ

0

w

h

Ill

c
z

<

►J

—
c

h

3
h

h

w

o

z
o
o

<
J

G

o

D
>

<

<
DQ

<
D
Z
<
>

z

o
o
o
<

o

h
<

<
—

-

fc

c

<

O

J

O

x

<

PC

h

2
W

o

C

<

<
>

D

>

z
g

-/.
o

W

w
z

<

o

7":

Z

PL

T3

a
<

Andrews Fiji.

Plate 30.

ANDREWS PHOTO.

CLIFF: FIVE LINES OF EROSION, VATU LEILE

Andrews Fiji.

Plate

3i

ANDREWS PHOTO.

THE HEUOTYPE PRINTING CO.. BOSTON.

CLIFF, VATU LEILE.

X

'J

<

<

G
'f
<
O
Z

OQ

6
>

X
b

w

z

o

[I]

<
■J

<

21

<r
J

h

DQ
W

r

53
W
Q

Z

<

IT.
Q

z
<

w

h
O

<

G
Z

<

<

r3

z
<
j

5

z

_

J]

Q
Z

<
J

m

<

o
z
o
h

D
<

>
<
>

Andrews Fiji.

Plate 37.

CLIFF, MANGO

Andrews Fiji.

Plate 38.

*:«»

I?

SilS

C«c

ANDREWS PHOTO.

THE HELIOTYPE PRINTING CO . BOSTON.

SOFT CLIFF FACE SINATOKA RIVER.

DC

<

D
3

The following Publications of the Museum of Comparative Zoology
are in preparation : —

Reports on the Results of I)re<lging Operations in 1877, 1878, 1879, and 188<>, in charge of Alex-
ander Agassiz, by the U. S. Coast Survey Steamer " Blake," as follows: —

E. EHLERS. The Annelids of the " Blake."

O. HARTLAUB. The Coiiiatulse of the " Blake," with 15 Plates.

H. LUDWIG. The Genus Pentacrinus.

A. MILNE EDWARDS and E. L. BOUVIEB. The Crustacea of the "Blake."

A. E. VERKILL. The Alcyonaria of the " Blake "

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in charge of
Alexander Agassiz, on the U. S. Fi?h Commission Steamer "Albatross," from Au&ust,
1899, to March, 190U, Commander Jefferson F. Muser, U. S. N., Commanding.

Illustrations of North American MARINE INVERTEBRATES, from Drawings by Bukk-
hakdt, SON R EL, and A. Agassiz, prepared under the direction of L. Agassiz.

LOUIS CABOT. Immature State of the Odonata, Part IV.
E. L. MARK. Studies on Lepidosteus, continued.

" On Arachnactis.

R. T. HILL. On the Geology of the Windward Islands.
W. McM WuGDWOKTH. On the Bololo or Palolo of Fiji and Samoa.
A. AGASSIZ and A. G. MAYER. The Acalephs of the Ea>t Coast of the United States.
AGASSIZ and WHITMAN. Pelagic. Fishes Part II., with U Plates.

Reports on the Results of the Expedition of 1*91 of the U. S. Fish Commission Steamer
"Albatross," Lieutenant Commander Z. L Tanner, U. S. N., Commanding, in charge of
Alexander Agassiz, as follows: —

A. AGASSIZ. The Pelagic Fauna. H. LUDWIG. The Starfishes.

The Echini. J. P. McMURRlCH. The Actinarians.

" The Panunie Deep Sea Fauna. E. L. MARK. Branchiocerianthus.

K.BRANDT. The Sagittas JOHN MURRAY. The Bottom Specimens.

" The Thalassicols9. ROBERT R1DG WAY. The Alcoholic Birds

C.CHUN. The Siphonophores. P. SCH1EMENZ. The Pteropods and Hete-

" The Eyes of Deep-Sea Crustacea. ropods.

W. II. DALL The Mollusks. THEO. STUDER. The Alcyonarians.

H.J.HANSEN. The Cirripeds. M. P. A. TRAUTSTEDT. The Salpidas and
W. A. HEliDMAN. The Ascidians. Doliolidse.

S. J. DICKSON. The Antipathids. E. P. VAN DUZKE The Halobatidse.

\V. E. 1IOYLE. The Cephalopods. H. B. WARD The Sipuncnlids.

G VON KOCH. The Deep-Sea Corals. H. V. WILSON. The Sponges.

C. A. KOFOID. Sole ■ ogaster. W. McM. WOODWORTH. The Nemerteans
R YON IjENDENFEU). The Piiospiio- " The Annelids,

rescent Organs of Fishes.

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletins Vols. I. to XXXV. ;
of the Memoirs. Vols. I. to XXIV.

Vols. XXXVI., XXXVII., and XXXVIII. of the Bulletin,
and Vols. XXV., XXVI. of the Memoirs, are now in course of
publication.

The Bulletin and Memoirs are devoted to the publication of
original work by the Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural History, and of work by specialists based
upon the Museum Collections and Explorations'.

"The following publications are in preparation: —

Reports on the Kesulis of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander 0. D. Sigsbee, U. S. N.,and Commander J. It. Bartlett, U. S. N.,
Commanding.

Reports on the liesults of the Expedition of 1801 of the U. S. Fish Commission
Steamer "Albatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Reports on the Scientific; Results of the Expedition to the Tropical Pacific, in
charge of Alexander Agassiz, on the U. S Fish Commission Steamer
"Albatross," from August. 1809, to March, 1900, Commander Jefferson F.
Moser, U. S. X., Commanding.

Contributions from the Zoological Laboratory, in charge of Professor E. L.
.Mark.

Contributions from the Geological Laboratory, in charge of Professor N. S.
Shaler.

Subscriptions for the publications of the Museum will be received
on the following terms : —

For the Bulletin, $4.00 per volume, payable in advance.

For the Memoiks, $8.00 " " "

These publications :ire issued in numbers at irregular inter-
vals; one volume of the Bulletin (£vo) and half a volume of the
Memoirs ( I to) usually appear annually. Each number of the Bul-
letin and of the Memoirs is also sold separately. A price list
of the publications of the Museum will be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

MAR 21 1901

bir*!

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 2.

THE STRUCTURAL RELATIONS OF THE AM iTD ALGID AL

MELAPHYR IN BROOKLINE, JTWTO.S. AND

BRIGHTON, MASS.

By Hexrt T. Burr.

With' Two Plates..

CAMBRIDGE, MASS , U. S. A. :

PRINTED FOR THE MUSEUM.

March, 1901.

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE.

Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 2.

THE STRUCTURAL RELATIONS OF THE AMYGDALOIDAL

MELAPHYR IX BROOKLIXE, NEWTON, AND

BRIGHTON, MASS.

By Hexry T. Bdrr.

With Two Plates.

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

PRINTED FOE THE MUSEUM.

March, 1901.

MAR 21 1901

No. 2. — TJie Structural Relations of the Amygdaloidal Melaphyr
in Brookline, Newton, and Brighton, Mass. By Henry T.
Burr.

The sedimentary rocks of Boston and vicinity are associated with
basic igneous rocks of various ages. Local geologists are accustomed to
divide these, roughly, into two broad groups, — the trap dikes and the
"melaphyrs" or " amygdaloids." The latter are the greater masses,
irregular in outline aud much altered in structure and texture. The
former are distinctly dikes. They still retain much of their original
crystalline texture and are unaffected by cleavage.

While there is seldom any difficulty in distinguishing the two types
in the field, it is not easy to define the points of difference. Under the
microscope the melaphyr is excessively decomposed. It has been
studied in the Brighton area by E. R. Benton ('80, p. 416), at Hough's
Neck, Quincy, by J. E. Wolff ('82, p. 232), and by T. G. White ('97, p.
140). These writers agree in describing the rock as an altered basalt.
It is made up, typically, of plagioclase feldspar, magnetite, epidote, and
a mass of calcite, chlorite and other more or less indeterminate alteration
products. I have nothing to add to this save that slides from various
portions of Newton and Brookline exhibit the same general characters.
The traps are likewise much decomposed, although in varying degrees.
Dr. T. A. Jaggar, Jr., in investigations not yet published, has found that
many of these are altered basalts. There seems to be uo means of dis-
tinguishing them, petrographically, from the melaphyr.

Microscopically, the exteut of alteration in the melaphyr is usually
sufficient to obliterate the crystalline texture. The traps almost invari-
ably appear, to the eye, crystalline. The melaphyr has generally
assumed a greenish to purplish tinge, difficult to describe, but quite char-
acteristic. The melaphyr is frequently amygdaloidal. The amyg-
dules have a peculiar habit of grouping themselves about centres.
They are seldom scattered uniformly through any considerable mass
of the rock. The traps are seldom amygdaloidal, and the amygdules,
where they occur, tend to be rather sparsely disseminated.

VOL. XXXVIII. — no. 2

54 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

The traps have the form of fairly regular dikes and sills. Even the
earliest of these gives evidence of having followed joint cracks more or
less perfectly developed. The melaphyr, on the other hand, where it
has come into igneous contact with the sediments, has torn its way
through without regard for their structures. It cannot be doubted
that portions of the sediments, at least, were but slightly consolidated
at the time that the igneous mass was forced into them. The melaphyr
shows evidence of having undergone nearly the same structural changes
as the rocks with which it is associated. It is cleaved, in many places,
quite as perfectly as the slates and conglomerates. It is cut by the
same joint systems. There is evidence for believing it to have been
tilted to somewhat the same extent as the associated sediments.

The melaphyr bears many inclusions of felsite x and of slate. In
many cases these appear to have been partially fused, sometimes to the
point of becoming intimately mixed with the melaphyr itself. These
inclusions are highly characteristic of this rock.

Melaphyr occurs, within the limits of the so-called Boston Basin, at
Nantasket, in Hingham, on Hough's Neck, Quincy, on the Neponset
Biver south of Mattapan, at Needham, and in Brookline, Newton, and
Brighton. It is with the latter area that this paper deals.

The melaphyr, although so much altered internally, is yet a fairly
resistant rock. It usually stands out in prominent ridges and is thus
readily traced. If it were known to be associated with a definite
horizon, it would be extremely useful in deciphering the very involved
structure of the region. Professor Crosby ('89, p. 10), and others who
have worked with him have regarded these rocks as contemporaneous
flows, occurring at or near the base of the conglomerate series. If such
were the case, the melaphyr would have a value as an horizon marker.
The evidence in the Brookline, Newton, and Brighton areas seems to
indicate that the melaphyr is intrusive into the conglomerate and even
into the overlying slate. If such is the case, it is not associated with a
definite horizon and thus loses the greater part of its value as a guide to
the structure.

The evidence to be presented is of three kinds : (1) The sedi-
ments which appear to overlie the melaphyr do not contain fragments of
it ; (2) The contacts, wherever observed, are of igneous character ;
(3) The melaphyr masses do not show that structural accordance with
the sediments which would be demanded of them as flows.

1 The ancient rhyolitic rocks of this region are known, locally, as felsites. It is
in this sense that the term is used in this paper.

BURR: MELAPHYR OF BROOKLINE, NEWTON, AND BRIGHTON. 55

1. The Lack of Melaphyr Pebbles in the Conglomerate. — The
melaphyr appears to be overlain by a considerable thickness of heavy
conglomerate. This is particularly true of that in the Brookline area,
where it occupies the centre of an anticline and is surrounded by the
coarse basal beds of the conglomerate, dipping away from it in all direc-
tions. Heavy sediments, such as these, must have been deposited in
shallow water, within the zone of wave action. It is not easy to see
how they could have been formed without including some fragments of
the underlying rock. With this in view the conglomerate has been
carefully examined. Few pebbles of basic character have been found,
none referable to the melaphyr.1 Such evidence is of negative charac-
ter, and is, therefore, not conclusive, but it has some considerable
value.

2. Evidence from Contacts. — The zone of contact between the
melaphyr and the sediments is evidently a zone of weakness, for its
place is generally occupied by swamps and other low places. Actual
contacts are, therefore, not often seen. On Newton Street, Brookline,
(Plate 2, Loc. 1), nearly opposite the end of South Street, is an irregu-
lar mass or tongue of melaphyr nearly surrounded by conglomerate.
This mass cannot be traced far in any direction, but is presumably con-
nected in some way with the main melaphyr mass to the west. It is
seen in contact with the conglomerate at several points. It cuts the
latter in a very irregular way, and without regard for bedding or
structure. It is certainly not a flow nor even an intrusive sheet. It is
best regarded as an irregular tongue or offshoot from the main melaphyr
mass. The rock at this point is the typical melaphyr and is not to be
confounded with the later traps.

Farther westward, on the south side of Brookline Street (Plate 2,
Loc. 2), the melaphjT is again exposed in contact with conglomerate.
The melaphyr at this point is not over one hundred feet wide and is
flanked by conglomerate on both sides. The only contacts exposed are
on the northern side of the melaphyr. This would be the lower side if
the rock represented a flow. This contact is unmistakably igneous.
The hard sediment has been fused, in contact with the igneous rock, to
an homogeneous red mass which has the appearance of felsite. It is
unfortunate that the southern contact is not shown. It is, however,
hardly possible that the melaphyr can be a flow, for it cuts the sedi-
ments vertically while the latter have a prevailingly gentle dip to the

1 Professor Crosby ('89, p. 10), however, states that the conglomerate contains
fragments of the melapliyr.

56 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

south. Still farther west, on the north side of Nahanton Street (Plate
2, Loc. 3), conglomerate occurs in a re-entrant angle between outcrops
of melaphyr. No actual contact is seen, but a large boulder, found
near to this conglomerate outcrop, shows both conglomerate and
melaphyr, the contact being apparently igneous.

No other contacts are known on the southern border of the melaphyr
area. On the northern border contacts have been seen in but one locality.
East of Hammond Street (Plate 2, Loc. 4), opposite the cemetery, the
melaphyr has been exposed in two trenches. In both cases it is in con-
tact with conglomerate on its southern side. The contact is practically
vertical, while the conglomerate dips north at a moderate angle. A red
zone is developed in the conglomerate along the contact, and the sedi-
ment is penetrated by little tongues of the melaphyr. The north side
of the melaphyr is not shown. It seems reasonable to regard this small
body as an offshoot from the main mass to the south. The deeper of
the two trenches is now filled, while the sides of the other have fallen
in, obscuring the evidence. The main mass of melaphyr in this area is
separated from the conglomerate by a belt of low land. A ditch has
been dug along this belt from Hammond Street to Heath Street. At
present there is no exposure of rock in this ditch, but the debris taken
from it has been piled along its banks. A number of fragments have
been found in this debris which show the melaphyr in contact with the
conglomerate. Several of these bear witness to the igneous nature of
the contact, and one, in particular, shows the sediment penetrated by a
tongue of the melaphyr. In the same specimen a stringer from the
sediment has been partially fused and included within the igneous rock.
Under the microscope the feldspars are seen oriented in striking con-
formity with the boundaries of this included fragment.

The melaphyr is seen in contact with the conglomerate in several
places on the bank of the Charles River north of Boylston Street, New-
ton Upper Falls (Plate 2, Loc. 5). It clearly penetrates the sediment
along the contacts. In two places it cuts, dike-like, across the bedding
of the conglomerate. Farther east, on Rockland Street (Plate 2, Loc. 6),
the two rocks appear together, the contact zone developed by the melaphyr
being quite evident. On the corner of Rockland Avenue and Boylston
Street the melaphyr, although not seen in direct contact with the sedi-
ments, is filled with fragments of sandstone, presumably derived from
the neighboring sedimentary rocks (Plate 2, Loc. 7).

On the north side of Commonwealth Avenue, a little east of Auburn-
dale (Plate 2, Loc. 12), a ledge of coarse sandstone is cut by several

BURR: MELAPHYR OF BROOKLINE, BRIGHTON, AND NEWTON. 5/

irregular, dike-like intrusions of the melaphyr. The two rocks look
much alike upon the weathered surface, so that it is difficult to differ-
entiate them in the outcrop. The main body of melaphyr does not
appear in contact with the sandstone, but is to be seen in an outcrop
fifty yards eastward. A half-mile farther east, on the corner of Prince
Street (Plate 2, Loc. 13), the melaphyr is in contact with similar sand-
stone. The contact is unmistakably igneous. The sandstone dips north
at an angle of 15°. The plane of contact also dips north, but at, a much
higher angle, approximating 50°. The effect of the igneous rocks upon
the sandstone is plainly visible. Structures are obliterated and a distinct
contact zone is developed, while the irregularity of the contact is such as
could hardly be produced except by igneous action. Under the micro-
scope the feldspars are seen to be aligned along the contact.

The melaphyr is not again seen in contact with sediments for three
miles toward the east. On Cambridge Street, Brighton, a short distance
west of Foster Street, a cliff of sandstone and conglomerate is capped by
melaphyr. The contact shows abundant signs of igneous action. This
would, however, be true if the melaphyr were a flow. In the Allstou
area, bounded approximately by Cambridge and Warren Streets and Com-
monwealth and Harvard Avenues, contact exposures are fairly numerous
and frequently satisfactory. The relations of the rocks in this area
have been discussed by several writers, notably W. 0. Ci'osby, E. R.
Benton, and H. G. Woodward. Benton concludes that the melaphyr is
intrusive into the sediments. Crosby and Woodward do not accept this
conclusion, their idea being that the melaphyr masses are made up of
successive flows. Wroodward has mapped the area with great care, and
has differentiated six flows. His work has not been published, and
is, therefore, not subject to criticism. His map, however, was printed
and distributed by Professor Crosby in connection with a course of lec-
tures given at the Boston Society of Natural History in 1889-90. As
this map shows very clearly the location of the six supposed flows, it
seems fair to discuss thern, not foi'getting that Woodward may have been
in possession of evidence which has escaped the writer.

The most southerly of these supposed flows is not well indicated in
the field. It is not seen in contact with the sedimentary rocks. The
second " flow " is indicated by three outcrops. One of these, lying north
of the elbow in Commonwealth Avenue, shows amygdaloidal melaphyr
overlain by patches of sandstone and slate. The contact is not striking,
as the melaphyr and the sandstone are of the same dull green color, and
not unlike in texture. It becomes clear, however, on examination,: that

58 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

the melaphyr has been forced into the sediment. Fragments of the
latter have been torn off and lie embedded in the melaphyr. In places
the igneous rock breaks across the bedding planes, occasionally develop-
ing a contact zone, within which the bedding is obliterated and the
sediment hardened.

The third " flow " is marked by a very irregular ridge of melaphyr,
extending, with a few interruptions, from Commonwealth Avenue across
Allston Street to Cambridge Street. Contacts with the slate are shown
in several places. The lower or southerly contacts need not be dis-
cussed. Whatever the relations of the melaphyr may be, those contacts
would necessarily be igneous. Contacts on the northern or upper side
are to be seen at the so-called carriage-house quarry on Allston Street
(Plate 2, Loc. 15), on the edge of the woods, two hundred yards north
of the elbow in Commonwealth Avenue, and in several exposures along
the northern face of the ridge at the end of Allston Heights (Plate 2,
Loc. 16).

The slate at the carriage-house quarry (Plate 2, Loc. 10) is seen
lying on the back of the melaphyr ridge. It is tilted at a high angle,
dipping north 40°-60°. It is warped as though powerfully twisted.
The melaphyr just below the contact is full of slate fragments which are
twisted and gnarled as though they had writhed under the influence of
the molten mass. The actual contact is irregular. Tongues of the
melaphyr penetrate the slate, cutting the bedding, while the contact
zone is marked by a deep red band, varying from a sixteenth to a quar-
ter of an inch in thickness. Under the microscope the slate is seen
(Plate 1, Figure 1) to be penetrated by numberless tongues of the mela-
phyr, the bedding is cut and twisted, and little molten particles of the
slate fill the mass of the igneous rock. The feldspar crystals are fre-
quently aligned along the contact. Often the amygdules are seen
clinging to the little blebby fragments of slate, suggesting that the
latter may have influenced their formation. The plane of contact
does not lie exactly in the bedding planes, but cuts them at a slight
angle.

The ledges on Allston Heights show the same phenomena. The
megascopic evidence is here even more striking. The slate is thoroughly
filled with fine tongues of the igneous rock. Frequently the two rocks
are mixed to such an extent that it is hard to say where the actual con-
tact is. In places the slate is baked to a hornstone, blotchy with red
and white, suggesting bedding lines which have become obscured in a
pasty rock. Here, again, the microscopic evidence is clear (Plate 1,

BUKR: MELAPHYR OF BROOKLINE, BRIGHTON, AND NEWTON. 59

Figure 2). - The cutting and distorting of the bedding planes is strik-
ingly shown.

The third exposure is intermediate between, but considerably south
of, the other two. The melaphyr mass is here not more than fifteen feet
thick, and is bordered by slate on both sides. The slate on the northern
or upper side lies in patches on the melaphyr. Here, again, a zone of
baking is developed along the contact, while the cutting of the slate
by the melaphyr is quite as conspicuous as it is in the other two
localities.

The two melaphyr bodies occurring next to the north make up a sin-
gle ridge, or rather a succession of melaphyr knobs. Slate appears in a
number of places on these knobs. These patches of slate are connected,
on Mr. "Woodward's map, to form a parting between the two supposed
flows. A narrow, wedge-shaped mass of slate occurs at Locality 26
(Plate 2), west of Webster Street. The melaphyr just below this slate
contains fragments of it. On the contact, the influence of the igneous
mass is apparent. The slate is broken by very many small faults. These
do not pass into the melaphyr, the line of contact between the slate and
the igneous rock being quite continuous. It is probable that these faults
were produced by disturbance attendant upon the intrusion of the igneous
rock, the effect being obliterated along the contact by the influence of
the molten mass. A second exposure of slate occurs on this line on the
west side of Webster Place. The evidence here is not so striking, but is
quite as good as in the other places. A third locality appears, on Wood-
ward's map, on Cambridge Street, behind the Couvent (Plate 2, Loc. 18).
This outcrop shows, in the most conclusive manner, the extremely com-
plicated way in which the melaphyr has been forced into the sediments.
In fact, the whole outcrop is practically a breccia of great slate and sand-
stone blocks, held in a matrix of melaphyr, which fills all the cracks and
penetrates the sediment in thousands of tongues. The sandstone blocks
themselves are shattered so that they have the appearance of a fault
breccia and these small displacements do not pass across the contact into
the igneous rock.

Another portion of this mass is shown in contact with slate and sand-
stone on Cambridge Street, opposite Saunders Street (Plate 2, Loc. 17).
Recent blasting at this point makes it possible to obtain very fine hand
specimens, showing the intricate way in which the melaphyr has pene-
trated the sediment. The fifth and sixth " flows " do not appear in con-
tact with overlying sediments.

No other contact exposures are known, excepting several in the Allston

60 bulletin: museum of comparative zoology.

area which arc beneath the melaphyr. These would be igneous in any
case. All of the known contacts, so far as they yield evidence, point
toward the intrusive nature of the igneous rock.

3. Evidence from Structural Relations. — On Professor Crosby's
map of the Boston area (The Geology of Eastern Massachusetts), the
melaphyr is made to consist of two long arms, extending in a generally
east and west direction and united in a rather broad area west of the
Charles River. Each of these arms is bordered by conglomerate, and
this, in turn, is rimmed with slate. Between the arms the conglomerate
occupies the greater portion of the area, but is parted by a wedge of slate
which broadens eastward till it encircles the two conglomerate masses,
and unites with the slate areas to the north and south. This distribu-
tion would seem to lend striking support to the idea that the melaphyr
is at the bottom of the series and is followed by the mass of the con-
glomerate with the slate at the very top. There are indicated two well-
defined anticlines with melaphyr exposed in the axes and wrapping about
the spoon-shaped end of the synclinal area of slate and conglomerate
(see Crosby, '89, pp. 9-12 a).

It may be assumed for the moment that this interpretation of the
general structure is correct. It then becomes necessary to ascertain
whether or not the distribution of the melaphyr accords with this struc-
ture. In Brookline the melaphyr occupies the middle of the southern
anticline. It does not, however, appear to lie symmetrically about the
felsite axis exposed in the southern part of Newton. In the consider-
able space between Brookline and Greenwood Streets the melaphyr does
not appear. Although it is possible that it does underlie a portion of
this area, the occurrence of five well-marked outcrops of conglomerate
and the presence of great numbers of large conglomerate boulders makes
it probable that the latter is the underlying rock. The conglomerate
found on the southern side of Greenwood Street is made up almost
wholly of felsitic debris. It closely resembles a conglomerate appearing
near the corner of Dedham and Walnut Streets. The strike of the
Greenwood Street conglomerate would carry it very near to the Dedham
Street locality. It is not impossible that they are continuous beneath the
intervening space. It is not at all certain that the melaphyr occurs be-

1 The position of these anticlinal areas is indicated, on the accompanying map
(Plate 2), hy heavy black lines (d d' and e e'). These are drawn, as nearly as pos-
sible, through the middle portions of the melaphyr areas as expressed on Crosby's
map. A heavy broken line (if) is drawn through the middle of the synclinal area
of slate.

BURR : MELAPHYR OF BROOKLINE, BRIGHTON, AND NEWTON. 61

tween Greenwood Street and Newton Highlands. It is not seen in the
area occupied by this anticline west of the vicinity of Dudley Street. On
the south side of Central Street, west of the Charles River, red felsite is
exposed within a very short distance of the conglomerate. There is not
room, at this point, for the melaphyr to pass between the felsite and the
conglomerate. Thus it appears that it does not completely encircle the
synclinal area.

Melaphyr occurs near Newton Upper Falls, on the south bank of the
reservoir, and on the Charles River north of Worcester Street. It is well
exposed in three ridges on Boylstou Street and Thurston Road, not far
east of Upper Falls. These localities are in the middle of the synclinal
belt, and trend diagonally across the supposed trend of the slate. At
Crystal Lake, Newton Centre, the melaphyr again appears on the border
of the slate belt. On Station Street, near Newton Centre, is a small ex-
posure. This is nearly east of the Crystal Lake locality and possibly
connected with it. It is much nearer the synclinal than the anticlinal
axis.

The northern belt of melaphyr is not nearly so continuous as the
southern. It does not appear to occupy an axial position. It seems,
from exposures in recent ditches, that the region south of Newton Lower
Falls is underlain by conglomerate and slate (Plate 2, Loc. 23). There
are excellent exposures on Chestnut Street, south of Commonwealth
Avenue (Plate 2, Loc. 24). Conglomerate appears on the shore of Chand-
ler's Pond, near Lake Street (Plate 2, Loc. 25). The melaphyr does not
appear south of these exposures. If the limits of the conglomerate area
be extended, as they should be, to include these outcrops, there will be
a much greater area of conglomerate south of the melaphyr than north
of it.

The distribution of the melaphyr does not, therefore, accord with the
structure, as interpreted by Professor Crosby. It occurs not only in
the anticlines, but in the axis of the syncline. In the northern conglom-
erate area it is bordered by a much greater thickness of conglomerate on
the south than on the north. In the southern area it is not symmetri-
cally arranged about the underlying felsite. It is not continuous about
the synclinal area of conglomerate. This irregularity of occurrence is
quite in accord with the idea that the melaphyr is intrusive.

A New Interpretation op the Structure. — Before settling upon
this conclusion it is necessary to consider the possibility of a different
interpretation of the structure. The broad conglomerate area of Brook-
line, Roxbury, and Dorchester is the one well-defined structural unit in

62 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

the Boston area. The conglomerate dips beneath the slates to the north
and to the south and pitches eastward beneath the slates which underlie
Boston Harbor. Outcrops in many parts of the area are abundant, and
all accord with the idea that this belt is a broad, flat-topped anticline.

The structure of the northern conglomerate belt is not so clearly
shown. On the southern side of the melaphyr in this area, the sedi-
ments dip toward it at prevailingly low angles. On Grant Avenue, New-
ton Centre, north of Commonwealth Avenue, the conglomerate beds are
practically vertical, but in no other instance has a dip of more than 25°
been seen. The average dip is probably below 20°. North of the mela-
phyr the dip is still toward the north, but usually at higher angles,
reachiug 60° to 65° on the railroad west of Newtonville. The attitude
of these rocks does not suggest the anticlinal relation. It is much
simpler to regard the structure as monoclinal, the dips becoming steeper
toward the north.

This view is strengthened by a study of the relation which the con-
glomerate bears to the slate belt to the south. If the two conglomerate
areas were anticlinal, the intervening slate would, of necessity, be syn-
clinal. The conglomerate to the south appears to pass with perfect
conformity beneath the slate. It is reasonably clear that the slate ac-
tually overlies this conglomerate. But the conglomerate of the northern
area invariably dips away from the slate, and is everywhere discordant
with it. The conglomerate dips northward at angles of from 10° to 25°.
The slate dips in the same direction, but at angles varying from 5° to
90°. It is clear that there is some sort of break along this line of con-
tact. Fortunately several exposures yield positive evidence of the
nature of this break.

On the northeast bank of Chestnut Hill Reservoir (Plate 2, Loc. 19)
the slate is seen interbedded with coarser sediments and clearly over-
thrust by a massive bed of conglomerate. The slate dips north at
about 80°, while the conglomerate above has a dip in the same direc-
tion of 30°. The occurrence of slate associated with bands of coarser
sediment is suggestive of the transition zone between the slate and the
conglomerate. The same conditions may be seen on the railroad near
the Pumping Station and also on the track a half-mile to the eastward.
It is quite possible that the thrust produced at this point a syncline
which finally yielded to the strain. The attitude of the rocks at the
thrust is quite in accord with this view (Fig. 1). The effect of the
thrust is observable in the slate on the neck of land separating the two
parts of the reservoir. The slate at this point is excessively crumpled.

BURR: MELAPHYR OF BROOKLINE, BRIGHTON, AND NEWTON. 63

For a mile west of the reservoir the slate seems to be quite lacking. It
next occurs in a limited exposure on Beacon Street, a half-mile east of
Newton Centre (Plate 2, Loc. 20). Here again there is strong evi-

Figurk 1. Section A A'.

dence of overthrusting. The conglomerate cuts diagonally across the
slate beds, and its under surface is deeply scored in the direction of the
thrust. The relations here differ from those at Chestnut Hill. The
slate dips north at a very low angle, while the overlying conglomerate
has a much steeper dip (Figure 2). This relation is not suggestive of
synclinal structure in the slate. If a syncliue was developed in the
eastward extension, it is safe to say that it did not extend to this point.
Farther west the. slate belt, if it exists, is covered beneath the glacial

Figure 2. Section B B'.

deposits of Newton. It seems extremely probable that it does underlie
a portion of the low laud extending along the line of Beacon Street. It
appears again on the east side of the Charles River. Outcrops of slate
and conglomerate occur near Chestnut Street, north of Newton Upper
Falls, and a considerable outcrop of slate has recently been exposed east
of Newton Lower Falls, near Waban Avenue (Plate 2, Loc. 23). In
all exposures the slate is contorted and crushed, indicating that it has
undergone something of the strains experienced by the rocks farther
toward the east. At Newton Lower Falls a ditch has disclosed fine

Figure 3. Section C C

conglomerate (Fig. 3), such as is frequently found immediately beneath
the slates. This seems to indicate that the fault was dying out toward
the west, the throw not being sufficient to bring up the lower members
of the conglomerate series. It is evident that the fault fades out in like

G-i BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

manner toward the east, for the conglomerate ridges disappear, and the
Chestnut Hill Slate belt apparently becomes continuous with that to
the north.

That faulting of the overthrust type is to be looked for in the region
is shown by the occurrence of a second thrust crossing Commonwealth
Avenue at Summit Street, half a mile north of the one described
(Plate 2, Loc. 22). This is well shown in exposures on both sides of
the avenue. The slate is here repeated, lying on top of the conglomer-
ate and again overthrust by conglomerate to the north. The outcrop
on the east side of the street shows particularly well the way in which
the conglomerate has been crumpled by dragging on the plane of the
thrust. This fault is seemingly not traceable westward. Toward the
east it is complicated by a normal fault of later date.

The discontinuity of the Chestnut Hill slate belt, the actual occur-
rence, in two instances, of faulting on the contact between the slate and
the conglomerate, the general discordance in structure between the two
rocks and the known occurrence of a fault bringing about similar rela-
tions a short distance to the noi'th, all point to the same conclusion, that
the conglomerate has been thrust over the slate. All of these facts are
very difficult to explain on the supposition that the conglomerate passes
in a syncline beneath the slate.

The stratigraphic succession shown in this conglomerate belt is not
such as to indicate anticlinal structure. South of the melaphyr the
rocks are prevailingly conglomerate, coarse, as a rule, but becoming
finer toward the igneous rock. Near the melaphyr sandstone bands
are frequent, and there are occasional interbedded slaty layers. This is
the character of the upper portion of the conglomerate in the adjacent
Brookline area and quite unlike the massive beds in the lower zones.
The sediments in actual contact with the melaphyr are seldom coarser
than sandstone. In the eastern extremity they are prevailingly slates.
Such rocks are out of place at the base of the sedimentary series, judg-
ing from the evidence elsewhere. North of the melaphyr, outcrops are
few. The greater number are of slate. Thin conglomerate bands occur
in a number of places in association with finer sediments. In two lo-
calities, only, are these conspicuous. A bed of conglomerate, perhaps
thirty feet in thickness, is exposed on North Beacon Street, Brighton.
It conforms in dip and strike to the slates on the north and on the
south, and is apparently interbedded with them. Similar beds of con-
glomerate are known to occur within the slate series in other parts of
the region. Conglomerate is exposed in a cutting near the Boston and

BURR : MELAPHYR OF BROOKLINE, BRIGHTON, AND NEWTON. 65

Albany Railroad, east of Auburndale. The dip is here 40° toward the
south. As this is discordant with the dips in all neighboring exposures,
it is evident that some irregularity of structure, probably a fault, has
brought the conglomerate to the surface at this point. The scarcity of
outcrop in this area is, in itself, nearly sufficient evidence of the small
importance of the conglomerate, for this is a resistant rock and has a
way of asserting itself wherever it occurs. It does not seem necessary
to regard these limited exposures of conglomerate as the equivalent of
the thick beds to the south of the melaphyr.

The series, reading from south to north, is made up of conglomerate,
growing finer northward, interbedded sandstone and conglomerate, sand-
stone, and slate associated with melaphyr, slate with interbedded con-
glomerate, and finally a broad area of slate stretching away to the
crystallines on the north. This section does not yield evidence of a
repetition such as would be demanded on the supposition of an anti-
clinal fold.

There are no details of structure which indicate anticlinal folding
about the melaphyr or synclinal about the slate. Folded strata are un-
known in the region, if we except some crumpled slates on the line of the
great fault. There is no changing of strikes at the eastern end of the
northern belt of conglomerate, nor about the western end of the inter-
vening slate belt. On Cedar Street, at the western end of the supposed
syncline, the conglomerate maintains its generally east and west strike
up to its last outcrop, within fifty feet of the melaphyr to the west.

The conclusion as to the structure is that the conglomerate of the
southern area is anticlinal, and dips conformably beneath the slate of
Chestnut Hill and Newton Centre, while the northern area is brought
to the surface by a fault on which it has been thrust over the slate and
tilted toward the north. The lower members of the conglomerate series
would occur in the centre of the Brookline area and on the southern
edge of the northern belt. In these places, therefore, the melaphyr
would be expected to occur if it were associated with the basal rocks.

It has been stated that melaphyr does occur in the axis of the Brook-
line anticlinal area. No melaphyr is seen, however, on the southern
edge of the northern area. The long arm which extends eastward from
Newton Lower Falls is associated with rocks which are considerably
above the base. The two areas of Newton Upper Falls and Newton
Centre are south of the slate belt and therefore not associated with the
northern conglomerate area. It is conceivable and even probable that
the throw of the fault has not been sufficient to disclose the lowest beds

66 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

of the conglomerate. If melapliyr were associated with these beds, it
also might remain unrevealed. Even if this be true, however, the occur-
rence of melaphyr in association with beds higher in the series remains
unaccounted for.

It appears, therefore, that neither under the former interpretation nor
under that which is offered here, is the melaphyr always associated with
the basal beds of the conglomerate, or, indeed, with any definite horizon.
On any interpretation, the occurrence of the melaphyr in association with
sediments of all kinds, from the coarse conglomerate of Brookline to the
slate of Allston, is suggestive of a considerable vertical distribution.

Evidence from Structural Details. — There are some details of
structure which have a bearing on the general question. At various
points about the periphery of the Brookline area, small, isolated ex-
posures of melaphyr are seen in contact with conglomerate. These do
not differ, lithologically, from the larger mass. It is reasonable to sup-
pose that they are connected with it. As the conglomerate dips away,
in all directions, from the central area, it is clear that the melaphyr in
these exposures is associated with sediments higher in the series than
those which surround the main mass. In all cases the smaller masses
are intrusive into the sediments. This suggests that the main mass also
is intrusive.

The conglomerate on the north bank of the reservoir, at »Newton
Upper Falls, dips away from the melaphyr on the south bank. The
rock walls are but a few feet apart. It is evident that the dip of the
conglomerate is not sufficient to carry the lower beds exposed above
the melaphyr cliff. The strike of the conglomerate varies by about 12°
from the trend of the parting between the two rocks, carrying it obliquely
toward the melaphyr. Conglomerate, outcropping near Eliot Station, on
the Boston and Albany Railroad, strikes toward melaphyr on the oppo-
site side of the tracks (Plate 2, Loc. 8). On Walnut Street, Newton
Centre, conglomerate, on the east side of the street, strikes directly into
brecciated melaphyr on the west. On the corner of Commonwealth Ave-
nue and Valentine Street is a small exposure of conglomerate which strikes
toward the melaphyr mass. In the grounds of the Golf Club, on Centre
Street, slate strikes directly across the trend of the melaphyr. On
Foster Street, Brighton, sandstones, on opposite sides of the melaphyr
mass, strike toward each other. These occurrences indicate deformation
in the sediments preceding or contemporaneous with the intrusion of the
igneous rock.

On the east side of Chestnut Hill Avenue, Brighton, the conglomerate

BURR: MELAPHYR OF BROOKLINE, BRICxHTON, AND NEWTON. 67

strikes N. 15° E. and dips west. On the opposite side of the street the
melaphyr rises abruptly and extends westward in a prominent ridge to
Foster Street and continues to Lake Street. Thus the conglomerate
dips directly beneath the end of the melaphyr ridge and strikes across
its trend. Considering this in connection with the attitude of the rocks
on Foster Street, it seems probable that the melaphyr at this point
actually cuts the sediments across their strike and is thus truly a
great dike.

In the Allston area, structural discordances are many. Some of these
are possibly due to faulting. In a number of instances, however, the
melaphyr may be seen actually cutting the sediments. Dikes are known
in the field r^ear the bend in Commonwealth Avenue, in the woods to the
north, on Cambridge Street opposite Saunders Street, and in three places
in the conglomei'ate between Allston Street and Harvard Avenue. Two
of these latter dikes are, however, doubtfully regarded as melaphyr.
They have been mapped by Woodward as trap.

Conclusions. — The facts set forth in this paper are believed to lead
to the following conclusions : 1. The melaphyr, in the region discussed,
is intrusive into the sediments. 2. The melaphyr is not associated with
any definite horizon, and is therefore of no value as a guide to the inter-
pretation of the structure.

The first conclusion depends upon the following facts : 1. The con-
glomerate, associated with the melaphyr, contains no fragments of it.
2. The contacts, wherever found, are igneous in character. 3. The
melaphyr is seen in contact with sediments varying from the coarsest of
the conglomerate to the finest of the slate. 4. The distribution of the
melaphyr shows it to be discordant with the structure of the sediments
under any interpretation of the latter that has been offered. The second
conclusion follows directly upon the first.

Incidentally, a new interpretation of the structure is offered for the
Chestnut Hill slate belt and the northern area of conglomerate.

tol. xxxvm. — no. 2 2

68 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

BIBLIOGRAPHY.

Bascom, F.

1900. Volcanics of Neponset Valley, Mass. Bull. Geol. Soc. Amer., vol.
11, pp. 115-126.

Benton, E. R.

'81. The amygdaloidal melaphyre of Brighton, Mass. Proc. Bost. Soc. Nat.
Hist., vol. 20, pp. 416-426, pi. 4.

Crosby, W. O.

'80. Contributions to the geology of eastern Massachusetts. Occas. Papers.
Bost. Soc. Nat. Hist., vol. 3, vii, 286 pp., 5 plates, map.

Crosby, W. O.

'89. Physical history of the Boston Basin. Boston, 22 pp.

Davis, W. M.

'81. Banded amygdules of the Brighton amygdaloid. Proc. Bost. Soc. Nat.
Hist., vol. 20, pp. 426-428.

White, T. G.

'97. A contribution to the petrography of the Boston Basin. Proc. Bost.
Soc. Nat. Hist., vol. 28, pp. 117-156, 5 plates.

Wolff, J. E.

'82. The great dike at Hough's Neck, Quincy, Mass. Bull. Mus. Comp.
Zool., vol. 7, pp. 231-242.

BURR: MELAPHYR OF BROOKLINE, BRIGHTON, AND NEWTON. 69

EXPLANATION OF PLATES.

PLATE 1.

Fig. 1. Microphotograph of thin section from upper contact of melaphyr at
carriage-house quarry, Allston. (Plate 2, Loc. 15.) Shows bedding of
slate cut and contorted by tongue of igneous rock.

Fig. 2. Microphotograph of thin section from upper contact of melaphyr at head
of Allston Heights. (Plate 2, Loc. 16.) Shows bedding of slate cut
irregularly by igneous rock and obliterated on contact. Detached frag-
ments and tongues of slate drawn into melaphyr.

PLATE 2
Map of a part of Brookline, Newton, and Brighton, Mass.

Amvgdaloidal Vltrlaph

Plate

THE HELIOTYPE PRINTING CO.. BOSTON.

UPPER CONTACT. BRIGHTON AMYGDALOID.

The following Publications of the Museum of Comparative Zoology
are in preparation : —

Reports on Hie Results of Dredging Operations in 1877, 1878, 1879, and 188", in charge of Alex-
ander Agassiz, by the U. S. Coast Survey Steamer " Blake," as follows: —

E. EHLEKS. The Annelida of the " Blake/'

C. HARTLAUB. The Comatube of the " Blake," with 15 Plates.

H. LUDWIG. The Genus Pentacrinus.

A. MILNE EDWARDS and E. L. BOUVIER. The Crustacea of the " Blake."

A. E. VERB ILL. The Alcyonaria of the " Blake."

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in charge of
Alexander Agassiz, on the U. S. Fish Commission Steamer "Albatross," from August,
1899, to March, 1900, Commander Jetferson F. Mcser, U. S. N., Commanding.

Illustrations of North American MARINE INVERTEBRATES, from Drawings by Buhk-
hardt, Sonrel, and A. Agassiz, prepared under the direction of L. Agassiz.

LOUIS CABOT. Immature State of the Odonata, Part IV.
E. L. MARK. Studies on Lepidosteus, continued.

" On Arachnactis.

R. T. HILL. On the Geology of the Windward Islands.
W. McM. WOODWORTH. On the Bololo or Palolo of Fiji and Samoa.
A. AGASSIZ and A. G. MAYER. The Acalephs of the East Coast of the United States.
AGASSIZ and WHITMAN. Pelagic Fishes. Part II., with 14 Plates.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission Steamer
"Albatross," Lieutenant Commander Z. L. Tanner, U. S. N., Commanding, in charge of
Alexander Agassi/., as follows: —

A. AGASSIZ. The Pelagic Fauna. H. LUDWIG. The Starfishes.

The Echini. J P. McMURRICH. The Actinarians.

" The Panamic Deep-Sea Fauna. E. L. MARK. Branchiocerianthus.

K. BRANDT. The Sagittre. JOHN MURRAY. The Bottom Specimens.

" The Thalassicolse. ROBERT RIDG WAY The Alcoholic Birds.

C.CHUN. The Siphonophores. P. SCHIEMENZ. The Pteropods and Heter

" The Eyes of Deep-Sea Crustacea. ropods.

W H. DALE. The Mollusks. THEO. STUDER, The Alcyonarians.

H. J. HANSEN. The Cirripeds. M. P. A. TRAUTSTEDT. The Salpidse and
W. A.: HERDMAN. The Ascidians. Doliolid.-e.

S. J. HICKSON, The Antipatl.ids. E. P. VAN DUZEE The Halobatidse.

W. E. HOYLE. The Cephalopoda. H. B. WARD. The Sipuneulids.

G 7QN KOCH. .The Deep-Sea Corals. H. V. WILSON. The Sponges.

C. A. KOFOID. Solenogaster. W. McM. WOODWORTH. The Nemerteans.
R. VON LENDENFELD. The Phospho- " The Annelids,

rescent Organs of Fishes.

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletins Vols. L to XXXV. ;
of the Memoirs, Vols. I. to XXIV.

Vols. XXXVI., XXXVII., and XXXVIII. of the Bulletin,
and Vols. XXV., XXVI., XXVII. of the Memoirs, are now in
course of publication.

The Bulletin and Memoirs are devoted to the publication of
original work by tlie Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural History, and of work by specialists based
upon the Museum Collections and Explorations.

The following publications are in preparation : —

Reports on the Results of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander C. D. Sigsbee, U. S. N., and Commander J. R. Bartlett, U. S. N.,
Commanding.

Reports on the Results of the Expedition of 1801 of the U. S. Fish Commission
Steamer " Albatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in
charge of Alexander Agassiz, on the U. S. Fish Commission Steamer
"Albatross," from August, 1899, to March, 1900, Commander Jefferson F
Moser, U. S. N., Commanding.

Contributions from the Zoological Laboratory, Professor E. L. Mark, Director.

Contributions from the Geological Laboratory, in charge of 'Professor N. S.
Slialer.

Subscriptions for the publications of the Museum will be received
on the following terms: —

For the Bulletin, $4.00 per volume, payable in advance.
For the Memoirs, $800 " " "

These publications ore issued in numbers at irregular inter-
vals; one volume of the Bulletin (8vo) and half a volume of the
Memoirs (-Ito) usually appear annually. Each number of the Bul-
letin and of the Memoirs is also sold separately. A price list
of the publications of the Museum will be soil on application
to the Librarian of the Museum of Comparative Zoology, Gam-
bridge, Mass.

MAR 29 1901

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXVIII.

GEOLOGICAL SEKIES, Vol. V. No. 3.

THE PHYSIOGRAPHY' OF ACADIA.

By Reginald A. Daly.

With Eleven Plates.

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

PRINTED FOR THE MUSEUM.

March, 1901.

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE.

Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 3.

THE PHYSIOGRAPHY OF ACADIA.

By Reginald A. Daly.

With Eleven Plates.

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

PRINTED FOR THE MUSEUM.

March, 1901.

MAR 20 1901

No. 3. The Physiography of Acadia. By Eeginald A. Daly.

TABLE OF CONTENTS.

Introduction : The Unity of the Ap-
palachian System ; Data of
the Investigation .... 73

The Uplands 75

The Southern Plateau ....

Structure

Form : an Old-mountain Pla-
teau ; Theories of Origin .
Extension of the Southern Pla-
teau Facet

The Cobequid Plateau ....

The New Brunswick Highlands .

North Mountain, Digby Neck,

and Long Island ....

The Bay of Fundy Trough in

Geological Time ....

75

SI

85

Warping of the Upland Peneplain

The Second Cycle : Development

of the Triassic Lowlands .

The Geological Dates of the

Peneplains

The Carboniferous Lowlands
Dissection of the Tertiary Pene-
plain ; Drowning and more
Recent History of the Ter-
tiary Peneplain

Summary

Homologies of Land-form and of
the Determining Structures
in Acadia and New England

Bibliography

Explanation of Plates

PAGE

87

92
93

96

97

99
103

Introduction.

The results of many independent workers in the old-mountain
Appalachian belt of eastern North America have shown that, as re-
gards axial trends, formatiorlal composition, and structure, the uplands
and lowlands from Georgia to Gaspe belong to one system. The unity
of the whole is the result of both orogenic and epeirogenic movements ;
the former leading to the fairly steady accumulation of Palseozoic sedi-
ments on a subsiding sea-floor, and to the filling with Triassic sand-
stones of basins produced in Nova Scotia, New England, New Jersey,
and Virginia during Permian warping; the latter inducing the great dis-
order referable to the Carboniferous " Revolution," and the simpler but
important block-faulting in Jura-Cretaceous time. Besides these periods
of wide-spread and pretty general mountain-building, other epochs of
more local deformation were characterized by synchronous folding in
limited parts of the belt. Extensive mountain-building occurred during
the middle Silurian folding of Acadian and of western New England
vol. xxxviii. — so. 3

74 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

rocks, — a period of important deformation that has also been de-
monstrated in Virginia. More localized areas of such upturning,
occurring at times not similarly occupied elsewhere, are represented
by the Acadian tract which was profoundly disturbed during the late
Devonian. These local foldings have, however, followed the same laws
of axial development that were obeyed in the more general mountain-
building, and have not interfered with the essential tectonic agreement
of the system.

Thus a great generalization has been built up, in largest part by
structural studies, aided, of course, by palseontological evidence, and
by the lithological investigation of the sediments involved. Of late
years a fourth efficient means of further developing the conception has
been wrought out, simply by the recognition of the fact that sedi-
mentation has its correlative in denudation, and that the later forms of
subaerial erosion on unsubmerged areas may be expected to show similar-
ities wherever the conditions of erosion were equivalent along the major
axis of the system. The remarkable parallelism subsisting among the
forms of wasted land-surfaces in New England, in the New Jersey-
Pennsylvania region and in the southern Appalachians, has been
broadly sketched by Davis * and further emphasized by Willis, Hayes,
Campbell, Keith, and others. As yet the definite statement as to how
far this last method of correlation may be extended to the Acadian
division of the belt has not been made ; in the following pages it is
proposed to describe briefly and illustrate some of the more important
erosion-forms in Nova Scotia and the adjacent portions of New Bruns-
wick and to inquire into their interpretation.

The field-work on which this sketch is based was . confined to a ten
days' tour on the main lines and branches of the Intercolonial and
Dominion Atlantic railways, supplemented by a second railroad trip
from Montreal to North Sydney and a cruise close inshore from Cape
Canso along the southern coast to Halifax, where the train was taken
for Yarmouth. Such rapid views of the country may possibly lead to
results of some value if they be checked and amplified by the study of
the detailed works published by governmental survey or by private
individuals.

The lack of good topographic maps of the inland areas renders
the discussion difficult, and many interesting questions must on that
account be left untouched. The best cartographic data are derived

1 The geological dates of origin of certain topographic forms on the Atlantic slope
of the United States. Bull. Geol. Soc. America, 1891, Vol. 2, p. 545.

DALY : PHYSIOGRAPHY OF ACADIA. 75

from the Admiralty charts of the coast-line (republished by the
Hydrographic Office at Washington), from the maps of the Canadian
Geological Survey so far as published, from the wall-map of Mackinlay,
and from the small-scale geological map accompanying the text of Daw-
son's " Acadian Geology." The physiographic literature treating of
Acadia is mainly concerned with glacial and post-glacial problems and
with recent crustal movements. The following pages will be restricted
almost entirely to the problems of bed-rock forms. The aim will be to
express certain conclusions regarding the pre-glacial denudation of this
region. Those conclusions do not demand an accurate knowledge of
pre-glacial drainage ; it is certain that the body of fact already deter-
mined will not permit of our constructing even a tolerable map of the
pre-glacial river-systems.

The Uplands.

The Southern Plateau. — Bather more than three-fourths of the
province of Nova Scotia (excluding Cape Breton Island) is occupied by
the largest topographic unit of which I shall have to speak in detail.
It may be called the " Southern Plateau." It is sharply defined in its
western half by the steep front overlooking the Cornwallis-Annapolis
valley, and by the ragged Atlantic shore-line ; in the eastern half, the
Atlantic bounds it on the south, the Truro-Pictou lowland, Northum-
berland Straits, and St. George's Bay on the north and northeast.

JStructtire. — The plateau is underlain by a complex of ancient rocks.
Most of the area exhibits the outcropping edges of a very thick series
of slates associated with a likewise extensive older group of quartzites ;
both of these series are presumed by the officers (Bailey, '98, p. 27 If.)
of the Canadian Geological Survey to be of Cambrian age. Van Hise
('92, p. 247) regards them as possibly Algonkian. Much less important
from the point of view of superficial extent but significant for the
unravelling of the history of the plateau, are the isolated patches of
fossiliferous Upper Silurian and Devonian sediments occurring in the
northern half of the long upland belt. Each of these three ancient
members has been involved in the late Devonian mountain-building,
which, more than auy other single event, has rendered the structures of
the plateau rocks complicated and difficult of interpretation. During
that revolution which must have erected over the region an alpine
chain of high mountains, the base of the resulting range was punctured
and displaced by enormous masses of exotic granites deeply buried at

76 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

that time, after the manner of granite just intruded. The Lower
Carboniferous rocks afterwards deposited among the Devonian ridges
in the east were, in their turn, folded, and they too compose a sub-
ordinate part of the plateau.

form. — But all this constructional topography due to folding has
been changed (Plates 1 and 2). The geologist now directs his exploring
canoe over a thousand square miles of granite moulded in the low re-
lief of gently undulating hills separated by shallow lakes. He finds but
little trace of the mountain-cover which blanketed the intrusions and
permitted the slow crystallization of the vast igneous bodies. As little
does he anywhere Hud the original surface of the ranges. There remain
to him only the edges of slate or quartzite band, that once, to right,
stretched up over the roof of a now vanished arch, and, to left, formed
half of a ruined trough, which he can still decipher, though tattered
and corrugated on its rims, the roots of an all but completely destroyed
mountain mass. Throughout these fifteen thousand square miles, the
observer finds no hill more than one thousand feet high, and ninety
per cent of the plateau is less than six hundred feet in altitude above
the sea. It will be well to note some of the readings of elevation which
I have compiled from the various sources already mentioned ; many of
them are merely barometric determinations, but no serious error is
believed to result from their use.

In addition to the fact that the coastal elevations are low, it was
speedily observed that they are systematically related. At the mouth of
St. Mary's Bay, the bed-rock surface of the hill-tops stands at from one
hundred and fifty to two hundred feet above the sea. Thence a more
rapid rise northeastwardly brings the average summit southeast of Digby
to the five-hundred-foot contour, and the long escarpment of " South
Mountain," running north-east by east, remains at that elevation for
some sixty-five miles. Along the roughly parallel line of the Atlantic
coast, the heights vary from seventy feet at Cape Sable to one hundred
and fifty feet at Halifax. Profile sections taken transverse to these two
belts of elevation show that there is a fairly gradual descent from one
to the other, in a southeast by south direction, a descent interrupted
only by local and faint reliefs to which reference will presently be made.
It will be further noticed that the greater height of the hills near Hali-
fax is related to a corresponding proximity of the Atlantic shore to
"South Mountain," proving an essentially equivalent angle of slope for
the surface throughout the western half of the plateau as far as the
town of Digby. The same angle and direction of slope characterizes

DALY: PHYSIOGRAPHY OF ACADIA. 77

the narrow portion of the plateau south of the Truro lowland, as well
as the slightly more broken northeastern third of the plateau which
gradually descends from the 600-1000 foot hills on the north to the
200-300 foot hills west of Cape Canso.

The granitic " axis " of Digby and Annapolis counties rises not more
than one or two hundred feet above the general surface of the broad
upland, an amount insufficient to interfere seriously with the general
law of a gentle southward slant for the whole plateau. But few individ-
ual hills break the even sky-line. Among these Mt. Ardoise (738 feet),
southeast of Windsor, and Aspatagoen (480 feet) at Mahone's Bay,
are the most prominent west of Halifax, while an unnamed summit
725 feet high lies south of Chedabucto Bay. The two last mentioned
are granitic and presumably project above the plateau surface because
of their comparatively great power of resistance to erosion ; Mt.
Ardoise also bears unmistakable evidence of having been, like the
neighboring low swells of quartzite, worn out of a larger mass, although
there is no direct evidence of differential hardness.

Thus the plateau is to be regarded as a single great topographic facet
above which rise local cameo-like reliefs in the form of a few residual
hills. It also presents the appearance of having been carved in intaglio.
Sunk beneath the facet are many deep, narrow, and steep-sided valleys
in the north, typified by the gorges of Bear River, Gaspereau River
(Plate 3), the upper Shubenacadie, and the East River; in the south
the bed-rock valleys are shallower and have been in large part filled by
drift, with the result of obscuring greatly their pre-glacial form. Bailey
('98, pp. 10, 11) notes the fact that they are commonly longitudinal,
lying on the softer slates between the quartzite ridges. All these
valleys are separated by broad interstream upland spaces and do not
furnish other than quite subordinate interruptions in the relatively
even plateau-top.

Is it expedient to leave, without decided emphasis in a description
of Nova Scotian physiography, this remarkable, and, in view of the
geological history, astonishing, flatness of the main body of the penin-
sula? Yet in the works already written on the subject, we read more
of the inequality in the land-surface than of its essential character of
low relief. " South Mountain " is spoken of as one of the two " ranges"
walling in the Annapolis Valley (S. E. Dawson, '97, p. 155) : surely a far
from satisfactory designation of the escarpment of an even-topped up-
land. The " dominant features " of Nova Scotia " consist of ridges run-
ning parallel to the length of the peninsula " (G. M. Dawson and A.

78 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Sutherland, '92, p. G9). In the attempt to do justice to the beauty of
his native province, many a Nova Scotian has kept our facet in the back-
ground and has laid an undue amount of stress on the hilly nature of
the plateau ; in certain instances he has literally made " mountains out
of mole-hills," apparently with the mistaken notion that the true lover
of nature cannot be especially interested in her land-forms when they are
subdued. Yet the marvel of Nova Scotian scenery lies in its flatness.

We have, then, to explain for this area several distinct facts : (1) The
actual low relief everywhere contrasts with a former strong relief of
earlier geological time. (2) The present surface is a consequence of the
truncation of the outcropping edges of stratified beds of various ages and
structures, — destruction of such magnitude as to lay bare certain of the
vast igneous cores within the Devonian mountain-chain. The different
rock-members are not only structurally diverse and chronologically dis-
tinct; they vary in hardness, although they are all absolutely resistant
to the weather on account of their well advanced consolidation during
the long stretches of geological time, and because of the intense crush of
mountain-building. (3) Our problem is even more special than to recog-
nize that ten thousand feet of elevation on serrate mountain-ridges has
been exchanged for maximum heights of from six hundred to one thou-
sand feet. We must also account for a general accordance of summit-
levels which fall into, or nearly into, the common plane of a topographic
facet gently inclined towards the south ; the fact of this attitude is just
as clear as the low absolute range of the elevations. (4) The isolated
knobs, cones, and swells of bed-rock rising above the facet, and the val-
leys incised beneath it, must find a place in our theory of the plateau.

Fortunately, we are not here compelled to break wholly new ground
in the interpretation of the Nova Scotia plateau. All will agree that it
is an old mountain mass worn down to far less than its original relief.
So limited, the problem resolves itself into the question as to how and
when the work of denudation and the truncation of the rock-members
was carried on.

Of the manner of erosion in such a region three hypotheses have been
proposed; they are of unequal value, but it is advisable to note them
all in this connection. The oldest of all would attribute the plain of
denudation mainly to marine action during an extremely slow but very
prolonged period of subsidence beneatli the level of the sea. The second
would regard it as a peneplain, the final product of a completed cycle of
subaerial erosion, and in our case would demand a southward tilt opening
a new cycle as explanatory of the present position of the peneplain.

DALY : PHYSIOGRAPHY OF ACADIA. 79

The third hypothesis would not necessarily posit more than one cycle
for the development of the existing land forms; one cycle advanced
to maturity, in addition to the episode of recent drowning, would be
required.

It must be confessed that we have not as yet for plains of marine
erosion, criteria either so numerous or so well worked out as those for
the peneplain. It has been often, and with justice, emphasized that the
theory of the peneplain must recognize the fact that there is no living
representative of these ancient plains of denudation. Among the many
peneplains so far described, no one of them has the attitude which it
should possess during and at the time of its completion ; that is, no one of
them stands so near the sea-level that the relief can no longer be seriously
diminished by its sluggish streams. It may be equally well advanced
against the theory of marine abrasion that no extensive submarine plain,
proved to be of this origin, can now be found on the flanks of the con-
tinents. In either case, explanation may be sought in the sugges-
tion that there has been, in late Tertiary time, a special amount of
epeirogenic movement. Absolute uniformity of crustal uplift and de-
pression throughout geological time can now no more be believed in
than absolute uniformity of climate or of the advance of organic life in a
given i-egion. The geological record shows that long periods of quiet
have been preceded and followed by others of uneasiness and of rela-
tively pai^oxysmal changes in elevation of continental masses. If the
brilliant generalizations of Suess regarding wholesale Tertiary subsidence'
of parts of the ocean-floor are shown to be correct, particularly those
referring to the North Atlantic, it is to be expected that plains of denu-
dation of either origin would be seriously disturbed from their original
position. This expectation would remain, although we might not feel
complete agreement with Suess's fundamental position as regards the
cause of the negative movements of sea-level. The future adjustment
of geological opinion with respect to these matters may some day remove
this i*eal difficulty which adheres to either theory of the gi'eater facets of
denudation.

Both theories demand of necessity a special behavior of the earth's
crust. The cycle of subaerial erosion can only be completed and a
peneplain produced if there has been average crustal stability through
an enormous period of time. The cycle of submarine erosion lead-
ing to the fashioning of a plain of marine denudation requires during
a period at least as long a more or less steady subsidence which will
permit the waves to cut inside the belt of aggraded sea-floor over

80 bulletin: museum of comparative zoology.

which they run. This subsidence must be so slow and so long continued
as to allow of the removal of alpine or subalpine relief over the region
attacked by the horizontal wave-saw. The oue hypothesis premises the
enduring constancy of a crust at rest or with but faint oscillations of
level; the other, the enduring constancy of a crust affected by tolerably
steady motion in one direction.

Partly on account of this difficulty in believing in the requisite
constancy of the base level during the cycle of subaerial erosion, the
inquiry has been recently made by Tarr as to whether the peneplain
explanation is correct when applied to the New England and New Jersey
uplands; he has proposed a third theory in its stead.1 The main idea
of his thesis has been enforced in a later paper by W. S. Tangier
Smith, who, however, seems to adhere to the peneplain explanation in
certain cases.2 A second objection to that explanation was founded on
Tarr's criticism of the method used to account for the disturbance
of the New Jersey and New England peneplains since completion.
These facets are supposed, on the hypothesis of peneplanation, to have
been tilted seaward, and more complex movements are believed by
Hayes and Campbell to have affected the peneplained surfaces of the
Southern Appalachians. In order to obviate the necessity of invoking
such warpings (albeit among the simplest and commonest of crustal
deformations), and to make more credible the denudational theory of the
uplands in these old-mountain areas, Tarr has suggested that essentially
the whole work of erosion has been accomplished in one cycle. In the
mature stage of that cycle, not only are the larger streams well graded,
but the slopes on the interstream spaces will be graded in sympathy
with the axial profiles of the streams. The line of divide belonging to
any interstream space will gently slope seaward, because the moun-
tainous terrane will have lost more material near the sea than farther
inland. On account of the greater volume of each stream near its
mouth, the lower part of the valley . carved by that stream will be
deepened nearly to baselevel before the upstream part. Hence the
valley-sides will waste faster near the mouth than toward the head-
waters. Smith has strengthened this hypothesis of "beveling" by
referring to the tendency at maturity toward a roughly equal spacing
of master-rivers in a region even of diverse structures;8 the result
thereof being a series of divides declining to the sea at about the same

i Amer. Geol., 1898, Vol. 21, p. 351.

- Bull. Depart. Geol. Univ. Cal., 1899, Vol. 2, p. 155.

8 Shaler, Bull. Geol. Soc. Amer., Vol. 10, p. 263.

DALY: PHYSIOGRAPHY OF ACADIA. 81

rate. These would show a common sky-line when viewed from some
commanding peak. Such a sky-line is supposed by Tarr and Smith to
be the one mistaken in some regions for the residual profile of a finished
peneplain, dissected after a new uplift. If the " beveling " hypothesis
be established, it is claimed that, in such cases, we are no longer forced
to believe in the repeated occurrences of land-surfaces that stood near
the sea-level through the enormous period of time occupied during an
old age sufficient for peneplanation. At the same time, tilting forms
a superfluous supposition.

But, even without questioning the conclusion that downstream wast-
ing will be so very much more rapid than upstream wasting, we may
believe that the " beveling " hypothesis cannot explain the topography
of the Southern Plateau. In the first place, the cross section of any
of the more important interstream areas indicates a striking discon-
tinuity of slope. Most of the sky-line profile is flat and is evidently
not in the position of a graded slope where the determinant of the
grade is either of the two adjacent streams. From this broad, flat
upland there is on each side a sudden transition to the steep slope im-
mediately overlooking the stream, — a sharp decline characteristic of a
young valley. The frequent repetition of such an arrangement of slopes,
the predominance of broad, undissected interstream facets developed
indifferently on diverse rocks and structures in no sympathetic relation-
ship with the existing stream-courses, is most simply explained by re-
ferring the land-form to two cycles. Secondly, the lack of accordance
between the slant of the upland facet and the trends of the larger
(pre-glacial) valleys seems to invalidate the idea of " beveling " in
almost as complete a manner. We cannot doubt that one of the largest
pre-glacial river-systems of Acadia lay in what are now the more or less
completely drowned valleys of the Bay of Fundy and of the Cornwallis-
Annapolis lowland, both of which trend at a high angle to the direction
of slant of the plateau. Yet, by the hypothesis of " beveling," we
should expect the northern edge of the plateau rock lying adjacent to
these master-valleys to be graded to at least as flat a slope as those in
the interior of the present upland, and we should look for a gradual fall
of the " South Mountain " divide from Wolfville southwestwardly.
We have already seen that neither of these conclusions corresponds
to the facts.

Field evidence, then, seems to exclude the newest of the three
hypotheses which have, up to the present time, been suggested for such
a case. Whether marine or subaerial denudation must be finally

82 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

appealed to for explanation of the upland facet of the Southern Plateau,
cannot as yet be asserted with the confidence of full proof; to attain it,
much additional field-work would be necessary. What is needed is
some such evidence as that of the subaerial origin of the inlaud lowlands
of Pennsylvania, which, on account of their enclosed nature, cannot
have been effectively reached by the sea. The apparently complete
lack of coastal sediments on the facet is suggestive of the subaerial ex-
planation, but it is open to question whether such an overlap of neces-
sarily weak rocks would not have been quite removed by general land-
erosion, during the long time required to excavate the Fundy lowlands
beneath the upland facet. The known inferiority of shore forces to
erode as compared with atmospheric agencies of wear is not, if time
enough and appropriate crustal conditions be granted, an antecedent
objection to the marine hypothesis.

Yet there are certain reasons why greater satisfaction may be felt
with the peneplain hypothesis. The occurrence of residuals just where
the waves would have especially great destructive power can hardly be
accounted for as a result of differential hardness. The very perfect
adjustment of drainage-courses to soft rock-belts in the Triassic area
hereafter described will be found to bear striking resemblance to similar
adjustment among the trap ridges of New Jersey, and to afford grounds
for a similar explanation of the phenomenon as the product of two
cycles of land wasting. So close are the similarities existing between
the Nova Scotian plateau, on the one hand, and the New England,
New Jersey, and Southern Appalachian plateaus, on the other, that one
cannot resist the conclusion that it is simplest to find a common expla-
nation for the greater denudation in the whole belt. One of the chief
recommendations for the peneplain theory as applied to the southern
province is the fact that, if we grant the truth of the theory, a more
complete and consistent account can be given of the multitude of land-
forms associated with the larger facets. The same is believed to be
true here. Bearing in mind the lack of complete demonstration for the
peneplain character of the Southern Plateau, we shall yet posit such a
character and proceed to inquire whether or not the topographic divi-
sions of Acadia are in accordant, organic relationship with that facet.

Extension of the Southern Plateau Facet. — In his definition
of the term " peneplain " Professor Davis noted a considerable area as
one of its necessary characteristics ; this for the reason that, during the
long lifetime of the cycle in which planation is established, denudation
will hold extensive sway even over harder rocks, and they will be re-

DALY: PHYSIOGRAPHY OF ACADIA. 83

duced to a lowland throughout a whole physiographic province. We
expect, therefore, that if the upland facet of the Southern Plateau be
of the nature of a true peneplain, it must have once extended far aud
wide over Acadia. It ueed not now be continuous, for the warping,
which we believe has occurred, may be expected to have revived the old
rivers of the former cycle, and to have developed new subsequent
drainage in the present cycle ; that, therefore, belts of lowland may be
looked for wherever weak rocks appear. This deduction is realized in
the facts. Where, for example, the older, harder Palaeozoics appear out-
side of the Southern Plateau, there uplands and fragments of the pene-
plain are found to occur, and it is the soft Carboniferous and Triassic
sediments that underlie the extensive lowlands. In fact, the various
topographic divisions of Acadia from the Southern Plateau to Gaspe are
perhaps best described in terms of this facet. The evidences for this
statement will be more clear after a brief recital of the facts relating to
each of the divisions. Since I have had the opportunity of making
but one excursion and that by railway, across Cape Breton Island, and
since the amount of information regarding the recent geological history
of that island is scanty, I shall not attempt to speak of it in detail.
From the nature of its terranes and from its geographical position, it
ought to include part of the peneplain ; and, indeed, from the descrip-
tions of Campbell1 and Fletcher ('84, p. 77), it would seem that the facet
is excellently represented in the northern half of the island.

The Cobequid Plateau. — The Cobequid mountain belt runs nearly
due east aud west along a major axis about seventy-five miles long ;
its average and fairly constant width is from nine to ten miles. The
opinion of the Canadian Survey officers has varied with respect to
the age of the rocks composing these old mountains. According to
the map issued by the Selwyn survey in 1886, they are pre-Cambrian,
but Ells ('97, p. 119) has recently expressed the conclusion that
they " may with propriety be regarded as more recent than the pre-
Cambrian," and this is true of the granite intrusions at least, which are
late Devonian or post-Devonian. In any case, however, we know that
the Yange is composed of highly crystalline, complexly folded series of
rocks that are now bounded by a rolling plateau-top from eight hun-
dred to one thousand feet above the sea, some "peaks " reaching eleven
hundred feet in altitude (Chalmers, '95, p. 9). The general surface
of this plateau now stands at an elevation which correlates it well with
the peneplain of the Southern Plateau. The extension of the latter facet
1 Dawson, Acadian Geology, 2nd ed , pp. 564, 685.

84 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

would carry it over much of the Cobequids just tangent to the surface
and into the rest of the massif, nowhere cutt-ng it more than four hun-
dred feet below its highest point. This close correspondence between
the two plateau surfaces is explained by the assumption of one continu-
ous peneplain as once having covered both plateaus and the intervening
space. At the same time, it is possible that long low residuals of the
" Unaka " type l overlooked the peneplain in the region of what is now
the Cobequid belt.

The New Brunswick Highlands. — The lack of good topographic
maps and of numerous accurately determined elevations is especially
felt when the attempt is made to carry the peneplain facet into New
Brunswick. Yet the grouping of facts summarized in the one-quarter
inch map of the Dominion Survey with those derived from written
reports seems to show that the facet is represented over considerable
areas of the upland. The strongly folded Cambrian, Silurian, Devonian,
and Lower Carboniferous beds wrap about the pre-Cambrian core which
runs parallel to Chignecto Channel and the Bay of Fundy shore. From
the eastern end of the range to a point about thirty miles northeast of
St. John, all these rocks stop off at a general level of from 1,000 to
1,200 feet. The surface is gently rolling and very comparable to that
of the Cobequid plateau. Still farther to the wrest it sinks to
the 600-foot contour, where it crosses the St. John River, only to
rise again to 1,200 feet and more on the northern arm of the massive
angle of the New Brunswick ancient crystallines running to Chaleur
Bay. Monadnocks like Bald Mt. dominate the peneplain. Below it
lie entrenched longitudinal valleys presumably of subsequent origin
(Bailey and Matthews, '72, p. 15).

North Mountain, Digby Neck, and Long Island. — Lastly, the
upland surface of the great lava ridge that stretches with interruptions
from Cape Blomidon to Briar Island, a hundred and twenty miles away,
admirably represents the peneplain facet (Plates 4 and 5). Extending
the latter in imagination from the Southern Plateau to the New
Brunswick highland, the nearly plane surface thus produced is found to
be tangent to the summit of the intervening " mountain." From
Blomidon to Digby Gut, the average elevation of the flat-topped and
truncated ridge is about 550 feet, matching well with the 500-foot
sky-line of South Mountain. Southwest of the Gut, the ridge-plateau
breaks up into two subordinate ridges separated by a long valley.
Their crest-liues accord in elevation as they sink to 120 feet on Briar

i Hayes, 19th Ann. Rept. U. S. Geol. Surv., 1899, Pt. II. p. 22.

DALY: PHYSIOGRAPHY OF ACADIA. 85

Island, while the surface of the Southern Plateau across St. Mary's
Bay similarly falls toward the southwest. Apart from these note-
worthy correlations, there are several arguments going to show that
the trap ridge has been once peneplained. These will be best appre-
ciated after a short account of the Fundy trough has been given.

The Bay of Fundy Trough in Geological Time. — The rocks now
exposed between the Southern Plateau and the New Brunswick High-
land are chiefly of Triassic (Newark) age. There is little doubt that
the original extent and detailed structures of these beds cannot be de-
termined with anything like the completeness that characterizes our
knowledge of the New England Trias. It is highly probable that rocks
of this age underlie the bay; but just what relation they bear to the
exposed members on the shore-belts has, of course, not been made out.
Even there our interpretation must be limited because of the drift-
covering about the Minas Basin and in the Annapolis Valley. I have
not been able to find a definite statement of what correlation can be
made of possible observations or of those already published. The facts
regarding the structures and history of these rocks may be noted. The
scarcity of these facts will indicate the evident need of careful field-work
in addition to that already done by Dawson, Bailey, and others.

The trough now occupied by the Bay of Fundy seems to have been
first delimited on the south during the late Devonian, the northern rim
having been defined at the beginning of the Cambrian (Bailey, '97,
p. 107). The estuary was filled with Lower Carboniferous sediments;
these were disturbed by folding in pre-Triassic times, denudation trun-
cated, the new structures, depression ensued, and thick Triassic sands
and gravels were laid unconformably on the Carboniferous. Observa-
tions on the most northerly and most southerly outcrops of the Tri-
assic show that these rocks (there chiefly of conglomeratic nature) mark
old shore-zones. The consensus of opinion is that the Trias never ex-
tended much farther than the present limits of that system.

One of the youngest, if not the youngest, accessible member of the
Triassic series is also one of the most important from the physiographic
standpoint. I refer to the lava-flows and volcanic breccias of North
Mountain and the adjacent islands including Partridge Island and the
prominent Isle Haute isolated in the bay. The lowest bed is an
amygdaloid conformable with the red sandstones and supposed by Daw-
son ('68, p. 93) to have been deposited beneath the sea. Above it is
the massive compact trap which crowns the ridge throughout its ex-
tent. Whether or not this thick flow was overlain by sediments in

86 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Triassic times cannot now be stated ; the ten-foot layer of limestone dis-
covered by Ells ('95, p. 416) on the lava near Scot's Bay is as yet an
unsolved mystery.

Now the present attitude of all these beds shows that they cannot lie
in their original position. From Kingsport to Blomidon and thence
many miles to the southwestward, the dip is toward the north and
northwest, averaging about 15°. On the opposite side of Minas
Basin, it is just as characteristically to the south, as shown in the
traps, sandstones, and conglomerates. This arrangement of dips and
the shape of the curve terminating in Cape Split suggest a synclinal
structure for this portion of the Triassic area : an imperfect " dishing"
of the trap sheet after the manner of the Totoket and other ranges in
Connecticut. That the same structure cannot be ascribed to the forma-
tion farther southwest is clear from Dawson's description of the occur-
rence at Quaco Head in New Brunswick. The soft red sandstones
and conglomerates aggregate 800 feet in thickness (Matthew) and dip
NNE. at angles varying from 25° to 45° (Dawson, '68, p. 108). Sim-
ilarly the dip is north 50° near the contact with the old rocks of the
Cobequid Mountains (Dawson, '68, p. 100). Such a persistent dip of
these shore-deposits toward the crystalline highlands doubtless indicates
profound faulting. It is interesting to note that Bailey and Matthew
('72, p. 218) early described strike-faults in the Southern New Bruns-
wick Trias with "extensive downthrows on the south." Excepting
the suspected low synclinal in the east, the Triassic bods nowhere
exhibit flexures on any but the most limited scale. They have, how-
ever, been so disturbed as to show typical monoclinal structure. No
important duplication of strata has yet been proved, but thei-e are indi-
cations that the general NW dip is due to block-faulting suggestive]}'
like that in the Connecticut Valley. In the long section of the " sea-
wall " at the head of St. Mary's Bay, Bailey ('98, p. 124) describes
many small faults with the upthrow constantly on the north. Bailey
('98, pp. 126, 127; '98, p. 358 ; '97, p. 114) further notes other faults
at Digby and at Granville, and also states his general conclusion that the
trough has been "faulted" in the direction of the axis of the bay.
It is further possible that one or more faults may explain Bailey's
('98, p. .357; '98, p. 12G) discovery of trap conglomerate lying ad-
jacent to the lava ridge of North Mt., the conglomerate having been a
contemporaneous deposit since faulted down to its present position.

The constructional topography resulting from these movements must
have been very different from that of the present time. The sandstones

DALY : PHYSIOGRAPHY OF ACADIA. 87

of the lowland belts probably extended along their planes of stratification
to altitudes higher than the existing summit-line of North Mountain.
The traps must certainly have stretched upward for some distance be-
yond the same line. One evidence for this is found in the independence
of the longitudinal ridge-profile and the dip of the trap ; at Blomidon it
measures 15°, at Digby Gut, 5° to 10°; near Bridgetown the beds lie
horizontal (Bailey, '98, p. 128). Yet we have seen that the upland level
remains very constant at about 550 feet throughout the whole distance
including these three points. The truncation of the constructional ridge
is even more clearly shown in the transverse profiles. A good example
appears in the extended view which one gets toward the southwest from
" Look-off," near Canning (Plate 8). There the upland surface is flat
and nearly level, without a decided slope to the north until the Fundy
sea-cliff is reached, while the dip is about 15° to the north. It is the
perfect level-topped sky-line of a typical plateau, not that of the back-
slope of a tilted lava-block. Such a sky-line would be extremely diffi-
cult to explain as belonging simply to the retreating escarpment of a
trap sheet, wasting during the first cycle following uplift. Nor could
the doctrine of " beveling " be applied here with any success ; more
probable would be the hypothesis of marine erosion.

The key to the problem is to be found in the sympathetic relationship
already described between the upland facet of the Southern Plateau and
this other one of North Mountain. The latter is a residual of the grand
peneplain which we have traced from Cape Sable to the Cobequids, the
New Brunswick Highlands, and beyond. The peneplain was tilted, its
streams invigorated, and thereby narrow valleys were cut in the harder
pre-Carboniferous rocks, and the lowland from Truro to the mouth of St.
Mary's Bay on the soft Triassic sandstones. In oi'der to understand
this and the adjoining lowlands of Acadia, it will be well to determine
as nearly as may be, and more in detail than heretofore, the amount and
directions of tilting which the peneplain has suffered, and thus get some
idea of the constructional l'elief at the beginning of the second cycle. In
thus treating of the deformations of a baselevelled land-surface, it is real-
ized that we are taking another step in the direction of pure theory ; but,
if the peneplain explanation be regarded as correct, the step is necessary.

"Warping of the Upland Peneplain. — As suggested on an earlier
page, the most significant displacement of the peneplain from its original
position near sea-level consisted in a tilt directed about S. 30° E., affect-
ing nearly uniformly all parts of the facet east of a line passing through
St. John and Digby. This differential movement must have been greater

VOL. XXXVIII. — no. 3 2

88 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

than is now registered in actual elevations. Recent studies have shown
that the whole coast of Acadia has been lately drowned to a fairly uni-
form depth of about 250 feet. Adding this figure to present heights
we find that the peneplain must have once stood above sea-level 1,250
to 1,450 feet in Southern New Brunswick, 1,050 to 1,250 in the zone of
the Cobequids, 800 feet along the axis of North Mountain, and at sea-
level some score of miles to seaward of the existing southeast Atlantic
coast-line. The maximum uplift occurred in the New Brunswick High-
land, for there is reason to believe that the facet was on the northwest
warped down toward the great Carboniferous basin. This may have
been in the nature of a revival of crustal energies acting along the same
axis of warp which located the original Devonian basin, and parallel to
which post-Pleistocene faulting has taken place (Matthew, '94, pp. 35, 3G).
A secondary warp, transverse to the first one mentioned, seems to have
affected the facet west of the St. John-Digby line, produced to Cape Sa-
ble. The relatively rapid fall in the crest-elevation of the Triassic trap
ridge from Digby Gut to Briar Island certainly suggests a direction of
tilt different from that of the eastern division of the ridge. Corresponding
thereto is the similar westward slant of the facet from a point twenty
miles east of the St. John River, and on the Southern Plateau west of
Digby. Finally, this westward tilt would explain the low position of the
ancient peneplained rock-surface at Passamaquoddy Bay and in the State
of Maine.

The Second Cycle : the Development of the
Triassic Lowlands.

One immediate result of these warps would be the revival of the rivers
that lay on the peneplain. This would mean the wearing out of valleys
beneath the facet, especially on the less resistant rocks. On account of
stream adjustment inherited from the long preceding cycle, the soft belts
would at the initial stage of the new cycle, be subject to particularly
rapid attack. Since those belts are longitudinally arranged, their re-
spective valleys and valley-lowlands would have Appalachian trends.
Such, in brief, is believed to be the origin, not only of the Cornwallis-
Annapolis Valley (Plate 6), but also of the valleys now drowned to form
the Minas Basin (Plate 10), Chignecto Channel, St. Mary's Bay, and
possibly the Bay of Fundy itself. It is evident that we cannot decide
these questions of origin with all desired fulness, for stream erosion has
here been strikingly supplemented by marine and glacial erosion, and the

DALY: PHYSIOGRAPHY OF ACADIA. 89

effects of all further complicated by possible down-warping along the axis
of the Bay of Fundy.

Of the four factors, glacial erosion may be safely excluded from the list
of important causes on account of its insufficient amount. On the other
hand, some sort of denudation has taken place. In the field it is clear
that the edges of the upturned sandstones and conglomerates in all
these embayed areas are truncated, and in such a way as to suggest
excavation of once filled troughs rather than a down-warping of the
general peneplain.

There can be no doubt that abrasion of the Fundy shores by wind
waves is enormously aided by tidal scour ; in the Annapolis Valley at
least, the wind wave could never have rivalled the tidal as an erosive
agent. It is but natural to inquire as to the value of such tidal action
in explaining the lowlands round about. If the present rate of retreat
of the cliffs on the western shore of the Minas Basin were maintained, it
is very possible, if not probable, that the sandstones of the Annapolis
Valley a'sfar as the source of the Cornwallis River would be reduced to a
surface below the level of low tide before atmospheric erosion could pro-
duce an ultimate plain of denudation over the same area (Plate 5). We
know as a matter of fact that the basin has been considerably enlarged
in Quaternary time. May we believe that the tidal currents working
here, in the Annapolis Basin, in St. Mary's Bay, in Chignecto Channel,
and in Northumberland Strait, would, at the completion of the " tide
cycle" already begun, succeed in producing a submarine plain of erosion
comparable to the once continuous floor of the Annapolis, Colchester,
and Cumberland lowlands 1 Might this be done before any serious
damage had been caused to the present form of the harder rocks of the
Southern Plateau 1 At present, we can see no reason why these ques-
tions should be answered in the negative. Hence, we must conclude
that, if the conditions of the time of excavation were like those of the
present, these lowlands, although in part the result of subaerial ero-
sion, might have been to the greater extent produced by tidal scour
operative at first in a number of rias, and then, at the close of the tide
cycle, in a long sound stretching from what is now the Bay of Fundy to
Northumberland Strait.

But it must be remembered that, if our theory of the upland facet be
correct, the constructional shore-line at the beginning of the second cycle
was far removed from the existing one. St. Mary's Bay did not exist
then. The Bay of Fundy extended no farther to the east than what is
now Digby Gut, and it was narrower than now. The amount of work to

90 BULLETIN : MUSEUM OF COMPAKATIVE ZOOLOGY.

be done in excavating a submarine platform of erosion in the sandstones
was thus vastly greater, both as to thickness and area, than the amount
required to complete the present " tide cycle." Moreover, the tidal cur-
rents would be much less powerful than now, both at the beginning and end
of the longer cycle ; they would probably be similar to the tides of the
Maine shore in the first stage and not much higher in the last. Chal-
mers ("98, p. 19) has calculated that, if the Chiguecto Isthmus were sub-
merged, the tides would not average more than 10 or 15 feet in range,
giving a ratio of scour for the corresponding currents not much greater
than in Massachusetts Bay. Yet the isthmus must have been submerged
if the Triassic and Carboniferous lowlands were in reality due chiefly to
tidal scour. Destroying the conditions of the present unusual tides,
niore normal flood and ebb currents would be called on to do work on a
scale that has not been paralleled elsewhere.

Such submergence obtained during the Saxicava Sand period. The
Fundy region was then depressed "at least" 120 feet according to
Chalmers ('95, p. 18), 220 to 225 feet according to Dana.1 Correspond-
ing to our deduction, the " Champlain " sea-floor, revealed in the exist-
ing deposits at Middleton, shows evidence of relative quiet. Their wide
extent seems to indicate that tidal scour was then powerless to remove
unconsolidated material from the floor of the Annapolis Valley, much
less excavate the Triassic sandstones. Even more striking is the occur-
rence of similar sands in Sandy Cove and on the slopes of the Petit Pas-
sage, in each of which there is a heavy tidal rip. Ebb and flood currents
must have been weak in those days. But, weak as they were, the con-
ditions were certainly more favorable to scour than they would have
been if there had been no open funnel-like Bay of Fundy to furnish
concentration.

Now, in addition to the fact that we do not know how much ability
for excavation on the part of the tides is at our disposal during the time
elapsing after the warping of the peneplain, we have the second great
difficulty that we do not know the exact form of the Bay of Fundy
when the warping was finished. We might imagine that during the long
period of the previous cycle, tidal currents had conspired with atmos-
pheric erosion to open a channel from the Bay to the Gulf of St. Law-
rence. The same result would be reached if the old subsequent valley
were slightly deepened by crustal warping. In either case, with the slow
uplift of the land, these currents might excavate the channel fast enough
to keep pace with the elevatory process, thus affording an example of an

1 Manual of Geology, ed. 4, p. 981.

DALY: PHYSIOGRAPHY OF ACADIA. 91

"antecedent" tidal run-way or sound of great dimensions. Or it might
be that the tilting of the peneplain was so rapid as nearly to obliterate
the Bay of Fundy of that time. Could the weak tidal currents then
deepen and lengthen the remnant of the bay as well as remove the large
amount of land-waste brought to the shore by the revived streams ']
Again, would the minor oscillations of level expected during the long
period following the last great uplift, occasion repeated drowning of the
river-valleys and therewith increased excavating power of the currents, a
power perhaps comparable to that of the present Fundy tides ? Similar
questions would arise in the discussion of the ideal " tide-cycle," which is
yet to be invented. Where so little has been done towards discovering
the criteria for a tidal plain of denudation as distinguished, from a pene-
plain or from platform of wave-erosion, it does not seem possible to get
very far in finding a complete history of the Bay of Fundy. There are
too many unknown elements in the problem to permit of its complete
solution. We are sure that tidal work cannot be ruled out ; we know
as well that certain facts lead to the conclusion that a part, perhaps the
major part, of the excavation has been done by subaerial processes.

Some of these facts have been already noted. In addition, it may be
questioned that the gates opening into the Annapolis-Corn wallis Valley,
are wide enough to permit of the entry of efficient tidal currents to any
considerable distance from the gates, even under the present conditions
of extraordinary tidal ranges. These openings are wider now than they
have ever been before ; yet scouring action is confined to the vicinity of
St. Mary's Bay, Digby Gut, and Minas Channel, and elsewhere deposi-
tion is taking place. Secondly, the development of the Trias at the
sea-wall, at Bridgetown, at Middleton, and at other points in the valley
seems to show a greater sympathy of the pre-glacial bed-rock topography
with pre-glacial St. Mary's, Annapolis, and Cornwallis rivers rather than
with the graded slopes towards St. Mary's Bay and Minas Channel ex-
pected if the tides did the excavation. At least one other large feature
of the unsubmerged Triassic suggests river-work, and it is fair to lay
emphasis on even a single instance of the kind where so much of the
physiographic record is drowned beneath the waters of the bay. I refer
to the long trench which has been hewn out of the amygdaloid lying
between the two trap ridges of Digby Neck. Here, as in the case of
the Annapolis Valley, tidal erosion would be very unlikely to produce a
submarine valley of such length and narrowness. The six-knot flood and
ebb past the southwest ends of Long and Briar Islands does not seem to
be appreciably lengthening, by differential erosion, the small inlets located

[)-J, BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

on the amygdaloid. The trench has, on the other hand, all the appear-
ance of a subsequent valley worn out in the second cycle by a river
already well adjusted in an earlier cycle. Finally, we shall see that the
structure and topography of the Acadian Carboniferous areas will sug-
gest a subaerial origin for the neighboring Triassic lowland. The broad,
low, gently rolling plains of the Colchester and Cumberland districts
flank the Cobequids, and are continuous with the great Carboniferous
lowland of New Brunswick on the northwest. Their discussion will be
postponed until the date of the upland facet has been fixed, for this
date will be found to have an important bearing on the problem of their
interpretation.

Granting a subaerial origin for the Triassic lowland, it is natural
to attempt a scheme for the pre-glacial drainage that conditioned the
erosion. It is evident that the picture must be incomplete. Much
of the drainage must have been longitudinal. Some of it was trans-
verse through courses still well preserved in the notches at Digby
Gut, Sandy Cove, Petit Passage, and Grand Passage (Plate 7). At
least two of these are located on faults causing dislocations in the trap
visible in the field.1 The offsetting of the trap ridge at all four notches
is most simply explained by as many upthrows on the west. It looks as
if these faults were older than the upland peneplain, that either a con-
sequent stream of the former cycle or a later subsequent stream occupied
each fault-zone, and that the notch in each case was deepened about as
fast as the adjoining lowlands were worn out. In like manner, the
Connecticut traversing the Holyoke range in Massachusetts has kept
open its notch, the product of two cycles. Since the recent drowning of
southwest Nova Scotia, tidal scour has somewhat widened the passes.
In passing, it may be noted that the fault at Digby Gut lies close to the
hinge-line about which I have posited the westward warping of the
peneplain.

The Geological Dates of the Peneplains. — In the absence of
the sediments which must have been deposited on the sea-floor during
the formation of the peneplain, it is not possible to deduce directly the
date of the facet in geological time. If I am right in correlating the
upland facets of North Mountain and the Southern Plateau, it follows
that the peneplanation must have occurred in post Triassic times. Beyond
this general fact, the Acadian record will not permit us to go, except as
that record is interpreted in terms of better known regions. Using the

1 Dawson, Acadian Geology, ed. 2, p. 9(3. Cf. Bailey, Geol. Surv. Canada, Ann.
Rep. M., 1896, Vol. 9, p. 131, and Dawson, op. cit., p. 95.

DALY: PHYSIOGRAPHY OF ACADIA. 93

latter principle, the completion of the Nova Scotian peneplain may be
provisionally referred to the close of the Cretaceous period. The grounds
for this reference will be readily appreciated by those conversant with
the similar land-forms of the southern Appalachians, where the dating
of the plains of denudation does not admit of essential doubt. The gen-
eral resemblance as to hardness among the rocks of North Carolina,
Pennsylvania, New Jersey, and Nova Scotia, the striking similarity in
Jurassic constructional topography in the four regions, and the equal
completeness of the planation, lead us to infer in Nova Scotia, as
in the other regions, that the peneplain was finished only after a
large part of post-Triassic time had elapsed. Again, from the study
of sediments in the southern Appalachian, it has been determined that
the disturbance of each peneplained area from its position near base-
level must have occurred in late Cretaceous or early Tertiary time.
Most of the destruction wrought on the peneplains by subaerial erosion
has been carried out during a Tertiary cycle that was not ideally com-
pleted. The opinion of those best qualified to speak on the subject is,
then, that these classic regions of the south have been wasting through
post-Triassic time. This long period has been almost wholly occupied in
the doing of two pieces of work, — the development of a Cretaceous
peneplain covering each region entirely save for local monadnocks, fol-
lowed by the etching out of Tertiary lowlands on the soft rocks and
narrow Tertiary valleys on the hard. In Acadia, we have a parallel
series of events. At present, I can discover no more trustworthy guide
for the dating of upland and lowland facets than in this strong analogy.
As is well known, the recognition of two cycles in New England and
their placing in the geological record have been influenced by similar
evidence and a like analogy. It is of course evident that in Acadia we
are far from our physiographic base of supplies ; and the relations of
these distant fields can bespeak no more than a strong probability
for the conclusions stated.

The Carboniferous Lowlands. — The Carboniferous of Cumberland
County has been described in great detail by Logan and Dawson
(Acadian Geology, ed. 2, p. 150), and we have added information in
the Canadian Survey Report for 1895. The long section on the
shore of Chignecto Bay, including the famous cliffs at South Joggins,
has furnished the key to the whole area as well as to the structures
of the Carboniferous elsewhere in Acadia. These beds aggregate nearly
fifteen thousand feet in thickness and represent all the larger sub-
divisions of the whole Carboniferous system. Iu Permian time, they

94 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

were strongly folded with a general synclinal arrangement along an axis
running roughly parallel to the Cobequid belt, the dips at South Jog-
gins averaging 19° and in other parts from 0° to 45° or more (Plate 9).
Strata characterized by such thickness and so diverse attitudes must
have been covered by a constructional topography of relatively great
strength. That topography is now lost, and a new one has taken its
place. From the railroad station at Wentworth, one looks down over
a vast expanse of this Carboniferous district and is struck by its even-
ness (Plate 9). A vast plain lies before the observer. It is highest at
the three hundred-foot contour where it abuts against the Cobequids.
Farther away, it sinks to an average elevation of two hundred feet, which
persists throughout most of the basin, except where narrow valleys are
incised below the plain. The plain is not absolutely level, but is gently
rolling, as is typically displayed on the railroad from Maccan to South
Joggins and along the Joggins shore.

The rocks of the Colchester district differ from those of Cumberland
in including only the Lower Carboniferous series and therewith a larger
proportion of limestones than occur in the northern trough. The struc-
tures are here more complicated. The beds stand at all angles to the
horizontal plane and are frequently interrupted by faults of large throw.
The great gorge at Truro affords an especially fine view of the steeply
inclined sediments. The present topography is again, however, quite
independent of structure. The lowland stands at an average height
above the sea of about two hundred feet. Gentle ridge-like swells on
the harder beds break the monotony of an otherwise nearly perfect
plain, in which youngish valleys are sunk, comparable to those of
Cumberland County.

Now, in accounting for this great amount of denudation in both dis-
tricts, we have not only to deal with a possible Jura-Cretaceous cycle and
a possible Tertiary cycle ; some place must be made for the enormous
Permian denudation implied by the unconformity between the Carbo-
niferous and the Triassic, and for Triassic denudation on unsubmerged
Carboniferous. Then, too, a mid-Carboniferous period of subaerial con-
ditions must have elapsed, since we find an unconformable relationship
between the Lower and Middle Carboniferous strata. However, not-
withstanding this latitude offered us in placing the dates of the various
erosion-periods which, from time to time, have worn down the construc-
tional Carboniferous relief, I have come to the belief that the existing
lowlands owe their existence essentially to the same Tertiary cycle of
wasting that explains for us the Annapolis Valley.

DALY: PIIYSIOGRArHY OF ACADIA. 95

If the lowland were existent in pre-Cretaceous time, it must have
been submerged to greater or less depth during the Cretaceous cycle,
and sediments of that age must have been laid in the trough. While
most of such a filling might have been removed during the Tertiary
cycle, we should hardly expect that every trace of it would have
vanished. Nor is it conceivable that the present surface of the low-
lands represents down-warped or down-faulted portions of the Cretaceous
peneplain. Neither faults nor warps would so faithfully occur only
where the soft Carboniferous rocks appear ; yet in this part of Acadia
the Carboniferous rocks and the lowlands are coextensive. No fact
regarding the lithology of these sediments is more strikingly brought
home to the field-observer than their similarity to those of Triassic age
in the Fundy trough, — a similarity which goes far to explain the long
delay in transferring the red sandstones of Prince Edward Island from
the Triassic to the Carboniferous division, where recent determinations
would place them. The rocks of both periods are about equally con-
solidated and equally resistant to atmospheric wasting. If the Annapolis
Valley be the product of erosion in the Tertiary cycle, we must agree to
the possibility of similarly profound dissection of the Cretaceous peneplain
in the Carboniferous tracts.

For these reasons, and on account of the agreement of summit-levels
on Triassic and Carboniferous rocks at the head of the Bay of Fundy, it
is highly probable that the lowland surface from St. Mary's Bay to
Northumberland Strait belongs to one great plain of denudation dating
from the end of the Tertiary cycle. This plain is a secondary peneplain,
and is believed to extend over Prince Edward Island and the Carbo-
niferous lowland of Central New Brunswick. Its longitudinal valleys
described by Sir William Dawson are regarded as the product of adjust-
ment. Surmounting the plain are monadnocks like Springhill (610 feet),
Claremout Hill (565 feet), Windham Hill (600 feet) in the Cumberland
district, Indian Mountain north of Moncton, and the 500-foot hills in
the interior of Prince Edward Island. The Cobequid Plateau is of the
nature of a "catoctin." The steep-walled gorges through the plateau
at Wentworth and at Halfway River are conceivably of the nature of
"shut-ins," and due to the erosive work of transverse streams let down
from the levels of the Cretaceous peneplain. The latter notch was low-
ered as fast as the soft Carboniferous rocks were removed in the lowlands
by a stream persistent in its course until glacial times. The rock-basin
of Folly Lake seems to show that the Wentworth notch early became a
wind-gap during the Tertiary cycle. Chalmers ('95, p. 13) suggested an

96 bulletin: museum of comparative zoology.

antecedent origin for the Halfway River Pass, but he has not brought
that hypothesis into relation with a coherent scheme of physiographic
development for the whole region.

Dissection of the Tertiary Peneplain; Drowning and More
Recent History of Acadia.

This peneplain of the second cycle was uplifted probably with differen-
tial movement during the later Tertiary. A new, third cycle was thus
introduced, and numerous valleys were sunk to considerable depths
beneath its surface. This cycle has been interrupted several times by
changes of level, the most significant of which was the positive movement
of sea-level whereby the lower courses of many of the rivers have been
drowned. In this way, the valleys of the Miramichi, Richibucto, Buc-
touche, Shubenacadie, Avon, and other rivers have been turned into
tidal estuaries. The trends of the valleys opening into the eastern end
of the Bay of Fundy seem to show that these valleys belong to one
great river-system. This has since been drowned probably by the same
wholesale down-sinking that explains the estuaries of eastern New
Brunswick, and of the Nova Scotian seaboard. In the light of present
knowledge, it cannot be definitely stated whether Northumberland Strait
represents part of a drowned river-valley or the locus of a northwest-
southeast trough due to crustal warping. Just how far such warping
has affected the Tertiary peneplain cannot be made out without better
topographic data, better maps, than we now have at command.

The general depression of Acadia developing the estuaries was followed
by an elevation of varying amount in different areas. The post-glacial
marine plain of Middleton dates from this last uplift. Much more
extended coastal plains of the .same age fringe Cbaleur Bay, the whole
eastern coast of New Brunswick, and the coast of Maine.

Still more modern are the famous tidal flats of the Fundy trough, the
sand-beaches and bars and the dune-belts of the coast, the fine palisade-
like cliffs which the waves, aided powerfully by tidal scour, are driving
inland, and the resulting marine benches of Minas Basin. Chignecto
Channel (Plate 8), and of the Bay of Fundy proper. Excavation by
tidal scour is now going on apace at Minas Channel, at Digby Gut,
and at the passes on the southwest, where all these gates are every day
opening wider.

DALY: PHYSIOGRAPHY OF ACADIA. 97

Summary.

The attempt has been made in the foregoing sketch to show, first,
that Acadian land-forms may be described in terms of two topographic
facets, each a nearly perfect plain of denudation, interrupted by in-
cised valleys and surmounted by residual hills ; secondly, that there
is evidence to show that the denudation was essentially subaerial and
referable to two chief cycles of geographic development. This evi-
dence, though not so complete, is of the same quality as that used in
the best extant treatments of similar facets in more southerly portions of
the Appalachian system. Finally, the following table will summarize
the very striking parallel which can be drawn between the physiographic
features of Acadia and New England. The similarity between the two
provinces is here expressed in terms of a theory of development, but the
homologies between the greater facets and the details of relief exist inde-
pendently of theory. Extending the comparison to the central and south-
ern Appalachians would from this standpoint of physiographic history
still further establish the organic unity of the whole system from Georgia
to the Gulf of St. Lawrence.

93

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

c- -= .

_. rs 'A

a

V

fc

g

<

o

j

o

o

(25

3

w

OS
CD

£

5

w

~z **"'

>-,

-S3 CD

Q

bxm

a

B -u

-tj

1

o

03 33

s

s

O sf

1*

*

.5

- CO

<

» cS

B 43 —

CO ~ tO

'- °te c

P.W

a

c5 CD

- <J

•* 3^

CD ST t*-(

«! Jp TJ cj CD

? 5 § w g

TO ° — '«

Qo a_s

• f CO 0) 3 C

" c. — -■+3 *

S3 t-1 S3 U O
CD TJ 'JO
"J 31 S3 S 3 .Jj

to > g)

C J?E

c J5.:: a

2o.

o?

= - sr » s

s.5Ph t3 ~.s

fe CO

C tO

►J 2

^- JSoo

" 41 *"

rt ^ -S3

-B S"C-=

oyr*

:_a

C£

Co

,B g ~£
.C,g co<;

^Stj*?

<3*

j:ro

K.SP

O 5 Is

<n CD co

^ _B

Q

<

«S Sw

£i

— <

TJ CO ^

J

-|S»

fci

,«»=a

tG 05 _

co tO.3
C. S3 —

C0_C0 ;2

S a o

O C CD '

F-. CD O

CO p- os
tptu • -;

IN*

to
o

S3 O.

c o

a t£'

to

o

c;-f-

o

-!

2 CD;

o

O - •

O

W

O^j

o _s

<U he 1C. —

S » a •
^"Si to

O fc0^3

C.J o o

o a s> tj

■; joo
opE =

S3 ^J 0J ?J
„ CD CD CD CD

(^ TJ O)

.Si

^

>JtJ

I -g is ?y5 h *

' •- "" B "" 1- w

. co . .-rj a) ? •«
: CJ "«tj - £ 5 u

; m S3 -S3 q; >-, CD

I'd a j) o t. j

j CD cS

::5K

•g trSo
b.SK

CD —

o —

) _2 ^

TjiS.S S.S S

fi< CO

k- : ^ w

2^°

ti g

TJ'^>

§TJ

co T^

j ^: *j CD

73 2 ft

3 o

t.3

bo

e B .J-S

•B.215 S

■^3 CT?
CD^ a S3
A 2 jorS

« CD"'

CO o

CD S3 _

CD CD ^

9 TJ

"5 5

S3^

'" >!

cr. JO
CD to

IS

o .

S3 Sj

g »

•5 <y

'H ^2-=

"S~r-r-r-H

S K a 2

O CD CO B.

h _ qj p

CD CD ^- tn

H

S „ 3

s —

£tj_2 2

« jS'.c H

^ ^h (N 53

J* *- 3 fe

* — 2 o

CD CJ CD CD

CD 3 ^

i>-5|3
i— Jj-s o

SSsLgS

- =0 c
O •> *

"S-s'tj

s«

DALY: PHYSIOGRAPHY OF ACADIA. 99

BIBLIOGRAPHY.

Bailey, L. W.

'82. On the Physical and Geological History of the St. John River, New
Brunswick. Trans. Roy. Soc. Canada, Vol. I. Sect. IV., pp. 281-284.

Bailey, L. W.

'84. Report of Explorations and Surveys in Portions of York and Carleton
Counties, New Brunswick. Geol. Surv. of Canada, Rep. of Prog, for
1882-3-4, Part G.

Bailey, L. W.

'84. On Geological Contacts and Ancient Erosion in Southern and Central
New Brunswick. Trans. Roy. Soc. Canada, Vol. II. Sect. IV., 1S84, pp.
91-97-

Bailey, L. W.

'86. Report of Explorations and Surveys in Portions of the Counties of
Carleton, Victoria, York and Northumberland, New Brunswick. Geol.
Surv. of Canada, Ann. Rep. (new ser.), Vol. I., 1885, Part G.

Bailey, L. W.

'89. On the Progress of Geological Investigation in New Brunswick. Trans.
Roy. Soc. Canada, Vol. VII. Sect. IV., 1889, pp. 3-17.

Bailey, L. W.

'95. Preliminary Report on Geological Investigations in South-western
Nova Scotia. Geol. Surv. of Canada, Ann. Rep. (new ser.), Vol. VI.,
1892-93. Part Q.

Bailey, L. W.

'91. Notes on the Surface Geology of South-western Nova Scotia. Trans.
Nova Scotian Inst, of Sci., Vol. VIII., 1890-94, p. 1.

Bailey, L. W.

'98. Report on the Geology of South-west Nova Scotia. Geol. Surv. of
Canada, Ann. Rep. (new ser.), Vol. IX., 1S96, Part M.

Bailey, L. W.

'97. The Bay of Fundy Trough in American Geological History. Trans.
Roy. Soc. Canada, Vol. III. (second ser.), 1897, Sect. IV., pp. 107-116.

100 bulletin: museum of comparative zoology.

Bailey, L. W.

'96. Some Nova Scotian Illustrations of Dynamical Geology. Trans. Nova
Scotian Inst, of Sci., Vol. IX., 1894-98, pp. 180-194.

Bailey, L. W.

'98. Triassic (?) Rocks of Digby Basin. Trans. Nova Scotian Inst, of Sci.,
Vol. IX., 1894-98, pp. 356-360.

Bailey, L. W., and Mclnnes, W.

'87. Report on Explorations in Portions of the Counties of Victoria, North-
umberland and Restigouche, New Brunswick. Geol. Surv. of Canada,
Ann. Rep. (new ser.), Vol. II. 1886, Part N.

Bailey, L. W., and Matthew, G. F.

'72. Preliminary Report on the Geology of Southern New Brunswick.
Geol. Surv. of Canada, Rep. of Prog, for 1870-71, pp. 13-240.

Bailey, L. W., and Matthew, G. F.

'73. Report of Observations on the Carboniferous System of New Bruns-
wick in the Counties of Queen's, Sunbury and a portion of York. Geol.
Surv. of Canada, Rep. of Prog, for 1S72-73, pp. 180-230.

Chalmers, R.

'84. Report on the Surface Geology of "Western New Brunswick. Geol.
Surv. of Canada, Rep. of Prog. 1882-3-4, Part GG.

Chalmers, R.

'85. Preliminary Report on the Surface Geology of New Brunswick. Geol.
Surv. of Canada, Ann. Rep., Vol. I. {new ser.), 1885, Part GG.

Chalmers, R.

'87. Surface Geology. Northern New Brunswick and South-eastern Quebec.
Geol. Surv. of Canada, Ann. Rep., Vol. II. (new ser.), 1S86, Part M.

Chalmers, R.

'88. Report on the Surface Geology of North-eastern New Brunswick. Geol.
Surv. of Canada, Ann. Rep., Vol. III. Pt. 2 (new ser.), 1887-88, Part N.

Chalmers, R.

'90. Report on the Surface Geology of Southern New Brunswick. Geol.
Surv. of Canada, Ann. Rep., Vol. IV. (new ser.), 18S8-S9, Part N.

Chalmers, R.

'95. Report on the Surface Geology of Eastern New Brunswick, North-
western Nova Scotia, and a portion of Prince Edward Island. Geol. Surv.
of Canada, Ann. Rep., Vol. VII. (new ser.), 1S94, Part M.

Chalmers, R.

'93. Height of the Bay of Fundy Coast in the Glacial Period Relative to Sea-
level, as Evidenced by Marine Fossils in the Bowlder-clay at St. John,
New Brunswick. Bull. Geol. Soc. Amer., Vol. IV., 1S93, pp. 361-370.

Dawson, G. M., and Sutherland, A.

'92. Elementary Geography of the British Colonies.

DALY : PHYSIOGRAPHY OF ACADIA. 101

Dawson, J. W.

'68. Acadian Geology, 2d ed.

Dawson, S. E.

'97. Canada and Newfoundland. Stanford's Compendium of Geography and
Travel, North America, Vol. I.

Ells, R. W.

'81. Report on the Geology of Northern and Eastern New Brunswick.
Geol. Surv. of Canada, Rep. of Prog. 1880-1-2, Part D.

Ells, R. W.

'84. Report on Explorations and Surveys in the Interior of the Gaspe Pen-
insula. Geol. Surv. of Canada, Rep. of Prog. 1882-3-4, Part E.

Ells, R. W.

'85. Report on the Geological Formations of Eastern Albert and Westmore-
land Counties, New Brunswick. Geol. Surv. of Canada, Ann. Rep., Vol.
I. (new ser.), 1885, Part E.

Ells, R. W.

'95. Notes on Recent Sedimentary Formations on the Bay of Fundy Coast.
Trans. Nova Scotian Inst, of Sci., Vol. VIII., 1890-94, pp. 416-419.

Ells, R. W.

'97. Notes on the Archaean of Eastern Canada. Trans. Roy. Soc. Canada,
Vol. III. (second ser.), 1897, Sect. IV., pp. 117-124.

Fletcher, H.

'81. Report on Part of the Counties of Richmond, Inverness, Guysborouuh
and Antigonish, Nova Scotia. Geol. Surv. of Canada, Rep. of Prog.
1879-80, Part F.

Fletcher, H.

'84. Report on the Geology of Northern Cape Breton. Geol. Surv. of
Canada, Rep. of Prog. 1882-3-4, Part II.

Fletcher, H., and Faribault, E. R.

'87. Report on Geological Surveys and Explorations in the Counties of
Guysborough, Antigonish, Pictou, Colchester, and Halifax, Nova Scotia.
Geol. Surv. of Canada, Ann. Rep., Vol. II. (new ser.), 18S6, Part P.

Ganong, W. F.

'98. Notes on the Natural History and Physiography of New Brunswick.
Bull. Nat. Hist. Soc. of New Brunswick, Vol. IV., 1S98, pp. 44-63.
Gilpin, E.

'83. The Folding of the Carboniferous Strata in the Maritime Provinces of
Canada. Trans. Roy. Soc Canada, Vol. I., 1SS3, Sect. IV., pp. 137-142.
Gilpin, E.

'86. The Geology of Cape Breton Island, Nova Scotia. Quart. Jour. Geol.
Soc, Vol. XLIL, 1886, pp. 515-26.

102 bulletin: museum of compakative zoology.

Hind, H. Y.

75. Lee 1'heuomena in the Bay of Fuudy. Canadian Monthly, Sept. 1875,

p. 189.

Jackson, C. T., and Alger, F.

'33. Remarks on the Mineralogy and Geology of Nova Scotia. Mem. Amer.
A rail. Arts and Sci., Vol. I., 1S33, pp. 217-330.

Mackinlay, A., and W.

Map of the Maritime Provinces of the Dominion of Canada. Halifax, n. d.

Mackinlay, A., and W.

Map of the Province of Nova Scotia, including the Island of Cape Breton.
Halifax, n. d.

Matthew, G. F.

'94. Movements of the Earth's Crust at St. John, N. B., in Post-glacial
Times. Bull. Nat. Hist. Soc. New Brunswick, No. XII., 1894, pp. 34-42.

Prest, W. H.

'92. Evidence of the Post-glacial Extension of the Southern Coast of Nova
Scotia. Trans. Nova Scotian Inst, of Sci., Vol. VIII., 1890-94, pp. 143-
147.

Roberts, C. G. D.

'91. Appleton's Canadian Guide-book, 1891, Vol. I.

Russell, I. C.

'92. Correlation Papers : The Newark System. Bull. 85, U. S. Geol. Surv.,
1S92.

Trueman, G. J.

'95. The Marsh and Lake Region at the head of Chignecto Bay. Bull. Nat.
Hist. Soc. New Brunswick, Vol. IV., pp. 93-103.

United States Navy, Hydrographic Office.
Charts of the Acadian Coast-line.

Van Hise, C. R.

'92. Correlation Papers: Archean and Algonkian. Bull. 86, U. S. Geol.
Surv., 1892.

Walcott, C. D.

'91. Correlation Papers: Cambrian. Bull. 81, U. S. Geol. Surv., 1S91.

DALY : PHYSIOGRAPHY OF ACADIA. 103

EXPLANATION OF PLATES.

PLATE 1.

View of the lake-covered Southern Plateau at Maitland, eighteen miles south of
Annapolis, Nova Scotia.

PLATE 2.

View of " South Mountain," the northern edge of the Southern Plateau, at Digby,
Nova Scotia; the Annapolis Basin in the middle ground.

PLATE 3.

Gaspereau Village and Vallej' of the Gaspereau River, which lies intrenched in
the even-topped Southern Plateau. Looking south.

PLATE 4.

North Mountain, seen across the drowned Tertiary lowland, from Wolfville, N. S.-
PLATE 5.

View of North Mountain at Cape Blomidon, showing the flat sky-line. In the
middle ground is seen the Tertiary peneplain level and the sea-cliff in tilted
Triassic sandstone.

PLATE 6.

The Triassic lowland seen from " The Look-off," near Canning, N. S. The even
sky-line of the Southern Plateau in the background. Compare the view of
the Connecticut Valley Lowland published in the 18th Ann. Rept., U. S. Geol.
Surv., Part II., Plate 1.

PLATE 7.
Digby Gut, looking south toward Annapolis Basin.

PLATE 8

View of North Mountain, looking west from " The Look-off." near Canning, N. S.

Sea-bench at South Joggins, N. S.. seen at half-tide, looking west.
vol. xxxvni. — no. 3 3

10+ BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

PLATE 9.

View of the Cumberland lowland from Wentworth Station on the Cobequid Pla-
teau. Looking northeast.

Tilted Carboniferous sandstones and shales overlain by drift at South Joggins, N. S.
Looking east.

PLATE 10.

Minas Basin and the Triassic lowland, seen from North Mountain, near Canning,
N. S.

PLATE 11.
Map showing the occurrence of Cretaceous and Tertiary peneplains in Acadia.

'f.

Q
<

h
<

h

b

<!

W

<;

Q.

2

K
W
X

D
O
W

—

J

W

►J
<

>

m

X
(X
U3

<

o

►J

-

w

X)

m

Daly. Acadia.

Plate 8.

NORTH MOUNTAIN FROM "LOOK-OFF."

THE HELIOTYPE PRINTING CO.. BOSTON.

SEA-BENCH, SOUTH JOGGINS. N. S.

Daly. Acadia.

Plate 9.

CUMBERLAND LOWLAND.

THE HELIOTYPE PRINTING CO., BOSTON.

STRUCTURE AT SOUTH JOGGINS. N. S.

g
<

<

3 -s O 5

,« ~ 0> P ~

"-> ^ -n ~ * a F

-5-5 hV i! - =3 .2 « W

C 3 T3 « « -5 C

C 3

5 S

o. a

CS s

c a

c

'S

a

.=

*

c

o

=

0.

a

B

c

< <
i i

r

■-

1

u

1

0

1,

'J.
1

OS

s

•f)

M OS

cq

_I

X

w

>

«

<! <

d

J

0

d

u

d ai

* %

M « c S.

2 -

2

(J T3

H

re

Q

The following Publications of the Museum of Comparative Zoology
are in preparation : —

Reports on the Results of Dredging Operations in 1877, 1878, 1879, and 1880, in charge of Alex-
ander Agassiz, by the U. S. Coast Survey Steamer " Blake," as follows: —

E. EHLERS. The Annelids of the " Blake."

C. HARTLAUB. The Coniatulffl of the " Blake," with 15 Plates.

H. LUDWIG. The Genus Pentacrinus.

A. MILNE EDWARDS and E. L. BOUVIER. The Crustacea of the "Blake."

A. E. VERR1LL. The Alcyonaria of the " Blake."

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in charge of
Alexander Agassiz, on the U. S. Fish Commission Steamer "Albatross," from August,
1899, to March, 1900, Commander Jefferson F. Moser, U. S. N., Commanding.

Illustrations of North American MARINE INVERTEBRATES, from Drawings by BuRJt-
hardt, Sonrel, and A. Agassiz, prepared under the direction of L. Agassiz.

LOUIS CABOT. Immature State of the Odonata, Part IV.
E. L. MARK. Studies on Lepidosteus, continued.

" On Arachnactis.

R. T. HILL. On the Geology of the Windward Islands.
W. MoM WOODWORTH. On the Bololo or Palolo of Fiji and Samoa.
A. AGASSIZ and A. G. MAYER. The Acalephs of the East Coast of the United States.
AGASSIZ and WHITAIAN. Pelagic Fishes. Part II., with 14 Plates.
J. C. BRANNER. The Coral Reefs of Brazil.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission Steamer
"Albatross," Lieutenant Commander Z. L. Tanner, U. S. N., Commanding, in charge of
Alexander Agassiz, as follows: —

A. AGASSIZ. The Pelagic Fauna. H. LUDWIG. The Starfishes.

The Echini. J. P. McMURRICH. The Actinarians.

" The Panamic Deep Sea Fauna. E. L. MARK. Branehioceriunthus.

K.BRANDT. The Sagittse. JOHN MURRAY. The Bottom Specimens.

" The ThalassicoljB. P. SCHIEMENZ. The Pteropods and Hete-

C. CHUN. The Siphoniiphores. ropods.

" The Kyes of 1 >eep-Sea Crustacea. THEO. STUDER The Alcyonarians.

W. II DALL. The Mollusks.
H. J. HANSEN. The Cirripeds.
W. A. HERDMAN. The Ascidians.

S. J. HICKSON. The Antipathids. E. P. VAN DUZEE The Ha1obatid«e.

W. E. HOYLE. The Cephalopods. H. B. WARD. The Sipunculids.

G. VON KOCH. The Deep-Sea Corals. H. V. WILSON. The Sponges.

C. A. KOFOID. Solenogaster. W. MoM. WOODWORTH. The Nemerteans.
R. VON LENDENFELD. The Phospho- « The Annelids,

rescent Organs of Fishes.

M. P. A. TRAUTSTEDT. The Salpidaj and
Doliolidse.

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletins Vols. I. to XXXV. ;
of the Memoirs, Vols. I. to XXIV.

Vols. XXXVI., XXXVII., and XXXVIII. of the Bulletin,
and Vols. XXV., XXVI., XXVII. of the Memoirs, are now in
course of publication.

The Bulletin and Memoirs are devoted to the publication of
original work by the Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural History, and of work by specialists based
upon the Museum Collections and Explorations.

The following publications are in preparation : —

Reports on the Kesults of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander C. I). Sigsbee, U. S. N., and Commander J. E, Bartlett, U. S. N.,
Commanding.

Reports on the Kesults of the Expedition of 1891 of the U. S. Eish Commission
Steamer "Albatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in
charge of Alexander Agassiz, on the U. S. Fish Commission Steamer
"Albatross," from August, 1899, to March, 1900, Commander Jefferson F.
Moser, U. S. N., Commanding.

Contributions from the Zoological Laboratory, Professor E. L. Mark, Director.

Contributions from the Geological Laboratory, in charge of Professor N. S.
Shaler.

Subscriptions for the publications of the Museum will be received
on the following terms : —

For the Bulletin, $4.00 per volume, payable in advance.
For the Memoirs, $8.00 " " »

These publications are issued in numbers at irregular inter-
vals ; one volume of the Bulletin (8vo) and half a volume of the
Memoirs (4to) usually appear annually. Each number of the Bul-
letin and of the Memoirs is also sold separately. A price list
of the publications of the Museum will be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXVIII.

"M^

GEOLOGICAL SERIES, Vol. V. No. 4.

AN EXCURSION TO THE GllANE CA;T1Y> :V TIT
COLORADO.

By W. M. Davis.

With Two Platks.

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

PRINTED FOR THE MUSEUM.

May, 1901.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 4.

AN EXCURSION TO THE GRAND CANYON OF THE
COLORADO.

Br W. M. Davis.

With Two Plates.

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

PRINTED FOR THE MUSEUM.

May, 1901.

MAY 29 1911

No. 4. — An Excursion to the Grand Canyon of the Colorado.
By W. M. Davis.

TABLE OF CONTENTS.

Introduction

Itinerary

Summary of Previous Work

Newberry

Powell

Dutton

The Rook Series

Local Names, Maps, etc. . . .

The Great Denudation

Two Cycles of Denudation
The Mature Valleys of the Kai-
bab and Coconino Plateaus .
The Landslides of Vermilion and

Echo Cliffs

The Migration of Certain Divides

Divide near Pipe Spring . .

Relation of the Pipe Spring

Divide to the Pipe Spring

Fault

The Cedar Ridge Divide under

Echo Cliffs

Migrating Divides in Arid Re-
gions

The Permian Scarps under the

Shinarump Cliffs ....

The Development of Talus .

Perched Boulders

Spurs and Ravines ....

The Stripped Plateaus . ....

Dates of Displacements ....

The Eastern Flexures ....

The Earliest Flexures . . .

The Kaibab and the Echo

(Paria) Flexures . . . .

The Crags of Echo Cliffs . .

The Western Faults

VOL. XXXVIII. — no. 4.

PAGE

108
110
110
111
112
113
115
115
118
118

120

121
126
126

128

129

130

132
133
134
134
136
139
139
139

140
141
142

PAGE

142
142

The West Kaibab Faults .
The Toroweap Fault . . .
The Sevier Fault Southwest of

Pipe Spring 143

Topographic Effects of Fault-
ing obliterated or reversed

by Erosion 145

The Hurricane Fault .... 146
The Grand Wash Fault ... 147
The Western Monoclinal Flexures 148
The Displacements of the High

Plateaus 150

Origin of the Drainage System . 151
General Explanation by Antece-
dence 151

Replacement of Antecedence by

Other Explanations . . . 152
The Smaller Streams of the

Grand Canyon District . . 153
The Streams of the San Rafael

Swell 154

The Summit Valleys of the

Kaibab 155

The Origin of the Colorado in

the Grand Canyon District . 158
The Geological History of the

Region 159

The Effect of the Flexures . 160
The Effect of the Faults . . . 161
The Bends of theGrand Canyon 164
Speculative Character of the
Preceding Sections . . . 160
The Erosion of the Grand Canyon 167

The Canyon Cycle 167

Comparison of Glen and Marble
Canyons 167

108

bulletin: museum of comparative zoology.

FAOB

Stage of Development of the
Canyon 168

Rapids in the Canyon . . . 168

Junction of Trunk and Branch

Streams 169

The Geological Section in the

Canyon Wall 171

The Two Unconformities . . 173
Correlation of Water Streams

and Waste Streams . . . 176

Cirques, Cusps, and Niches . 178
The Esplanade 180

Two Theories of the Esplanade 181

Comparison of the Kaibab and
the Kanab Sections . . . 181

Eastward Fading of the Es-
planade 182

Relation of the Inner and Outer
Canyons 183

PAGE

Relation of the Esplanade to

the Toroweap Fault . . . 184
Conclusion as to the Origin of

the Esplanade 185

Hints for a Visit to the Canyon . 18(5
Former Climates of the Grand Can-
yon District 187

Diverse Opinions of Early Ob-
servers 187

Moist Miocene and Arid Plio-
cene Climates 188

The Toroweap 189

The Pluvial Equivalent of the

Glacial Period 192

Volcanic Phenomena 192

Summary 195

Bibliography 197

Explanation of Plates 201

Introduction.

In June, 1900, it became possible for me to visit the district of the
Grand Canyon of the Colorado, and to see upon the ground the wonder-
ful features of a region that had long been familiar from the reports
of our governmental surveys. Our party consisted of Prof. R. E.
Dodge of Teachers' College, Columbia University, Prof. H. E. Gregory of
Yale University, Mr. R. L. Barrett of Chicago, Mr. Richard Wetherill
of Pueblo Bonito, N. M., Dr. Tempest Anderson of York, England, and
the writer. We reached Flagstaff, Arizona, by the Santa Fe Western
Railroad on June 3, spent twenty-three days travelling irregularly across
country, and went out from Milford, Utah, to Salt Lake City by a
branch of the Oregon Short Line on June 26. Our itinerary is shown
on the accompanying outline map, Figure 1, with dates of camps, and
in the list of camps given below. We travelled partly in wagon, partly
on horseback, and averaged about twenty-five miles a day. The clouds
of thunder showers were frequently seen in the distance, but we had
rain only twice ; first a few drops in the canyon, June 7, and next a
brisk shower near the Little Colorado crossing on the morning of June
10; the centre of this shower passed north of us, and the muddy
streams from its short-lived down-pour met us as we were ascending
a dry arroyo, or " wady." Many days were almost cloudless and
oppressively hot over noon. The nights were cool, with the exception

DAVIS : THE GRAND CANYON OF THE COLORADO.

109

of one that we spent near the bottom of the canyon, which was un-
pleasantly warm. A brief report upon our trip has already been
published in the "American Journal of Science," for October, 1900.

Figure 1.

Route-map of Grand canyon district. The dotted belt represents the weak lower Triassic
and Permian strata separating the mesozoic area on the northeast from the palaeozoic
area on the southwest. The several blocked plateaus are separated by faults (con-
tinuous lines) or flexures (broken lines). The route followed is marked by a fine
broken line, with numbers to indicate dates of camps in May, 1900. F, Fredonia;
H, H, Hurricane ledge and fault; K, Kanab; M, Mt. Trumbull; P, Pipe spring;
Q, Toquerville; T, Toroweap valley; Y, recent lava mesa. Outline taken from
Dutton's Atlas.

-

110 bulletin: museum of comparative zoology.

Itinerary. — June 4th, 1900, Flagstaff northward to Stokes spring,
at northwest base of San Francisco mountain ; 5th, northward to
the Coconino forest, within four miles of Hance's on the canyon rim ;
Gth, descended from Cameron and Berry's Hotel by Grand View trail
into canyon and spent the night at the level of the lower Tonto shales ;
7th, returned to Camerou and Berry's ; 8th, southward to Hull's spring
on road to Flagstaff; 9th, northeastward to Little Colorado river at
crossing of road from Flagstaff to Tuba; 10th, northward to Tuba;
11th, northward along base of Echo cliffs to Cottonwood tanks; 12th,
still northward along Echo cliffs to Tanner's tanks ; 13th, still north-
ward along Echo cliffs, crossing the Colorado river at Lee's Ferry ;
14th, southwest to Jacob's pools under the Vermilion cliffs of the
Paria plateau ; 15th, west to Jacob's lake on the Kaibab plateau ,
16th, northwest to Fredonia ; 17th, westward to Pipe spring and
south westward to Yellowstone spring near Antelope valley; 18th,
south westward to Trumbull spring at southern base of Mt. Trum-
bull;1 19th, ascended Mt. Trumbull, then southward to Oak spring;
20th, southward to Vulcan's throne in the Toroweap, and back to Oak
spring; 21st, northward to Clay holes; 22nd, northward to Gould's
(Workman's) spring ; 23rd, northward past Toquerville to Kelsey's
ranch ; 24th, northward past Cedar City to Push lake ; 25th, north-
west to Minersville ; 26th, northwest to Milford ; night train to Salt
Lake City.

Summary of Previous Work. — An account of new observations
made in such a district as that of the Grand Canyon of the Colorado,
already well studied by the explorers of our western surveys, naturally
lavs more emphasis on novel interpretations of former observations or
on subordinate matters newly observed, than on the great structural
features of the region or on the principal events of its history. But
whatever of novelty is now to be gleaned in that marvellous region must
rest so immediately on the work that has been already done there that
1 wish at the outset to express the great indebtedness that all of our
party felt to the pioneer work of Newberry, Powell, Gilbert, Dutton,
and Holmes, whose labors have transformed a desert wilderness into
classic ground for the geologist, and whose reports are quoted whenever
it is desired to illustrate all that is marvellous in the way of displace-

1 Trumbull spring is on the slope of the mountain several hundred feet above
its base : at the time of our visit it gave very little water. The place is not to be
recommended as a camping ground. Oak spring, four miles further south, is
much better.

DAVIS: THE GKAND CANYON OF THE COLOKADO. Ill

merit and denudation. The topographical maps prepared by Bodfish
and Eenshawe in 1879 are also of great service to the traveller. The
main conclusions of the earlier explorers are not to be disputed. The
great unconformities at the base of the plateau series, the enormous
volume of nearly horizontal and conformable strata from lower Palaeozoic
to Tertiary, the division of the region into great blocks by displacements,
either faults or flexures, trending about north and south, the great
denudation by which the plateaus bordering the canyon have been
stripped of thousands of feet of strata, the sharp erosion by which the
canyon has been incised in the plateaus, and the superb development of
volcanic phenomena, — all these great features are standard examples
for citation. There are, however, certain subordinate conclusions an-
nounced in the earlier reports which seem open to question, and it is
chiefly to a consideration of these debatable points that the present
essay is devoted.

The following brief summary of certain aspects of the work of three
earlier observers may be of service to the reader.

Newberry, geologist of the Ives expedition to the Colorado river of
the west in 1857-58, ascended the Grand Wash cliffs to the plateaus
from the deserts among the Basin ranges on the south of the river,
descended northward into the Grand canyon near its western end by
the side canyon of Diamond creek, and, ascending again, traversed the
southern plateaus past San Francisco mountain from west to east. He
recognized the fundamental crystalline rocks beneath their heavy un-
conformable cover of palaeozoic strata (pp. 54-58) ; he perceived the
importance and efficacy of ordinary erosive processes not only in the
excavation of the narrow canyons beneath the plateaus by the larger
and smaller streams (pp. 45, 46), but also in the broad recession of the
cliffs upon the plateau (pp. 45, 62), indeed he regarded the opening of
the broad upland valleys on the plateaus, such as that of the Little
Colorado, as " a much grander monument of the power of aqueous
action than even the stupendous canon of the Colorado" (p. 86). He
noted a " slight arching of the strata " in passing from what we may
now call the southern Shivwits to the southern Uinkaret plateau
(p. 58), and a "curve of the underlying rock " on descending from the
Coconino plateau (south of the Kaibab) to the platform of the Little
Colorado valley (p. 61) ; but he denied the occurrence of other dis-
placements, not only in the canyons but also along the north-south
escarpments, saying that "the strata of the table-lands are as entirely
unbroken as when first deposited" (p. 46) ; and this is not unreasonable

112 bulletin: museum of compakative zoology.

when it is remembered that his route led him across the southern
plateaus where the great displacements weaken and disappear as they
come down from the north. He did not demand two periods of erosion
for the sculpture of the plateaus and the narrow canyon ; difference of
resistance in the upper and lower strata seemed to him to account for
these contrasts in the amount of destructive work (p. 62), but he inferred
a more active erosion in. former times than at present; "everything
indicates that the table-lands were formerly much better watered than
they now are " (p. 47, also pp. 62, 76).

Powell in his adventurous expedition down the canyon (1869) and in
his journey over the northern plateaus (1870), discovered the double
unconformity in the Kaibab section of the Grand canyon (a, pp. 212,
213), gave many new details concerning the rock series, and em-
phasized the production of the canyons by erosion in his announce-
ment of the "antecedent" origin of certain rivers (p. 163). He
presented a clear account of the great displacements by faults and
flexures which divide the Grand canyon district into huge " blocks,"
trending north and south (a, pp. 185-190, Figure 73), as well as of the
great cliffs of erosion or retreating escarpments, north of the canyon,
facing south and trending irregularly east and west (a, pp. 190, 191,
Figure 74) ; " the cliffs of erosion are very irregular in direction, but
somewhat constant in vertical outline ; and the cliffs of displacement
are somewhat regular in direction, but very inconstant in vertical
outline" (a, p. 191). Powell does not seem to have felt the necessity
of supposing an uplift of the region between the great denudation of the
uplands and the incision of the narrow canyons (pp. 206, 213), but he
states that " the carving of the canons ... is insignificant when com-
pared with the denudation of the whole area, as evidenced in the cliffs
of erosion " (a, p. 208). The date of the displacements is not very
sharply defined ; when the great denudation began " there were no
faults and no benches " (a, p. 200). The first displacements occurred
after the erosion of valleys had been begun, the displacements were
long continued, and must have been slower than the erosion of valleys
by the principal streams, for the displacements did not modify the
stream courses (a, p. 201). " Though the entire region has been folded
and faulted on a grand scale, these displacements have never determined
the course of the streams. . . . All the facts concerning the relation
of the water-ways of this region to the mountains, hills, canons, and
cliffs lead to the inevitable conclusion that the system of drainage was
determined antecedent to the faulting and folding " (a, p. 198). The

DAVIS: THE GRAND CANYON OF THE COLORADO. 113

recession of cliffs is greater in the plateau blocks of greater uplift than
in those of less uplift (a, p. 191). Volcanic action began with the
period of displacement and continued till after the Grand canyon had
been eroded {a, pp. 201, 94, 131) ; a close relation is inferred between the
lines of fracture and the location of the volcanic vents (a, pp. 6, 94, 196).
An arid climate is thought to have long prevailed, for otherwise the
cliffs and canyons could not have maintained their sharp forms (a, pp.
204, 209, 211). Some additional details concerning the sti'atigraphy of
the plateau series are presented in Powell's " Geology of the Uinta
Mountains " (b, pp. 43, 54, 62, 70).

Button, first presented his conclusions in a report on the "Geology of
the High Plateaus of Utah " (1880) ; they were afterwards elaborated in
his " Tertiary History of the Grand Canyon District " (1882). Reference
is here made chiefly to those points which carry the study Of the region
beyond the stage reached by Dutton's predecessors. Several periods of
uplift and displacement are given geological dates, but in the absence of
the later Tertiary deposits, such terms as Miocene and Pliocene are
used only in a general way (c, p. 192). One of the oldest displacements
is the "VVaterpocket flexure, involving all strata up to and including the
Cretaceous ; it is crossed by the Colorado in Glen canyon, and is
unconformably buried in the northwest by the horizontal Eocene strata
of Thousand lake mountain, one of the high plateaus (a, pp. 44, 288, c,
215). The whole region began to rise in the early Tertiary (a, pp. 14,
c, 219), and several broad gentle swells were at the same time locally
elevated above their surroundings ; one of these was the San Rafael
swell (Miocene) north of the Henry mountains, and another of much
greater size was the Grand canyon district (late Eocene, a, p. 19).
Faults and flexures in the Grand canyon district were of later date.
The whole uplifted region was greatly eroded in Miocene time, and more
slowly in Pliocene time (a, pp. 18-21). The first chapter of erosion,
frequently named " the great denudation " and placed in Miocene time,
witnessed the removal of strata having an average thickness of six
thousand feet from a broad area in the Grand canyon district, reducing
the surface to moderate relief, "a very flat expanse " (c, p. 77, also pp.
119, 224). On such a surface, the volcanic eruptions which had begun
at an earlier period poured forth great flows of lava; eruptions con-
tinued still later, the vents being generally independent of fault lines,
even though new vents were made after the time of faulting (c, pp. 105,
107). The high plateaus of Utah are residual table-lands capped by
Tertiary strata and lavas, remnants of the great denudation that

114 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

stripped the lower plateaus on the south and east (a, p. 22). It was
after the great Miocene denudation that the Kaibah and the Echo cliffs
monoclines began to have a separate existence (a, p. 42, c, pp. 192, 205),
and with these local uplifts there was also a wide-spread and sudden up-
lift, ing of the whole district in virtue of which the deep and narrow
canyons of to-day have been cut (a, pp. 38, 45, c, pp. 100, 191) ; it is by
this uplift of late date that the high plateaus of Utah have reached
their great altitude (a, p. 23). The system of north and south faults,
extending from the Grand canyon northward into the high plateaus, was
closely associated with the Pliocene uplift of the region, although some
of them may have begun earlier and continued later, even as late as the
glacial period (a, pp. 27, 28, 35, c, pp. 94, 116, 187, 191, 226). It was
during a pause in the Pliocene uplift that the esplanade of the "outer
gorge" in the Kanab and TJinkaret plateaus was opened (c, pp. 121,
226, 227), while the inner gorge of the Grand canyon in these plateaus
has been eroded since the later faulting took place (a, p. 37, c, pp. 94,
227). Although some superposed streams are found in the Water-
pocket flexure, where the Tertiary strata have been stripped from it
(a, p. 288), the water-ways are thought to be antecedent in nearly
all cases (a, pp. 16, 17, c, pp. 73, 187, 204, 219) ; San Rafael river and
Curtis creek, both of which cross the San Rafael swell, are instanced as
particularly good examples of their class (b, p. 63). Most of the great
Miocene denudation was performed in a humid climate; then there was
a gradual desiccation, as a result of which a number of tributary
streams disappeared in Pliocene time, leaving untrenched valley floors
high above the present canyon bottom (c, pp. 99, 194, 201) ; to the sur-
viving streams of this dry period belongs the erosion of the narrow can-
yons and the steep cliffs (a, pp. 22, 23, c, pp. 223, 227). A temporary
return to a humid climate occurred in the glacial period, and certain
ravines on the Kaibab and Paria plateaus were then carved (c, pp. 196,
202).

In brief, the work of denudation in the Grand canyon district seems
to require two cycles of erosion : the first, initiated by a broad uplift,
had advanced well towards old age under a humid climate when
the second, characterized by an arid climate, was introduced by another
broad uplift complicated by certain dislocations ; the second was sub-
divided into two episodes by a pause in the uplift, the first episode having
approached maturity, while the second has not passed beyond youth.

The work of other observers, chiefly Gilbert, Howell, and Walcott,
will be referred to further on.

DAVIS: THE GRAND CANYON OF THE COLORADO. 115

The Rock Series. — The following figure of the rock series is here
introduced for the convenience of the reader. It is compiled from
reports by Dutton (a, c) and Walcott (d, p. 50), to which the brief
article by Freeh may be taken as supplementary.1

The strata exposed in the canyon walls reach up to the top of the
Aubrey ; those in the terraces of the High plateaus reach down nearly
to the base of the Trias. The weak beds of the lower Trias and the
Permian, with the resistant Shinarunip sandstone between them, occupy
the border between the base of the great terraces and the stripped sur-
face of the upper Aubrey which extends over so large an area of the
plateaus adjoining the Grand canyon.

Local Names, Maps, etc. — Our habit of accenting the antepenul-
timate syllable of Indian words, whose original accent is on tho penult,
leads to a mispronunciation of a number of names for local features
adopted by Powell, Dutton, and others in the Grand canyon district.
The following list compiled from various sources may serve a useful
purpose in preserving something of the original sound of these words,
as well as in giving their meaning, along with that of some English
names : —

Aubrey: The name of an army officer, given to a valley southeast of
the Shivwits canyon, and extended by Gilbert to the Hue of cliff's north
of the valley and to the upper Carboniferous strata in the cliffs
(or, p. 177).

Coconino : (Variously spelled Cocanini, Coanini, etc.). Name of a
forested plateau south of the Kaibab (Merriam, p. 35).

Grand Wash : " In this wide valley are several lines of drainage, of
which the main one ... is called the ' grand wash.' These washes gen-
erally have the canon form — flat, gravelly bottoms, sloping talus, and
steep escarpment above" (Marvine, p. 196). The same name is applied

1 The name, Vishnu, given to a spur on the Kaibab wall by Dutton (c, p. 148)
and applied to the crystalline schists in the bottom of the canyon by Walcott
(d, p. 50), is to be regretted as out of place in the occidental world. Unkar and
Chuar, as subdivisions of the Grand canyon series, Shinarump as a dividing mem-
ber between the paljeozoic and mesozoic series, Kaibab, Kanab, and the rest for the
blocked plateaus, Toroweap and Paunsagunt for valleys, have much local flavor,
however barbarous and ungraceful they may sound to our ears. But Shiva's and
Vishnu's temples as names for pinnacled spurs of the Kaibab in the canyon wall,
and Vishnu as a name for the fundamental rocks, buried by the plateau-making
strata, are unnaturalized foreigners. The Vishnu schists in the desert bottom of
the Grand canyon are as inappropriately named as Diana's baths in the cold White
mountains of New Hampshire.

116 bulletin: museum of compaeative zoology.

Tertiary

Cretaceous

Jurassic

Triassic

Permian

Carboniferous

Unkar and Chuar Algonkian

Archean ?

Figure 2.
Columnar section of formations in the Grand canvon district.

DAVIS: THE GKAND CANYON OF THE COLOKADO. 117

to the fault beneath the valley floor and to the great cliffs of the Shivwits
plateau bordering the valley on the east.

Hurricane Ledge: "It is related that a storm overtook a party of
Mormon officials while attempting to explore a route for a wagon road
up a gulch which comes down from the upper country, and hence its
name " (Powell, a, p. 187). The same name is given to the fault which
determines the " ledge " or cliffs along the western border of the Uinkaret
plateau.

Kaibab : " Mountain lying down " (Powell, a, p. 185). The highest
plateau bordering the Grand canyon.

Kanab : " Willow " (Powell). Name of a tribe, a creek, a plateau, a
canyon through its middle, and a town on the ci'eek.

Paria : " The Ute name for elk " (Button, a, p. 253). A Triassic
plateau east of the Kaibab, and a creek and canyon in it.

Shindrump : " Capping the cliffs, we find conglomerate, over which
are scattered many fragments of silicified wood, known to the Indians
as the arrows of Shin-au'-av, or Shin-ar'-ump" (Powell, a, p. 190).
"The weapons of Shinav, the wolf-god " (Dutton, a, p. 147). A con-
glomeratic sandstone at the bottom of the Trias.

Shi'-vwits: Shiv signifies spring ; Shivwits, the people of the springs
(Powell). This word is spelled Shiwits, Sheavwits, Scheavwits, in various
reports; the spelling here adopted being given by Powell (a, p. 128).
The westernmost of the blocked plateaus.

Tonto : Spanish name of an Indian tribe, meaning " fool ; " given by
Gilbert to the basal sandstones of the Palaeozoic series.

Toro'weap : " A clayey locality " (Dutton, a, p. 30). The valley be-
tween the Uinkaret and Kanab plateaus, probably so named from the
fine silt deposited on the valley floor back of Vulcan's throne.

Uinka'ret: "Pine mountains" (Powell, p. 199). One of the blocked
plateaus, capped with many volcanic cones and flows.

The chief features of the district are shown on eight sheets of the
topographical map published by the United States Geological Survey :
St. Thomas and Camp Mohave, Nev., Mt. Trumbull, Kaibab, Echo
cliffs, Diamond creek, Chino, San Francisco mountain, Ariz.

Geological maps are given in Dutton's reports on the High plateaus
and on the Grand canyon.

The drawings from which the illustrations of this article are repro-
duced can make no claim to accuracy of detail. They are made up
from hasty sketches in which but little more than an outline was re-

118

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

CO

o

a
O

E

s-
d

1c
00

G

c

o

E

s-

corded. They err, as a rule, towards too great
regularity of structural features, as in tbe view
of Vermilion cliffs of the Paria (Figure 5). As
a whole, they are diagrams rather than pictures.

The Great Denudation.

Two Cycles of Denudation. — Some time be-
fore our excursion I had heard doubts expressed
by a competent and critical observer as to the
necessity of postulating pauses in the uplift of
the region in order to explain the production
of the existing topography of the Grand canyon
district. These doubts were based on the close
association of the general surface of the plateaus
bordering the canyon on the north and south, and
of the floor of the esplanade, with certain resistant
strata ; the first with the upper Aubrey, the second
with the Red wall, as indicated in Figure 3. Omit-
ting consideration of the esplanade for the present,
let us consider the possibility of producing the
broadly denuded plateau and the sharply en-
ti*enched canyon in one cycle of erosion, introduced
by an essentially continuous uplift, without sig-
nificant pause during the movement or supplement
after its close.

It is here assumed that the normal progress of
erosion in a single cycle demands the relatively
rapid deepening of the main valleys by corrasion,
and the correspondingly early attainment of a
graded slope along the valley bottoms ; a weath-
ering of the valley walls, slower at first than the
deepening of the valleys but faster afterwards, yet
always at a decreasing rate ; an associated but
on the whole a slower series of changes along the
minor water-courses ; and with the progress of all
these processes, a gradual advance of rapidly awak-
ening activities to a phase of maximum develop-
ment, followed by a much longer phase of relaxation
and a very gradual attainment of the remote and
ultimate phase of rest. The assumption of a still-

DAVIS: THE GRAND CANYON OF THE COLORADO. 119

stand of the land during all these systematic changes is accepted tem-
porarily, but abandoned as soon as need be in order to meet whatever
other conditions may be demanded in any particular case. The question
here at issue regarding the sculpture of the Grand canyon district is
between an essentially single uplift, rapid or slow but continuous, on
the one hand, and two uplifts separated by a long period of denudation
on the other.

It is evident that a broad denudation of the upper members of a strati-
fied series can be contemporaneous with a narrow trenching of the lower
members only if the former are relatively weak and the latter resistant.
The plateau series is not so simply arranged, for among the strata that
have been broadly denuded are the heavy and resistant Triassic sand-
stones which still stand forth in strong and steep cliffs along their line
of outcrop north and east of the Grand canyon ; while in the walls of
the canyon, especially in the Kaibab section, there are two heavy series
of relatively weak strata, the upper Tonto and the lower Aubrey, which
have already retreated to partly graded slopes. The upper Aubrey
limestone and sandstone and the Red-wall sandstone and limestone,
to whose strength the maintenance of the plateau would have to be
credited on the hypothesis of a single cycle of erosion, do not appear to
possess any extraordinary resistance in excess of that of the heavy strata
which make the retreating cliffs north of the canyon ; and the weak
members of the higher series, occupying the slopes between the retreating
cliffs, do not seem to be notably weaker than the weak strata between
the cliff faces in the canyon, — although exception to this statement
should be made with respect to the unusually feeble blue clays of the
lower Trias. Hence the canyon ought to be much wider than it is,
or the northern "terraces" (cliffs) ought not to have retreated as
far as they have, if the whole erosion had been accomplished in
one cycle. We are thus held to the conclusion that the broad denu-
dation of the plateaus must have been far advanced in an early cycle
before the incision of the canyon was begun in a later cycle of
erosion.

These considerations lend direct support to Dutton's view that there
were two periods of uplift in the Grand canyon district, but it is desir-
able that some additional evidence of the verity of this view should be
found in the shape of facts that are less immediately related to postulated
conditions than are those thus far mentioned. Four groups of facts of
this kind may be noted here, and a fifth will be presented in the account
of the canyon in a later section.

120 bulletin: museum of comparative zoology..

The Mature Valleys of the Kaibab and Coconino Plateaus. —
The upper Aubrey limestone caps the Kaibab plateau and the high
ground opposite to it south of the canyon, to which the name, Coconino
plateau, has been given. Powell refers to the southern highland as a
" companion, or twin plateau " of the Kaibab, but a separate name is
needed for it not only because the Grand canyon is cut down between
the two plateaus, but also because they are separated by a strong mono-
clinal flexure with dip to the northeast, of which some account will be
given in a later section.

The limestone capping of these plateaus is maturely dissected. Broad-
floored, well-graded valleys with gently sloping sides ramify through the
uplands in the most perfect manner, presenting a maturely developed
form even to their heads ; and this in spite of the fact that they are
nearly always dry, for the wash of waste down their sides and along
their floors is accomplished only during the rains and thaws of winter
and the occasional showers of summer. It was of the southern or
Coconino plateau that Newberry wrote : " The surface has been con-
siderably modified by erosion, and now presents many broad and shallow
excavated valleys " (p. 58). Dutton describes the Kaibab plateau as
undulating " with rolling hills and gently depressed vales." The valleys
are open-floored with gentle descent ; they are innumerable, covering
" the whole broad surface of the plateau with an infinite network of
ramifications" (c, pp. 131, 134). Our party saw these well-established
drainage ways both north and south of the main canyon, and we were
much impressed with the maturity of their graded sides and floors, in
contrast to the youthful expression of the precocious canyon. Plate 2
shows one of these mature valley floors in the Coconino plateau : a hill
and farm of the farmer ant are in the foreground. If the upper Aubrey
were so resistant as to have prevented the widening of the canyon
while the heavy Triassic sandstones were stripped away for scores of
miles on the north and south, no such mature valley system should be
carved in the stripped Aubrey surface.

Furthermore, the contrast between the rapid wasting of the cliff in
the canyon walls and the slow change of the mature valleys on the
plateaus strongly suggests that the two processes represent different
cycles of erosion. Dutton notes that, on the Kaibab rim of the canyon,
" we often find an old ravine suddenly cut off on the brink of an abyss,
and the continuation of the same ravine is seen upon the other side of the
amphitheatre " (c, p. 170). A similar encroachment of the canyon upon
the mature valleys of the Coconino plateau may be seen in the neighbor-
hood of Hance's camp. "We could also see, when looking up the canyon

DAVIS: THE GRAND CANYON OF THE COLOEADO. 121

from the southern rirn at Cameron and Berry's, that the upper Aubrey
cliffs which descend into the canyon from a detached part of the
Kaibab plateau that lies south of the canyon are notched by wide-open
mature valleys which descend eastward with the monoclinal flexure that
terminates the plateau in that direction. The valleys evidently once
had a greater extension head ward, but they have now lost some of their
upper length by the widening of the steep-walled canyon; thus recalling
the relation existing between the mature centrifugal valleys of Mt.
Mazama and the sharp cliffs of the caldera that holds Crater lake in
southern Oregon. It is normal enough for the deep-cut canyon to
encroach upon the heads of valleys that descend away from it ; but
if the upper Aubrey had been resistant enough to hold the canyon
narrow while the Triassic sandstones were worn away for scores of miles
on either side, we should hardly expect to find beheaded valleys of a
mature form ; yet such a relation would be easily explained on the sup-
position that the beheaded valleys still retain, little changed, a mature
form that they acquired in an earlier cycle of erosion, while the main
river has destroyed its mature valley of the earlier cycle by incising the
canyon in its floor.

The Landslides of Vermilion and Echo Cliffs. — If the erosion
of the plateaus and the canyon had been accomplished in a single cvcle,
there should be at the present advanced stage in the cycle no signs of
revival in the processes of denudation on the retreating cliffs of the
plateaus. If, on the other hand, there have been two cycles, revivals
should be of frequent occurrence and of systematic distribution. The
wonderful series of landslides that occur along the base of the Vermilion
and Echo cliffs, on the western and eastern sides of the great notch in
the Triassic escarpment that heads up at Lee's Ferry, seem to be features
of this kind. It is noticeable that these slides occur relatively near the
river, while further away the cliffs have a much more mature profile,
entirely without slides. Inasmuch as landslides are characteristic of
the earlier and more energetic stages of a cycle of erosion, it seems
probable that those here described should be associated rather with the
reviving activities in the youthful stage of a second cycle than with the
fading activities in the advanced stage of a first and single cycle.

The neighborhood of Lee's Ferry is especially interesting in this con-
nection. Here the Carboniferous strata of the Marble platform1 dip

1 It seems permissible in the interest of brevity to speak of this plateau as the
"Marble platform," instead of as the "Marble canyon platform." In the same
way, the Echo cliffs monocline will be called the Echo monocline.

VOL. XXXVIII. — no. 4 2

122

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

gradually northeastward and descend below the level of the river, and
the removal of the weak Permian and lower Triassic strata causes the
valley to widen thereabouts, so that it loses the canyon form that is so
marked further up-stream in Glen canyon, where the river cuts the

Figure 4.

Sketch map of the landslide district, southwest of Lee's Ferry. Triassic escarpment,

hacliured, above Permian belt, dotted. Reduced from Dutton's Atlas.

heavy Triassic sandstones, and farther down-stream in Marble canyon,
where the river has reached the resistant Carboniferous sandstones and
limestones. Had the river cut down its chanuel continuously in a single
cycle of erosion, the retreat of the Trias on either side of the notch in

DAVIS : THE GRAND CANYON OF THE COLORADO.

123

the Triassic escarpment should by this time have given the bordering
cliffs — Vermilion on the west, and Echo on the east — a relatively
stable profile, such as they usually have elsewhere ; but as a matter of
fact, it is precisely in the notch of the escarpment that the cliffs are
most unstable and that landslides are most numerous. The cliffs have
not retreated here far enough to allow the weak underlying strata —
especially the blue clays of the lower Trias — to be concealed beneath
a graded slope ; it is because the cliffs are sapped by the rapid removal

Figure 5.

Landslides of Vermilion cliffs. The Triassic cliffs rise to the rim of the Paria plateau.
Monoclinai slides lie on the Shinarump bench, one of whose promontories is seen in the
centre and another on the right, half smothered in tumultuous slides that descend to
the plain of Marble platform in the foreground. Drawn from rough sketch.

of the clays that the landslides result. Yet twenty miles northwest of
the river, in House-rock valley between the Kaibab and Paria plateaus,
a graded basal slope is well established beneath the cliffs and no slides
occur. The same is true along the foot of the Echo cliffs, twenty miles
south of Lee's Ferry ; the weak blue clays are there concealed under
a well-graded monoclinai valley floor, and slides are wanting. It is only
as the river is appi-oached that the clays are laid bare, and that, with

124 bulletin: museum of comparative zoology.

their appearance, huge landslides become abundant. Just the reverse
relations should obtain if the' erosion of the region had been accom-
plished in a single cycle.

Imposing as these landslides are to one who follows the road along
the base of the cliffs, they must be considered only as subordinate
details among the colossal cliffs and platforms of this large-featured
11. for no mention of them is found in the reports of Powell, Gilbert,
and Button, although all these observers describe the Triassic escarp-
ment from which the slides have slipped down. Some further account
of them may therefore be given here.

In the Vermilion cliffs of the Paria plateau bordering the northern
or middle part of House-rock valley, the upper members of the Triassic
sandstone retreat somewhat, leaving the lower members to stand forth
in a bench, and thus making two steps of mature expression in the
descent from the plateau to the valley. Here the blue clays of the
lower Trias and the underlying weak Permian layers are all con-
cealed beneath the washed alluvium of the graded valley floor. If one
looks down on the valley from the heights of the Kaibab on the west,
an axial arroyo is seen to separate the wash of red Triassic waste on
the east from that of gray Carboniferous waste on the west. In the
southern part of the valley near House-rock spring, the blue clays be-
gin to appear, and at once the lower bench of the cliffs breaks down
here and there in normal landslide form; that is, the gentle dip of the
cliff-making strata in the Paria plateau is steepened in the slide, so
that the fallen mass takes the form of a ridge, more or less shattered
and disordered, yet with a monoclinal structure apparent enough ; the
back slope of the ridge descends to an uneven depression along the base
of the refreshed cliffs, while the outcropping face of the ridge descends
to a confused mass of jumbled mounds in which the weaker basal strata
are greatly disordered. The monoclinal slides become more nearly
continuous and gain a formidable volume near the southern corner of
the Paria about Jacob's pool; the harder strata that once formed the
lower bench of the cliffs have here dropped down several hundred
feet ; but their forward movement is much greater, often being the
better part of a mile. Passing around to the southeastern face of
the Paria. the lower part of the cliffs consists of Permian strata, and
the Shinarump sandstone (base of Trias) caps a conspicuous basal bench
and cliff, splendidly carved. Here the slides are of two types. If their
forward movement does not carry them over the edge of the Shinarump
i. they preserve the relatively orderly monoclinal form already

DAVIS : THE GRAND CANYON OF THE COLORADO. 125

noted ; but when the slides advance far enough to fall over the bench,
they flow down its froutal slope and sprawl out on the platform beneath
like gigantic lava floods, as in Figure 5. The streams of rock-waste
sometimes descend by a re-entrant notch in the front of the bench, like
the lava flows in the notches of the upper Aubrey cliffs on the west side
of Toroweap valley ; they sometimes advance in large volume and
cascade over the front of the bench, smothering salient as well as re-
entrant of the Shinarump. The different colors of successive strata are
normally arranged in ordei'ly bands on the face of the cliffs ; reddish
gray sandstones at the top, then red-and-yellow-brown sandstones —
vermilion seemed to us too strong a name for these cliffs, and " freight-
car red" was suggested as a substitute — blue in the lower Triassic clays,
chocolate brown in the Shinarump cliff, and banded creamy gray in the
upper Permian slope. Something of banded colors may still be seen in
the slides that have moved down and forward in orderly fashion to form
the monoclinal ridges, but the color bauds in the slides of greater forward
advance become greatly confused when the slides sprawl down the Shina-
rump bench, and a most curious patchwork of reds, yellows, blues, grays,
and browns is the result. Most of the slides seem to be old enough to
have suffered some erosion since their descent, yet they are young and
large enough to be very conspicuous elements in the landscape. They
are easily recognized when seen at a distance of eight or ten miles from
the summit of the Kaibab, where the road crosses it westward from
House-rock spring.

If we now cross the river and go some twenty miles south of Lee's
Ferry, the blue clays are concealed under the evenly graded floor of
a monoclinal valley that follows the foot of the Echo cliffs ; the Shina-
rump sandstone rises on the western side of the monoclinal valley, so as
to overlap the eastern part of the Marble platform. The Triassic sand-
stones are, however, bolder in this part of the Echo cliff's than in the
Vermilion cliffs of the northern part of the House-rock valley; they
here expose a larger proportion of bare rock cluttered over with scanty
and very coarse waste, not of orderly enough arrangement to be called
a talus. Passing northward towards the river, the Shinarump sandstone
crosses the valley obliquely as a low ridge, thus shifting from a cuesta-
like attachment on the border of the Marble platform to a basal bench
and cliff at the foot of the Echo cliffs ; at the same time, the monoclinal
valley shifts from the blue clays of the lower Trias to the weak Permian
strata. Thus the blue clays come to outcrop on the slopes beneath
the red sandstones of the Echo cliffs, and at once the landslides begin.

126 bulletin: museum of comparative zoology.

They are not so large or so numerous as those by Jacob's pool in the
cliffs of the Paria, but they are of perfectly similar development when
they occur. At first only forming monoclinal ridges above the Shina-
rump bench, they soon sprawl forward and downward on the broad plat-
form beneath the bench. Among many others, one sprawling slide that
advances at least a mile is a very striking feature to the traveller who
follows the monoclinal valley road northward towards Lee's Ferry. When
first seen at a distance of eight or ten miles, the slide seems to be a vari-
colored ridge or spur, stretching forward in abnormal position from the
base of the Shiuarump bench ; on a nearer view, its origin is evident
enough, and when seen from in front, so that its full breadth is recog-
nized, its lower part takes the form of a huge paw, below a wrist that is
narrowed at the fall over the bench. We did not recognize any wrinkles
ploughed up in front of the slides, such as Russell has described in asso-
ciation with certain great landslides in Washington (b. 47).

All this seems inconsistent with the scheme of one cycle of erosion,
but perfectly consistent with the scheme of two cycles : the first having
witnessed the development of a floor of weak Permian and lower Trias
over the northeastern part of the Marble platform, with a border of
maturely stable cliffs of Triassic sandstones on the north and east • the
second having witnessed the incision of the Marble canyon, along with
which came a rapid removal of the weak Permian strata down to the
more resistant. Carboniferous members, and with this an accelerated
recession of the Triassic cliffs in the Lee's Ferry angle. It may be well
to say again that the " scheme of one cycle" is an essentially single
and continuous uplift, during and after which the denudation of the
a has been accomplished in a single and persistent effort of the
destructive forces; while the "scheme of the two cycles" involves two
uplifts separated by a pause of relative quiescence, during which the
great denudation of the plateaus was well accomplished ; the incision of
the canyons, the removal of weak strata from the denuded plateaus,
and the revival of certain activities, such as have just been described,
being the result of the second uplift.

The Migration of Certain* Divides. — Among the minor evidences
of a modern revival of erosion, mention may be made of two examples
of migrating divides in localities where a stable water-parting might
have been expected if the erosion of the region had been accomplished
in a single cycle.

Divide near Pipe Spring. — One example of this kind lies in the weak
lower Triassic area west of Pipe spring and south of the Vermilion

DAVIS: THE GRAND CANYON OF THE COLORADO.

12'

cliffs. It is crossed by the road leading from Kanab and Fredonia to
St. George. The divide does not here separate well-defined valleys, but
marks the boundary between two long and broad graded platforms of
gentle declivity, one slowly descending eastward and southeastward to
Kanab creek, the other westward to a dry branch of Virgin river.

The eastern platform heads westward in a sharp slope or " wall " of
weak lower Triassic shales, two or three hundred feet high and locally
carved into bad-land spurs, forming a great westward curve, as if a large
bight had been gnawed into the head of the western platform, whose
border is appropriately concave. Evidently the eastern slope is gaining
area by undercutting the western. When we saw first this wall on
looking west from Pipe spring, it was perplexing, as there was no struc-
ture known in the lower Trias by which it could be explained ; and
moreover its even profile, descending gradually to the south, bevelled

Figure 6.

View from east of Pipe spring, showing "wall" of migrating divide in the background.
P, Pipe spring; T, Triassic plateau; L, landslide; S, Shinarump mesa; V-R, Sevier-
Toroweap f.iult-line. Constructed from map and sections.

the gentle northern dip of the strata in a curious fashion (Figure 6).
But later in the day, when we looked northward to the divide from a
point on the road between Pipe spring and Mount Trumbull (S, Fig-
ure 6), the origin of the wall was easily discovered. The two graded
platforms were then seen in opposing profiles, the eastern heading under
the western, as in Figure 7. Indeed, the northern part of the western
platform could be seen to extend several miles eastward around the
head of the eastern platform, forming a bench that obliquely ascended
the front of the Triassic cliffs independent of structural guidance, un-
til it must have been a good thousand feet above the eastern plat-
form. I have never seen so fine an example of the relation between
two drainage slopes, one encroaching on the other. It was a disap-
pointment that our hurried movement made closer study of this area
impossible.

123

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

It is noteworthy that here, as in the neighborhood of Lee's Ferry,
landslides have taken place where the blue clays have been uncovered
by the encroachment of the eastern platform on the base of the Triassic
cliffs. This was the case just west of Pipe spring, where the former
head of the western platform has been deeply undercut. The blue clays
were seen here and there along the base of the refreshed cliffs, and
several landslides were noted in the same small district (L, Figure 6).

Relation of the Pipe Spring Divide to the Pipe Spring Fault. — An-
other peculiar feature of the eastern platform near Pipe spring is that
it crosses the line of the Sevier-Toroweap fault (see that section) with
unbroken grade, and that its descent is against the heave of the fault ;
that is, from the relatively depressed block of the Uinkaret to the rela-
tively uplifted block of the Kanab plateau. No topographic indication

Figure 7.
View north near Pipe spring. Vermilion cliffs of Triassic plateau in background: the
eastward dip of strata toward the Sevier-Toroweap fault-line is seen on the right, Pipe
spring being near the end of the cliffs. The "wall" of the migrating divide sweeps
through the middle of the landscape. Drawn from a sketch.

of the fault is preserved at this locality, because the weak lower Trias-
sic shales on the west are brought against the weak Permian shales on
the east. The obliteration of the fault, as far as the relief of the sur-
face is concerned, is an unusual feature of the plateau region, according
to Datton ; for while he noted the obliteration of the Shinarump escarp-
ment in the neighborhood of Pipe spring (c, p. 80) he did not connect
its disappearance with any dislocation, and he elsewhere remarked that
" every fault in the district is accompanied with a corresponding break
in the topography" (c, p. 130). It should be stated that we identified
tbf geological formations mentioned here and elsewhere entirely by
their location, their lithological features, and their succession, as de-
scribed in Dutton's reports, which in this way, as in so many others,
were of the greatest service at every turn. Most of the formations are

DAVIS: THE GKAND CANYON OF THE COLORADO. 129

astonishingly bare of fossils ; we saw only an occasional Productus in
the limestones of the Kaibab.

The migrating divide between the two platforms near Pipe spring
seems to indicate a recent aud active revival of a contest that should,
in a single cycle of erosion, have been settled long ago, but that might,
on the other hand, be still in progress in the present youthful stage of a
second cycle following the late maturity or the fine old age of the first.
It is as if the news of the uplift that enabled the Colorado to incise its
canyon beneath the plateaus had but lately reached the Pipe spring
district. We may, therefore, suppose that a stable divide between the
western branches of Kanab creek and the eastern headwaters of the
westward drainage once existed somewhere to the east of the present
unstable divide. If the stable divide were reconstituted in the position
indicated for it by the remnants of the west-sloping graded platform,
the east-sloping platform would necessarily stand above the level that it
occupies to-day. This would demand the reconstitution of a graded
Permian platform several hundred feet above the level of the Kanab
plateau southeast of Pipe spring, a position that is entirely consistent
with certain expectations as to the altitude of the ancient lowland of
denudation produced in this neighborhood by the long-continued denu-
dation of the first cycle of erosion, as will be shown further on. A
reason fur the present advantage of the eastern drainage may be found
in some slight tilting associated with the uplift by which the second
cycle was introduced ; or it may be found in the lower stand of the
resistant upper Aubrey strata in the nearly level structures of the
Kanab plateau than of the same strata on the somewhat upturned west-
ern side of the Uiukaret plateau ; but no decision can be announced
without much more field work. The evidence furnished by this exam-
ple of a migrating divide as to the date of the Sevier-Toroweap fault
will be considered in connection with the fault lines.

The Cedar Midge Divide under Echo Cliffs. — A second example of a
migrating divide occurs in the monoclinal belt of the weak blue lower
Trias clays already described as running along the base of Echo cliffs.
The belt is followed by a wagon-road from Tuba to Lee's Ferry, first
slowly ascending the graded floor of a south-discharging monoclinal
valley, then slowly descending the graded floor of a north-discharging
valley ; the divide between the two valleys being known as Cedar
ridge. The divide is, however, not a ridge in any proper sense. The
southern valley undercuts the head of the northern one, and the two
are separated by an abrupt southward slope, exposing the blue clays

130

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

that are elsewhere completely hidden under alluvium (Figure 8). It is
very clear that the southern valley is in process of being lengthened by
beadward growth or retrogressive erosion at the expense of the northern
valley. Southward from the divide, there are terraces on the side of
the deeper valley, attesting the once greater height of its floor. It was
not possible for us to determine surely any local or general cause for
the change that is here so actively in progress. It may, perhaps, be in-
dependent of the uplift in response to which the larger streams have
cut down their canyons ; nevertheless, we seem to have here as before a

■■■" t ' r^'T^n^w;t-!-r°rvTr

SS^^^g^^gggfe^

/

.* \

Figure 8.

Migrating divide at "Cedar Ridge" under Echo cliffs. The geological section here seen
includes weak Permian strata under the Shinarump sandstone in the foreground; weak
Triassic strata in the monoclinal valley at mid-distance, and the heavy Triassic sand-
stones of Echo cliffs in the background. The valley with shaded floor on right (south)
is encroaching upon higher-floored valley on left (north). Constructed from field notes.

revival of activities that should have long since become quiescent if the
erosion of the region had resulted from a single uplift.

It may be that a similar migration is in progress at the divide be-
tween the northward and southward drainage of the monoclinal House-
rock valley, but as this divide lay several miles north of our road we
could not examine it.

Mirjratliitj Divides iii Arid Regions. — It should be recalled in this con-
nection that the migration of divides in general depends only indirectly
on the action of streams ; it is directly the work of weathering, creep-

DAVIS: THE GRAND CANYON OF THE COLORADO. 131

ing, and washing on the slope above the head of some deep-cutting
stream. Water acts on such a slope only at time of rains and thaws.
All this is strikingly illustrated in the plateau region, for here even the
stream beds are habitually dry, and the whole process of rain and
stream washing is intermittent and exceptional. But when the streams
work, they work furiously, as is proved by the coarse stony fans of
gentle slope at the ravine mouths on the flanks of the Kaibab and else-
where. A very striking example of this kind was seen at the mouth
of a ravine near the road from Flagstaff to Tuba, about ten miles south-
west of the Little Colorado crossing. The ravine was sharply eroded
in the sloping monoclinal surface by which the uppermost Aubrey
layers descended eastward under the red shales and sandstones of the
Permian ; the lower ground opposite the mouth of the ravine was
strewn over with huge limestone boulders that had been swept forward
at time of flood.1

On the bare clay slope of Cedar ridge and on the red shale wall west
of Pipe spring, erosion is accomplished by weathering and rain-washing
rather than by stream corrasion. In spite of the aridity of the plateau
climate, rain-washing is important ; for the general absence of vegeta-
tion allows a single shower to effect a significant amount of work. The
bare slopes retreat speedily, and the gentle grade of the valley floor
or platform at the base of the slope is actively prolonged headward.
The processes here at work resemble very closely those by which the
recession of cliffs is accomplished : little destructive work is done on the
graded platforms above or below the bare cliffs; nearly all the erosion is
effected by the erosion of the cliff face and under-slope, as Dutton has
pointed out (c, pp. 64, 221). But while cliff faces are determined by
the outcrops of resistant strata, the bare slopes at the divides above
described are independent of structure and are determined only by the
difference of altitude of two graded valley floors or platforms. The
bare slopes retreat rapidly while the graded platforms are hardly
changed.

There is no reason for thinking that the average declivity of the bare
slopes at these divides is notably different to-day from what it long has
been or what it long will be. Hence there should be some systematic
relation between the descent of waste from a bare slope and the re-

1 It may be noted that if the transition here noted from limestone to yellowish
and red shales be taken as marking the division between upper Aubrey and Per-
mian, then the boundary of these formations is placed about five miles too far east
in this locality on sheet XXIII. of Dutton's Atlas of the Grand canyon district

132 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

moval of waste cm the graded platform beneath ; and the relation must
evidently be one of equality. The slopes have been described as
" bare," and so they are in a first general view, but in reality they are
cloaked with somewhat loosened clay or shale, a little displaced from its
normal position, still revealing the general attitude of the strata, yet
masking the finer details of bedding. It must be supposed that the
creep and wash of loosened waste from the slopes just suffices to keep
the wet-weather streams busy when they are running in the channels on
the graded floor; and hence that we here have one of the many exam-
ples of a natural equilibrium between capacity for work and work to be
done. Similar examples of equilibrium probably occur between the
descent of waste on certain dissected mountain slopes and the removal
of the waste across the graded rock-floor platforms that stretch forward
from the mountain base, as described by Penck for the subarid moun-
tains north of Madrid in central Spain (p. 132) and by McGee for the
arid mountains of the Sonoran district of Arizona and Mexico (6, p. 91,
Plate 12). The basal angle between the slope and the platform seem
to be more sharply defined in arid than in humid climates, probably
because of the dilferent values of the various agencies for soil produc-
tion and removal in the two cases.

The Permian Scarps under the Shinarump Cliffs. — The compara-
tively small quantity of waste on the Permian scarps beneath the
Shinarump cliffs of certain fine mesas in the northern part of the
Uinkaret plateau suggests a recent revival in the process of sapping ;
for it is difficult to understand how cliffs that have receded as far as
these — forty miles from the canyon — should not by this time have
their under-slopes well sheeted over with waste, if the retreat had been
in a single cycle. The Permian scarps are not always bare : for exam-
ple, those enclosing a valley opened in the Shinarump mesa ten miles
southeast of Toquerville, and extending from Sheepti'ough to Workman
spring (see Atlas of Grand Canon district, sheet XX.), are cloaked with
a considerable covering of waste, which here and there bears a mask
of vegetation ; but in this case the processes of erosion have been much
delayed by a recent flow of lava on the valley floor (not shown on
Duttou's map). The southern scarps of the same group of mesas, only
about ten miles further south and without lava at their base, are very
bare, exhibiting many delicately colored strata that maintain a horizon-
tal course across beautifully varied spurs and ravines. The contrast
l)i 'tween these two examples re-enforces the suggestion that the bare
scarps may have been much better clothed with waste during some

DAVIS : THE GRAND CANYON OF THE COLORADO. 133

former pause in the activity of the erosive forces, and that their fresh
exposure to-day results from a revival of these processes, such as is else-
where manifested in landslides and in migrating divides.

The Development of Talus. — It must be confessed, however, that the
development of talus-covered slopes in the arid region is a complicated
problem. The quantity of talus material will vary (in so far as it
comes from the capping cliff) with the relation between the thickness
and resistance of the hard cliff-making strata above and the weak slope-
making strata below. The texture of the talus will vary chiefly with
the manner of retreat of the cliff-maker. The completeness of the talus
cloak will vary with the stage of the cycle of erosion. Certain mesas,
perhaps of Permian strata, seen from the Santa Fe Western Railroad,
east of Winslow, have a very coarse and discontinuous talus of lar^e
slabs, derived from the strong but thin cliff-maker, which is there
broken into great fragments by the rapid sapping of the weak strata in
the slope. A lava-capped mesa of Permian under-slope, seen near the
road between Flagstaff and Tuba about ten miles southwest of the
Little Colorado crossing, has a very coarse talus. On the other hand,
certain cliffs of massive red sandstone, on the line of the railroad east of
Gallup, descended to the alluvial floor of their valley without any talus
at their base. Dutton (c, p. 228) has suggested that such a relation
may be explained by the burial of a normal talus under recently
aggraded alluvium, thus assuming the talus to be an essential accom-
paniment of the cliffs. In the example east of Gallup, the absence of
weak basal strata and the habit of massive sandstones to weather by
crumbling, rather than by breaking into blocks, seemed to be at least
as important factors as the accumulation of alluvium. That some cliffs
are habitually free from talus is recognized by Dutton in his account of
the Jurassic escarpment flanking the high plateaus : here " a notable
feature is the absence of talus ; or, if it be present, its very small
proportions " (c, p. 36). What has been said in the section on the mi-
gration of divides regarding the retreat of bare slopes does not particu-
larly apply here, because in the examples now under consideration, the
scarps of weak strata are capped by harder strata; and the habit of
such strata is to conceal themselves under a cloak of waste from their
capping cliffs. Even in scarps of so small a measure of retreat as those
in the Grand canyon, the weaker strata are often largely cloaked with
talus, as is the case with the great slopes of the upper Tonto series
in the Kaibab section, and in the slopes of the lower Aubrey that
descend to the esplanade in the western part of the Kanab section. It

134 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

is true that the Permian scarps of the northern Uinkaret are long,
bein"- from three to five hundred feet in relief, and that the Shinarump
can may be only fifty or a hundred feet thick. Nevertheless, the oc-
currence of considerable sheets of tains on the scarps in the valley east
of Workman spring serves to show first that the Shinarump capping
yields sufficient material to form a cloak; secondly, that the cloak will
quickly accumulate if erosion is retarded at the base of the slope;
and thirdly, that aridity of climate cannot account for the absence of
talus on the bare scarps a few miles further south. Hence it seems
that the bareness of these Permian scarps must be exi^lained by the
recent removal of a talus that once covered them.

Perched Boulders. — There are certain details connected with the
problem of retreating escarpments that are finely illustrated southwest
of Lee's Ferry, where the full height of the Permian scarps with their
Shinarump capping is revealed. Here, as on the Uinkaret, the scai'ps
are bare ; the bareness is appropriate to the active recession of the
scarps already indicated by the landslides of this neighborhood. An-
other item was noted. We frequently saw large blocks of sandstone, ten
to thirtv feet in diameter, evidently derived from the cliff above, and now
perched on pedestals of weak shale a little distance forward from the
base of the scarp, as in Plate 2. The perched blocks correspond to
rock tables on glaciers. The pedestals are from three to fifteen feet
in height ; some of the blocks have lately fallen to one side, and the
pedestals that once supported them are not yet wholly worn away. The
top of the pedestals represents the height of the graded platform at
the base of the scarp when the block rolled down from above. Since
then the platform has been degraded by the height of the pedestals ;
at the same time, the scarp has been pushed back a decidedly greater
amount, and at a rapid rate compatible with the absence of a talus cloak
on its slope.

Spurs and Ravines. — The minute morphology of the carved scarps
affords a pleasing study; it is a subject of subordinate importance,
truly, yet perhaps as deserving of careful attention as the morphology
■ if the leaves of plants. "Where the mesa scarps have a straight front,
the ravines descend in relatively direct and parallel courses with few
branches. Each main ravine receives the waste from but a small length
of cliff front above. The spurs are correspondingly simple, each one de-
scending with little variety of form from top to bottom of the scarp,
although with increasing relief downwards. But where a curved re-
entrant occurs in a mesa front, the ravines converge more or less dis-

DAVIS: THE GRAND CANYON OF THE COLORADO.

135

tmctly towards the base of the scarp ; several upper branches may unite
in a single lower trunk, and thus the lower ravine receives the waste
from a considerable length of cliff face ; at the same time, many spurs
fail to reach the base of the scarp, but terminate acutely between the
forking ravines, as in Figure 9. On the other hand, where a projecting
scarp or promontory stands forth, the ravines diverge, and a single spur
at the top of the scarp may broaden and repeatedly divide in its descent,
assuming a sprawling paw-like form, and including many short ravines
that begin within it. Variations of this kind are repeated systematically
over and over again in the Permian scarps. When the entire height of
the spurs is only some four hundred feet, it may be hardly worth while
to develop any special terminology for their different forms ; but when

•wi-x

Figure 9.

Diagram of Permian spurs under Shinarump cliff. Tapering spurs on the left, split spurs
on the riffht. Constructed from rouirh sketches.

one finds that very similar spurs with heights to be measured in thousands
of feet occur on the dissected slopes of Mt. San Francisco, the value of
a simple terminology becomes more apparent.

There seems to be a curious alternation in the generation of spurs and
ravines where a scanty or insufficient supply of resistant talus fragments
from the cliff-maker is associated with an active erosion of the strata in
the under-slope. As the coarse waste weeps and washes down from the
cliff, it tends to avoid the spurs and to accumulate in the ravines. The
bare spurs are then eroded more rapidly than the protected ravines, and
thus a new ravine may come to be worn down the axis of what was a
spur, while a spur stands forth where the ravine was before. A new
supply of waste then gathers in the new ravine, while the older waste is

136 bulletin: museum of comparative zoology.

in time removed from the new-made spur. Then the process is reversed.
Several examples that seemed to be explainable in this way were noted
on the Permian scarps east of Hurricane ledge ; they are shown in
diagrammatic fashion in Figure 9.

The Stripped Plateaus. — The details presented in the four preced-
ing sections concerning mature valleys, landslides, migration of divides,
and bare scarps, all support Dutton's conclusion that the plateau surface
north and south of the Grand canyon was " planed down to a com-
paratively smooth surface " (c, p. 119) before the erosion of the canyon
was begun ; yet they can hardly be regarded as giving altogether inde-
pendent demonstration of the conclusion, for some of them at least
might all have arisen from some minor uplift of recent date. It is only
when all of the elements of the problem are considered together, as
Button often remarks, that a consistent explanation of the history of
the region can be inferred ; and then its erosion in two cycles seems
essential. It is not desired to assert that two simple uplifts, separated
by a long interval of perfect rest, constitute the diastrophic history of
the region ; but that a long period of relative rest and persistent denu-
dation with respect to a comparatively constant baselevel occurred be-
tween earlier and later periods of uplift and dislocation. Each period of
movement may have been of a considerable duration, complicated by
pauses and intermittent displacements. Some of the earlier movements
may have been separated by long pauses, of which no record now re-
mains ; but nothing less than a long period of relative quiescence before
the last period of uplift seems sufficient to account for the great denuda-
tion of the plateaus and the narrowness of the rapidly widening canyon,
along with the correlated details above described. The date of the faults,
by which the plateau district is divided into great blocks, with respect to
the earlier and later periods of denudation will be considered later.

The original uplift of the region, by which deposition was stopped and
denudation begun, has been with good reason assigned by Dutton to
Eocene time (a, p. 21, c, p. 217). The same author places the " period of
quiescence " in " late Miocene or early Pliocene time" (c, pp. 77, 221) ;
but in the absence of local fossiliferous deposits, by means of which the
greater and lesser denudations can be strictly correlated with the geo-
logical time scale, I shall for the present use the terms, plateau cycle
and canyon cycle, to name the greater and lesser periods in which
the work of erosion has been performed ; and so far as the events in
the history of the district can be arranged in order, they will be thus
dated.

DAVIS : THE GRAND CANYON OF THE COLORADO. 137

Reasonable as the hypothesis of two cycles of erosion seems to be
when all the facts are viewed together, it is difficult to point out rep-
resentatives of the " very flat expanse " that was produced when the
plateau cycle was interrupted by the uplift that introduced the canyon
cycle. Dutton states that the older lavas of the Uinkaret and Shivwits
plateaus cover considerable areas of Permian strata, which at the time
of the early eruptions must have " constituted the general platform [of
the district] in much the same way as the upper Carboniferous now
does" (c, p. 107). Details concerning the Permian platforms are at
present wanting ; but if future observation shall show that they have
no cap of resistant Shinarump sandstone, and that they are as level as
is implied in Dutton's descriptions, then it may be fairly inferred that
they were lowlands at the time of their burial under the lavas ; for the
Permian strata are too weak to permit of the production of a plain of
erosion within their mass at a signi6cant height above baselevel, such as
might well enough occur on the upper surface of a more resistant formation.
Accepting this conclusion provisionally, as the most probable one now
obtainable, it is then reasonable to infer that the Permian lowland once
extended far and wide, and that it was in fact part of the peneplain to
which much of the region had been reduced at the close of the plateau
cycle. To-day the peneplain on the Permian is preserved only where it
was sheeted with lava ; the Permian floor elsewhere visible north of the
broad Carboniferous platform must be referred to the canyon cycle, as in
the district southeast of Pipe spring. It will be an interesting matter
for some future observer to inquire into the stratigraphic relation of the
floors on which the older lavas rest in adjacent plateaus, in order to
determine how they bear on the date of the faulting by which the
plateau blocks are separated. Much more work on the ground will be
needed before this elusive question can be definitely settled.

Since the uplift by which the canyon cycle was introduced, sufficient
time has elapsed for an extensive removal of the weaker Permian strata
from the plateau surface, wherever they lay within easy reach of the
"wash" that must have been actively revived at the opening of the
new cycle : " considerable masses of the Permian were then remaining,
which have since been eroded," as Dutton puts it (c, p. 120). Even
the resistant upper Aubrey strata, revealed by the stripping of their
Permian cover, early in the canyon cycle, have suffered a significant
amount of dissection, as seems to be the case over much of the Kanab
plateau ; but the dissection here is not so mature as that by which the
higher Kaibah is characterized, as will be further considered later.
vol. xxxvm. — no. 4 3

138 bulletin: museum of comparative zoology.

The graded Permian surface in the northern part of the Kanab plateau
to-day may, therefore, be better explained as having been in large part
newly developed with respect to existing local baselevels (such as sills
of upper Aubrey at the heads of branch canyons) instead of as having
been preserved with little modification since the close of the previous
cycle. I am consequently inclined to dissent from Dutton's opinion
that " the broad and slightly varied expanse " of the surface of denuda-
tion which " cuts the strata in such a way that the [low] hills usually
consist of lower Permian strata lying horizontally, while the shallow
valleys expose the Carboniferous" (c, p. 118) is a close representative
of the broad lowland of denudation produced at the close of the plateau
cycle. Truly, " the mean position of the surface of denudation is very
nearly coincident with the dividing horizon between those two forma-
tions " : and if " we successively visit districts considerably apart, we
shall find in one of them [the southern] that the mean position of the
surface of denudation is below that horizon so far that no Permian
rocks appear in the hills ; in the other [the northern] it is so far above
it that no Carboniferous rocks appear in the valleys" (c, p. 118) ; truly
also the gentle bevelling of slightly inclined strata over large areas is
often accepted as evidence of peneplanation, as by Campbell and Men-
denhall for the plateau of West Virginia (p. 483), and by Philippson for
the plains of Russia (p. 40) ; but in the plateaus of the Grand canyon
district another interpretation seems permissible. It should be noted
that only two formations are referred to in the above-quoted extracts
from Dutton's report ; the lower one (Carboniferous) resistant ; the
upper one (Permian) relatively weak. These formations rise gently
southward. The great Carboniferous platform (upper Aubrey) on either
side of the canyon in the Kanab district, here chiefly under considera-
tion, is free from Permian remnants over all those large areas whose
drainage is actively tributary to the main river ; while the continuous
Permian cover, apart from certain large patches that are protected by
lava sheets, is found only some distance north of the canyon, where the
dip of the strata brings the Permian down to a level that is safe from
erosion for the present, and that can be attacked only as the wet-
weather streams cut deeper into the Aubrey sills on their way down
branch canyons to the Colorado. Now, if the Permian were the resist-
ant and the Carboniferous were the weak formation, or if the next over-
lying and resistant formation further north (Triassic) were also bevelled
down to the Permian level, the bevelled surface would necessarily be
regarded as a peneplain, at present uplifted and in process of dissection ;

DAVIS: THE GRAND CANYON OF THE COLORADO. 139

but no such conclusion seems compulsory in a case where only two
formations are involved, and where the upper of the two is the weaker.
It therefore seems legitimate to say that a peneplain, so far as one was
developed at the close of the first cycle, lay in the Permian formation at
some unknown height above the present plateau surface in the Kanab
district ; and that the Carboniferous platform as now exposed in the
Kanab plateau is a stripped and somewhat dissected plain, with refer-
ence to whose northern margin the Permian plain of to-day is graded :
the stripping and moderate dissection of the Carboniferous and the new
grading of the Permian being the work of the canyon cycle. The Uin-
karet and the Shivwits plateau seem to be susceptible of similar inter-
pretation. The Marble platform also is probably to be regarded, like
the Kanab and other western blocks, as a stripped structural surface.
It must have been covered by baselevelled Permian and lower Triassic
strata at the close of the plateau cycle, but these have been now for the
most part stripped down to the upper Aubrey. In brief, the evidence
for two cycles of erosion in the evolution of the existing topography of
the Grand canyon district is not to be found so much in the general
evenness of the great Carboniferous platform, beneath which the narrow
canyon is cut, as in the great recession of the Triassic and other cliffs
compared with the small width of the rapidly widening canyon, and in
the minor phenomenon therewith correlated.

Dates of Displacements.

The Eastern Flexures. — Flexures appear to preponderate in the
eastern part of the Grand canyon district, while faults are the chief
form of displacements in the western part, as shown in Figure 10. It
will be shown that there is much probability that the flexures are older
than the faults as a whole, and that the displacements of both classes
seem to deserve an earlier date than has been assigned to them in
Dutton's reports.

The Earliest Flexures. — The "Waterpocket and the Escalante flexures
lie northeast of the field of onr excursion, but as they are the only dis-
turbances which have been regarded as pre-Tertiary, it is important to
make some mention of them here. They involve Cretaceous strata,
whose eroded edges are unconformably buried by the Eocene of the High
plateaus on the northwest (Dutton, a, p. 43, pp. 280, 288, 294 ; c, p. 215 ;
Gilbert, c, p. 10). The San Rafael swell, still further northeast (Dutton,
c, Atlas, sheet 11), is described as of Miocene date, probably because of

140

bulletin: museum of comparative zoology.

X

-iO

CO

the roughly concentric attitude of the escarpment
of horizontal Eocene on the northwest with the
escarpments of the swell itself (a, pp. 19, 20) ;
hut it nevertheless seems permissible to class this
dome-like uplift with the "Waterpocket and Esca-
lante flexures as having been formed in what ma}'
be called the Cretaceous-Tertiary interval, and
as having been, like the flexures, greatly eroded
before it was buried by Eocene deposition. True,
the unconformity thus implied is nowhere pre-
served, for the Eocene strata have now receded
to a distance of about forty miles from the centre
of the swell ; but some indirect evidence for their
original unconformable extension over the eroded
swell is found in the small amount of adjustment
of two transverse streams to the domed struc-
tures, as will be more fully stated further on.

The Kaibab and the Echo (Paria) Flexures have
been described as about contemporaneous with
the broad uplift that introduced the erosion of
the Grand canyon (c, pp. 192, 205), but they seem
to be of older origin if one may judge of the date
of their deformation by the recession of the Trias-
sic cliffs that stand in association with them. This
is shown as follows : The Triassic (Vermilion) cliffs
have retreated at least twenty-five miles around
the northern end of the Kaibab since its uplift-
ing; and the same Triassic cliffs have retreated
to close coincidence with the line of the Echo
(Paria) flexure since its production, from what-
ever irregular line of front they had previously
when the strata were horizontal. But, on the
other hand, the upper Aubrey cliffs, enclosing the
outer gorge of the canyon in the Kanab and Uin-
karet plateaus, have retreated only two or three
miles since the beginning of the canyon cycle.
Surely then the Kaibab arch and the Echo mo-
nocline must be older than the beginning of
the canyon cycle. Moreover, Walcott has called
attention to the flexure of the east Kaibab mono-

DAVIS: THE GRAND CANYON OF THE COLORADO. 141

cline without fracture, and has concluded from this that a considerable
load of overlying strata must have covered the Kaibab when the flexure
was produced (d, pp. 60, 64) ; in other words, that the flexure occurred
sometime before the close of the plateau cycle.

If this conclusion be correct, the crown of the Kaibab arch must
have risen a thousand feet or more above a denuded lowland on the
east and west at the close of the plateau cycle. The strikingly even
skyline of the Kaibab, as seen from any distant point, is therefore to be
regarded as indicating a stripped structural (upper Aubrey) plain, and
not as a part of a once baselevelled surface, afterwards uplifted. The
exposure of the arch to denudation during the later stages of the
plateau cycle, as well as through so much of the canyon cycle as has yet
elapsed, is indicated by the much greater dissection of its surface than
that of the Kanab, which, as explained above, is best regarded as a
stripped structural plain of the current cycle. The difference between
the greater and less dissection of the two plateaus is indicated in
Dutton's descriptions, for one is there spoken of as diversified by valleys
which " cover the entire surface " (c, p. 134), while the other is said to
be (except for deep canyons) " no more uneven than the rolling prairie
of Iowa " (c, p. 124) ; and this contrast of form may be easily recognized
on comparing sheets VIII. and XI. of the Grand Canon Atlas. Another
corollary of the early uplift of the Kaibab arch is that an open valley
must have been eroded across it in the later stages of the plateau cycle.
If this be true, then it must be further supposed that the rapid widen-
ing of the new and deep cauyon in the current cycle has consumed all
traces of the earlier valley.

The Crags of Echo Cliffs. — A peculiar feature of the Echo cliffs
deserves brief mention. In looking along this great escarpment from
the south, its crest at several points between Tuba and Lee's Ferry is
seen to rise in serrated peaks with sharp slopes on the east as well as on
the west, thus departing from its ordinary tabular form. We at first
took the sharp serrations to indicate a local increase of dip, and on
coming abreast of them were much surprised to find the dip practically
unchanged from its uniformly moderate measure. The front view of
the escarpment at certain points between the serrated peaks disclosed
sags in the crest line that could not be explained by any local thinning
of the cliff-making sandstones, for they seemed to maintain sub-
stantially a uniform thickness for tens of miles together. It was there-
fore suggested that the peaks and sags resulted from the occurrence of
strike faults with small throw, torn on the slope of the monocline. As

142 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

the retreating cliff face approaches such a fault, the crest becomes sharp
and serrated, and strong eastward slopes locally replace the ordinary
gentle eastward descent where faults are absent. Powell noted the
sharp crags and relates them to the " line of displacement," but he did
not specify the aid that local faults would give (a, p. 192). When the
cliff face retreats still further, so as to remove the sandstones from in
front of a fault plane, the crest line of the remaining cliff sags to a
lower level than usual. It would be necessary to follow the crest of the
escarpment for some distance in order to determine the measure of
verity in this suggestion.

The Western Faults. — Attention may next be given to the fault
lines of the Grand canyon district west of the Kaibab. These fall into
the class to which Dutton gave a very late date : " None of them will go
back of the Pliocene in age, and I think it probable that none of them
will go behind the middle Pliocene " (a, p. 43). " Their formation
seems to have been incidental to the uplifting of the platform which
took place about the time the present Grand Canon began to cut. . . .
Before this epoch we know nothing of them" (c, p. 22G). Some of the
faults may now be considered in more detail.

The West Kaibab Faults. — Of the three faults indicated in Dutton's
map along the western border of the Kaibab, we identified only the
eastern one. It forms a tine wall, of generally graded slope and much
scored by ravines. Its date is not easily determined, as it does not
cross the cliff-making formations on the north until, according to
Dutton, it joins the east Kaibab flexure. The recession of the cliff
from the fault line here, as elsewhere, affords an uncertain measure of
the antiquity of the faulting ; for if the faulting occurred before the
opening of the canyon cycle, much of the height of the present cliff may
then have been fronted by weak Permian strata, since removed.

The Toroweap Fault. — The next displacement west of those which
limit the Kaibab arch is the Toroweap fault. This is regarded by
Dutton as for the most part of very recent date; at least part of
its movement being later than the erosion of the broad-open flat-floored
upper part of the canyon, called the esplanade, because the floor of the
esplanade drops down to the west on passing the fault line; and later
also than the eruption of a certain lava flow that was poured into the
esplanade, because the lava also is dislocated by the fault (c, p. 94).
On the latter point, I am unable to offer any evidence ; but as the prob-
lem now presents itself, the dislocation of the floor of the esplanade
does not seem to bear on the date of the fault ; for if we regard the

DAVIS: THE GRAND CANYON OF THE COLORADO. 143

esplanade as dependent on structure, as will be considered further
on, the break in its floor would occur at the fault line whether the fault
was earlier or later than the erosion of the canyon. The small reces-
sion of the Red-wall cliffs in the Toroweap valley near Vulcan's throne
(c, p. 94) is explainable by their relatively recent exposure in the
canyon cycle, by the failure of erosion as yet to disclose any weak un-
derlying beds, and by the protection of the base of the cliffs by lava
floods, as will be more fully described further on in the section on the
Toroweap. The location of this valley in close association with the
fault line suggests that the fault originated before the valley was
begun, that is, before the initiation of the canyon cycle ; and this view
is borne out by the strong recession of the Triassic (Vermilion) and
Shinarump cliffs on the eastern side of what appears to be the same
fault line in the neighborhood of Pipe spring. As my interpretation of
the stratigraphy here differs somewhat from Button's, the case must be
presented in detail.

The Sevier Fault Southwest of Pipe Spring. — The great Sevier fault
is described and mapped by Dutton as coming southward from the High
plateaus and ending near Pipe spring (c, p. 20), while the Toroweap
fault is said to end northward at the head of the valley of the same
name, twenty miles from the canyon (c, pp. 20, 93). A longer exten-
sion of the Toroweap fault is indicated by Powell, who briefly stated
that where it " crosses the Vermilion cliffs " its throw " is only about
two hundred feet ; " but he does not specify that the crossing is at Pipe
spring (a, p. 186). It happened that on our excursion we had good
opportunity of seeing the general features of the Pipe spring district,
where we noted that the cliffs formed on the Triassic and Shinarump
sandstones were distinctly displaced with respect to what may be called
the Sevier-Toroweap fault line. Hence, instead of curving the Sevier
fault to the southeast and ending it near Pipe spring, as it is drawn on
Dutton's map, it seems better to continue it to the south-southwest till
it joins the known Toroweap fault. Where the fault-line thus ex-
tended crosses Antelope valley, a shallow depression in the broad
plateau surface south of the Shinarump cliffs, its occurrence is not so
easily proved as in the neighborhood of the cliffs. As indicated in
Figure 11, both the Triassic and the Shinarump cliff-makers stand
about ten miles further north on the heaved (eastern) than on the
thrown (western) side of the fault near Pipe spring, thus indicating not
only an important dislocation, but a relatively long time since its pro-
duction. This is the rule of the region. Similar large measures of

144

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

erosion since faulting are generally indicated by a break in the align-
ment of the Triassic and Shinarump cliffs wherever the fault lines cross
them. Powell (a, p. 191), Gilbert (a, p. 51) and Button (c, p. 200)
all recognize the greater recession of the escarpments of the heaved side
of the faults as a general occurrence, but they do not explicitly connect
the amount of recession with the date of faulting. The dislocation of
the cliff front is, as Gilbert phrases it, " not due to any horizontal

. -.Pipe Spr. /

_UTAH
ARIZONA »k&

^^at•"",",,•"'*'

9''n*....Jy.k

*'wi i <*•

t^xi'ito- I'"'""*' FYe.donia

'^mi/%

Antelope fi
Valley

: O

s

10M

Figure 11.

Sketch map of the Pipe spring district, showing the Sevier-Toroweap fault. Weak lower
Triassic strata, dotted; Triassic and Shinarump cliffs, hachured. Constructed from
Dutton's map and original field notes.

displacement along the line of fault, but merely to the fact that the
eastern portions, being lifted higher than the western, became subject
to different conditions of denudation" (a, p. 51). Near Pipe spring,
not only have the cliffs of the eastern (Kanab) block receded ten miles
since the faulting occurred : but ten miles is the excess of the recession
of the eastern cliffs over that of the western ; hence the chief movement
of the fault here must certainly be of earlier date than the beginning of

DAVIS: THE GRAND CANYON OF THE COLORADO. 145

the canyon cycle, however recent the latest of its movements may have
been on the brink of the canyon.

Topographic Effects of Faulting obliterated or reversed by Erosion. —
The continuously graded platform, sloping gently eastward across the
fault line near Pipe spring and obliterating the topographic effects of
faulting, has already been described. The fact that this platform is
now gaining drainage at the expense of the west-sloping platform,
strongly suggests a remote date for the displacement ; had a recent
movement taken place here, with uplift on the east, the westward mi-
gration of the divide could hardly be explained. A little further south,
the Shinarump cliffs in the Uiukai'et block, west of the fault line, over-
look lower ground east of them, thus furnishing an excellent example of
topographical relief standing in opposition to the displacement of the
fault ; evidently the result of the long action of erosion upon the unlike
strata brought together by the faulting. Indeed, this part of the fault
line offers many instructive illustrations of the difference between the
geological and the geographical values of a fault. Geologically con-
sidered, the fault has a fixed value ; geographically considered, the
fault has a changing value. When the fault attains its full displace-
ment, its geological history is completed, and its geological measui-e
thenceforward remains constant. When the fault attains its full dis-
placement, its geographical history is just begun, and thereafter its
measure perpetually changes. The topographic features initiated by
displacement may be more or less modified by contemporaneous erosion,
but the young fault cliff will still be a tolerably regular feature, scored by
growing ravines. Movement ceasing, erosion continues and the fault-cliff
becomes lower and less regular. Where hard beds in the heaved block
confront weak beds in the thrown block, the cliff consequent on fault-
ing will last longest ; where these hard beds are not underlain by weak
beds (uuless beneath baselevel), the cliff face will retreat most slowly
from the fault line. Where weak beds are brought together, all topo-
graphical indication of the fault will soon disappear. Where hard beds
in the thrown block confront weak beds in the heaved block, the topo-
graphic relief may come to contradict the throw of the fault, as is the
case with the Shinarump cliffs above mentioned. But in time, all indi-
cation of the fault will be obliterated in the general reduction of the
region to a peneplain, strong and weak rocks alike being laid low. If
a general uplift of the region then occur, erosion may renew the fault-
cliff topographically wherever strong beds adjoin weak ones, although no
new movement takes place upon the fault plane. In such cases, the small

146 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

retreat of the revived cliff" from the fault line has sometimes given rise
to the belief that the fault is of recent date, but it is evident that no
such conclusion can hold. The revived cliff may be in the thrown-
block, if the arrangement of the rocks favors such a result. A striking
example of this kind may be seen in Tennessee (see Briceville, Tennes-
see, topographic sheet, United States Geological Survey), where the front
of the Cumberland plateau is cut across by a northwest-southeast fault,
the northeastern block being relatively uplifted ; but to-day the surface
is worn low on the weaker strata of the uplifted block, so that it is over-
looked by the uplands of Carboniferous sandstones on the depressed
block.

If these considerations are accepted, there can be little doubt that the
Sevier fault continues in strong force beyond Pipe spring, as if to join
the Toroweap fault, some thirty miles further to the southwest. Some
topographical expression of the fault ought to be found in the broad
floor of Antelope valley, beyond the Shinarump cliffs of the Uinkaret
plateau, where the weak Permian of the Uinkaret block lies opposite
the resistant upper Aubrey of the Kanab block. It is perhaps in this
way that one may explain a west-facing bluff which extends about ten
miles south of Antelope valley (Dutton, c. Atlas, sheet XXII.), but
further study on the ground is needed before this can be assured. It
is interesting to note, however, that the wet-weather drainage here is
from the weaker beds of the thrown block eastward through a gorge in
the bluff that is determined by the more resistant beds of the heaved
block ; a condition that is again entirely inconsistent with a recent
movement on the fault plane.

The Hurricane Fault is not definitely dated in Dutton's report, but its
southern part is thought to be of " comparatively recent origin," be-
cause " the amount of recession by erosion of the cliff of displacement is-
very small" (c, pp. 116, 117). Its northern part near Virgin river
must, however, be old enough to have allowed a strong recession of the
Triassic escarpment since the faulting ; for the two parts of the Trias
east and west of the fault line are now fifteen miles apart (c, pp. 42,
200); and as in the Sevier-Toroweap fault, the displacement of the
cliff's is a measure not merely of their erosion since faulting, but of
the excess of erosion in the heaved block over that in the thrown. The
small amount of recession of the Hurricane ledge (Aubrey) from the
fault line in the southern part of the Uinkaret plateau, as stated in
the last quotation from Dutton, may be explained not so much by the
recent date of the fault, as by the recent date at which erosion had

DAVIS: THE GRAND CANYON OF THE COLORADO. 147

opportunity of attacking the strata here concerned ; for, as in the case
of the Toroweap fault, until renewed uplift introduced the canyon cycle,
it must be supposed that the strata in the Hurricane ledge of to-day
were close to or even beneath baselevel, and hence out of reach of
erosion. On the other hand, a branch of the main Hurricane fault cuts
certain lava beds, — not the most recent ones (c, pp. 116, 117), — and
this branch must be younger than the lavas, which in turn must at the
oldest have been erupted late in the plateau cycle.

The Hurricane fault seemed to me to fade away just north of Virgin
river. The line of great cliffs that comes up from the south here gives
place to a monoclinal flexure, dipping to the west in sympathy with the
throw of the fault ; and a few miles further north the flexure loses im-
portance, as far as we could interpret its structure in the view from the
escarpment south of the river. A new displacement seemed to begin
two or three miles to the west, rapidly increasing its throw northward
and so continuing from Toquerville past Belleview, Kanarra, and beyond,
where it sharply separates the Plateau province from the Basin range
province. The southern termination of this displacement was not cer-
tainly determined. As a fault, it ends just south of Toquerville, but it
may turn southwest and find a continuation in the anticline at the
eastern base of Pine valley mountain. In any case it seemed entirely
independent of the Hurricane fault proper. The course of Virgin
river hereabouts is perhaps consequent upon the overlap of these dis-
placements, for it lies close to the sag between them. It may be added
that, although we had a fine view northward into the valley of Le
Verkin creek from the rim of Hurricane ledge, we did not detect the
lapse of the Permian and the unconformable overlap of the Trias on the
Carboniferous, as noted by Howell (p. 285) ; but to assert the non-
occurrence of these features would be going too far. The neighborhood
of Toquerville offers an unrivalled field for special study. The village
serves as a good base of supplies, the surrounding district is traversed
by many roads and trails, stratified rocks range from Carboniferous to
Tertiary, igneous rocks are present in good variety, the structural
features are of unusual interest, denudation has been enormous, and
the half-desert scenery is superb in form and color.

The Grand Wash Fault is said by Dutton to be no older than Pliocene
(c, p. 191) ; yet the work of erosion after the fault had attained a strong
measure of displacement would give it a much greater antiquity, for
while a monoclinal slice of the Trias is preserved along its thrown (west-
ern) side, there is no Trias on the heaved Shivwits block for fifty miles

148 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

(Dutton, c, p. 200, Atlas, sheet II.). The Trias must have once ex-
tended eastward beyond the line on which the fault was broken ; and
tin' uplifted eastern extension of the formation must have been worn
away after the faulting. On the other hand, a late movement on the
same fault line has been thus far tacitly postulated, for the uplift by
which the canyon cycle was introduced in the Grand canyon district
does not seem to have extended with equal strength into the Basin
range province west of the Grand wash fault line. But as I have not
seen this district, it must be passed only with brief mention.

The Western Monoclinal Flexures. — The above review makes it
probable that, however modern certain minor movements on the fault
planes may be, the chief movements, by which, as Gilbert said, the
adjacent plateau blocks were made subject to " different conditions of
denudation," are of considerable antiquity. They must antedate the
beginning of the canyon cycle. Yet the first deformation of the Grand
canyon district is of still earlier date, for the faults with heave on the
east were in a number of cases preceded by monoclinal flexures with
heave on the west. The few exceptions to this rule of unlike move-
ments do not seriously invalidate it. The east Kaibab monocline is torn
into an east-throwing fault near the canyon ; and similar small faults
are suspected along the Echo monocline, as noted above. The Escalante
flexure dips to the southwest, but its displacement is gradual compared
to that of the Waterpocket, Echo (Paria), and east Kaibab flexures,
which all dip eastward. A flexure with strong throw on the west is
indicated for the western border of the Kaibab in Powell's general sec-
tion of the district (a, p. 190, Figure 73) ; the displacement given to it
is as great as that of the upper of the two east Kaibab flexures. Two
west-dipping flexures on the west side of the Kaibab are shown by Gilbert
(a, p. 51) ; but in the more detailed descriptions and sections given in
Dutton's report (c, pp. 183-186, Figures 3, 4, and Plate II.), nearty all
of the displacement on the west side of the Kaibab is accomplished by
two faults, with hardly a tract of flexure ; the gentle westward dip in the
western half of the Kaibab highland — more pronounced to the north
where faulting is changed for a west-dipping flexure — may suffice to
warrant the use of the term " Kaibab arch," but it seems to be even
less pronounced than the broad Escalante flexure. Where our party
descended westward from the Kaibab by one of the greater ravines, west
of Jacob's lake, the horizontality of the strata all the way to the main
limiting fault was in strong contrast to the pronounced flexure of the
eastern border. It may be, however, that both the western faults be-

DAVIS : THE GRAND CANYON OF THE COLORADO. 149

come flexures uear their northern end, as is noted for one of them,
above, and as is the case with the Hurricane fault proper, just north of
Virgin river. But the only flexure on the western side of the Kaibab
mentioned by Dutton is one by which the thrown beds are turned down
as they approach the fault plane from the west (c, p. 128).

The northeast dip of the flexure that separates the Coconino plateau
from the Kaibab is strongly marked. Strong eastward dips are observ-
able on the lines of the Grand wash fault and of the great fault (almost
in line with the Hurricane fault) north of Toquerville. The rule that
the faults throw to the west and that the flexures throw to the east is
therefore very generally obeyed.

The frequent association of flexures and faults on the same line is
noted by several observers. Dutton described the downward flexure of
the thrown beds in connection with the Hurricane, Sevier, and west
Kaibab faults (c, pp. 41, 113, 114, 115, 128, 185, 186) ; while Gilbert
(or, p. 54) and Marvine (p. 196) give diagrams of the Grand wash fault
which exhibit the same feature. Dutton suggests that flexing pre-
ceded faulting (c, p. 115), and Walcott discusses a remarkable pre-Cam-
brian fault with throw to the west, on the line of which the east Kaibab
torn flexure was formed in Tertiary time, with throw to the east
(d, pp. 49-64). It seems probable that the prevailing coincidence of
flexures and fractures may elsewhere, as well as in the east Kaibab
example, be associated with faults of ancient origin, possibly in the
pre-Cambrian foundation of the palaeozoic and mesozoic strata, although
it is only in the Grand canyon that this relation is open to study.

It is certainly remarkable that the distinct flexures of the Grand
canyon district dip eastward so generally, while the faults have their
throw to the west with almost equal regularity. In all cases where this
relation obtains, the later movement by faulting was of greater measure
than the earlier movement by flexing. It is further noteworthy that
the unfaulted or least faulted flexures, such as the Waterpocket, Echo,
and east Kaibab, lie to the east ; while the distinctly faulted flexures
lie to the west. It may also be remarked that, if the difference of date
here inferred for flexures and faults holds true, it will be inappropriate
to use the term " Kaibab structure " in the sense given to it by Powell
(b, pp. 14, 22), and adopted by Gilbert (b, p. 86); namely, as a designa-
tion for a plateau that " primarily " surmounts lower ground on both
sides. The greater part of the altitude by which the Kaibab surmounts
the Kanab on the west is probably not a primary feature, but a second-
ary one, due to faulting after flexing.

150 BULLETIN: museum of comparative zoology.

The faults have already been referred to the plateau cycle ; the flex-
ures must therefore belong still earlier in that cycle. Hence the period
of the great denudation, thus far undifferentiated, should now be divided
into a pre-flexure cycle, an inter-flex ure-and-fault cycle, and a post-fault
cycle. It is believed that a complicated scheme of this kind is much
nearer the truth than the simple scheme of time division thus far postu-
lated ; but it still remains true that the post-faulting quiescent period
must have been long enough for a strong excess of cliff-recession to occur
in the heaved blocks, before the relatively modern erosion of the canyon
was excited by a broad uplift of the region.

The Displacements of the High Plateaus. — Although our excur-
sion did uot lead us into the district of the High plateaus, it seems
necessary to examine what has been written about them in order to see
how far what has now been inferred as to the date of the faults and
flexures in the Grand canyon district may find application in the adjoin-
ing district on the north. Our source of information is again chiefly in
one of Dutton's reports, from which it appears in the first place that the
faults of the two districts are to be considered as a single group of dis-
placements ; and, in the second place, that the uplift by which the canyon
cycle was introduced probably affected the district of the High plateaus
also (n, pp. 27, 28, 45). The faults of the plateaus are dated as younger
than the youngest formations that they dislocate ; namely, younger than
the middle Eocene sediments and heavy lava sheets and conglomerates
of a somewhat later epoch ; and as old enough to have allowed a certain
moderate amount of erosion since their production. The erosion on the
borders of the faulted High plateau blocks seems small compared to that
by which the recession of the bordering cliffs on the south has been
accomplished, and Dutton therefore decides to " place the age of the
principal displacements in a period which had its commencement in the
latter part of Pliocene time, and extended down to an epoch which, even
in a historical sense, may not be extremely ancient, and which certainly
falls on this side of the glacial period " (a, p. 35). It seems, however,
that in reaching this conclusion, no explicit consideration was given to
the possibility that the faults might have occurred during a former cycle
of erosion, when the district stood much lower than now; that the forms
then initiated by faulting may have been much reduced or even nearly
obliterated by the erosion of this earlier cycle ; and that the erosion on
the borders of the blocks, by which the faults have been dated, has
taken place only since a general uplift has revived the erosive processes.
There is some evidence that such is the case. Certain sections of the

DAVIS: THE GRAND CANYON OF THE COLORADO. 151

district (Dutton, a, Atlas, sheet 6, sections 1, 2, 3) show the Sevier
plateau to consist of a heavy body of early Tertiary sedimentary strata,
covered by a heavy volcanic series. This pair of rock masses ascends
by an eastward rising monocline to the Fish-lake plateau, where the
volcanic series has been largely removed, and even the underlying Ter-
tiaries are reduced to much less than their normal thickness. So un-
equal a measure of erosion in two adjoining plateaus of similar structure
suggests that the rocks of the higher one were greatly consumed during
a former lower stand of the laud, when the destructive processes halted
at a horizon that is now found by drawing a line over the plateau tops ;
and that the existing valleys are chiefly the work of a revival of erosion
after a broad uplift had introduced the present cycle ; this broad uplift
being perhaps associated with that by which the canyon cycle was intro-
duced in the Grand canyon district. True, the displacement between
the two plateaus here referred to is a flexure, not a fault ; but if a former
cycle of erosion is indicated by the unequal erosion on the flexed masses,
it is possible that the faults which elsewhere dislocate the High plateau
blocks may, as well as the flexures, have been produced in that cycle,
and not in the present one. In this case, it is manifestly impossible to
date the faults by the erosion that has occurred in the present cycle ;
they may therefore be as old in the High plateaus as in the Grand
canyon district.

The relation of flexures and fractures appears to be the same in the
two districts. Speaking of the High plateaus, Dutton instances three
faults that are younger than the flexures which they traverse. He
adds : " It is a most curious circumstance that where we find this two-
period displacement the motion of the fault1 is often reversed, — the
lift of the first period is the throw of the second. It is not always so,
but I believe it to be true in a majority of cases where the double move-
ment has been detected " (a, p. 43). But it is of course not intended
by this to imply that the High plateau flexures, which affect the Eocene
beds, are as old as the Waterpocket flexure, which was eroded and uncon-
formably buried by the Eocene (a, pp. 43, 44).

Origin of the Drainage System.

General Explanation by Antecedence. — Powell classified all streams
as consequent, antecedent, or superimposed (now generally called super-

1 The word "fault" is here, as in some other passages, used to include both
flexure and fracture. See, for example, Dutton's account of the "east Kaibab
fault" (a, p. 32).

152 bulletin: museum of comparative zoology.

posed). In a district where superposition seemed impossible, all streams
that did not follow the dip of the strata in consequent fashion were
classified as antecedent. Writing of the Colorado basin, Powell said :
" All the facts concerning the relation of the water-ways of this region
to the mountains, hills, canons, and cliffs, lead to the inevitable conclu-
sion that the system of drainage was determined antecedent to the fault-
ing, and folding, and erosion, which are observed, and antecedent, also, to
the formation of the eruptive beds and cones " (a, p. 198). Even certain
minor streams that follow monoclinal valleys along the northern flank
of the Uinta mountains were for this reason explained as having been
there before the mountains were raised, and the uplift of the mountains
was thought to have been too slow to displace them, in spite of their
small volume (a, pp. 159-166). To-day there can be little question
that these monoclinal streams are not antecedent but subsequent ; that
is, they have gained their position by headward erosion along the strike
of the weak strata in which their valleys are eroded — as first explained
by Jukes, who wrote nearly forty years ago, " the longitudinal valleys
are of subsequent origin" (p. 400) — from time to time capturing and
diverting the upper courses of such consequent streams as they encoun-
tered, and thus bringing about that remarkable adjustment of streams
to structures which characterizes in so high a degree all deeply denuded
regions of strong deformation.

Like Powell, Gilbert recognized three classes of streams, but he seems
to have felt some doubt as to the generally antecedent origin of the in-
consequent streams. He wrote : " A large share of the drainage of the
plateaus is not consequent. How much is super-imposed and how much
antecedent remains to be determined " (6, p. 102).

Dutton was as deeply impressed with the antecedent origin of the
Colorado system as was Powell. Not only the trunk river, but most
of its branches were thus explained. A consequent origin is ascribed to
the lateral ravines which descend the structural slopes of the Kaibab
arch (c, p. 195), but an antecedent origin is announced for nearly all the
tributaries of the Colorado ; for the San Juan, Little Colorado, and Cata-
ract on the south (c, p. 219), and for the San Rafael, Curtis (b, p. 63),
Fremont (a, p. 282), Paria, and Kanab on the north (c, p. 188), as well
as for the streams that are thought to have once occupied the now dry
Summit-valley depressions of the Kaibab (c, p. 193), and the House-rock
valley between the Kaibab and Paria plateaus (c, p. 188).

Replacement of Antecedence by Other Explanations. — The nat-
ural history of rivers is to-day better understood than when Powell and

DAVIS: THE GRAND CANYON OF THE COLORADO. 153

Button maintained the antecedent origin of the Colorado drainage sys-
tem, trunk and branch. It is very probable that these geologists might
now modify in some degree the statements that they made thirty and
twenty years ago as to the indifference of the streams, especially of the
smaller streams, to the displacements of the plateau region. Other
origins may be suggested for several of these waterways.

The Smaller Streams of the Grand Canyon District. — Cataract creek
(called Cascade river by Newberry, pp. 62, 66, and Coanini creek by
Powell, a, p. 197) follows in a general way the gentle northward dip of
the strata that it dissects, and may well be classified as a consequent
stream, revived with every uplift. A possibly consequent origin for
that part of Virgin river which passes between the two parts of the
Hurricane fault has been suggested above. The Little Colorado fol-
lows a monoclinal belt of relatively weak Permian and lower Triassic
strata for a hundred miles, and in this part of its course it may be
plausibly regarded as a subsequent stream ; such was certainly its habit
where we crossed it in the Permian belt on the Flagstaff-Tuba road ;
but for forty miles northwest from this point to its junction with the
larger river, it runs obliquely against the gentle structural slope of the
Marble platform and enters the main canyon just east of the Kaibab
monoclines, a highly significant fact which will be referred to further
on. Paria creek has, according to the geological maps of the district
(Dutton, c, Atlas, sheets II., XXI., XXII.), an anticlinal course in its
upper, and a monoclinal course in its lower part. Although the lower
part is now deeply incised in resistant Triassic strata on the northeast
border of the Paria plateau, its position there may have been gained by
headward growth along the once-overlying weak strata of the gently
dipping monocline during a lower stand of the land ; for the stream
seems to be accordant with the strike of the beds. The upper part of
the Paria drains the denuded district in which the Kaibab arch fades
away to the north (Dutton, a, pp. 253, 297); the lateral branches of
the stream are here to all appearances normal obssquent streams,
whose length increased as the denuded area widened ; the trunk stream
is merely the axial member of the obsequent group, longer than the
laterals because the dip of the strata to the north is gentler than to the
east and west. The whole length of the creek may therefore be reason-
ably explained as an example of spontaneous adjustment of drainage to
structure, and not as of antecedent origin. House-rock valley is un-
questionably subsequent, as has been implied already on page 124, and
as will be more fully considered below. Kanab creek has every appear-

vol. xxxviii. — no. 4 4

154 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

ance of being an obsequent stream, of unusual length, it is true, but
associated with receding cliffs of unusual number and strength, in a
region of extraordinary denudation. The growth of this creek and its
branches, like that of the Paria headwaters, must have been at the
expense of many pre-existent consequent streams. The essential prin-
ciples of the development of such streams were stated by Powell in a
o-eneral way. Ho said, speaking of a series of receding cliffs, facing
southward : "As the cliffs are undermined, . . . the area with a south-
ern drainage would be increased, the area with a northern drainage cor-
respondingly diminished" (a, p. 210). Extensive changes of this kind
must have gone on during the great denudation of the plateau cycle,
and the growth of long obsequent streams is a natural, almost a neces-
sary accompaniment of the great recession of the cliffs that flank the
High plateaus on the south.

The Streams of the San Rafael Swell. — Curtis creek and San Eafael
river were not within our field of observation, but they gain importance
from having been described as typical antecedent streams by Dutton,
who thus explained them because they run across the San Rafael swell
without regard to its structure (b, p. 63). They may, however, be
equally well regarded as superposed through overlying Tertiary strata
which may have once covered the denuded swell unconformably. They
would thus be associated with a part of Fremont river — called Dirty
devil river by Powell (a, p. 67), and Gilbert (a, p. 130, Plate I.), —
which Dutton explains as having been superposed on the mesozoic strata
of the Waterpocket flexure (a, p. 288), although regarding it as ante-
cedent in its original course on the Eocene (a, p. 282). It has already
been pointed out that the swell and the flexure are neighboring struc-
tures, involving the same series of strata. Dutton demonstrates that
the flexure is of pre-Tertiary date (a, p. 288. c, p. 215), and Gilbert
comes to the same conclusion (c, pp. 10-12) ; for Cretaceous strata
are involved, and their bevelled surface is unconformably covered by
the horizontal Eocene. It is eminently possible, as has already been
suggested, that the swell is of the same date, and that its truncated
uplift was buried by the Tertiary strata, which certainly once stretched
over it. Either superposition or antecedence would locate the two
streams on the swell without regard to its structure ; but of these two
processes the latter seems to me much the less probable for the reason
that, if the whole series of strata had been domed, and if the antecedent
streams had had to cut down through the successive alternations of
strong and weak strata from Tertiary to Carboniferous, a greater amount

DAVIS: THE GRAND CANYON OF THE COLORADO. 155

of adjustment to weak structures thau is now seen would have been
almost inevitable. This opinion is fortified when it is noted that the
two streams in question are of moderate size, and still more when it is
seen that they head to the northwest against the retreating Tertiary
escarpment, thus suggesting that they are now longer and larger than
they were formerly. Their antecedent origin seems improbable, to say
the least.

The Summit Valleys of the Kaibab. — The irregular longitudinal de-
pressions, including Summit valley and De Motte park, in the highlands
of the Kaibab plateau, have been regarded as of especially obscure origin.
Dutton thought them all as the work of a single south-flowing antece-
dent stream whose waters had been withdrawn in the change from the
moist Miocene to the dry Pliocene climate, and whose bed had been
deformed by the local uplift of the Kaibab (c, pp. 193-195). This view
seems open to question because of certain improbabilities that it in-
volves and of certain possibilities that it omits. The association of the
Summit valley depression with the axis of the Kaibab uplift, a broad
flat arch, with stronger flexure on the east than on the west, seems too
close to be the work of chance, — as would necessarily be the case if
the depression were the work of a stream whose origin antedated the
uplift. The axial line of the longitudinal depression does not now de-
scend continuously towards the canyon ; a long northern stretch (Sum-
mit valley proper) slopes northward and discharges to the east into
House-rock valley and thence to Paria river ; a more southern por-
tion (De Motte park) also slopes northward, and, except for shallow
sinks on its floor, discharges eastward by a deep ravine down the east
Kaibab monoclines. Two other shorter portions also slope northward.
Certain intermediate parts slope southward, but they are much shorter
than the parts just mentioned. Dutton concludes that all these parts
once had a continuous southward slope, and that the present discon-
tinuity of slope is due to a reversal of grade by the uplift of the
Kaibab.

The continuous southward slope assumed for the depression seems
open to question, especially when it is remembered that the recession of
the Triassic cliffs demands a more remote date for the Kaibab arch than
early Pliocene. The retreat of the heavy Triassic strata around the
north end of the Kaibab suggests that at least some of the mesozoic
strata once stretched partly over the uplifted area : they may indeed
have stretched all over the Kaibab when the uplift was formed, as sug-
gested by TValcott on account of the absence of fractures in the east

156

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Kaibab flexure, as has already been indicated. If so, there should have
been a time in the denudation of the uplifted strata when their lateral
slopes were deeply gashed by consequent streams ; and when subse-
quent longitudinal branches of these streams were developed in anticli-
nal valleys, opened north and south from the head of each consequent
stream on the weak Permian and lower Triassic beds that were discov-
ered along the axis of the uplift. The longitudinal valleys would be
then enclosed by the retreating escarpments of the Triassic sandstones
on the east and west ; the Jura mountains of to-day offer many exam-
ples of this kind. In the early stages of this development there would

be many separate anticlinal valleys along
the axis of the uplift, each drained by
its own outflowing consequent stream ;
but as time went on, the branches of the
deeper-cut consequents would captui'e
the anticlinal drainage area of the shal-
lower-cut consequents, and as the axis of
the uplift descends to the north, these
captures would generally be made in a
southward direction ; thus the drainage
areas of the successful consequents would
become unsymmetrical ; each one would
receive a longer anticlinal branch from
the south than from the north. As
between east and west flowing conse-
quents, the former would generally be
more successful than the latter, because
the eastern slope of the Kaibab de-
scends to a much lower level than the
western ; the difference of altitude of cor-
responding strata in the Kanab plateau and the Marble platform being
nearly two thousand feet (Dutton, c, Plate II.). If the rule regarding
the earlier date of flexures than of faults prevailed here as well as else-
where, the eastern descent from the Kaibab during the pre-faulting period
would have been three thousand feet or more in excess of the western
descent ; and all the anticlinal valleys would thus come to discharge
eastward from near their northern ends, as in Figure 12.

The more important of the longitudinal streams thus established on
the Kaibab might have worn down their channels into the resistant
Aubrey strata before the Triassic sandstones had retreated so far on the

Figure 12.

Diagram of self-adjusted drainage on
the K;iibab.

DAVIS : THE GRAND CANYON OF THE COLORADO. 157

east and west as to allow their infacing escarpments to shed water into
the lateral subsequent valleys opened on weak inonoclinal strata along
the flanks of the uplift. But when a sufficient retreat of the Trias had
been accomplished and the lateral subsequents had been developed, the
axial or anticlinal streams would be reduced to small volume ; for
much of the drainage of the uplift that used to enter them will at this
later stage flow away from them down the stripped structural slopes on
either side. The drainage of the stripped slopes forms a new series of
lateral waterways ; they are not strictly consequent streams which have
persisted since the Kaibab was uplifted, but are regenerated successors
of the original consequents. It may be noted that the valleys of such
regenerated consequent streams will have been eroded first in their
upper parts and afterwards in their lower parts, thus reversing the
iisual order of progress in which erosion acts in a more retrogressive
fashion.

The topographical maps of the Kaibab (Dutton, c, Atlas, sheets XXL,
XXII.) allow us to compare these deduced conditions with the facts.
The lateral subsequent valleys have now shifted to the lower ground
bordering the Kaibab, by reason of the far and wide retreat of the
Trias ; the most representative example being House-rock valley, well
enclosed by the Trias of the Paria plateau which still stands near the
Kaibab on account of the relatively low level of the Marble plat-
form. The regenerated consequents have scored the flanks of the
plateau with deep ravines. Four axial valleys all discharge eastward
from near their northern ends. Xear the northern termination of the
Kaibab Carboniferous area, there is a canyon that cuts directly across
the uplift, draining a portion of the western Permian monoclinal valley
to the corresponding eastern monoclinal valley (House-rock valley) at
Adairville ; this probably being an example of structural superposition,
that is, of a stream whose course was determined when the weak over-
lying Permian strata still covered the area during the plateau cycle
and whose course has been maintained through the resistant Carboni-
ferous strata. The only exception to the rule of eastern discharge is
in the case of the southern end of De Motte valley ; bat this lies
beyond the area of greatest altitude on the Kaibab (nine thousand feet)
and discharges, as might have been expected, southward to the canyon.
It is therefore not necessary to conclude that the "Summit valley"
depressions of the Kaibab were ever drained by a single antecedent
stream, and it seems advisable to regard the Kaibab uplift as having
taken place long before the mesozoic strata were stripped from its

158 bulletin: museum of comparative zoology.

crown : thus the evidence from drainage confirms that derived from
structure and erosion.

The Origin of the Colorado in the Grand Canyon District. — In
view of the various origins other than antecedent that may be ascribed
to the branches of the Colorado in the Grand canyon district, it does
not seem legitimate to adduce these streams in support of the antecedent
origin of the trunk river. That question must be settled by itself, and
it is by no means free from difficulty. The antecedent origin of Green
river in its passage through the Uinta mountains has been seriously im-
pugned by Emmons (1877), who maintains that it is a superposed river
(a, pp. 194, 205; b). The antecedent origin of certain parts of the
Colorado in the Grand canyon district has later been questioned by
Jefferson (1897), who points out the southward bends of the river
around the Kaibab and Shivwits plateaus, and suggests that these
deflections may be consequent on local uplifts instead of regardless of
them. To these doubts must be added a whole series of considerations
which had no place in the discussions of Powell and Dutton regarding
the spontaneous rearrangement of watercourses during the dissection
of a region which has suffered repeated movements and heavy denuda-
tion. It is true that, as the Colorado runs for the most part on nearly
horizontal strata and in a general way transverse to the displacements
that its basin has suffered, it cannot be classed with subsequent rivers,
for they always follow the strike of a weak stratum in a series of tilted
rocks. But the studies of Hayes and Campbell on the migration of
certain divides in a region of uearly horizontal strata that has repeatedly
suffered slight tilting during the development of its rivers deserve se-
rious consideration in the plateau province, where they have as yet found
no exponent. Until this new aspect of the problem shall have been dis-
cussed by some one who has a wide acquaintance with the region, it
does not seem safe to regard even the trenchant Colorado in its course
through the Grand canyon as a purely antecedent river.

In the mean time I cannot resist the temptation of speculating some-
what freely as to the possible development of the great river across the
Grand canyon district, especially in view of what has been said in
previous sections as to the successive movements that the region has
suffered, and in view of the many ways in which drainage lines may be
modified during and after such movements. Certain considerations that
I wish to bring forward in this connection concern not only the Grand
canyon district, but also the adjacent district of the High plateaus and
the province of the Basin ranges (or the Great basin) which have not

DAVIS: THE GRAND CANYON OF THE COLORADO. 159

come especially under my observation. In referring to these areas un-
der the names of the Grand cauyon district, the High plateau district
(these two together making a large part of the Plateau province), and
the Great basin province, it must be understood that the existing
topographical features are of relatively modern origin, that altogether
different topographies prevailed in earlier periods, aud hence that the
names here used will serve chiefly to designate areas and not forms.
The Great basin province was for a time a lofty mouutaiu region ; the
Grand canyon district was a district of broad plateaus ; and the High
plateaus were part of a great interior basin. Most of the statements
concerning these districts are, it is believed, well assured, but some of
them are open to the serious criticism of departing from the conclusions
reached by the original observers whose observations are quoted. This
section as a whole must therefore be largely speculative ; yet the
speculations have some recommendation, in that they do not contradict
recorded observations, and that they combine to form a mutually con-
sistent scheme of geological events. These speculations may at least
serve as targets toward which discussion may be aimed, even if they
do not present any ultimate truth.

Tlie Geological History of the Region. — The Basin range province
has been disturbed by post-Jurassic plication and by late Tertiary fault-
ing. King wrote that the Great basin " was a region of enormous and
complicated folds, riven in later time by a vast series of vertical dis-
placements. . . . The Great basin . . . has suffered two different types
of dynamic action : one, in which the chief factor evidently was tangen-
tial compression, which resulted in contraction and plication, presumably
in post-Jurassic time; the other of strictly vertical action, presumably
within the Tertiary " (pp. 735, 744). Dutton makes a similar state-
ment : " These [Basin range] flexures are not . . . associated with the
building of the existing mountains. . . . The flexures are in the main
older than the mountains, and the mountains were blocked out by faults
from a platform which had been plicated long before, aud after the in-
equalities due to such pre-existing flexures had been nearly obliter-
ated by erosion" (a, p. 47). Several passages in Gilbert's first western
report are of interest in this connection. He concluded " that the
Plateau [region] is not a unit in history and origin, and that the only
criterion by which it can be distinguished from the [Basin] range
country, is the . . . superficial one of table and ridge. . . . The
whole phenomena [of displacement] belong to one great system of moun-
tain formation, of which the ranges exemplify advanced, and the pla-

160 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

teau faults the initial, stages." He goes on to say that the Basin range
province was disturbed at the close of the Jurassic and of the Eocene
periods ; while the faults in the plateau province are of Tertiary date.
In the border land between the two provinces, the later disturbances
not merely run parallel to the Jurassic upheavals, but in places actually
coincide with them (a, pp. 58, 59, Gl).

The Effect of the Flexures. — During the earlier elevation and denu-
dation of the Basin ranges, in post-Jurassic times, much waste seems to
have been carried from them towards the east and northeast. While
the Cretaceous strata of Utah and Arizona are marine deposits, the
Tertiaries are continental ; and from this it may be inferred that the
movements of the Cretaceous-Tertiary interval (such as the Water-
pocket flexure) formed enclosed basins, from which the sea was excluded,
and into which the waste was gathered as lacustrine or fluviatile depos-
its. The great volume' of the successive Tertiary formations implies
that the highlands of the southwest long maintained a considerable
altitude. It is possible that their height was intermittently renewed
by movements which gave. rise, in the Grand canyon district, to the
monocliual flexures ; for these flexures had, as has been s"hown, prevail-
ingly an uplift on the west or southwest. It is postulated that the
initial effect of a series of flexures would be to form a flight of very
broad steps, such as certainly must have been the case with two well-
defined members of the series, the east Kaibab and the Echo flexures ;
but as the surface of each " tread " may have departed from a level
attitude, it is not safe to assert that the height of the top of the
flight (southwest) above the bottom (northeast) was equal to the sum of
all the " rises." It is, however, here assumed that the top was higher
than the bottom, and that the flexures were effective re-enforcements
of the mountain-making upheavals in maintaining the Basin range
province at such an altitude that it could shed waste abundantly to the
northeast. In this connection, it should be noted that the Aubrey cliffs
on the south of the plateaus are determined by a north (northeast1?) dip-
ping flexure (Gilbert, a, p. 46, section), and that the east-dipping strata
which are now found close along the fault lines by which the Basin
ranges are separated from the plateaus, show some of the strongest dips
of the region.

Through all the time during which the mountains of the Basin range
province stood higher than the plateau area, the lines of river-flow may be
assumed to have followed the direction in which the mountain waste was
so abundantly transported ; namely, east and northeast ; and through

DAVIS: THE GRAND CANYON OF THE COLORADO. 161

the second chapter of this time, after the later flexures had been formed
(the inter-flexure- and-fault cycle), the destructive processes which had-
previously deuuded only the mountains of the Basin range province
are believed to have extended to the Grand canyon district also ; but
deposition still continued in the northeast, interrupted only by minor
unconformities. The northward retreat of the great escarpments by
which the High plateaus are now bordered on the south, must have
then begun ; the retreat was probably greater on the southwest, where
the stepping flexures are supposed to have caused the greatest renewals
of uplift. At the close of the cycle that began with flexing and ended
with faulting, the Triassic escarpment, for example, may have receded
to a line that is roughly marked by three points : the lava mesa on
the Shivwits plateau, the junction of the Great and the Little Colo-
rados, aud the eastern side of the lava fields around Mt. San Francisco ;
but something of a northward bend must be made in this line as it
crosses the flat arch of the Kaibab. The amount of denudation accom-
plished during this cycle is indeterminable ; yet if any are disposed to
limit it to a small measure, their attention should be called to the local
instances of great erosion in the Cretaceous-Tertiary interval, when the
Waterpocket flexure was essentially baselevelled. As complete a con-
sumption of the up-flexed blocks in the southwest may have been accom-
plished before faulting took place, although no unconformities remain
there to prove it. But although the denuded southwestern area from
which the Trias had been stripped may have been reduced to moderate
relief, it need not have been a lowland with altitude but little above
sea-level, for its streams discharged into interior continental basins,
very probably without outlet, on whose floors the Tertiary sediments
were accumulating. Some of these basins may have held lakes, inter-
mittently at least ; but the occurrence of mountain i*anges for hundreds
of miles to windward (west and southwest) must have tended to produce
a dry climate in the lower continental area to leeward ; many of the
basins may have been nearly or quite dry, gathering (as I have else-
where pointed out (b)) fluviatile rather than lacustrine deposits at con-
siderable altitudes above the sea.

The Effect of the Faults. — The faulting of the region is the next
important occurrence. The faults have their uplift in nearly all cases
on the east ; hence the eastern area, from having long served as a seat
of deposition, became in turn the seat of extensive denudation. At the
same time, the western area was greatly depressed from its long resi-
dence at a hi"h mountainous altitude. It is not desired to assert that

162 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

all the faulting of the Basin ranges occurred at the time when their
province was dropped to a less altitude than that of the plateaus by the
production of great faults along the boundary of the two provinces ; the
faulting of the ranges was probably a complicated process, but on the
whole it is believed to be associated with the lowering of the Basin range
province, rather than with the earlier stage of its general elevation. It
may be noted that in the southwest (southeastern California) some of
the Basin range blocks have been greatly worn away (Fairbanks, p. 70),
while in the northwest (southern Oregon) some of the blocks are as yet
very little eroded (Russell, a, p. 444:). In general, however, the post-
faulting erosion of the ranges has produced a mature dissection without
altogether destroying the block outlines (Powell, b, p.. 198; Gilbert, d,
p. 341). The most recent studies of the ranges by Spun*, reported at
a recent meeting of the Geological Society of America, indicate that the
faulting has been more complicated and longer continued than had form-
erly been supposed. In certain cases, streams flow through instead
of around the ranges of to-day, from which it may be inferred that at
least some of the existing drainage of the province had been established
before the range blocks were tilted up.

The drainage consequent on the new topography produced by fault-
ing is assumed to have taken the best course that it could find to the
southwest. The chief discharge may have here and there followed pre-
existent valleys, but with a reversed direction of flow (see extract from
Powell, below). Many short-lived lakes may have been formed by the
somewhat adverse eastward tilting that has been supposed to have
accompanied the faulting of the blocks ; but there does not appear to
be any positive proof that the tilting may not have been produced at
the earlier time of flexing instead of at the later time of faulting :
certainly the strata of the Marble platform dip gently eastward
between the east Kaibab and the Echo flexures, and here the dip and
the flexures seem to be of the same date. In any case, the drainage
of a large area of northeastern country, floored with Cretaceous and
Tertiary strata, seems to have been gathered at a favorable point of
discharge ; namely, near the point where the Kaibab uplift had been
greatest, and where the exposure of the weak Permian beds had caused
che greatest break in the strong Triassic escarpment by which the
interior country was enclosed. Westward from the Kaibab, the escap-
ing drainage presumably followed the lowest line that it could find
between the Triassic escarpment on the north and the general ascent of
the stripped Carboniferous country to the south. Perhaps in some such

DAVIS : THE GKAND CANYON OF THE COLORADO. 163

manner the Colorado found its way across the Grand canyon district to
the broken-down mountains of the Basin range province, and thence
southwestward to the sea. The river would thus seem to have been for
the most part consequent on the form and slope of the surface at the
time of faulting, yet it may have been antecedent to various subordi-
nate movements of the crust here and there ; and there is a possibility
that it may have been in some greater or less degree developed after the
faulting by the retrogressive erosion of west-flowing streams that had
been encouraged by a favorable tilting of their courses. McGee has
suggested the latter process to explain certain streams in " Papagueria "
(Arizona and Sonora), whose growth is thought to have been thus
accelerated duriug a time corresponding to what is here called the
canyon cycle (c, p. 352), and this process certainly deserves deliberate
consideration.

It should be pointed out that a consequent origin for the Colorado is
indicated by the following passage in Powell's report on the Uinta
mountains: "At last the movements which began at the commence-
ment of Tertiary time succeeded in bringing the whole [Plateau] area
not only above the level of the sea, but above the general level of the
Basin province itself ; so that while the Basin province was drained into
the Plateau province in earlier Tertiary time, in late Tertiary time the
drainage was reversed, and the streams of the Plateau province found
their way to the sea by passing through the Basin province. ... It is
the opinion of Mr. Howell, and I believe also that of Captain Dutton,
that this drainage was in some cases reversed along the very channels
occupied by the ancient streams which ran from the Basin province into
the Plateau lakes" (b, p. 35). There is in this quotation much more
suggestion of a consequent origin of the Colorado than is elsewhere to
be found in the reports of the observers whose names have been so
often cited here : nevertheless, the published reports give no indication
of the manner in which the denudation of the upflexed area on the
southwest may have prepared the way for the production of a south-
westward slope when the district was afterwards broken and heaved by
faulting.

If there be any truth in the suggestion that the Colorado did not
cross the Grand canyon district until after the faulting of the plateau
blocks, then it was in the cycle of erosion thus initiated (the post-fault-
ing cycle) that the great denudation was essentially completed, leaving
little more than the stripping of weak strata and the canyon-cutting for
the canyon cycle. The retreat of the escarpments on the faulted blocks

164 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

during the last chapter of the great denudation must have been over
a considerable distance, for each line of cliffs should have been, before
the faults were made, worn back further on the west of a flexure than
on the east ; while now, in consequence of the faults, the reverse
relation has been brought about. The total recession since faulting
has therefore probably been decidedly greater than the distance by
which the cliffs on the east of a fault-line now stand north of their
fellows on the west ; and thus an even greater antiquity for the faults
is suggested than has been thus far suspected.

The Bends of the Grand Canyon. — There are three peculiar features
in the course of the Grand canyon : one is the southward bend around
the Kaibab, the second is a northward bend toward the mouth of
Kanab creek, and the third is a southward bend around the Shivwits
plateau. Jefferson has suggested that the first and third of these bends
may be consequent on the form of the surface given by the flexures ;
but according to the analysis of events here presented, the drainage
consequent on the flexures ran to the east and northeast, and the west-
ward course of the Colorado was not assumed until after the flexed
plateaus had been greatly denuded, and until the denuded surface had
been raised on the east by the block faulting. Something later than
the initial slopes of the flexed surface should therefore be found to
guide the river, if an antecedent or initially consequent origin is not
accepted for it. It has seemed to me that certain details in the flexing
of the Kaibab and Coconino plateaus may here be appealed to.

The eastern lobe of the Coconino is a striking feature as seen from
the lower plateau on the south (Figure 13) and from the valley of the
Little Colorado on the east. The surface of the district at the time of
flexure was presumably covered by some of the mesozoic formations.
Just before the time of faulting, the Trias may have been pushed back
to some such outline as would reveal the Permian in the valleys on the
south and west of the Kaibab, and perhaps even some of the Aubrey
was laid bare on the highest part of the Kaibab. Up to this time, the
drainage hereabouts was eastward down the slopes of the Kaibab and
Echo flexures, but many longitudinal subsequent streams must have
been developed along the weak lower Triassic and Permian strata west
of and underlying the Triassic escarpment that was then retreating
eastward towards its present position on the axis of the Echo flexure.
Now as the drainage of the interior basins of the northeast was turned
sonthwestward by the upheavals associated with the faulting — long
after the close of the Eocene — the most available point for its escape

DAVIS : THE GRAND CANYON OF THE COLORADO.

16;

was, as has been stated, near the centre of greatest local upheaval (the
Kaibab) where the revelation of weak underlying strata had reduced the
level of the Trias by sapping, and had thus produced an amphitheatre,
open on the southwest. At the same time, the subsequents under the
Triassic escarpment were gathered into a single stream to form the
Little Colorado, discharging northward ; but this result may have been
in part the effect of spontaneous interaction among the streams them-
selves, after their habit in such cases, and only in part the effect of
upheaval at the time of faulting. Once reaching the Kaibab amphi-
theatre, the new-born river followed southward along the subsequent
Permian valley which must have been opened along the east Kaibab
flexure, until the slopes of the northeast-dipping flexure that limits the

Figure 13.

The eastern lobe of the Coconino plateau, as seen from near Lockett's tank. The fore-
ground shows a ravine in the upper Aubrey limestone, once filled with lava, and now
partly re-excavated. The "tank " is a waterfall pool. Drawn from rough sketch.

Coconino plateau were encountered. The river then turned northwest-
ward along a trough of weak and low-lying Permian strata that occu-
pied a depression between the two uplifts ; and it is notable that
the small but sharp flexure which bounded this trough on the south
may now be traced through the spurs of Carboniferous strata on the
southern canyon wall (Dutton, c, p. 185). The flexure comes from the
southeast, where it forms the northern border of the eastern lobe of
the Coconino plateau ; it trends northwest, as if to join the now-faulted
flexure at the western base of the Kaibab. We had a good view of it
from the southern rim and from the bottom of a side canyon. Thus
interpreted, it appears that a part of the true Kaibab uplift lies south
of the canyon, where it slopes into what is locally known as " the

1G6 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Basin," and ends on reaching the strong flexure by which the strata are
turned up into the Coconino ; but it is probable that in so far as the
terms Kaibab and Coconino have a geographical application, their areas
will be separated by the canyon rather than by the flexure that
separates their uplifts.

The further northwestward course of the river may have been made
inviting if a relative depression or failure of uplift occurred in the
eastern Kanab plateau when the west Kaibab faults were formed, and
Dutton gives some evidence that this was the case when he describes
the northward slope of the Kanab along the southwest border of the
Kaibab (c, p. 184) ; but I am at loss to find any conclusive proof of
such a condition, apart from the behavior of the river itself. The
south beud on the Shivwits may plausibly be explained as a displace-
ment from a once more direct course by reason of the volcanic out-
pourings which now culminate in Mt. Dellenbaugh.

Speculative Character of the Preceding Sections. — All this is avowedly
speculative ; and it is presented rather as a combination of various
possibilities than as embodying the only permissible explanation for the
course of the Colorado. Other stages of the great denudation, now
unsuspected, may yet be discovered in the development of the Grand
canyon district ; new processes of river development, as imperfectly
known to-day as was the growth of subsequent streams at the time of
the earlier exploration of the canyon, may be added to the growing
resources of physiographic study ; and all of these elaborate stages and
processes will deserve as careful consideration as has been given to the
simpler conception of the antecedent river, marked out by the deeper
lines of the Eocene lake floor, and remaining unaltered, save for deep-
ening, ever since the ancient lake became dry land.

Nothing less than extended studies over a large area of the Cordil-
leran region will suffice to determine what value should be given to the
various possibilities thus suggested. It is not my intention to discount
such studies by attempting to announce their result at once ; but only
to emphasize the opinion that the facts now on record, combined with
such knowledge of the region as our party was able to gather last sum-
mer, warrant the consideration of at least one hypothesis alternative to
the theory of antecedence, as an explanation for the origin of the drainage
lines in the Grand canyon district. I do not on the one hand consider
the antecedent origin of the Colorado disproved, but, on the other hand,
such an origin does not seem compulsory. The chief objection to the
theory of antecedence is not that rivers cannot saw their way through

DAVIS: THE GRAND CANYON OF THE COLORADO. 167

rising mountains, for there are some well-attested examples of such a
process, in which its successive steps are more or less fully traced ; but
rather that this theory makes a single stride from the beginning to the
end of a long and complicated series of movements and erosions, over-
looking all the opportunities for drainage modifications on the way.
The simplicity of the theory is certainly attractive in comparison with
the rather tedious length of the considerations that are involved in the
attempt to analyze the processes of spontaneous river adjustments; but
it should now be generally recognized that nothing less than a deliberate
analysis of all movements (of which the uplift of the present plateaus is
the last) and erosions (of which the cutting of the Grand canyon is the
least) will suffice to discover the actual origin of the Colorado. The
preceding paragraphs are offered as the beginnings of such an analysis.

The Erosion of the Grand Canyon.

The Canyon Cycle. — The general uplift that introduced the canyon
cycle in the Grand canyon district seems to have been accompanied by
a strong renewal of movement on the faults that divide the Basin range
province from the plateaus. The northward increase in the heave of
the bounding faults between the Basin range province and the plateaus
suggests that the uplift of the High plateaus at this time was greater
than that of the Grand canyon district. Dutton says : " Until near
the close of the Pliocene the High plateaus were not only the theatre of
an extended vulcanism, but those portions which never were sheeted
over by lavas were low-lying areas, where alluvial strata tended to
accumulate. They remained, in fact, base levels of erosion during the
greater part of Tertiary time" (a, p. 23). It is possible that some
movement was renewed on other than the bounding fault-lines at the
time of the general uplift, but it has been pointed out above that there
is good reason for thinking such movement to have been insignificant on
the fault line near Pipe spring. Still, in the south Toroweap valley
(Dutton, c, p. 94), in the Great Basin (Gilbert, d, p. 341), and in the
High plateaus (Dutton, a, pp. 250, 277), there are various indica-
tions of faulting at a much later date than the beginning of the canyon
cycle.

Comparison of Glen and Marble Canyons. — Chief among the fore-
going considerations that point to the division of the Tertiary history
of the Grand canyon district into at least two cycles of erosion is the
similar resistance to weathering manifested by the Triassic and the

168 bulletin: museum of comparative zoology.

Carboniferous strata. It would be difficult to determine by ordinary
tests which one of these two formations is the more resistant, so strong
does each appear ; but a very good natural test of their relative strength
is found in a comparison of the upper part of Marble canyon, cut in the
upper Carboniferous, with Glen canyon, cut in the Triassic rocks them-
selves, where they descend to the level of the lower plateaus east of the
Paria-Echo monocline. The two canyons are of similar date, for the
monoclinal flexure that separates them is much older than either canyon;
they are of similar width (Dutton, c, Atlas, sheet XXII.) ; and hence
the resistance of their walls must be similar^ It is true that the widen-
ing of Glen canyon has been retarded by the failure of the river as yet
to cut down to the weak blue clays that underlie the heavy red sand-
stones of the Trias; while the stripping of the Trias from the great area
of the plateaus south of the Grand canyon, and the recession of the
Vermilion cliffs for fifty miles or more north of the canyon have been
greatly aided by the sapping of the underlying clays ; but even with
this aid it does not seem possible to explain the great denudation of the
plateaus in contrast to the narrowness of the canyon by a single cycle of
erosion.

Stage of Development of the Canyon. — The day has passed when it
was necessary to ask whether the canyon is the work of the river ; but
in the renewed attention lately given to the relation of trunk and
branch valleys, certain features of the canyon serve as important wit-
nesses. In spite of the youth of the canyon, its branch streams gen-
erally enter at accordant grade with the main river, and thus testify to
the promptness with which side valleys are cut down to the depth of
the main valley at their points of junction. This feature must be
considered in some detail.

Rapids in the Canyon. — The canyon is so young that the great river
at its bottom has not yet established a completely graded channel.
True, there are no leaping waterfalls now remaining : " Throughout the
canons there are no cataracts ; that is to say, at no place does the river
fall from a ledge of rock into the pool below " (Gilbert, a, p. 75). But
there are still many rapids, especially in those parts of the canyon
where the fundamental crystallines are trenched. When these resistant
ledges are rasped away, the upper canyon will be significantly deeper
than it is now. On the other hand, the corrasion of the canyon must at
present be proceeding at a slower rate than at some earlier time, before
the development of the graded stretches that now constitute the greatest
part of the river. It was the existence of these graded stretches, where

DAVIS: THE GRAND CANYON OF THE COLORADO. 169

the river is hardly more than a transporting and comminuting agent,
that brought Powell's boat journey within the limits of possibility, even
while the lingering rapids made it terribly dangerous. The occurrence
of rapids on the down-stream side of boulder heaps that have been
washed into the main canyon by the flooded streams of side chasms,
also indicates a retarded rate of corrasion by the main river. Powell
mentions several such obstructions (a, pp. 82, 83, 97) ; Gilbert gives a
more explicit account of them and points out their probably temporary
character. He says that many of the tributary canyons " are of very
rapid fall, and are occasionally traversed by powerful torrents, which
sweep down bowlders of great size — in some instances 10 or 15 feet
in diameter — and heap them in the main canon in dams, that must
often be of great depth. Over each of these the water finds pass-
age at the edge opposite the tributary, and descends the lower slope
with swift current and broken surface ; and thus arise the great major-
ity of the rapids." It is further pointed out that there are times " when
each of these dams in turn is removed ... so that, while the dams will
recur at the same localities and with the same characters, they can-
not be regarded as strictly permanent" (a, p. 71 ; see also Dutton, c,
p. 241).

It may therefore be concluded that although the canyon is still to be
deepened, the river as a whole approaches and for considerable parts of
its course actually reaches the delicate balance between degradation and
aggradation which characterizes a maturely graded stream. After such
a condition is reached, further deepening of a valley is possible only with
the decrease of the load furnished to the main stream by its side streams,
in the later maturity of the region.

Junction of Trunk and Branch Streams. — It is now to be deter-
mined how the side streams join the main river at this relatively early
stage in the present cycle of the Colorado. The importance of this
question turns on the prevailingly discordant junction of side and main
valleys in strongly glaciated mountains, as has recently been pointed
out by De Lapparent and Richter for the fiords of Norway, by Gannett
for Lake Chelan in the Cascade range, and by Penck for the valleys of
the Alps. I have lately prepared a general review of this problem in
connection with my own observations on the valley of the Ticino in
southern Switzerland and elsewhere in Europe (c, p. 310; d, p. 138).
It has been contended hy some that the discordant junction of side and
main valleys in the Alps is the normal result of the more rapid deepen-
ing of the main valley by its larger stream, even to the point of thus

VOL. XXXVIII. — no. 4 5

170 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

accounting for lateral valleys standing five hundred or more feet over
main valleys whose flood plains are ten or twenty times as broad as their
rivers. On the other hand, the observers above named conclude that
the discordance is due to glacial erosion. If the former view were cor-
rect, then surely the discordance of side and main valleys ought to be
very strong in the Grand canyon, where there is no question of glacial
erosion, where the disparity of volume between trunk river and side
streams is notoriously great, and where the main valley is still so young
that no significant widening of its floor has been yet accomplished. Yet
singularly enough, the side canyons of the Colorado join the main canyon
at accordant levels in nearly all cases. The views from the southern rim
of the canyon at various points near Hance's and Cameron and Berry's,
and still better from the great Red-wall spur that advances northward
from the "copper mine" on the Grand view trail, show repeated in-
stances of dry or nearly dry lateral canyons, only five or six miles long,
which are nevertheless cut down at their months as deep as the main
river, if their lower course lies on the stratified rocks. It is only where
the main canyon narrows on entering the resistant crystallines that the
side canyons are held up at a discordant level ; and even there the large
lateral canyons seem to enter closely at grade (see detailed map, by
Bodfish, Dutton, c, p. 258, Plate XLIL). However discordant the side
and the main canyons may have been during a still earlier stage of the
present cycle of erosion, Playfair's law is already exemplified to-day ;
and if, under conditions of peculiar difficulty, some of the smaller
streams have not yet reached a normal relation to the master stream,
they only prove the verity of the law by the need of their being
excepted from it.

As my own views of the canyon were only in the middle of the Kaibab
section and from Vulcan's throne in the Toroweap, the following infer-
ences to records made by others may prove pertinent in this connection.
Powell describes his ascent of a side canyon west of the Kaibab, and
mentions waterfalls in its course, one of which was one hundred and
fifty feet high ; but as no fall is mentioned at the mouth of the side
can von, it probably unites with its master in accordant fashion («, p. 92).
A little further down the main canyon, a stream from the north leaps
into the river " by a direct fall of more than a hundred feet " over a
"bed of very hard rock, . . . thirty or forty feet in thickness" (a, p. 93).
Before reaching the crystallines of the Shivwits canyon, " a little stream,
with a narrow flood plain, comes down through a side canon " from
the north. An Indian settlement was found there, with fields of corn

DAVIS: THE GRAND CANYON OF THE COLORADO. 171

and squash (a, p. 96). Even within the crystallines of the Shivwits
block, small side streams of " gentle slope " are reported by Wheeler as
entering the main river and forming boulder rapids, but no mention is
made of falls at the stream mouths (pp. 166, 167). Plate XVII. , in
Button's Monograph, reproduced from a photograph, shows two side
canyons entering the narrow inner gorge of the main canyon just east
of the Toroweap, and joining the main river in accordant fashion. Evi-
dently, then, hanging valleys have no important place in the Grand
canyon ; and the hanging lateral valleys of the Alps, whose floors are
five hundred feet or more above the open flood plains of their main
valleys cannot be explained by normal river erosion.

Some striking examples of hanging valleys in a very narrow canyon
are, however, described by Gilbert in the case of the North fork of
Virgin river in southern Utah, where it cuts through the massive
Triassic sandstones. This narrow defile is many times deeper than
broad ; its walls are nearly vertical and parallel for the greater part
of their height, but they depart sufficiently from the vertical, now to
this side, now to that, to hide the sky from the adventurous observer
who follows the narrow, boulder-strewn stream bed. " The side canons
all partake of the chai*acter of the main, but, being worn by smaller
streams, are narrower, and their bottoms are of steeper grade. Many of
them at their mouths are not cut so deep as the one we followed, and
discharge at various heights above the river" (a, p. 79). Gilbert's
figure of this canyon has become well known from being copied on the
binding'Of Leconte's " Elements of Geology," as if in witness of the effi-
ciency of erosive processes ; but it may be noted that a plain of denu-
dation, truncating the edges of upturned strata, is a much more impressive
though less outspoken witness to this conclusion.

The Geological Section in the Canton Wall. — The excavation
of the Grand canyon is properly regarded as a colossal work. Standing
on the southern rim, the view of the chasm is overwhelming; yet the
prospect includes four other records of erosion, and suggests two more
still, in comparison with any one of which the excavation of the canyon
is but a small matter. This has all been pointed out by Powell and
others ; but it deserves repeated statement.

The geological section exposed in the northern side of the canyon in
the Kaibab, as seen from any of the promontories in the neighborhood of
Cameron and Berry's or Hance's hotels, includes the fundamental crystal-
lines, the inclined strata of the Grand canyon series (Algonkian), twelve
thousand feet thick, and the palseozoic series, over four thousand feet thick.

172

bulletin: museum of comparative zoology.

The strata of the Grand canyon series are seen in the form of a wedge
with its apex pointing westward, as in Figure 14 ; its lower members
rest unconformably on an inclined floor of denuded schists, while the
basal members of the palseozoic series rest unconformably on a horizontal
floor denuded on the schists continuously with the bevelled upper sur-

■■:! ■' .' ■ • ' ' m

Figure 14.

The Algonkian "wedge" between the crystalline foundation and the palaeozoic series;
looking north across the canj-on from nearHance's on the Coconino river. Constructed
from rough sketch.

face of the wedge of the Grand canyon series.1 The skyline of the
Kaibab is, as has been pointed out above, the edge of a stripped struc-

1 The view of this structure given by Powell (a, p. 212, Figure 79) is somewhat
misleading, for it represents the inclined members of the Grand canyon series with
a horizontal base, as if they were examples of cross bedding on a gigantic scale,
The correct relation is shown in figures by Walcott (b, p. 551 ; e, p. 507 — this
from a drawing by Gilbert — and c, p. 553) and by Freeh (p. 477). It may be noted
that a sheet of basalt or diabase, which occurs near the base of the Unkar series
and which has been regarded as contemporaneously interbedded by Walcott
(-, p. 508) and doubtfully described as either an intrusion or a surface flow by
Freeh (p 177), seemed to us to be an intrusive sheet or sill, because it appeared to
step from one bedding plane to another by distances of a score of feet or more at
one or two places; but this opinion is based only on field-glass observation at a
distance of over a mile.

DAVIS : THE GRAND CANYON OF THE COLORADO. 173

tural plain, from which more than six thousand feet of strata have been
removed. There have, then, been two periods of essentially completed
erosion, marked by the floor of the Grand canyon series and by the floor
of the palaeozoic series, and one period of far advanced erosion, marked
by the skyline of the Kaibab ; and three alternate periods of enormous
erosion elsewhere to supply the strata locally deposited in the Grand
canyon series, the palaeozoic series, and the mesozoic series ; and the
work done in any one of these six periods far outranks that thus far
accomplished in the erosion of the canyon. Excepting the periods of
mesozoic deposition and erosion, all this history is recorded in the canyon
with the clearness that one ordinarily sees only in colored diagrams on a
blackboard, and with infinitely greater detail and impressiveness.

The Two Unconformities. — The double unconformity associated with
the " wedge " of the Grand canyon series (Figure 14) deserves special
attention. The floor on which these strata rest is of remarkable even-
ness, in spite of the great deformation of the fundamental schists. It is
exposed in the section on the northern wall of the canyon for the greater
part of a mile, dipping under the river at the lower (eastern) end and
terminated at its upper (western) end by the surface of the second un-
conformity. The line here seen may be taken as a fair sample of a large
area of the floor on which the Grand canyon series rests, because it is
exposed by the chance section made by the river, whose course was
originally selected with no regard whatever for the then deeply buried
crystalline foundation of the region. The whole surface must have been
much larger than the part seen in the canyon ; for if the crystallines
were evenly truncated here, they must have been similarly worn down
over a large extent of adjoining territory ; and, moreover, a formation
that measures ten thousand feet in thickness, like the Grand canyon
series, cannot be of merely local development. The conclusion seems
compulsory that before the deposition of the Unkar strata (the lower
members of the Grand canyon series, Walcott, e, p. 506) the crystalline
rocks were reduced to a plain of admirable evenness, either by marine or
by subaerial forces ; and however many cycles or partial cycles of erosion
were devoted to this ancient task, the last cycle must have been undis-
turbed until it was very far advanced.

The floor on which the palaeozoic strata lie was formed by extensive
erosion after the tilting of the crystalline schists with their heavy cover
of the Grand canyon series, the compound mass being planed down to
an almost even surface. The Kaibab section of this floor is over fifty
miles in length along one side of the river, or about forty miles in a

174

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

direct line. Through very nearly all this distance, the floor is covered
by the basal (Tonto) sandstones of the palaeozoic series. When the
actual contact is seen, it is found to be ragged, but in the twenty-mile
stretch that I saw, the vertical inequalities of the floor were trifling in
comparison to its horizontal extent. A few miles west of the Grand
view trail into the canyon, a broad flat mound of the crystallines rose
high enough in the northern wall of the canyon to interrupt the whole

Figure 15.

The canyon in the crystallines, looking down the river (northwest) from Coppermine spur,
Grand View trail. The crystalline floor is generally covered by the Tonto sandstone;
but a low crystalline monadnock rises through the Tonto in the distance. Drawn from
a sketch.

thickness of the Tonto sandstones (Figure 15), but this was the only
significant unevenness of the surface between the Tonto and the schists
visible in that direction. A short distance east of the apex of the
" wedge," the more resistant members of the older series are not so well
planed down as are the crystallines just west of the same point, and the
Tonto sandstone is there greatly reduced in thickness or altogether
wanting fur large fractions of a mile. But these inequalities in the

DAVIS : THE GRAND CANYON OF THE COLORADO. I/O

palaeozoic floor are of small value compared to the strong relief that
must have been developed in the mature stage of erosion on the tilted
Grand canyon series and their underlying schists. The ancient floor
was certainly a topographically old surface, in a far advanced stage of its
erosion cycle.

The extension of the palaeozoic floor further west may be briefly
sketched. Powell noted " patches of granite, like hills thrust up into
the limestone " in his passage through the Kanab section of the canyon
(a, p. 92). Dutton describes the floor in detail, emphasizing its small
relief, and Holmes made an admirable drawing of the southern wall
in the Kaibab section of the canyon (c, pp. 178-181, 207, 209, Plate
XXXV.). Newberry wrote, when describing the section in the Shivwits
plateau : " The erosion of the caiion has beautifully displayed the ancient
surface of the granite, and shows it to have been extremely irregular ;
hills several hundred feet high, many of which have precipitous sides, and
deserve the name of pinnacles, have been exhumed from the sediments
in which they were enveloped. The sandstones and shales are seen to
have been deposited quietly around them ; their strata, nearly horizontal,
abutting against their sides " (p. 58). Gilbert says that " all along the
southwestern border of the plateau region in Arizona, the Archaean schists
and granites are seen beneath nonconforming members of the Grand
canon rock system; usually the Tonto sandstone" (a, p. 186), and
his diagrams represent the contact of the two systems by a straight
line. My lamented classmate, Marvine, records similar observations
(p. 199). A photograph of a point in the western part of the Grand
canyon, near Peach spring, Ariz. (View No. 173, taken by W. H.
Jackson and Co., Denver, Colo.), shows the crystallines as evenly capped
by the Tonto as they are in the Kaibab section. The palaeozoic floor is
thus traced for over a hundred miles from the Kaibab section, and in
spite of its inequalities, it is nearly everywhere capped by the lower
palaeozoic strata. It was certainly a surface of moderate or small
relief.

The slope by which the crystallines descend to the river under the
stratified rocks has different pi-ofiles east and west of the apex of
the " wedge." Under the Unkar, the slope is uniformly steep from top
to bottom. Under the Tonto, a bench of crystallines stands forth for
several hundred feet, and then bends by a strong curve to a slope as
steep as that beneath the Unkar (Figures 14 and 15). The persistence
of this feature in the view down the river from a point over the edge of
the " wedge " is remarkable, and strongly suggests that the crystallines

176 BULLETIN: museum of comparative zoology.

under the Tonto are weaker in their upper than in their lower portion.
It occurred to me that this might be the result of deep weathering of
a pre-Tonto peneplain, before it was buried by the palseozoic series, and
that the incursion of the Tonto sea over the peneplain was too rapid to
allow the waves to abrade anything more than the superficial soils ;
thus leaving the weakened rocks to reveal their weakness in the canyon
profile to-day. The sub-Unkar floor, on the other hand, seems to have
been smoothly worn down to firm rock by marine abrasion before it was
covered by Unkar sediments; but whether the advancing Unkar sea
destroyed great mountains of schists or merely planed off the soils and
weathered rocks of a pre-Unkar peneplain is an unsolved question.

Correlation of Water Streams and Waste Streams. — The pa-
laeozoic walls of the canyon in the Kaibab exhibit a combination of cliffs,
slopes, and platforms in great variety, appropriate to the horizontal
structures on which they are eroded. The widening of the canyon here
has not yet produced broad benches or platforms, although a systematic
beginning of their development is repeatedly observed ; but in the
Kanab, Uinkaret, and Shivwits sections of the canyon the upper sur-
face of the Red-wall group has been denuded to form a broad platform,
called the esplanade by Dutton, to which fuller reference will be made
below. Returning to the Kaibab, a pleasing comparison may be made
between the path of a running stream of water along the valley bottom
and that of a creeping stream of waste down the valley side ; the
simplest examples for comparison being in horizontal strata such as are
so magnificently displayed in the canyon walls. In both cases, graded
slopes are first established in discontinuous stretches on the less re-
sistant strata, while cliffs remain ungraded on the outcrops of the
stronger strata. The close analogy of the two cases is brought to mind
by comparing a plunging fall of water, leaping down the nearly vertical
face of a resistant stratum in the valley bottom, with the discontinuous
fall of rock waste over the cliffed outcrop of a resistant stratum on
the valley side. The evenly graded reach of the water stream on the
weaker strata above the fall corresponds with the flat platform in the
valley walls above the cliff; and the " cave-of-the-winds " in the weak
strata just behind the waterfall corresponds with the hollow or " rock-
house " beneath the overhanging base of the cliff; the scanty heap of
rock fragments beneath the waterfall corresponds with the sheet of
rock waste (with the coarsest fragments lowest down) that cloaks the
under slope of weak strata from the rock-house down to the back or
inner margin of the next platform below. It is natural that the pro-

DAVIS : THE GRAND CANYON OF THE COLORADO. 177

portiouate development of the several elements, cliff, rock-house, talus-
slope, and platform on the valley side, should vary from that of the
corresponding elements, cliff, cave-of-the-winds, rock-heap, and reach
along the valley bottom, on account of the difference cf behavior
between a sluggish stream of waste and a nimble stream of water. In
the latter, the vertical element of form is reduced to a small fraction
of the horizontal at an early stage of the cycle; while in the former,
the vertical dimension remains strong until the stage of maturity is
past.

Graded reaches make the greatest part of even a young river, but
graded platforms widen slowly on valley sides and attain no great
breadth until late maturity. It is along these graded reaches that the
finer waste from the rock-heap, as well as from further up-stream, is
steadily carried forward ; comminution of rock waste is active, but
erosion of the rock bed is here extremely slow. Rock-heaps under
waterfalls are so unimportant topographically that they are seldom in-
cluded in geographical descriptions; yet they are really characteristic
elements of a young river's course in structures of the kind here con-
sidered. Talus slopes on valley sides are of much greater dimensions ;
so great indeed that they form a systematically graded surface. Their
grade is steep compared to that of the platform because their waste
is coarse. These accumulations of coarse waste, both rock-heap and
talus, consist chiefly of fragments broken from the cliff above ; the fine
waste that is weathered from them is carried forward or down-stream,
while the surface on which they rest retreats backward or up-stream.
Bock-heap and cave-of-the-winds vary inversely; the sum of their
vertical measure together with that of the face of their fall ledge
(a constant) makes the height of the fall : talus-slope and rock-house
are similarly related. The cave-of-the-winds is larger but less habitable
than the rock-house ; but in the wet as well as in the dry cave, air-
movement is an important factor of erosion. The cliffs in the two cases
are so much alike that they need no further comment.

The strongest fall-and-cliff-makers endure longest, while the less
strong ones are sooner worn back until they disappear under a graded
reach or platform (Gilbert, b, p. 100), or retreat into the cave-of-the-
winds or the talus-slope under a master cliff. Among the disappearing:
falls and cliffs, those which are down-stream from a very resistant ledge
•will be the first to vanish, because their more rapid recession will soon
push them back under the slowly retreating master ledge ; the canyon
walls give many examples of this kind. On the other hand, the dis-

178 bulletin: museum of comparative zoology.

appearing falls and cliffs will remain in evidence for a much longer time
if they are up-stream from the master ledge, for there they must recede
until they are concealed beneath the backward extension of the gently
sloping graded reach or platform below them. Thus a master ledge
promotes the development of a high fall or long talus-slope beneath it,
and of a long reach or broad platform above it ; and when this relation
is established, the number of separate elements of form in valley
bottom or on valley side is much less than it was in earlier youth.
The great escarpments of the High plateaus exhibit the concentration
of all the large and small cliffs of youth in a relatively small number
of great cliffs in late maturity ; and thus still again confirm Dutton's
opinion that the erosion of the plateaus and the erosion of the canyon
took place in different cycles, separated by a strong uplift of the region.

The general principles here reviewed are of wide application, but in
regions of moderate relief and moist climate, the forms of valley sides
are seldom analyzed : certainly they are frequently overlooked in
geographical descriptions. In the Grand canyon district, where the
relief is on a huge scale, and where the arid climate lays bare every
topographical detail, the elements of form assumed under moving
streams of waste on valley sides are conspicuous ; they are glorified by
mere magnitude so that one is tempted to treat them as a new class of
topographic forms, until it is recognized that they are only new varia-
tions on an old class, examples of which are to be found in all regions
of horizontal structure. The canyon walls in the Kaibab and the great
mesozoic " terraces " that overlook the plateau from the north exhibit
the earlier and later stages of all these forms with great clearness.

Cirques, Cusps, and Niches. — The many variations in the horizontal
pattern of the cliffs in the canyon walls have been briefly described by
Dutton (c, pp. 258, 259). The cliff outline, as seen in plan, has two
expressions (Figure 1G). In some cases, the re-entrants of a cliff are
sweeping concave curves, but little notched, while the intervening
salients are sharply attenuated cusps. In other cases, the re-entrants
are narrow and acute notches, while the salients are broad and rounded
spurs. The difference between these two cases seemed to depend partly
on the drainage area of the uplands, whose waters are shed over the
cliffs from higher levels, and partly on the amount of erosion that the
cliffs have suffered ; these two factors being indirectly connected.

Where an upland sheds a stream over a cliff, the cliff will be cut back
much faster by the stream than it will be weathered back on the inter-
stream front; here the re-entrant must be an acute notch; while the

DAVIS : THE GRAND CANYON OF THE COLORADO.

179

adjoining salients will be spurs with rounded front. But where the
upland is of so small an area that it cannot gather streams, then
the head of the axial line of a re-entrant has no more importance as a
stream way than the lines that join it from either side; as a result the
retreat of the walls will be of about uniform value for a considerable
length of front, and the head of the re-entrant will here assume a con-
cave or cirque-like pattern. At the same time, the widening of the re-
entrant will narrow the adjacent spurs to mere skeletons of their original
size and sharpen their salients into cusps.

Figure 16.

Notches and cusps in cliff patterns. Diagram of two cliff-makers; the lower one showing
acute notches between rounded spurs; the upper showing rounded cirques between acute
cusps. Slightly modified from Bodiish's map of part of the canyon wall in the Kaibab.
(Dutton, c, Plate XLII.)

If this analysis be correct, rounded spurs and sharp re-entrant notches
should prevail in the cliffs near the base of the canyon walls, because
streams will generally descend into such re-entrants from the higher
slopes ; unless, indeed, time enough has elapsed for the strata overlying
a low-level cliff to have been swept away, and the area of its platform
reduced by the widening of adjoining side-canyous, so as to imitate
conditions that would prevail only at higher levels in an early stage of
erosion. On the other hand, cirque-like re-entrants should prevail in
the high-level cliffs that rim the sides of the great spurs of the canyon

180 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

walls, and the spurs between these cirques should be acutely sharpened
cusps. Cirque-like curves should occur in the rim of the plateau only
where no significant area sheds drainage from it, and notched re-en-
trants should occur in the high-level cliffs chiefly at the head of side
canyons where back-country drainage is delivered to them. As a conse-
quence of all this, a i-ather systematic relation should frequently be
found between the two sets of forms; the curved re-entrants of the
higher cliffs should frequently stand back of and above the sharp re-en-
trant notches of the lower cliffs; while the sharp spurs of the upper
cliffs should project forward along the axis of the rounded spurs of the
lower cliffs.

As far as this scheme was tested on the ground, it seemed to give
reasonable explanation to a good number of examples ; but unfortu-
nately it was not reduced to formal statement until after leaving the
canyon ; hence, as so often happens, the observations made on the
ground were less critical than they might have been if they had been
immediately accompanied by analysis.

The niches of the massive Red-wall limestone, described but left un-
explained by Dutton (c, p. 260, Plate XLI.) seem to exemplify a spe-
cial case of a problem that McGee has discussed in connection with the
"origin and hade of normal faults" (a, p. 290). The niches all occur
over the heads of subordinate ravines or gulches, and are, therefore, to
be associated with the sapping by the underlying weaker strata and the
falling away of the basal part of the massive limestone. In the absence
of numerous planes of bedding and jointing, the upward breaking of the
rock may be compared to the upward propagation of a fault ; and McGee
shows that in such case the fracture must be a curved surface with
decreasing hade upwards, so that the broken face may eventually be-
come vertical or even overhanging. The overhanging arch by which the
niche is covered seems to correspond to the upper part of such a fracture.

The Esplanade. — Although the Kaibab and the Uinkaret are only
forty miles apart, the canyon in these two plateaus exhibits very unlike
cross-profiles. In both, the double cliffs of the upper Aubrey are re-
peated with similar outlines. In the Kaibab, a platform of moderate
width is worn on both the Red-wall group and the Tonto sandstone, the
lower one being rather wider than the upper, and the two being sepa-
rated by the huge Red-wall cliff and the long gray waste-covered slope
of Tonto shales beneath it. In the Kanab ana Uinkaret sections, as
seen from Vulcan's throne at the mouth of the Toroweap valley, the
Red- wall platform is greatly widened ; it becomes a broad floor, stretch-

DAVIS : THE GRAND CANYON OF THE COLORADO. 181

ing far up and down the canyon, and fully deserving the name, esplanade,
given by Dutton ; but the Tonto platform is wanting ; the cliff's from
the top of the Red-wall descend with their steepness little diminished
nearly to the river, enclosing the narrow inner canyon or chasm ; while
the Aubrey walls of the upper or outer canyon are five miles or more
apart. It is significant that the floor of the esplanade everywhere lies
on the same stratum, namely, a calcareous sandstone ; the upper mem-
ber of the Red-wall group.1

Two Theories of the Esplanade. — Dutton explained the esplanade
as a mature valley, eroded during a pause in the uplift by which the
broadly denuded plateaus gained their present altitude. During this
pause, " the river sought and quickly found a new base-level at the bot-
tom of the great esplanade of the Grand canon. . . . The cliffs . . .
receded away from the river, gradually developing the broad avenue of
the outer chasm. . . . Again the country was hoisted, this time more
than before. . . . Swiftly the inner gorge was scoured out, and the
chasm assumed its present condition. At present the uplifting force is
inactive . . . and the river has nearly but not quite reached another
base-level" (c, p. 121).

It had been suggested to me before our trip to the canyon that it
was not necessary to assume a pause of this kind in the later uplift of
the plateaus in order to account for the esplanade, for its relation to
the resistant Red-wall group was such as to indicate its dependence on
structure rather than on a former baselevel. Several reasons for adopt-
ing this view presented themselves.

Comparison of the Kaibab and the Kanab Sections. — In the first
place, if the form of the canyon in the Uinkaret, Kanab, and Kaibab

1 In Gilbert's section of the canyon at the mouth of Kanab creek, a heavy
cross-bedded sandstone of the lower Aubrey is placed at the floor of the espla-
nade, and no such sandstone is shown in the wall of the outer canyon above
the esplanade (a, p. 70, Figure 38) ; while the section of the Shivwits canyon
sets the floor of the esplanade on the Red-wall group, and places the cross-bedded
sandstone of the Aubrey as a cliff in the wall of the outer canyon. Gilbert explains
tli is variation by the statement that the order of the hard and soft beds in the
lower Aubrey is not constant (a, p. 177) ; but Dutton says : "The cliff formed out
of the upper and lower Aubrey series is very remarkable for the constancy of its
profile throughout the entire extent of the great chasm." As far as my own ob-
servations went, they agreed with the latter view. It is possible that the cross-
bedded sandstone which Gilbert placed on the floor of the esplanade at the entrance
of the canyon of Kanab creek into the Grand canyon was really the calcareous
sandstones which as a rule forms the upper member of the Red-wall group, capping
its great cliff.

132 bulletin: museum of comparative zoology.

plateaus were dependent on recent sub-cycles or episodes of erosion,
rather than on ■structure, some indication of the esplanade ought still to
be apparent in the Kaibab section at an altitude corresponding with
that so well maintained further west ; but inasmuch as the uplifts of
the Kaibab are believed to be of much earlier date than the erosion of
the canyon, the esplanade ought to stand at a lower geological horizon
in the Kaibab than in the Kanab. No trace of an esplanade is, however,
to be recognized in the Kaibab ; the descent from plateau to river is
there accomplished by a general incline, broken only by the benches of
the Eed-wall and Tonto groups, as stated above. On the other hand,
the slopes of the Kaibab walls and the cliffs and esplanade of the Kanab
walls are easily explained if they are dependent on local structural con-
ditions ; and there is independent evidence that the appropriate condi-
tions actually occur. It is evident that if the Tonto shales that form
the slope between the Red-wall and the Tonto cliffs in the Kaibab should
become more resistant as they pass westward into the Kanab, their
face would steepen, and the Red-wall cliff would come to stand more
immediately over the Tonto; at the same time the Eed-wall platform
would broaden to form the esplanade. Observations are not lacking to
support this supposition. Gilbert noted that, at the entrance of Kanab
creek into the main river, the Tonto shales are of " firm texture " (a,
p. 70). Button says more explicitly : " In the Kanab division the
whole series of nearly 2,500 feet thickness [Red-wall and below] is won-
derfully massive, and the partings of the strata are comparatively few.
In the Kaibab the great 750-foot limestone is as solid as ever, but most
of the other members have become laminated much more minutely than in
the Kanab and Uinkaret, and are of more perishable texture" (c, p. 257).
As these values of resistance are precisely such as would undermine the
Red-wall cliff-makers in the Kaibab and support them further west,
a structural explanation for the esplanade seems to be sufficient.

Eastward Fading of the Esplanade. — In the second place, if the east-
ward extension of the esplanade be traced on the topographical maps, it
gradually loses definition by decrease in its breadth and by increase in
the number of ravines that break its front ; and in the easternmost part
of the Kanab it can hardly be recognized. This is best seen by tracing
the southern side of the canyon, for on approaching the Kaibab. the
form of the northern canyon wall is much complicated by the displace-
ments that occur there, and bedded structure has not alone a full control.
A gradual change of form of this sort is an appropriate consequence of
the structural origin suggested for the esplanade.

DAVIS : THE GRAND CANYON OF THE COLORADO. 183

Relation of the Inner and Outer Canyons. — In the third place, the
position of the inner canyon along the middle of the esplanade is, as
was suggested to me several years ago by Mr. C. H. White (then one of
my graduate students, now instructor in mining at Harvard), singularly
significant of a single period of erosion. This may be understood by
considering the two alternatives in order. If the esplanade repi'esented
a mature valley floor, broadened by the lateral swinging of the river
during a lower stand of the land, rather than by the rapid wasting of
the weak lower Aubrey layers in contrast to the persistence of the
strong Red-wall group, then when renewed uplift revived the process of
canyon cutting, the river must have occupied an irregular path along
the esplanade, and the inner canyon would have been cut at one place
near the northern wall of the outer canyon, and at another near the
southern wall. This condition is actually illustrated in the valley of
the Rhine ; here a narrow inner gorge is incised in the flat floor (espla-
nade) of a broad trough which in turn is eroded beneath the bordering
uplands ; a structural origin for the trough is inadmissible because the
rocks are greatly deformed ; and, moreover, river gravel and silt still
cover the floor of the trough. The narrow gorge is intrenched irregu-
larly along the trough floor, sometimes turning so far to one side that a
continuous descent leads from the high upland directly to the river, while
a correspondingly broad mid-level floor remains on the other side. The
same relation occurs in the valley of the Moselle, whose young gorge
meanders conspicuously from side to side in the floor of the mature
trough. The valleys of the Lot and the Dordogne in southwestern
France exhibit similar features, except that the younger or inner mem-
ber of the composite valleys are here opened wide enough to have scrolls
of incipient flood plain on one or the other side of the river, while flood
plains are as yet only just begun in the gorges of the Rhine and the
Moselle.

If, on the other hand, both the outer and the inner canyons of the
Colorado are the work of a single cycle of erosion, and the esplanade is
of structural origin, then it is necessary that the walls of the outer
canyon should retreat symmetrically on either side of the inner canyon,
and that the inner canyon should bisect the floor of the esplanade thus
produced. This case would correspond to that of all one-cycle valleys
in horizontal strata, where the symmetry of the benches and slopes on
the two walls results from their essentially equal retreat under the
weather. Innumerable illustrations of such symmetry may be seen in
the side canyons of the Kaibab section, where the Red-wall cliffs are

184 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOCxY.

1 symmetrically with respect to the narrow gorges in the Tonto
sandstones below and between them ; the detailed map of part of the
canyon in the Kaibab by Bodlish already referred to (Button, c, Plate
XLIL), gives good illustration of many examples of this kind.

Now returning to the facts, there seems to be little doubt that the
inner canyon bisects the esplanade. Its medial position is very striking
in the majestic view that we had of both these features eastward from
Vulcan's throne ; a view made famous by Holmes's wonderfully effective
drawing (Button, c, Atlas, sheet VI.). All the published maps and
sections of the district exhibit this arrangement in the most systematic
manner through the Kanab and Uinkaret plateaus, and as well in the
lateral canyons of Cataract and Kanab creeks as in the main canyon.
Even the first accounts of the canyon made mention of the medial
course of the inner chasm. Ives, descending from the plateau south of
the Kanab section by a branch canyon, reached a floor (a branch of the
esplanade) which, when seen from the enclosing cliffs, looked smooth,
but which was really covered with hills thirty or forty feet high ; " along
the centre we were surprised to find an inner canon, a kind of under cel-
lar " (p. 107). Powell's first mention of the esplanade, as seen from the
cliffs near the border of the Kaibab and Kanab plateaus, is as follows :
" The walls seem to rise very abruptly, for two thousand five hundred or
three thousand feet, and then there is a gently sloping terrace, on each
side, for two or three miles, and again we find cliffs, one thousand five
hundred or two thousand feet high. From the brink of these the pla-
teau stretches back to the north and south, for a long distance. . . .
The effect of the terrace is to give the appearance of a narrow winding
valley, with high walls on either side, and a deep, dark, meandering
gorge down its middle" (a, p. 92 ; also p. 196). It should be noted,
however, that the topographical maps show the Shivwits section of the
canyon to have walls of less symmetrical form than prevails further
east ; but not having seen this part of the canyon, I shall not venture
to discuss its complications.

R( lotion of the Esplanade to the Torotveap Fault. — Finally, it may be
noted that the early date given on page 14G for the Toroweap-Sevier
fault is not at all inconsistent with the discontinuity of the structural
esplanade where it crosses this fault-line ; for if the floor of the esplanade
is of structural control, it would be opened at whatever level the upper
surface of the Red-wall group occupied at the time of the erosion of the
canyon. If, on the other hand, the esplanade had been controlled by a
former baselevel, its discontinuity at the Toroweap would demand a

DAVIS: THE GRAND CANYON OF THE COLORADO. 185

very late date for the fault, as Dutton concluded when he wrote : " It
seems very plain that the outer chasm had been formed and attained
very nearly its present condition before the [Toroweap] fault started "
(c, p. 9-1). So recent a date of faulting seems inconsistent with the
evidence presented in the section on the Pipe spring fault.

Conclusion as to the Origin of the Esplanade. — In view of these various
considerations, it seems necessary to conclude that while many partial
cycles of erosion may have preceded the long pause during which the
broad denudation of the plateaus was completed, only a single uplift
and a single down-cutting are recorded in the canyon. It should be
noted that Dutton considered this supposition, but rejected it. Speak-
ing of the esplanade, he said : " We might explain it by assuming the
rocks of the inner gorge to be much more obdurate than those above.
This is true in part, but still the difference in this respect is insuffi-
cient. A much more satisfactory explanation is found in the sup-
position that the broad esplanade of the canon between the upper
palisades was an ancient base-level of erosion" (b, p. 121). In an-
other place he explains the difference between the Kaibab and Kanab
sections of the canyon in the following manner: "The causes which
have produced in the Kaibab a topography differing so widely from
that which is seen in the other divisions of the chasm may be readily
explained. The Kaibab is now, and throughout the period of evolution
of the chasm it always has been, higher than the other plateaus. Cor-
rasion has, therefore, penetrated there more deeply than elsewhere. It
has, moreover, laid bare the edges of the softer beds underlying the Red
Wall, and the rapid decay of these lower beds has undermined and wasted
the Red Wall to a great extent. In the other divisions of the chasm
corrasion has only at a very recent period cut below this great series of
hard limestones. . . . Besides the greater altitude leading to deeper
corrasion, the climate of the Kaibab is moister, and the degrading forces
are, therefore, more efficient " (c, pp. 257, 258). From the other point
of view, it seems as if these several considerations might legitimately
be adduced to account for the esplanade as of structural origin ; for if
differences of structure, altitude, and climate all lead to differences of
form between the Kaibab and Kanab sections, there seems to be all the
less need for explaining the esplanade in the Kanab plateau by a pause
during the uplift. On the other hand, the earlier date of the Kaibab
arch must have led to the erosion of the Aubrey limestones on its crest
much earlier than the same strata were attacked in the erosion of the
canyon further west. The Kaibab must have been trenched, as has
vol. xxxviii. — no. 4. 6

186 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

been already suggested, to a depth of one thousand or one thousand five
hundred feet (probably revealing the lower Aubrey group, but not cut-
ting through it) befoi'e the introduction of the canyon cycle, and a broad
floor, an esplanade, must have been opened through it, indifferent to
structure, at the level of the denuded plateaus on the east and west.
Since then there has been time for the deep cutting and rapid widening
of the new canyon to destroy all traces of this floor and bring about the
correlation of form and structure that to-day so fully characterizes this
section of the canyon.

Hints for a Visit to the Canyon. — A few words as to the descent
into the canyon from the Coconino rim opposite the Kaibab, with special
reference to a sight of the " wedge," may not be out of place. The best
geological map of this district is that made by Walcott (e, Plate LX.).
The Grand View trail, descending from Cameron and Berry's, gives a
convenient approach to the apex of the " wedge." Horses for riding and
donkeys for carrying a small pack of blankets and provisions can be
hired at the hotel ; but if the visitor goes on foot, an early morning
start should be made, before sunrise, if possible, as the temperature in
the canyon is very high over the noon hours during clear summer
weather. It would be certainly a great advantage to plan a visit at the
time of full moon, so as to be able to use the night hours for at least
part of the climb out of the canyon. The Coppermine spur may be
easily reached in the morning before the greatest heat sets in, and noon
may be passed at its end on the Red-wall platform, studying the
" wedge," as exposed in the northern wall, and examining the canyon
to the east and west. A light shelter of canvas to keep the sun off
would be a great comfort in this work, as shade is scanty and the glare
from the bare rocks is fatiguing while one is attentively studying details
of structure and form. A field-glass greatly aids and extends observation
in all parts of the canyon, for distances are large. A camp for the night
may be made in the side canyon next west of the Coppermine spur,
where a small stream is found just above the level of the basal Tonto
sandstone ; but the descent to this spot will be more enjoyed if it is left
till late in the day, as the trail down a gulch in the Red-wall cliff is
directly exposed to the afternoon sun. A second day can be well spent
in descending to the river and in examining the palaeozoic floor ; or in pass-
ing around the Tonto platform at the base of the Coppermine spur so as
to enter the next ravine on the east, where the " wedge " seemed to be
exposed on the south side of the canyon. When viewed from above, the
various slopes and platforms seem very smooth, as if one could walk over

DAVIS: THE GKAND CANYON OF THE COLORADO. 187

them without difficulty ; but when actually upon them, the smoothness
vanishes; the dimensions of every item of form are magnified far beyond
expectation ; everything is rough and rugged, and walking is tiresome
work. A third day can be given to various problems regarding stratified
structures which are here open to observation in a larger way than almost
anywhere else in the world. The mere sight-seer may be content with a
brief trip down and back, but the geologist who can go as far as the
canyon ought to spend several entire days in its depths. We were
unfortunately so hurried as to have only an afternoon, a night, and a
morning for the canyon.

If the canyon is visited from the north, the trip may be made on
horseback or with wagon from Belknap station, Rio Grande Western rail-
road, southward to Kanab, and thence either southeast to the Kaibab
or southwest to the TJinkaret plateau and the Toroweap valley. The
former trip repeats the views seen from the Coconino rim, but it is
without the advantage of good trails by which the descent may be made
to the river. The latter takes the observer to a great field of volcanic
phenomena as well as to the brink of the canyon at a most interesting
point. After recent rains, water-pockets can be depended on for camp-
ing near Vulcan's throne ; but if the season has been dry, water must
be brought from Oak spring on the TJinkaret. Although this is trouble-
some, it is entirely feasible. Two nights and a day are the shortest
time that should be wisely allowed to this wonderful spot : it was a real
hardship that our party had only a day for the ride from Oak spring to
Vulcan's throne and back.

Former Climates of the Grand Canyon District.

Diverse Opinions of Early Observers. — The several opinions ex-
pressed by Newberry, Powell, and Dutton as to the former climates of
the region under discussion seem to me difficult to maintain, so many
are the doubtful elements iu problems of this class. Newberry inferred
a former greater rainfall (p. 47), apparently because of the immense
amount of denudation that has been accomplished ; but until more is
known as to the time occupied in the great denudation, it is impossible
to make inferences as to its rate, and hence as to the strength of the
eroding and transporting agencies, and the rainfall that excites them.
Powell inferred a long-maintained arid climate, because " in a region of
country where there is a greater amount of rainfall, the tendency is to
produce hills and mountains, rather than plateaus and ridges, with

188 bulletin: museum of comparative zoology.

escarpments " (a, p. 204). It may be urged against this view that
escarpments of a strengtli appropriate to their capping layers occur in
better-watered regions, as in the Catskill mountain front of eastern
New York, and in the Swabian Alp of southern Germany; and that
the cause of the great cliffs of the Grand canyon district is to be found
chieflv in the massive thickness of the resistant strata that guide them,
and in the weakness of the underlying strata that sap them. Truly,
the sharpness of the cliffs is highly suggestive of aridity, but a relatively
short arid period would suffice to sharpen the cliff profiles, even if they
had been somewhat dulled by a previous humid pex-iod.

The narrowness of the canyons, so generally explained by the aridity
of the plateaus through which the vigorous Colorado flows, certainly
finds a large part of its explanation also in the recency of the uplift by
which the canyon cutting was initiated, and in the massiveness of the
resistant layers by which the canyon walls are defended. The side
canyons by which the plateaus are dissected are not insufficient in num-
ber even for a moist climate ; they are much more numerous than the
perennial side streams. The latter are notoriously rare : the former are
present in good number, and in wet weather they are actively washed
by their temporary floods. It is true that the side canyons are of so
steep a descent along their floors that they lose much of their depth at
a moderate distance back from the main river ; but this may be a conse-
quence of recent uplift as well as of slow corrasion. The side canyons
branch frequently enough to satisfy an active drainage system, and cer-
tainly there is no deficiency of ramifying valleys on the higher uplands
where the surface is so thoroughly and maturely dissected. It is the
steep grade of the waste on the floors of these valleys that suggests a
development under an arid climate, and such a grade would be soon
acquired under a dry climate even if the valleys had once been cut
somewhat deeper under a moist climate. Hence, even in the more
recent past of the canyon cycle, a humid climate seems no more impos-
sible than unnecessary, and in the more distant past of the plateau
cycle, climates of any and all kinds might have prevailed, as far as the
present topography of the Grand canyon district is concerned.

Moist Miocene and Arid Pliocene Climates. — In contrast to Powell,
Dutton concluded that the Miocene (plateau cycle) was humid. This
opinion seems to have been based partly on the occurrence further north
of extensive fresh-water deposits of Miocene age, usually interpreted as
hiving been laid down in large lakes, whose existence pointed to a good
supply of rainfall (c, p. 223), and also on the apparent desiccation near

DAVIS: THE GRAND CANYON OF THE COLORADO. 189

the beginning of the Pliocene (canyon cycle) of certain branches of the
Colorado, whose previous existence implied the continuance of a good
water supply, just as their later disappearance implied its cessation
(c, pp. 99, 201, 223) ; the arid climate thus introduced continued to
the present day, except for the occurrence of an inferred pluvial period
corresponding to the glacial period elsewhere.

Other interpretations seem to me possible for several of the facts here
brought forward. I have elsewhere presented some considerations re-
garding the possible explanation of at least some of the fresh-water Ter-
tiary deposits of the Rocky mountain region as fluviatile rather than as
lacustrine formations (b, p. 360), and in so far as this alternative explana-
tion is found applicable, the evidence above quoted for the humidity of
the plateau period will weaken. The ai'gument for a change from humid
to arid climate, based on the apparent disappearance of certain streams,
seems to me open to serious question ; it is considered in the next sec-
tion. The ravines that are adduced by Dutton as the evidence of a
brief Pleistocene pluvial period appear to be open to explanation under
as arid a climate as prevails to-day, as will be specially considered in
the next section but one. Thus the climatic history of the district
becomes extremely uncertain. Whatever conclusions are reached as to
the climatic conditions of any period of past time — the glacial period,
for example — in surrounding regions may be permitted for the Grand
canyon district, but no special climatic conditions are demanded by
local evidence. Yet among the several opinions quoted above, the long-
enduring arid climate suggested by Powell seems to me the most prob-
able, because for a long time the Grand canyon district has lain to
leeward of a mountainous area, and during much of this time the dis-
trict stood lower than it does now, while the mountains to windward
were higher.

The Toroweap. — The Toroweap (T, Figure 1) has already been
referred to as a valley that has been eroded along or near the southern
part of a long fault-line that comes from the northeast past Pipe spring ;
it is part of the boundary between the Uinkaret and Kanab plateaus.
The peculiar feature of the valley now in evidence is its shallowness ;
for its broad floor leads forward to the level of the esplanade and shows
in its surface form hardly any sign of incision to a deeper level. Dut-
ton infers from this that the stream by which the valley was originally
eroded became extinct about the time of the latest upheaval of the
region, after which the inner canyon was cut down beneath the espla-
nade by the main river, whose water supply in distant mountains had

190 bulletin: museum of comparative zoology.

not been cut off (c, p. 99). There are two arguments against this view.
First, a number of side canyons in the neighborhood of the Toroweap.
of similar or smaller drainage area and equally dry, are cut down to
essentially accordant junction with the bottom of the main canyon. A
Bmall valley of this kind is seen to enter the main canyon from the
south at a short distance west of the South Toroweap ; 1 a view of its
upper part is included in a drawing made by Holmes to illustrate the
dikes in the canyon wall (Dixtton, c, Plate XVIII.). Although the
drainage area from which the wet-weather stream is here gathered is
but a small fraction of that which supplies the floods of the Toroweap,
this short side valley is cut deep below the esplanade level for several
miles back from the inner canyon wall, and as its stream line descends
with the steep grade appropriate to short lateral canyons, it makes an
essentially accordant junction with the Colorado. A view of the inner
canyon from near Vulcan's throne (Dutton, c, Plate XVII.) has al-
ready been referred to as showing the accordant entrance of two side
canyons into the main canyon in the neighborhood of the Toroweap.
The same relation obtains in the case of an unnamed side canyon, com-
ing from the north twelve miles east of the Toroweap : its drainage area
is similar to that of its high-floored neighbor, yet it is cut down so that
its temporary floods may join the main river in perfectly accordant
fashion, as far as one may judge from the topographic map of the local-
ity. Many other examples of this kind might be given, as could be
inferred from what has already been said in a previous section as to the
generally accordant junction of side and main canyons. It, therefore,
does not seem possible to ascribe the failure of erosion in the Toroweap
to the desiccation of a once permanent stream ; for in that case all the
neighboring streams in small side canyons must also have been desic-
cated, and should have high floors ; yet as a matter of fact their streams
have not ceased to erode effectively.

Secondly, some local obstacle to erosion in the Toroweap would suffice
to explain its peculiar form, and such an obstacle certainly occurs there,
for the broad floor of the valley has been heavily and repeatedly sheeted
over with floods of resistant lava, supplied by the magnificent lava cas-
cades from the Uinkaret, as pictured in one of Holmes's most effective
drawings (Dutton, c, Atlas, sheet V.). The occurrence of the lava
flooring is fully recognized by Dutton, who wrote : " There is reason to
believe that at some prior epoch it [the Toroweap] was cut a few hun-

1 I have used this name for a valley corresponding to the Toroweap and appar-
ently structurally continuous with it but south of the canyon.

DAVIS: THE GRAND CANYON OF THE COLORADO. 191

dred feet deeper than its present floor, and was subsequently built up
by many floods of basalt coming from the cones on the Uinkaret and by
considerable quantities of alluvium washed from its cliffs and overlook-
inw mesas" (c, p. 92). We may therefore conclude that in the early
stages of the can}ron cycle, when the intermittent side streams had cut
down their canyons into the Red-wall group and the canyons had al-
ready widened somewhat through the sapping of the upper Aubrey cliff-
makers by the weak red beds of the lower Aubrey, the Toroweap was
flooded with lava, while most of its neighbors remained free from vol-
canic interference. The streams in the latter continued to deepen their
courses, while deepening was practically stopped in the Toroweap. But
just as the main canyon has widened above the resistant floor of the
esplanade, so the Toroweap continued to widen, hence it is to-day nor-
mally broad but abnormally shallow. It is a " hanging valley " because
of local volcanic interruption of the normal work of canyon-cutting.

Directly opposite the Toroweap is a similar high-floored valley
(c, p. 99), splendidly exhibited from Vulcan's throne, and already re-
ferred to as the South Toroweap. Like its northern fellow, it is floored
with lava near the main canyon, on which it opens close to the es-
planade level (Dutton, c, Plate XVIIL). Hence its deficiency of depth
again seems to be best accounted for by the difficulty of wearing away
its lava sill, an accidental and purely local detail, rather than by a
failure of rainfall, which must have been general.

Dutton mentions several other high-floored valleys, which he
classes with the Toroweap as indicating a decrease of rainfall at the
beginning of or early in the Pliocene (canyon cycle). One of these is
the Queantoweap, which follows the Hurricane fault along the boundary
between the Uinkaret and Shivwits plateaus (c, pp. 99, 115). Not
having seen this valley, I shall not venture to express an opinion about
its origin, yet it may be noted that a small flow of lava near its mouth
is marked on the geological map (Dutton, c, Atlas, sheet VIII.) ; and
judging by the great lava cascades that plunge into the valley (ibid., c,
p. 116), some lava may lie concealed beneath the alluvium of the
valley floor. A third high-floored dry valley mentioned by Dutton is
that on the summit of the Kaibab uplift (c, pp. 194, 197, 223) already
discussed and explained as a series of independent anticlinal valleys,
and thus seeming to be within reach of explanation without recourse to
climatic change. A fourth example is House-rock valley (c, p. 201),
which has already been referred to as a normal subsequent valley worn
on weak monoclinal strata. Taken altogether, the lava-floored valleys,

192 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

the anticlinal valleys, and the monoclinal valley seem to be indifferent
witnesses as to changes of climate in the past.

The Pluvial Equivalent op the Glacial Period. — There is much
probability that the rainfall over the Grand canyon district was in-
creased at those times when ice sheets and glaciers were formed over
other parts of the continent ; yet it seems to me difficult to find in-
dependent proof by purely local evidence that such has been the case.
Button, however, suggests that evidence of this kind is to be found
in the ravines of the Kaibab plateau, concerning which he finds " no
conjecture so satisfactory as that which supposes that during the glacial
period the rainfall was sufficient to sustain living streams in these
ravines and that they were then carved by running water" (c, p. 196).
The same conjecture is thought to hold true for the ravines of the Paria
plateau (c, pp. 202, 228). Nevertheless, it is difficult to find valid
reason for ascribing the erosion of these ravines only to the glacial
period, for it seems admissible to conclude that they can have been in
large part eroded by the intermittent streams of a climate as arid as
that of to-day. The maturity of many of these ravines in resistant
strata indicates that, besides glacial time, some of preglacial and all of
interglacial and of postglacial times must have been needed for their
erosion; hence instead of ascribing theui to the glacial period alone,
their beginning may be carried back even into the later part of the
cycle of the great denudation, as has already been suggested on page
141. Moreover, existing processes do not seem to be inoperative.
The stream beds in the lower ravines of the Kaibab bear every mark of
being strongly flushed by occasional wet-weather floods. Well-defined
fans of coarse and fine waste extend forward from the mouth of the
ravines at the eastern base of the Kaibab into House-rock valley,
as has been already stated ; and a torrent fan was seen at the western
base where the road from Jacob's lake descends toward Fredonia.
However slow the erosion of the ravines may now be, it is still going
on ; indeed, the scantiness of vegetation makes the removal of waste so
easy that it may be questioned whether rain and stream work is not
advancing here at a comparatively rapid rate, — a rate that very likely
far exceeds that which prevails in the peat-covered uplands of much
moister regions.

Volcanic Phenomena.

Our interest was frequently attracted by magnificent displays of vol-
canic phenomena in various parts of the Grand canyon district, but we
had no time to turn from our route to study them. A few notes on

DAVIS: THE GRAND CANYON OF THE COLORADO.

193

the more striking things are here presented in the hope that they may
prove of service to more leisurely travellers.

The dissection of Mt. San Francisco, as seen from the wagon road
that runs northward from Flagstaff along the western base of the

Figl-re 1

Diagram to illustrate a proposed terminology of spurs and ravines. Constructed from
several sketches of the western side of Mt. San Francisco, a dissected volcano.

mountain, suggested a simple terminology for spurs and ravines that
might perhaps be serviceable in detailed descriptions of mountain forms.
The scheme is illustrated
in the accompanying fig-
ure, 17, which repeats some
of the features illustrated
in the diagram of the Per-
mian scarps (Figure 9). I
shall hope to return to
this phase of morphology
at some future time. The
young ash cones, still hold-
ing unbreached craters,

were numerous in the volcanic field north of Mt. San Francisco (Fig-
ure 18). About ten miles west of Hull spring (on the Flagstaff-canyon
road) a large ash cone was seen with a great breach in its eastern

Figure 18.

An ash cone and crater, north of Mt. San Francisco
and east of Hull spring.

194 bulletin: museum of comparative zoology.

slope, a promising specimen for the study of the structure of such
cones. One of the northernmost of the eruptions in this field (Y, Fig-
ure 1) seems to be also one of the youngest ; it lay several miles south
of the rough trail that we followed eastward from Hull spring to the
Flagstaff-Tuba road, and as seen at that distance it seemed to form a
ragged black mesa, two or three hundred feet high. The various rela-
tions of lava flows to small valleys were beautifully exhibited in the same
district. In one case a flow banked against the scalloped escarpment of
the upper Aubrey, and a little waterway has since been eroded along
the line of junction, with yellow-gray limestone on one side and black
lavas on the other ; this was well seen on the canyon road, a few miles
north of Hull spring. Further east, we followed valleys eroded to
a depth of two or three hundred feet in the Aubrey layers, and then
floored with streams of lava from some neighboring cone. In one nar-
row valley of this kind, the wash of waste from the walls seems to have
prevented the formation of new waterways along the lava margin ; here
the present stream bed lies for a distance on one side of the valley, then
trenches obliquely across the lava and follows its other side for a
stretch. In a broader valley, the lava surface was untrenched, and new
waterways ran persistently along its margins, this being evidently the
incipient stage in the formation of a lava mesa. At Lockett's tank
(probably several miles north of Black tank of the topographic map) a
narrow canyon of moderate depth in the Aubrey limestone has been
filled nearly to its brim with a slender but heavy lava flood, but at
present the stream has re-excavated part of its valley, consuming the
terminal part of the lava flow for half a mile or more, although leaving
scraps of lava here and there, frozen to the walls ; and at the head of
the new valley is an abrupt fall from the surface of the lava that still
remains (Figure 13). Here the wet-weather floods have scoured a
basin, in which water remains long after the supplying storm has
cleared away ; cattle tracks converge on the dry upland from all sides
towards this tank.

A narrow dike was noted on the west-facing slope of the Triassic
escarpment, just south of the valley of the Moencopie and not far from
Tuba ; the dike formed a sharp ridge, quite unlike the more tabular
forms normally associated with the escarpment.

The great lava cascades that descend from the Uinkaret into the Toro-
weap and even into the canyon itself are among the most magnificent
geological phenomena of the region. Our camp by Oak spring, south of
Mt. Trumbull, was at the scarped edge of one of the younger lava flows

DAVIS: THE GRAND CANYON OF THE COLORADO. 195

of the Uinkaret ; and the youngest flow of that volcanic field lay a mile
to the east (Powell, a, p. 131 ; Dutton, c, p. Ill, Atlas, sheet IX., lower
view) ; we crossed it on the way down into the Toroweap, and came
back by a trail that ascended one of the great lava cascades. A young
ash cone stands on the tabular top of Mt. Trumbull, whose mass is but
the remnant of a once much larger lava mesa ; the young cone was pecu-
liarly placed near its southeastern end of the mountain. A series of
cones and black lava flows were passed in the northern part of the Uin-
karet block, between the western scarps of the Shinarump mesa and the
Hurricane ledge. The lava-floored valley above Workman spring has
already been mentioned. The canyon of the Virgin just after the river
emerges from the Hurricane ledge is cut in lava for several miles.
A side canyon, about two miles south of Toquerville, revealed a fine
section of a lava flow lying on a coarse gravel which partly filled a small
valley, all shown in a single section. These latter items, along with
many othei's, would be included in the special study of the Toquerville
district, which may be so highly commended as a sixbject for a summer's
field work.

Summary.

A high degree of consistence among the consequences of various
associated theories is justly regarded as strong proof of the correctness
of the theories themselves. Each of the earlier students of the Grand
canyon district probably reviewed his conclusions to see that they were
mutually consistent, and published only those that survived this test.
As the progress of exploratiou and the advance of earth-science furnished
more and more items in the sequence of events that make up the history
of the district, the interlacing of more and more elements seemed to give
at once greater difficulty to the problem and greater certainty to its
solution. Dutton in particular appeals this aspect of the method of
proof, and truly the successive steps in his thesis are strongly enchained.
Yet it now seems possible to arrange a new sequence for some of
the events by which the present order of things has been produced:
a sequence which differs from that announced by previous observers in
several particulars, as presented in the concluding summary below, al-
though still holding to the main outlines of the geological history of the
region as presented in the several governmental reports. The question
then arises, where shall the permanent truth be found among all these
suggested possibilities'? Further exploration, and more particularly
detailed study of certain special problems, will in due time review all

196 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

the explanations now offered and select the survivors from among them.
The work of new observers is especially needed in the examination of
certain structures, as in the neighborhood of Toquerville, and between
Pipe spring and the Toroweap ; and in the discussion of dates of certain
past phenomena, such as flexing, faulting, and uplifting; and in the dis-
covery of the origin of the drainage system and the character of the
climate of past periods. Whether the conclusions here announced shall
then stand or fall in whole or in part, it would be a great reward to the
writer if they might afford later students of the region even a small
share of the aid that he has derived from the earlier work of Newberry
and of Powell, of Gilbert and of Dutton.

The results reached in the foregoing pages may be summarized as
follows. There is some probability that the San Rafael swell, like the
Waterpocket flexure, is of pre-Tertiary origin. The other deformations
of the region, both flexures and faults, are almost exclusively of much
earlier date than the canyon cycle, and they may have been formed
relatively early in the erosional history of the district. The total denu-
dation of the region thus far accomplished may be considered in two
parts, of which the first — the great denudation — was far advanced
before the general uplift by which the second — the erosion of the
canyon and the stripping of weak strata from the plateaus — was intro-
duced. But the great denudation was complicated by repeated move-
ments, after each of which the processes of erosion may have reached an
advanced stage before the occurrence of the next series of disturbances.
It is only by an analysis of these repeated movements and revived
erosions that the origin of the drainage system can be determined.
As far as this analysis can be attempted at present for the Grand
canyon district, the side streams seem to be of various origins, except
that none of them appear to be antecedent. The Colorado itself may
be in part antecedent to some of the many dislocations that the district
has suffered, but it seems to be for the most part consequent on the
displacements caused by faulting in the later part of the great denuda-
tion, and on the form that the surface had assumed at that time.
The floor of the Toroweap valley is higher than the neighboring valley
floors, because it is sheeted with heavy lava flows which have effectively
withstood the intermittent erosive efforts of wet-weather floods. The
past climate of the region cannot be safely determined ; a change from
a humid to an arid climate at the close of the Miocene does not appear
to be demanded by the facts that have been appealed to in its support.

DAVIS: THE GRAND CANYON OF THE COLORADO. 197

BIBLIOGRAPHY.

Campbell, M. R.

Drainage Modifications and their Interpretation. Journ. Geol. (Chicago),
Vol. IV., 1896, pp. 567-581, 657-678, (571-581).

Campbell, M. R. and Mendenhall.

Geologic Section along the New and Kanawha Rivers in West Virginia.
17th Ann. Rep. U. S. Geol. Surv., Part II., pp. 479-511.

Davis, W. M.

a. Notes on the Colorado Canyon District. Amer. Journ. Sci., Vol. X., 1900,

pp. 251-259.

b. The Freshwater Tertiary Formations of the Rocky Mountain Begion.

Proc Amer. Acad., Vol. XXXV., 1900, pp. 345-373.

c. Glacial Erosion in France, Switzerland, and Norway. Proc. Boston Soc.

Nat. Hist., Vol. XXIX., 1900, pp. 273-322.

d. Glacial Erosion in the Valley of the Ticino. Appalachia, Vol. IX., 1900,

pp. 136-156.

Dutton, C. E.

a. Report on the Geology of the High Plateaus of Utah, with Atlas. U. S.

Geog. and Geol. Surv. Rocky Mountain Region, Washington, 18S0.

b. The Physical Geology of the Grand Canon District. 2d Ann. Rep. U. S.

Geol. Surv., 18S2, pp. 49-166.

c. Tertiary History of the Grand Canon District, with Atlas. Monogr. II.,

U. S. Geol. Surv., 1882.

Emmons, S. F.

a. U. S. Geological Exploration of the Fortieth Parallel. Vol. II., Descrip-

tive Geology, Washington, 1877-

b. The Origin of Green River. Science, Vol. VI., 1897, pp. 19-21.

Fairbanks, H. W.

Notes on the Geology of Eastern California. Amer. Geol., Vol. XVII. , 1S96,
pp. 63-74.

Freeh, F.

Section in Congress Canyon opposite Point Sublime. Compte-rendu 5th In-
ternal Congress of Geologists. New York, 1894, pp. 476-481.

198 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Gilbert, G. K.

a. Report on the Geology of Portions of Nevada, Utah, California, and Ari-

zona, examined in the years 1871 and 1872. Engineer Dep't U. S.
Army. Report upon Geog. and Geol. Expl. W. of 100th Merid. Vol.
III., Washington, 1875. Part I. pp. 17-187-

b. The Colorado Plateaux Province as a Field for Geological Study. Amer.

Journ. Sci., Vol. XII., 1876, 16-21, 85-103.
c Report on the Geology of the Henry Mountains. Dep't of the Interior,
U. S. Geog. and Geol. Survey Rocky Mountain Region, Washington,
1877.

d. Lake Bonneville. Monogr. I., U. S. Geol. Suit., Washington, 1890.

e. Flagstaff to Grand Canyon. Geol. Guide-book for an Excursion to the

Rocky Mountains. Compte-rendu 5th Internat. Congr. of Geol. New-
York, 1894, pp. 472, 473.

Hayes, C. W.

Physiography of the Chattanooga District in Tennessee. 19th Ann. Rep.
U. S. Geol. Surv., Part II., 1899, pp. 1-58, (58).

Hayes, C. W., and Campbell, M. R.

Gemorphology of the Southern Appalachians. Nat. Geogr. Mag., Vol. VI.,
1894, pp. 63-126, (105-107).

Holmes, W. H.

The Drawings by which the Colorado Canyon is so finely illustrated in
Dutton's Reports (b. c), are from the pencil, pen, and brush of this
artist-geologist.

Howell, E. E.

Report on the Geology of Portions of Utah, Nevada, Arizona, and New
Mexico, examined in the years 1872 and 1873. Engineer Dep't U. S.
Army, Rep. upon Geog. and Geol. Expl. W. of 100th Merid., Vol. III.,
Washington, 1875. Part III., pp. 227-301.

Ives, J. C.

Report upon the Colorado River of the West. Senate, 36th Congress, 1st
Session, in Washington, 1861, Part I. General Report, pp. 13-131.

Jefferson, M. S. W.

The Antecedent Colorado. Science, Vol. VI., 1897, pp. 293-295.

Jukes, J. B.

On the Mode of Formation of Some of the River Valleys in the South of
Ireland. Quart. Journ. Geol. Soc, Vol. XVIIL, 1862, pp. 378-403.

King, C.

U. S. Geological Exploration of the Fortieth Parallel. Vol. I. Systematic
Geology. Washington, 1S78.

DAVIS: THE GRAND CANYON OF THE COLORADO. 199

McGee, W.J.

a. On the Origin and Hade of Normal Faults. Amer. Journ. Sci., Vol.

XXVI., 1883, pp. 294-29S.

b. Sheetfiood Erosion. Bull. Geol. Soc. Amer., Vol. VIII., 1897, pp.

87-112.
c Papagueria. Nat. Geogr. Mag., Vol. IX., 1S98, pp. 345-371.

Marvine, A. R.

Report on the Geology of Route from St. George, Utah, to Gila River, Ari-
zona. Engineer Dep't U. S. Army, Rep. upon Geog. and Geol. Expl.
W. of 100th Merid., Vol. III., Washington, 1875, Part II., pp. 189-225.

Merriam, C. H.

Results of a Biological Survey of the San Francisco Mountain Region and
Desert of the Little Colorado, Arizona. N. Amer. Fauna, No. 3, U. S.
Dep't of Agriculture, Washington, 1890.

Newberry, J. S.

Report upon the Colorado River of the West. Senate, 36th Congr., 1st Session.
Washington, 1861. Part III., Geological Report, pp. 1-154.
Penck, A.

Studien iiber das Klima Spaniens wahrend der jiingeren Tertiar-periode und
der Diluvialperiode. Zft. Ges. Erdk. Berlin, Vol. XXIX., 1894, pp.
109-141.

Philippson, A.

Das russische Flachland. Zft. Ges. Erdk. Berlin, Vol. XXXIIL, 1898, pp.
37-68.

Powell, J. W.

a. Exploration of the Colorado River of the West and its Tributaries, ex-

plored in 1869, 1870, 1871, and 1372, under the Direction of the Sec-
retary of the Smithsonian Institution. Washington, 1875.

b. Report on the Geology of the Eastern Portion of the Uinta Mountains

and a region of country adjacent thereto, with Atlas. Dep't of the In-
terior. U. S. Geol. and Geog. Survey of the Territories. Washington,
1876.
c Canyons of the Colorado. Meadville, Pa., 1895. (A general account of
the Colorado River region, with reprint of certain portions of the origi-
nal report and with the addition of many illustrations from various geo-
logical reports.)
d. The Canons of the Colorado. Scribner's Monthly, Vol. IX., 1S75, pp.
293-310, 394-409, 523-537-
Russell, I. C.

a. A Geological Reconnaissance in Southern Oregon. U. S. Geol. Surv. 4th

Ann. Rep., Washington, 1884, pp. 431-464.

b. A Reconnaissance in Central Washington. Bull. No. 108, U. S. Geol.

Surv., 1S93.

200 bulletin: museum of comparative zoology.

Walcott, C. D.

a. Pre-Carboniferous Strata in the Grand Canon of the Colorado, Arizona.

Amer. Journ. Sci., Vol. XXVI., 1883, pp. 437-442, 484.

b. The Fauna of the Lower Cambrian or Olenellus Zone. 10th Ann. Rep.

U. S. Geol. Surv., 1890, pp. 509-760.

c. The North American Continent during Cambrian Time. 12th Ann. Rep.

U. S. Geol. Surv., 1891, pp. 523-568.

d. Study of a Line of Displacement in the Grand Canon of the Colorado, in

Northern Arizona. Bull. Geol. Soc. Amer., Vol. I., 1890, pp. 49-64.

e. Pie-Cambrian Igneous Rocks of the Unkar Terrane, Grand Canyon of the

Colorado, Arizona. 14th Ann. Rep. U. S. Geol. Surv., 1S94, pp. 497-
519.

Wheeler, G. M.

Itinerary of Colorado Grand Canyon and River Trip of 1871, with Map.
Engineer Dep't U. S. Army, Rep. upon Geog. Surveys W. of 100th
Merid. Vol. I. Geog. Report, Washington, 1SS9, Chap. II., pp. 147-
171. (A brief account of all previous explorations of the Colorado
River is here included.)

DAVIS: THE GRAND CANYON OF THE COLORADO. 201

EXPLANATION OF PLATES.

PLATE 1.

General view of the Colorado Canyon, looking about north from O'Neill's point to
the Kaibab plateau. Copyright photograph by Detroit Photographic Co.

PLATE 2.

A. A shallow valley floor in the Coconino forest, south of the Colorado canyon ;
a hill and drying-ground of the farmer ant in the foreground. Photo-
graph by Dr. H. E. Gregory.

R. A boulder of Shinarump sandstone perched on a pedestal of weak Permian
shales at the base of the Shinarump bench, beneath Vermilion cliffs,
Paria plateau, southwest of Lee's Ferry. Photograph by Dr. H. E. Gregory.

VOL. xxxviii. — no. 4.

■■■'!., \ i? ?

; tram

*, ff

The following Publications of the Museum of Comparative Zoology
are in preparation : —

Reports on the Results of Dredging Operations in 1877, 1878, 1879, and 1880, in charge of Alex-
anduk Agassiz, by the U. S. Coast Survey Steamer " Blake," as follows: —

E. EHLERS. The Annelids of the " Blake."

C. HARTLAUB. The Comatuhe of the " Blake," with 15 Plates.

H. LUDW1G. The Genus Pentacrinus.

A. MILNE EDWARDS and E. L. BOTJVIER. The Crustacea of the "Blake."

A. E. VERRILL. The Alcyonaria of the " Blake."

Reports on the Scientific Results of the Expedition to the Tropical1 Pacific, in charge of
Alexander Agassiz, on the IT. S. Fi>h Commission Steamer "Albatross," from August,
1899, to March, 1900, Commander Jefferson F. Moser, U. S. N., Commanding.

Illustrations of North American MARINE INVERTEBRATES, from Drawings by BuRK-
hakdt, Sonrkl, and A. Agassiz, prepared under the direction of L. Agassiz.

LOUIS CABOT. Immature State of the Odonata, Part IV.
E. L. MARK. Studies on Lepidosteus, continued.

" On Arachnactis.

R. T. HILL. On the Geology of the Windward Islands.
W. McM. WOOD WORTH. On the Bololo or Palolo of Fiji and Samoa.
A. AGASSIZ and A. G. MAYER. The Acalephs of the East Coast of the United States.
AGASSIZ and WHITMAN. Pelagic Fishes. Part II., with 14 Plates.
J. C. BRANNER. The Coral Reefs of Brazil.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission Steamer
"Albatross," Lieutenant Commander Z. L Tanner, U. S, N., Commanding, in charge of
Alexander Agassiz, as follows: —

A. AGASSIZ. The Pelagic Fauna.

" The Echini.

'* The Panamic Deep-Sea Fauna.

K. BRANDT. The Sagittse

" The Thalassicolas.

C. CHUN. The Siphonophores.

" The Eyes of Deep-Sea Crustacea.

W H DALL. The Mollusks.
H. J. HANSEN. The Cirripeds.
W. A. HERDMAN. The Ascidians.
S. J. HICKSON. The Antipathids.
W. E. HOYLE. The Cephalopoda.
G. VON KOCH. The Deep-Sea Corals.
C. A. KOFOID. Solenogaster.
R. VON I,ENDENFELD. The Phospho-
rescent Organs of Fishes.

H. LUDWIG. The Starfishes.
J. P. McMURRICH. The Actinarians.
E. L. MARK. Branchiocerianthus.
JOHN MURRAY. The Bottom Specimens.
P. SCHTEMENZ. The Pteropods and Hete-

ropods.
THEO. STUDER. The Alcyonarians.
M. P. A. TRAUTSTEDT. The Salpidaj and

Doliolidse.
E. P. VAN DUZEE The Halobatidse.
H. B. WARD. The Sipunculids.
H. V. WILSON. The Sponges.
W. McM WOOD WORTH. The Nemerteans.
" Tlie Annelids.

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletins Vols. I. to XXXV. ;
of the Memoirs, Vols. I. to XXIV.

Vols. XXXVI., XXXVII., and XXXVIII. of the Bulletin,
and Vols. XXV., XXVI., XXVII. of the Memoirs, are now in
course of publication.

The Bulletin and Memoirs are devoted to the publication of
original work by the Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural History, and of work by specialists based
upon the Museum Collections and Explorations.

The following publications are in preparation: —

Reports on the Results of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander ( \ 1>. Sigsltee, U. S. N., and Commander J. R. Bartlett, U. S. N.,
Commanding.

Reports oil the Results of the Expedition of 1891 of the U. S. Fish Commission
Steamer " Albatross," Lieut. Commander Z L Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in
charge of Alexander Agassiz, on the U. S Fish Commission Steamer
"Albatross," from August, 1899, to March, 1900, Commander Jefferson F.
Moser, U. S. N., Commanding.

Contributions from the Zoological Laboratory, Professor E. L. Mark, Director.

Contributions from the Geological Laboratory, in charge of Professor N. S.
Shaler.

Subscriptions for the publications of the Museum will be received
on the following terms : —

For the Bulletin, S4 00 per volume, payable in advance.
For the Memoirs, $8.00 "

These publications are issued in numbers at irregular inter-
vals; one volume of the Bulletin (8vo) and half a volume of the
Memoirs (4 to) usually appear annually. Each number of the Bul-
letin and of the Memoirs is also sold separately. A price list
of the publications of the Museum will be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge; Mass

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 5.

THE GEOLOGY OF THE NORTHEAST COAST OF LABRADOR.

By Reginald A. Daly.

With Thirteen Plates.

CAMBRIDGE, MASS , U. S. A. :

PRINTED FOR THE MUSEUM.

February. 1902.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 5.

THE GEOLOGY OF THE XORTHEAST COAST OF LABRADOR.

By Reginald A. Daly.

With Thirteen Plates.

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

PRINTED FOR THE MUSEUM.

February, 1902.

F£6

I9Q2

No. 5. — The Geology of the Northeast Coast of Labrador.
By Eeginald A. Daly.

TABLE OF CONTENTS.

Introduction

Itinerary of the cruise

Observations on Topography and

Bed-rock Geology . . .

From the Straits of Belle Isle to

Paul's Island

General Topography . . .

The Geology

Sedimentary rocks at Pomi-

adluk Point

Sedimentary rocks at Aillik

Bay

The Nain gabbros ....

The Kiglapait

The Kaumajet Mountain group
The Torngat Mountain range

Topography

The Geology

The Ramah Sedimentary

Series

Observations in and about
Nachvak Bay ....
Topography and scenery

PAGE

PAGE

205

233

207

The General Structure of the

Coastal Belt

234

209

Observations on the Surface Geol-

ogy

236

210

Glacial Markings ; Direction of

210

212

Glacial Lunoid Furrows . .

237

The Glacial Deposits ....

245

214

The Glaciation of the Torngats .
The Zone of Postglacial Emer-

245

215

gence in Northeastern Lab-

216

rador and in Newfound-

218

252

219

Location of the Highest ele-

223

vated Shore-line ....

254

223

Boulder Barricades ....

260

225

Continuance of Elevation . .
The Scenery of the Emerged

261

225

262

The Need of Further Exploration .

267

226

269

226

{Bearings refer to the true Meridian.')

Introduction.

In the late summer of 1899, Mr. Huntingdon Adams, of the class of
1901 at Harvard, paid a flying visit to the coast of northern Labrador.
He was so impressed with the beauty of the fiords and mountains of the
country that he conceived the idea of organizing a party which should
spend the following season exploring in the same region, with intent to
bring back more definite information regarding its general geography
than had yet been obtained. "With the advice and aid of Prof. E. B.
VOL. xxxviii. — no. 5

20G BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Delabarre of Brown University, the scheme was successfully carried out,
and a most profitable and enjoyable summer spent by those who par-
ticipated in the expedition. The party also included Messi-s. H. B.
Bigelow, L. B. McCornick, and H. \V. Palmer, undergraduates of Har-
vard. I was asked to accompany them and record any geological obser-
vations that might be possible on a trip of the kind proposed. Bather
more was accomplished in this direction than was hoped for at the
outset, and the results arc believed to be of sufficient interest to war-
rant some degree of detail in their recital. The following pages are
intended to afford a brief treatment of the more important problems
met with during the summer; some of these have found solution, others
will, it is hoped, be more clearly defined for future visitors to a little
known shore.

The equipment of the expedition was quite modest but sufficed for
most needs, excepting that for rapidity of travel. The vessel selected
for the journey was the "Brave," a forty-ton fishing schooner, just re-
built and specially fitted up to be the summer home of the party and
of the four seamen employed to take the larger share of the navigation
required. Thoroughly staunch and clean, and commanded by a skip-
per of unusually wide experience and close knowledge of the thousand
dangers of this coast, the craft was found to be both safe and comfor-
table. The instrumental outfit included five aneroid barometers ; two
ordinary thermometers ; a corrected thermometer lent by the Superin-
tendent of the United States Coast Survey, from whom there were also
obtained two sets of salinometers and a hydrometer cup; four Negretti-
Zambra deep-sea thermometers, and five hundred fathoms of piano-wire,
lent b}' Mr. Alexander Agassiz ; a sextant, lent by Brown University ; and
a prismatic compass. The tools in trade of a botanist, an ornithologist,
and a geologist, were likewise represented on board. Professor Dela-
barre worked assiduously on the flowering plants encountered on the
route, and collected freely from the cryptogamic vegetation which is so
striking a feature of the coast. Mr. Bigelow's former experience enabled
him to seize quickly the peculiar ornithological features of the coast, and
he has added several species new to its list of birds.1 The writer,
though chiefly occupied with the general geology ashore, spent some
time in a limited study of the hydrography of the Labrador Current.
Before proceeding to the discussion of results, it will be well to give a
brief account of the itinerary. By so doing, the conditions under which
the observations were carried on may be most easily understood.

1 Pf. Auk, 1002, xxvii. pp. 24-31.

DALY : GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 207

Itinerary of the Cruise.

On June 25th the " Brave " weighed anchor at St. John's, Newfound-
land, and sped rapidly northward with a fair wind to about Lat. 49° 20'
north. The next day a strengthening gale decided our captain to run
for shelter at Greenspond. Compelled to remain there until July 2d,
we had leisure to examine the hummocky islands of Laurentian porphy-
ritic granite and to study the hard conditions of life in a town of New-
foundland fisher-folk. We came to a full understanding of the fact that
the dangerous sea more readily gives them livelihood than the soilless
ledges which need imported earth before they yield a useful vegetable.
A quick run without incident, save a halt at Change Island, brought us
to Cape Bauld at noon on July 4th. Here we were destined to disap-
pointment in the hope that the Straits of Belle Isle might be crossed
and " the Labrador " attained without delay. The northern half of the
strait was found to he impassable on account of a broad stream of pan-
ice ; the " Brave " beat back and anchored at nightfall in Kirpon Harbor
which lies in the " tickle "' east of Kirpon Island.

The next day, from a hill north of the harbor, we had a remarkably
striking view of drift-ice streaming west-southwest into the strait and
southward past Belle Isle along the east coast of Newfoundland. The
Labrador current was a vivid reality to us as we watched the truly
majestic procession of these dazzling migrants from a polar sea. That
day closed with no indication that our snug harbor would be threatened
with an invasion, but we came on deck the next morning to find our-
selves in an Arctic landscape. The schooner was firmly wedged in
among the heavy pans of ice which during the night had quietly drifted
into the " tickle " under the impulse of north wind and flood current.
The unusual thickness and quantity of the ice, coupled with the steady
character of the north or " up " wind, caused our detention in this har-
bor until the morning of the 13th. These seven days were spent among
the picturesque hills and valleys in the vicinity and among the exqui-
sitely tinted ice-floes. Among the sedimentary rocks which compose
Kirpon Island, Jacques Cartier Island, and the mainland roundabout,
an interesting boulder basalt, with pillow structure, was discovered. It
is hoped that a description of this typical occurrence will be given on
another occasion.

Along with more than a hundred other schooners from the many
anchorages of this indented coast, the passage of the straits was finally
made. Not the least memorable scene of the summer was this bril-

208 bulletin: museum of compaeative zoology.

liantly sunlit expanse of water covered with the great fleet and with a
long train of icehergs, two of which were probably the loftiest seen
during the cruise.

Headed by the wind, immersed in thick fog, and trapped once more
by ice-floes, we lay at anchor in Assizes Harbor on the north side of the
straits until the 1 7th, when we escaped again and ran sixty-five miles to
Seal Island. Three days' delay by ice and head winds here, and three
more at West Bay Head on the south side of Hamilton Inlet, completed
the first month of the cruise and evoked many remarks from our skipper
on the extraordinary difficulty of " getting down " the coast this season.
The failure of westerly and southerly winds and the massive character of
the drift-ice so late in the summer were alike unparalleled in his thirty
years' experience on this coast.

" The extraordinary smoothness of the sea covered by drift-ice, even when
the pans are widely spaced, is truly astonishing to one making his first voyage
in such waters. His sailing ship may be favored with a fresh breeze and yet
the ocean surface be quite level, save for the minute rippling characteristic of a
small pond ruffled by a summer breeze ; ground-swell does not exist. It is a
matter of common knowledge among the fishermen of the Atlantic Labrador
coast that the Labrador current, or ' tide,' as they invariably express it, often
shows high velocity, although its surface, for. a length of a thousand miles and
a breadth variable but at times as much as three hundred miles, is covered
with loose pan-ice. At such times, the wind is, or has just been, strong and
from a northerly quarter. We are justified in believing that the pans act as
the sails which, in ice-free waters, are represented by wind-waves. Floes and
pans project above the surface from one to twenty feet or more. They may be
expected to exert a coercive force on the film of relatively fresh water derived
from the melting of the ice in contact with the heavier salt water beneath.
According with the behavior of such ' dead water,' as described by Nansen and
others, the light surface layer will tend to move en masse and in the direction
of common pull exercised by the wind-driven masses of ice. By reason of
friction the motion will be communicated to lower layers of the sea. This
cause of surface currents is of importance to the theory of movement of those
polar waters which, for several months after the winter ice begins to break up,
are free from larger wind-waves. Deprived of its chief sails, the Labrador cur-
rent, always sensitive to wind conditions and at times subject to temporary
reversal with contrary winds, yet preserves and perhaps exceeds during the
period of ice-drift, the average velocity of current-flow for the year."1

On the 24th, the mouth of Hamilton Inlet was crossed. Ice afforded
little trouble henceforth, but head winds prevailed ; so that, at frequent

1 Science, Nov., 1900, vol. 12, p. 688.

DALY : GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 209

intervals, we were compelled to drop anchor and wait for the short-lived
favorable winds on which our progress against the south-flowing current
depended. The longest run made between the straits and Hebron, a
distance of eight hundred miles, was only fifty-three miles in length.
Twenty-two halts of greater or less duration were made en this part
of the journey. Nachvak Bay, eleven hundred miles from St. John's,
and the objective point of the expedition, was not reached until August
21st. Thus but a small portion of the summer remained for the explo-
ration of the high mountains in the north. At the end of two weeks we
were forced to weigh anchor and begin our homeward journey.

Disappointing as our rate of progress was in this one respect, there
yet remained the advantage that, with so many opportunities to land in
southern Labrador, we were able to sample, with fair continuity, the geol-
ogy of a coast-line which is in every part in need of investigation. In fact,
some of the most interesting problems of the summer would have been
necessarily left untouched, if our early ambition to make a rapid north-
ward run had been satisfied. The return to St. John's was accomplished
in four weeks, during which time, several gaps in the required series of
observations were filled up. We dropped anchor for the last time in the
early morning of Oct. 3d, having been out a few hours less than a hun-
dred days. In that period, we had been thirty-nine days at anchor against
our will, but there was, at each detention, always the consolation of an
opportunity to get a view, however hurried, of a region full of novelty
and at times no less interesting than the goal of our endeavor, the
Torngats of the north. At the same time, it is clear that nothing
more than a reconnaissance could be made at any of the anchorages.

Observations on Topography and Bed-rock Geology.

The general form and composition of the old-mountain plateau of
Labrador have already been admirably treated by Packard,1 Bell, and
Low, and by the writer of the article " Labrador " in the Encyclopaedia
Britannica. These and earlier writers agree that the northeastern coast
of the peninsula marks the edge of the great Archaean shield of North
America; and, further, that, if exception be made of the "Domino
quartzite," the Eamah slates, and certain occurrences of sedimentary
rocks in Nachvak Bay, the bed-rock of the coastal belt is throughout

1 A bibliography relating to works on Labrador is published in " The Labrador
Coast" by A. S. Packard, Jr. New York and London, 1891.

210 bulletin: museum of comparative zoology.

crystalline. Gneisses, intrusive granites, syenites, gabbros, and traps
prevail from the Straits to Cape Chidley.

The results of last summer confirm this general view, but it was found
that sedimentary formations not heretofore described appear in great
development on the shield border. At Pomiadluk Point, at Aillik Bay
and in the Mugford region, as well as in the long stretch from Saeglek
Bay to Ramah, the crystallines form the foundation to stratified series
of very diverse character. These merit particular notice in a sketch,
however brief, of the general geology of the coast. The extrusive lavas
of the Mugford series, the intrusive traps which occur in astonishing
profusion in the 700-mile belt, and the gabbros of Paul's Island and
vicinity will also claim attention. It will be shown that a correlation
of the strike-directions of schists and sediments indicate in the coastal
border a decided N.W.-S.E. trend which corresponds rather closely with
the average trend of the shore-line. Finally such observations as have
been made on the topography and physiography will be described in
connection with the bed-rock geology.

From the Straits of Belle Isle to Paul's Island.

General Topography. — As far north as Cape Mugford, over five
hundred miles from the Straits, the edge of the plateau is in plan
extremely ragged. Numerous fiords, ria-like bays and a vast archi-
pelago of outlying islands or skerries form a coastal fringe. The simi-
larity of landscape is so great that Forbes's description of the coast of
Norway on the route from Troudhjem to Bergen may be repeated for
this portion of Labrador. A sei'ies of inlets penetrates " in all directions
a low, bare, rocky land, partly island, partly continent, nowhere rising
but to a very small height above the sea, and so monotonous in charac-
ter", and destitute of long reaches, or natural landmai'ks, as to seem to
require an almost superhuman instinct for its pilotage." 1

The contours of the islands are repeated in the hills of the low
plateau of the mainland ; the inlets, sounds, and narrow channels among
the islands (the "tickles" of the fishermen) represent the submerged
equivalents of the valleys on the mainland. From any command-
ing hill on island or mainland, the eye ranges far and wide over a
surface showing everywhere the evidence of universal and profound
glaciation. Unobscured by forest, soil, or thick drift, and singularly ex-
panded because of the crystalline clearness of the atmosphere, the view

1 J. D. Forbes, Norway and its Glaciers, Edinburgh, 1853, p. 104.

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 211

typifies that which may be had in the Laurentian Highlands of Canada
or in the Archaean of the Scottish Highlands. It is a great wilderness
of innumerable rounded, ice-worn hummocks generally gneissic in com-
position. Among the roches moutonnees lie equally countless ponds and
bogs connected by the small streams of a most disordered drainage.

In sheltered places and at the higher altitudes the snow lies through-
out the year. On Pomiadluk Point, two ravines, running from about the
four hundred-foot contour to the eight hundred-foot, a distance of four
hundred yards in each case, wei-e found to be occupied with snow to a
depth of from fifty to seventy-five feet. Tj-pical transverse crevasses
ten to twenty feet in depth and yawning widely at their mouths indi-
cated by their attitude and down-stream curvature that the snow was
moving, glacier-like, down the slope. No blue ice was seen, but it is
possible that such existed in the deeper parts. Though many such
banks were found on the coast, especially on the higher mountains of
the north, none of them so closely approximated a true glacier in look
as these on Pomiadluk Point.

The existence of the coastal fringe suggests at once that the plateau
has been drowned by recent subsidence. Yet, in view of the evidence
derived from the study of hanging valleys and fiords in Norway, Alaska,
and elsewhere, it is open to question whether the same glacial erosion
which has so conspicuously moulded the mainland may not be as well re-
sponsible for the depressions on the plateau-border now occupied by salt-
water. In other words, the differential erosion of an ice-sheet extending
out beyond the island-zone may have excavated these depressions which
would permit of the entrance of the sea when the ice had left the coun-
try. It is significant that where, on this coast, north of Hebron, glacial
erosion was confined to the main valleys, the fiords corresponding to the
latter are magnificently developed but the islands largely fail. The
sounds and tickles do not now present the systematic relations of
drowned river-valleys ; such irregularity as is known to characterize the
ei-osive activity of valley glaciers is without doubt represented beneath
an ice-cap and might leave just such channels below sealevel as a
memorial of that fact. It cannot yet be asserted that differential.glacial
erosion has contributed more to the present outline of the Labrador coast
than simple drowning, but it has certainly affected the original pre-
glacial form of the plateau to a great extent. In any case, it is import-
ant to recognize in a systematic discussion of shore-lines, a type wherein
glacial erosion, independently of crustal movement, may produce an island-
zone similar to that of a "drowned coast."

212 bulletin: museum of comparative zoology.

"While this glaciation has given a monotonous though not unpleasiug
character to the plateau, the headlands and islands of the coast exhibit
in detail a great variety of form. Sea-cliffs, benches, sea-chasms, fretted
ledges, beaches, gravel-bars, spits, and narrow coastal plains, all belong-
ing to the zone of postglacial emergence, add much to the scenic quality
and interest of a voyage along the coast. In this rugged soilless belt,
it is literally true that he who runs may read, and if he be a student of
geological processes, will find perpetual instruction from the coastal
views.

The Geology. — South of Hamilton Inlet, the bed-rock seems to be-
long entirely to the crystalline complex (Plate 11). Assizes Island and
the mainland at the mouth of the St. Charles River, are underlain by
greatly contorted biotite and hornblende gneisses, biotite schists and
amphibolites ; quartz and pegmatite veins occur in large numbers.
Similar rocks were found at St. Francis Harbor on Granby Island.
The geology is further much complicated by the frequent occurrence
of large areas of massive and gneissoid granites and diorites and by an
equally persistent appearance of trap-dikes. The coarse gneissoid granite
of Great Caribou Island on the south side of St. Lewis Sound, was re-
marked for its development of ilmenite in anhedrons from one eighth to
one half an inch in diameter. The mineral showed a perfect develop-
ment of the octahedral parting which was again seen in the ilmenite
crystals an inch in diameter that are plentifully distributed in a vein of
graphic granite at Rogue's Roost, Seal Island. This vein is twelve feet
in average width and is beautifully exposed for some three hundred feet
of length. Seal Island is chiefly composed of coarse hornblende granite
cutting coarse diorite and is itself cut by pegmatites, fine-grained aplites
and numerous wide dikes of hornblende biotite diorite.

Drift-ice prevented our anchoring in Domino Harbor and thus I had no
opportunity of seeing in its classic locality the " Domino Gneiss " of
Lieber,1 nor the quartzitic rock described by Packard2 and interpreted
by Bell 8 as representing a remnant of the Huronian in eastern Labrador.
In a private communication to the writer, Packard states that he is
inclined to agree with Bell that the latter rock is in reality an arkose.
At Pottle's Cove, West Bay, which Packard's map places in the zone of
the Domino Gneiss, a peculiar and striking rock-type has large develop-
ment. It is a medium-grained to fine-grained gneiss weathering under

1 0. M. Lieber, Rep. U. S. Coast and Geodetic Survey, 1860, p. 402.

2 A. S. Packard, Jr., The Labrador Coast, 1891, p. 286.

3 R. Bell, Scot, Geog. Mag. 1895, Vol. XL, p. 349.

DALY : GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 213

the almost universal covering of peat, to a grayish-white color highly
suggestive of weathered quartzite. The structure is linear-parallel and
is formed with extraordinary perfection. The gneiss is penetrated hy
dikes and more irregular intrusions of diorite which are probably related
as to age and origin with the trappean mass of Tub Island.

Nowhere on this coast are trap-dikes so massive or so influential in
determining the topographic form of the country, as in the region about
Ice Tickle on the north side of Hamilton Inlet. Rodney Mundy Island
and Ice Tickle Island are each about three parts composed of medium
to coarse grained honi blende granitite, granitoid gneiss and fine-grained
gneisses. The remaining surface is occupied by a multitude of great
dikes, thick sheets, pods, or stock-like bodies of diabase which, on
account of their black color, contrast very strongly with the other crys-
tallines round about. The trap wherever examined is plainly intrusive.
No vesicular structure was to be found. Both contacts of a typical pod
showed the chilled and porphyritic zone characteristic of intrusion. A
fine example of the dikes is seen on the ridge at Indian Harbor. It has
parallel walls, is some two hundred feet in width, and is visible as a steep
black wall a quarter of a mile along the Tickle. Very commonly the
dikes are seen to swell out into thick trap-bodies three hundred to one
thousand feet in diameter, and again, a stock will project through the
gneisses without visible connection with the trap of the neighboring
hills. The average strike of the intrusive masses is N. E.-S. W. The
trap forms all the higher elevations, which thus assume the varying out-
lines of palisade, ridge, or dome, according to the shape of the intruded
rock-body. Between them the schists and granites, on account of their
inferior power to resist erosion, now underlie valleys that open sea-
ward on the deep bays and tickles of a "drowned " coast-line. One will
go far to discover so fine an example among wholly crystalline rocks, of
such control over land-forms — the control of differential resistance to
weathering.

The same association of gneiss-valleys and trap-hills extends at least
ten miles to the westward of Ice Tickle. At Sloop Harbor, though still
plentiful, the dikes (here hornblende biotite diorite with accessory augite)
are too narrow to produce a great effect on the topography. Webeck
Island, a soilless, driftless, because wave-swept, granitic swell a half-mile
or more in length, affords a splendid exhibition of twenty-two huge
dikes cutting across the island in parallel fashion. On the mainland,
opposite Jigger Island near Webeck Island, an interesting group of dikes
in granitite was studied. One of these is a handsome biotite diorite

214 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

porphyrite ; the other, a slightly porphyritic alkaline rock, with pheno-
crysts of microcline and a grouudmass composed of microcline and
quartz in micropegmatitic intergrowth and an amphibole with the char-
acters of riebeckite.

Massive granitic rocks were found at each anchorage between Hamil-
ton Inlet and Hopedale. At Sloop Harbor the hills are composed chiefly
of a coarse hornblende granite ; Jigger Island exhibits granitites similar
to that of the mainland; thirty miles farther north, on the mainland
opposite Conical Island, a flow-breccia of granitite cutting diorite proved
to be extensive. Associated with similar rocks are the only sedimentary
formations that were seen to the southward of Hopedale.

Sedimentary rocks at Pomiadluk Point. — Pomiadluk Point forms the
extremity of a bold peninsula which projects northeastwardly from the
mainland in about 55° N. Lat. Stretching along the southeast side of
the peninsula for a distance of some five miles, is a broad bench from
two to three hundred feet in height and from one and a half to two
miles in breadth. On the southeast the bench falls into the sea quite
abruptly ; on the northwest it ceases at the foot of a steep ridge of
granitite that composes the main part of the peninsula. Barometric
readings on two different days accorded well in giving eleven hundred
feet as the elevation of a prominent summit of the ridge. Its average
height is not less than one thousand feet; toward Cape Strawberry it
rises to twelve hundred feet, and is thus the highest land encountered
on the coast between the Cape and the straits of Belle Isle.

The glaciated ledges of the flat though hummocky bench have been
wave-swept during postglacial submergence and thus one can study the
rock-composition and structure with exceptional ease. The bench is
conterminous with a well exposed mass of metamoi'phic conglomerate.
Along four cross-sections about a half a mile apart, the sedimentary baud
was proved to be very homogeneous. In a silicious matrix, often highly
schistose, pebbles and boulders up to one or more feet in diameter are
embedded. They are composed of granitite, quartzite, vein quartz,
granite porphyry and metamorphic sandstone. They are almost always
considerably flattened, though the granitite pebbles have oftentimes so
far resisted the shear as to present still well-rounded outlines. The
schistose structure strikes steadily N. 15° W. ; its dip is variable but
steep and generally to the westward. Not allowing for duplication, the
thickness of exposed conglomerate measured on this secondary structure-
plane was estimated at eight thousand feet. On the west the conglomer-
ate becomes finer-crrained and for a distance of three hundred feet from

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 215

the granitite contact, is replaced by a band of metamorpbic sandstone.
The structural features are here identical with those of the conglomerate.
Part of the cement of both rocks is made up of epidote, which colors the
rock an intense green, especially marked where myriads of epidote
crystals have formed along the joint-planes.

The limited time at our disposal, the extensive covering of elevated
beach deposits and of land-wash at the inner edge of the bench, and, in
an important degree, the disturbing influence of clouds of mosquitoes,
prevented the discovery of the actual contact between sandstones and
granitite. The granitite^pebbles in the conglomerate are very similar in
character to the rock of the great ridge and suggest that they were
derived from it. The deformation of the conglomerate is represented in
the granitite by the appearance, especially near the contact, of a schistose
structure, with a trend common to that of the conglomerate ; the two
rocks have evidently been squeezed together.

A sharp lookout was kept for fossiliferous bands, but no organic remains
were discovered. Thus no conclusion could be made as to the age of
the sediments. They have a close superficial resemblance to the
Archaean metamoi-phic conglomerate series of Finland which the writer
had already seen in the field. These Labrador sediments are cut by
diorite, granite porphyry and diabase dikes and by many pegmatite and
quartz veins.

Sedimentary rocks in Aillik Bay. — Aillik Bay opens at a point about
fifteen miles northwest of Pomiadluk Point. It is a picturesque inlet
some five miles in length and three-quarters of a mile in width. Its axis
is directed north and south. The bay is rimmed about with massive
rocks; diorites cut by granite porphyry dikes on the west; diorite intru-
sive into amphibolites and hornblende granite on the south ; and coarse
hornblende granite on the east. These various rocks compose high
encircling ridges ; at the base of these a narrow and interrupted belt of
variegated banded quartzites outcrops on all sides of the bay. On the
west, to north and south of our anchorage at Summer Cove, the quartz-
ites could be traced at least two miles along the shore, but they were
never found more than about one hundred yards from the beach. The
white, red, and purplish layers often exhibited the cross-bedding of a
typical sandstone. The strike of the beds is variable, changing from
X. G0° W. to N. 20° W., the dips remaining low and constantly directed
toward the land. The total thickness of the beds exposed was measured
at one hundred feet. The belt terminates at the mouth of the bay in a
veritable museum of rock-types, the quartzites being cut by an inter-

216 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

lacing network of granitic and trap dikes as well as by pegmatitic, aplitic
and quartz veins. On the east side of the bay opposite Summer Cove,
similar quartzites dipping 50° S. 70° E. occupy a second belt along the
shore of about the same dimensions as those of the western shore-belt,
■ting that the sediments are here exposed for only three hundred
yards. At either end the belt is cut off by hornblende granite and
hornblende gneiss on which the quartzite seems to have been deposited.
All these rocks are cut by numerous wide and conspicuous dikes of
augite porphyrite and squeezed amphibolitic trap. At the head of the
bay, quartzites dipping 55° S. 35° W. were seen on a low spur projecting
into a boulder-covered tidal flat.

In this case, as at Pomiadluk Point, diligent search failed to disclose
fossiliferous evidence as to the horizon of the stratified deposits ; at
both localities there is the same general relation between eruptives and
sediments, suggesting that the latter are, roughly at least, contempora-
neous deposits.

Threading the maze of islands extending from Aillik Bay to Hope-
dale, no halt was made and no sampling of the coast geology could
be carried on. From the somewhat monotonous character of the
island-zone both as to color and form, it seemed to be composed of
the coarse biotite gneiss which characterizes the rugged hills about
Hopedale. Trap dikes were always in the view. Striped Island,
twenty miles E. S. E. of Hopedale, owes its name to the curious con-
touring habit of twelve great dikes that have penetrated the granitic
boss-like island along a nearly horizontal system of master-joints.
They are now visible all around the clean-swept island but are par-
ticularly conspicuous on the west side for a distance of a half mile.
At Hopedale, the wonderfully contorted gneiss, which is penetrated
by numerous diabase dikes, has an average strike of N. 55° W.
Thence northward, the archipelago of ragged sub-conical islands of the
same gneissic composition extends with northwest trends to Ford
Harbor on the east end of Paul's Island. Anchoring at Quirk Tickle,
fifty miles north of Hopedale, wc found the banded biotite gneiss with
northwest strike and indistinguishable from the rock at Hopedale.
Extending as it does a distance of at least one hundred and twenty
miles along the coast, the Hopedale gneiss is one of the most im-
portant members of the whole crystalline complex.

The Nain gabbros. — The Hopedale gneiss underlies the eastern end
of Paul's Island, but a few miles west of Ford Harbor, it comes in con-
tact with the famous anorthosite and allied gabbro whence is derived the

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 217

schillerizing labradorite. (Plate 11). One may collect specimens of the
feldspar from the numerous erratic boulders sprinkled over the hills about
the harbor. The direction of glacial striation shows that these must
have been carried to their present resting-places from the west and
southwest. It was likewise made clear to us on the northward journey
that the gabbro must be in great development in order to furnish such
an immense amount of erratic material as we saw. Skirting the north
shore of Paul's Island, the "Brave" was headed for Nam, passing
through a long tickle walled in on either side by high cliffs of massive
gabbro for fifteen miles before the mission station was reached. At
the station we were still about twelve miles from a quarry where
" precious " labradorite in place has been opened by Mr. R. G. Taber.
The desire to see the mineral in place decided us to risk the schooner
among the dangerous channels of the island-labyrinth. At Nain, how-
ever, we had the good fortune to find Dr. Wilfred Grenfell, superin-
tendent of the Deep-sea Mission, then in command of his fine new
steamer, the " Stratbcona." With great kindness he invited our
party to accompany him on the steamer to Mr. Taber's quarry ; we
thus spent one of the most enjoyable days of the summer in the
accomplishment of what would have taken the schooner, by reason of
the calms and baffling winds then prevailing, two or three days to
effect. The quarry is situated on the southwest side of a small island
called by the Eskimo, " Napoktulagatsuk." It lies within one fourth
of a mile from the mainland.

Napoktulagatsuk is elliptical in shape and has a major axis of about
six hundred yards. The opening has been made in a steep glaciated
slope and covers an ai'ea of about twenty-five hundred square feet. At
the foot of the slope a thirty-foot raised beach beai's a ruined derrick,
and a tramway which showed by their dilapidation that work has been
discontinued for some years. While the whole island is composed of
anorthosite, the schillerizing feldspar occurs only in the form of isolated
patches up to fifty feet in diameter. These are generally, though not
always, coarser in grain than the surrounding rock, and are pegmatitic
in look. In the sunlight, the fresh surface of the rock presents a rich
and beautiful appearance. The dominant color is the familiar blue,
but it is associated with vivid green, bronze, orange, and red phases.
These pocket-like schillerizing masses are clearly contemporaneous with
their country-rock. Both types are characterized by the well-known
composition of this gabbro. Both are cut by aplitic dikes and peg-
matite veins.

218 bulletin: museum of compakative zoology.

The loftv dome-shaped islands situated between Nain and Napok-
tuLvatsuk as well as all the mainland visible thereabouts are composed
of the "abbro. There is every probability that the schillerizing phases
may be found sporadically throughout this great area. They are
known to occur on the eight hundred-foot Mt. Pikey, southwest of
Ford Harbor. Peculiarly sombre in hue, profoundly glaciated and
almost entirely devoid of vegetation, these great hummocks afford a
scene of complete desolation almost without parallel even on the barren
coast of Labrador.

Gneisses similar to those at Hopedale compose the outer islands
north of Ford Harbor, but it is probable that the Nain gabbro is
continuous with another great area that we first met with on Newark
Island, and afterwards found extensively developed on the mainland
at Port Manvers. At Black Island Harbor the gabbro is coarse ; at
Port Manvers it is finer-grained. In both cases the composition is
identical with that characteristic of the Nain occurrence. The rock
of this northern area has the habit of concentric weathering, boulders
of disintegration forming great cyclopean walls on the glaciated ledges
at Port Manvers. The rock seems to be much less resistant to the
weather than the gneisses. The floors of glaciated valleys in the
gabbro are invariably occupied with screes of the crumbling rock. A
truly imposing example is seen in the long sweeping curve of waste
that covers the lower fifteen huudred feet of the northern face of Mt.
Thoresby.

The Kiglapait.

Fifteen miles north of Port Manvers the eastern end of the Kiglapait
springs directly out of the sea. The name of this mountaiu-group is
:>ii Kskimo word meaning " The Great Sierra" and refers to the very
ragged sky-line and general outlines. The axis of the range runs due
east and west, parallel to the coast-line which here has an exceptional
trend. The sierra is not more than thirty miles in length, but, on
account of its conspicuous position on the shore, is strikingly pictur-
esque. Ten distinct and individual summits from two thousand five'
hundred to four thousand feet in height could be counted from the
schooner. No one of these, so far as the writer has been able to deter-
mine from missionaries, fishermen, or from the literature, has as yet
received a name. Here, as in the higher mountains of the north,
there is abundant opportunity for systematic field-work on the part
of such an organization as the Appalachian Mountain Club.

DALY : GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 219

We had hoped to spend some days, if not weeks, in the study of
those interesting mountains, but the lateuess of the season forbade our
dropping anchor within reach of the noble range. Judging again
simply from the peculiarly dark color of the bare rock-surfaces, it
seems probable that the gabbro seen at Port Man vers makes up most
of the Kiglapait, which will thus represent the Cooliu type of gabbro
mountains in Scotland.1

On the other hand, it was discovered that banded and much con-
torted gneisses compose the numerous low islands lying between Ford
Harbor and Mugford Tickle, a distance of fifty miles. At the Tickle,
the undulating platform of the complex disappears beneath the quite
different formations of the Kaumajets.

The Kattmajet Mountain Group.

For a distance of fifty miles to the southward we had marked the ma-
jestic pile of the Bishop's Mitre with the associated mountains of the main-
land. Their summits were at the time covered with a fresh fall of snow ;
the brilliancy of the crests recalled the etymology of the name which
again illustrates the Eskimo's feeling for natural scenery. "Kaumajet"
signifies ''shining" ; the range is the Himalaya of Labrador.

As indicated by its position, composition, and topographic character,
the island of Ogua'lik really forms the southern extremity of the
Kaumajets. Mugford Tickle separates it from the mainland. It was
in this nai*row channel that our anchorage was chosen. Again we had
occasion to mourn the slowness of our northward progress, for it would
have been of the highest interest to devote a fortnight at least to the
exploration of this region ; in order to be certain of reaching Nachvak,
however, we allowed but two days in which to secure information con-
cerning the nature of the massifs immediately surrounding the vessel.

The nine-hundred-foot scarps of Ogua'lik would have been impressive
among the tamer landscapes of southern Labrador, but they were dwarfed
beside the mighty walls of the opposing mouutains only a mile or two
distant. We had entered the tickle late at night, and in the brilliant
starlight had discerned the huge piles looming up in solemn and form-
less grandeur. Their mystery became in part dispelled as a bright
sun disclosed a scene in its way unrivalled in Labrador. Due north
in the centre of the view two gracefully rounded knobs, estimated, by
the aid of barometric readings half-way to their summits, to be

1 A. Geikie. Scenery of Scotland. Ed. 2, 1887, p. 215.
vol. xxxviii. — xo. 5 2

220

bulletin: museum of comparative zoology.

two thousand five hundred feet in height, lay close to the verge of
an almost vertical precipice from one thousand to one thousand two
hundred feet high. Below this a series of lesser cliffs, separated by
steeply sloping screes of rock-waste stepped downward to the uneven
floor of a deep N. E.-S. W. valley. On the southeast the valley is
hounded by a similar arrangement of cliffs and taluses. It ends as a
great cul-de-sac two miles in length in a thousand-foot head-wall over
which there cascades a large brook.

Figure 1.— Dissected plateau of Mugford sedimentary series, viewed from Mugford
Tickle. Each of the two summit knobs was estimated to be about 2500 feet in elevation.
Looking north. Drawn from a photograph. Heavy black line represents upper surface
of the basement Complex.

On landing, we found that the first and natural impression, that this
systematic array of scarps and taluses signified a stratified structure for
the massif, was justified (Fig. 1 and Plate 11). At the foot of the great
slope the basement of all the other rocks is represented in an irregular
floor of contorted and faulted gneisses and amphibolites. The surface of
the familiar complex appears at all elevations above the sea up to about
six hundred feet. Next above the light colored zone of the schists comes
a series of black slates fifty to one hundred feet thick, indurated at the
contact by a conformable three hundred-foot sheet of apparently intrusive
diabase. The edge of this sheet forms one of the lowest strong cliffs
above the basement. The diabase in its turn is overlain by a great thick-
ness of slates, quartzites, sandstones, quartz breccias, volcanic agglomer-
ates, and trap layers. The inaccessibility of the cliffs and the shortness
of the time allowed for examination rendered it impossible to determine
the absolute strength of the various members. The enormous size and
number of the blocks found in the screes seemed to show that the greater

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 221

part of the thousand-foot cliff is made up, as they are, of a volcanic breccia.
The basalt agglomerate, which is also a significant part of the stratified
series, is typical and encloses numerous bombs with bread-crust structure.

These sediments and intercalated traps everywhere have low dips ;
therein we have an explanation of the plateau-like character of the
relief, as dissimilar to any other on the coast, as the residuals of Tor-
ridonian sandstone are unlike the rolls of gneiss in the Scottish High-
lands. Combining the observations made in the gorge below the waterfall,
with those made in the Tickle and outside to the southeast of the Bishop's
Mitre, the structure of the rocks on the northwest side of the Tickle is
that of a flat syncline with a N. W.-S.E. axis, crossed, west of the water-
fall, by a transverse warp. On the southwest, the effect is that of an
eroded low half-dome. The steepest dips occur near the middle of the
Tickle, where a magnificent section of the whole stratified series can be
easily studied as the beds plunge into the deep water (Plate l). The
total thickness of the stratified series is about twenty-five hundred feet.

On Ogua'lik at the southwest end of the Tickle, the gneisses are over-
lain by an intrusive sheet of diabase, about fifty feet in thickness, upon
wdiich are piled slates, quartzites, limestones, and sandstones with inter-
bedded traps. Whether the latter are intrusive or extrusive we had not
time to determine, but the apparently complete absence of vesicular
structure and of other signs of volcanic activity would speak for the
latter interpretation. The limestone is rather crystalline, and is sti'ongly
charged with pyrite. A shallow opening and a few rusty tools showed
that attempts had been made to test the rock for either gold or copper.

The whole of Ogua'lik corresponds in composition with these cliffs of
the northwestern shore of the island. At the eastern end, black cliffs,
fifteen hundred feet in height, form the edges of a capping massive sheet
of trap estimated at seven hundred feet in thickness, overlying some nine
hundred feet of sediments again unconformable on the gneissic complex.
While there exists a general similarity between the formations on either
side of Mugford Tickle, their failure to match on opposite sides of the
strait seems to imply a fault coincident with it in direction.

The unconformity of the crystalline basement and the sediments is
extremely well shown at both ends of the Tickle, especially at Cape
Mugford itself (Plate 2). The gneisses have participated in the fold-
ing of the sedimentary cover ; it is owing to that fact that they disap-
pear in the middle cliffs of the strait. The surface of unconformity
would, if produced, rise to seaward so as to pass completely over the
fantastic peaks of Nanuktut Island. This island thus forms a strangely

222 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

1 outlier of the crystallines from which the sedimentary cover lias
been swept by denudation. The island has been called by the "New-
foundland Pilot" "the most remarkable and unmistakable laud on the
Labrador coast."1 Nothing could exceed the contrast between its char-
acter and that of the Mugford massifs. The light gray precipitous
two-thousand-foot peaks of the one oppose the black, tabular, greatly
dissected piles of the other.

At Cape Mugford, which forms a nearly vertical sea-cliff about two
thousand feet in height, we were struck with the highly ferruginous
character of the sediments, broad bands of variously tinted iron-rust
enriching and enlivening the color effects in a marked degree. Numerous
waterfalls and extensive banks of snow leut welcome relief to the dark
cliffs, the black recesses of huge sea-chasms, and the savage gorge-like
iulets that opened one after another as our schooner slowly forged
through the "tide" around the cape.

Fine as this scenery was, still greater magnificence awaited us as we
came face to face with the Bishop's Mitre toward the close of a memor-
able day of sailing. Seen from the northeast, the mountain, estimated
to be considerably more than three thousand feet in height, displays a
symmetry which is most remarkable in view of the fact that the pres-
ent profiles are everywhere the result of erosion. As the name im-
plies, there are two peaked summits. They are separated by a sharp
notch about five hundred feet in depth. This breach is but the upper-
most part of a gigantic ravine that cleaves the mountain to its base at
the shore more than two miles from the notch. Occupying the bottom of
the ravine an uninterrupted snow-bank still marked, in the month of
August, the line of symmetry of the whole mountain. From either peak
of the Mitre a rugged razor-back ridge descends, each gradually diverg-
ing from the other across the widening intervening trench. With essen-
tially equivalent transverse and longitudinal profiles, the two spurs
further match as each terminates at an elevation of about one thousand
feet, in a bold rock-tower. Each tower rises eight hundred feet or there-
abouts above the ridge-crest and, on the east, drops suddenly the full
eighteen hundred feet into the sea. The matching of the right and left
halves of the mountain does not stop with the form. Each of the sen-
tinel towers is composed above of black Mugford sediments reposing on
five hundred feet of the light gray gneissic complex. The architectural
quality of these great buttresses and of the Mitre itself is greatly en-
hanced when a fresh fall of snow brings out the nearly horizontal structure

1 Newfoundland and Labrador Pilot (ed. 3, Hydrog. Office, London, 1897, p. 680).

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABEADOK. 223

of the whole massif and is such as to make one believe that the Mitre
is the most beautiful single mountain on the coast.

About four miles to the westward of the Mitre is the summit of the
highest mountain in the Mugford region ; it was estimated to be in the
vicinity of four thousand feet in elevation. A cone of simple yet effec-
tive outline, it is easily recognizable from a ship for a distance of
seventy-five miles either north or south of Cape Mugford. So conspicu-
ous is it that one cannot but wonder how it has been left so long
neglected and left unnamed among the landmarks of the Admiralty
charts. On board our own craft we fell into the habit of calling the
peak " Brave Mountain," after the name of our doughty little schooner.

Pursuing our northward way tolerably close to land, it could be seen
that for at least fifteen miles from Cape Mugford, the heavy sedimentary
cover and its crystalline cover continued. Not the least important ele-
ment in the imposing panorama was the " Finger Hill " of the charts, a
lung narrow plateau, perhaps two thousand feet in height, that stands
close to the shore, and is evidently composed of the Mugford sediments.
Its name is derived from a large number of tors or rock-pinnacles result-
ing from the dissection of the edges of nearly horizontal strata. Beyond
Finger Hill, the coast turns sharply to the westward, and we could see
no more of the Kaumajet. It is known that this range extends to
the west and northwest of Cape Mugford, and it is presumably of sedi-
mentary or volcanic origin over most of its extent. No fossils were
found in any part of the Mugford series. Here, as in the Kiglapait,
exploration is urgently needed.

From Finger Hill to Hebron we saw but little of the land. Going
north, the " Brave " traversed this section of the coast during the night ;
returning, she remained at an average distance of five miles or more
from the shore. Mainland and islands are relatively low, altitudes of
one thousand or fifteen hundred feet being rarely surpassed. It was
clear from the color of the rocks that the Mugford series does not com-
pose this stretch of country ; it is highly probable that it belongs entirely
to the gneissic complex.

The Torngat Mountain Range.

Topography. — The triangular peninsula east of Ungava Bay is com-
posed of two distinct topographic belts. On the west and southwest the
land is caribou country, low, flat, grass- or moss-covered, with a consider-
able amount of stunted timber growing upon it. Rising abruptly out of
this little elevated belt (charted as the " Kangiva ") is the long, serrate

224 bulletin: museum of comparative zoology.

chain of mountains, or strongly-dissected mountain plateau, extending
oue hundred and fifty miles north-northwest from Kangerdluksoak to
Cape Chidley. In the southern part, the range has an average width of
ahout fifty miles, but it narrows in the north. On the east, it is bordered
throughout its extent by the Atlantic. Technically, the chain is closely
related to the entire gneissic border of eastern Labrador, but its superior
elevation and peculiar topography early marked it out as an orographic
individual. Owing to the wild, forbidding, and awe-inspiring aspect of
the mountain- wall, it is called by the Eskimo the home of the " Torngat,"
or " bad spirits." x Kohlmeister and Kmoch mention the name " Torn-
gets " for the X. W. extremity of the ranges,2 which was mapped as such by
Wei/.3 So far as the writer has been able to discover, the only other
name for any part of the system is that given to the east-central
portion, the "Nachwak Mountains" of Steinhauer,4 or " Nachvak
Mountains " of Kohlmeister and Kmoch.5 There seems to be no good
reason for regarding the whole highland belt as other than a structural
and orographic unit. It is therefore proposed that the ancient name
" Torngat " be extended so as to include all the belt from Hebron to
Cape Chidley.

For a summary of what little is known concerning the Torngats, the
reader is referred to " The Labrador Coast " of Packard.6 Lieber, Bell,
and Koch have respectively made local studies at Eclipse Harbor, Nach-
vak, and Eamah ; all agree in emphasizing the wild, ragged, alpine
nature of the relief. From end to end of the range, razor-back ridges
and horns abound. These are separated by lower rounded hills and yet
more conspicuously by numerous deep fiords and glaciated valleys or
glens, the near relatives of the fiords. All three observers came to the
conclusion that an alpine character has been preserved from preglacial
times because the continental ice-cap did not cover the Torngats. Bell
placed the average elevation of the local valley-glaciers of the ice-period
at about two thousand feet above the present sealevel. As noted more
fully below, this summer's observations correspond very closely with his
estimates. It would be a mistake, however, to attribute a glacial origin
to the rounded profiles of many of the dome-shaped mountains that
alternate with the horns. The former are to be regarded as the result

1 Translation due to Rev. A. Stecker.

2 Journal, 1814, p. 50.

8 A. S. Packard, The Labrador Coast, p. 226.
* H. Steinhauer, Trans. Geol. Soc. London, 1773, vol. 2, p. 488.
8 Journal, p. 20.
. e Pp. 3, G, 19, 226.

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 225

of atmospheric erosion and their slopes as the graded surfaces of moun-
tains normally subdued to relatively tame form by that agency. The
same stage of development awaits their more acuminate neighbors. Gla-
cial erosion has thus not only not reduced the higher summits to flowing
outlines ; by reason of the fact that glaciation has been confined to the
valleys, it has even greatly steepened many slopes and given a more rugged
aspect to the landscapes than belonged to them in preglacial times. One
must ascribe a good share of the wild picturesqueness of the range to its
profound trenching by valley-glaciers. Daring the glacial period, the
Torngats seem to have formed a great dam facing the central neve of Lab-
rador which thus lay on the Kangiva side. It was only here and there
that ice-tongues crept over the low transverse passes, overflowed into the
larger longitudinal valleys, and reached the Atlantic through the corre-
sponding valleys on the east. At that time, the range appeared in the
form of a large number of gigantic nunataks projecting from one to per-
haps five thousand feet above the ice. There resulted among the salient
features of the mountain-belt, the long east and west fiords, of which
Nachvak Bay is doubtless the fmest example.

The Geology. — Crystalline schists form the principal constituents
of the range. Near Hebron, the Johannesberg (2200 feet) is the loftiest
of the high points which begin the chain on the south. At this mission
station, the rocks were examined and found to be common biotite gneiss
and amphibolites, intersected by trap dikes. The schists here trend due
northwest, fairly conforming in attitude with the general trend of the
southern Torngats. Bear Island, a half dozen miles to the eastward,
exhibits a splendid exposure of the dikes. Similar, though much larger,
ones can be traced on the cliffs all the way to Nachvak, a distance of
seventy miles. They are usually vertical, often as much as three hun-
dred feet in breadth and always in contrast with the schists into which
they have been intruded. They are particularly developed on the flanks
of Mt. Blow-me-down (ca. 3500 feet). The continued prevalence of
these intrusives along the coast for the seven hundred miles from the
Straits is, indeed, one of the most notable phenomena of its geology.

The Ramah Sedimentary Series. — The enterprise of Prof. Delabarre
and Mr. Adams permits of the introduction at this place of some interest-
ing data regarding a very extensive stratified cover which appears to
bear the same relation to the crystalline complex as that of the Mugford
sediments. They left the schooner at Hebron and walked over the
mountains a distance of one hundred miles to the Hudson's Bay Post
in Nachvak Bay, where they boarded the vessel again. They crossed

226 bulletin: museum of comparative zoology.

the peninsula between Kangerdluksoak and Saeglek Bay and were
carried by Eskimos across the latter inlet from the Pangnertok (see
Weiz's map in Packard's " The Labrador Coast, p. 226). Thence they
walked to Ramah Bay, were most hospitably received by the mission-
aries and after a much needed rest, went once more overland to their
destination. They report that until they had reached a point three or
four miles north of Saeglek Bay (at about 04° W. Long.), they passed
over gneisses similar to those seen at Hebron. Then bands of black
slate alternating with quartzite were met with. Soon continuous slates
with interbedded sandstones and quartz breccias were crossed, and these
persisted to a point some four miles north of Ramah Bay. Thence the
route carried them over schists equivalent in character to those found
on both sides of Nachvak Bay. The sediments are highly indurated,
and often somewhat metamorphosed. Neither the quartzites nor the
greatly cleaved and pyritiferous slates yielded fossils to the travellers,
who were constantly on the lookout for them. What seems to be a con-
tinuation of the same sedimentary terrane was seen by the writer, though
at a distance, in the form of ragged black cliffs fifteen hundred feet or
more in height, lying southwest of Gulch Cape. The massifs correspond-
ing are tabular, with very low dips and are plainly stratified. Kohl-
meister and Kmoch mention the Ramah slates in the narrative of their
journey in 1811.

Observations in and about Nachvak Bay. Topography and Scenery.
— At dawn on August 21st, the "Brave" was lying-to about six miles
from Gulch Cape waiting for daylight before the not particularly safe
passage of the Nachvak Narrows could be attempted. As we drew
nearer the shore, a brilliant sun flooded with light a scene most impres-
sive to us who were fresh from the softened outlines of the southern
coast and from the yet more featureless landscapes of southern New
England. In front stood the bold fifteen hundred-foot headland whose
many ravines have given the cape its name. Just west of it rose a
curiously regular hill of about the same height, as typical a cone as one
could well imagine, yet apparently an erosion form derived from the
destruction of crystalline schists. To the right of the cone, a long east
and west ridge, estimated at twenty-five hundred feet in height, claimed
instant attention, as it represented, better than any part of the Torngats
yet seen, that serrate topography which Bell and others have emphasized
in the descriptions of the region. It is a knife-edged sierra, trenched
on either flank and from top to bottom by a score of deep transverse
ravines. In the background, visible through two U-shaped notches,

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABKADOK. 227

appeared the stratified tables already referred to. Still farther to the
north, the view included a lofty wall four miles long, interrupted by
deep clefts and by one strong valley-notch. It ends abruptly at the
Narrows of Nachvak Bay. About five hundred yards from the Narrows
we lay some time becalmed close beside this cliff (the eastern slope of
Karmarsuit), one of the grandest on the coast. It was estimated at
approximately two thousand feet in altitude. With an angle of slope
of about eighty degrees, with neither vegetation nor talus, it surpasses in
its wild severity many more celebrated precipices of the world. About
three miles northeast of the Narrows, a very perfect imitatiou of a
volcanic cone was photographed and is here reproduced in Plate 5. Esti-
mates gave the upper lip of the " crater " a height of two thousand feet
above the sea; the depression itself a depth of one thousand feet or
thereabouts. The amphitheatre seems to be an unusually good speci-
men of the glacial cirques which we were to find in great numbers
wherever we had a chance to view the Torngats. Lieber noticed the
frequency of similar forms on Aulatsivik Island. Finally, on the ex-
treme right of the view, some fifteen miles from Gulch Cape, the coast
was outlined by a sierra, " Mt. Razorback " (4000 feet, est.), one of the
most rugged and most alpine, though not among the loftiest, members
of the mountain-system.

Among the maps of Nachvak Bay which have yet been published,
that of Weiz * is the best. While on board the schooner or in the skiff
used in sounding, the writer made a rough sketch of the inlet which is
believed to represent still more closely its true form. (Plate 12). For
purposes of orientation, the Eskimo names for the more important
landmarks have been inserted. These names we owe to Mr. George
Ford, the Hudson's Bay agent at Nachvak, who, in this as in all other
matters, was unwearied in affording us information about the country.
The spelling of the names furnished by Rev. A. Stecker, originally in
German, is phonetic after the sound of English vowels and consonants,
and hence differs in a few instances from that adopted by Weiz.2

i See Packard, " The Labrador Coast," p. 226.

2 The following is a glossary of names appearing in the sketch-map of
Nachvak Bay, with their translations by Rev. A. Stecker of the Moravian
Society.

Idyutak, "lever," referring to the form of the mountain.

Ivitak, "red-colored," land with a reddish color; red ochre; bricks.

Kaputyat, "place for trout-spearing," or (better) "place for spearing."

Karmarsuit, " wall." Some lands in the shape of a wall are called " Karmarsuk "
(little wall) or " Karmarsoak " (great wall). " Suit " is a plural suffix. [over]

228 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

It will be seen from the map that the width of the bay averages about
one and a half miles. Its length is twenty-five miles when measured to
the head of either of the two arms, the Tallek, or the Tessyuyak. Weiz's
map exaggerates both dimensions considerably. It is important to note
that the trench filled with the salt water of the bay is continued at
the head of each arm by a wide and deep glaciated valley. That cor-
responding to the Tallek runs to the south, its floor rising slowly, until,
at a distance of about twenty miles from the Bay, it terminates in a flat
divide adjacent to the valleys drained into Nullatatok and Saeglek Bays.
The upper end of the Tessyuyak has been dammed across by a splendid
alluvial fan that is growing vigorously out from the mouth of a hanging
valley, the floor of which stands about one hundred feet above sealevel
on the south wall of the inlet. The result has been the formation of a
fresh-water lake, the Tessersoak, three miles long, which is drained by
a rapid stream slightly trenching the fan. Beyond the lake to the
westward, the deep valley extends with but little abated strength as far
as the eye can follow it when viewed from the top of the fan. Mr.
Ford informed us that the valley leads directly across the Torngats to
the flat country on the west. The total length of this whole valley is
thus about forty-five miles ; rather more than half its floor lies beneath
the sea.

On both sides the bay is walled in by very precipitous cliffs varying
in height but averaging nearly two thousand feet above the sealevel.
The highest point immediately overlooking the water is unquestion-
ably the summit of the Idvutak, where its western face dominates the
Tallek (Plates 3 and 4). Its height (3400 feet, bar.) was long since

Kipsimarvik, meaning unknown ; old Eskimo word perhaps, of which the mean-
ing is lost, as in the case of many names of places.

Kogarsulc, " small brook."

Korlortoaluk, " big water-fall."

Kutyautak, " wedge," referring to the shape of the mountain.

Nachvak, "found, found at last." The first Eskimo, coming from the west in
search of the salt water, cried out "Nachvak " when he reached the head of the
bay.

Naksaluk, " great valley."

Sennerkitte, "tributary, branch stream."

Sittorutsit, " mountain side where avalanches are common."

Tallek, " arm," referring to the shape of the hay.

Tessersoak (also Tesseksoak), "large pond."

Tessyuyak, " like a pond."

Tinutyarvik, a basin at the head of a cove, barred off by a continuous wall-like
shoal, so that at low tide fish are trapped as in a weir.

DALY : GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 229

measured by Captain Bolton of the British Navy. On the opposite side
of the Tallek is the slightly lower but equally majestic Kutyautak. The
twenty-five-huudred-foot cliffs of the latter have slopes of from fifty to
eighty degrees. Yet more imposing, perhaps, are the extremely rugged
cliffs of the Tessyuyak near its head. They were estimated at twenty-
five hundred feet in height on both shores. Their striking character is
due not only to altitude and the narrowness of the bay, here but a half
mile in width, but also to the greatly variegated color of the rocks.
The usual neutral tones of the cliffs in the Torngats is exchanged for
an irregular association of browns, reds, greens, yellows, slaty blue,
grays and even white, according to the nature and condition of the
schists or of the products of efflorescence.

^Yhile the fiord walls, varying thus from fifteen hundred to thirty-
four hundred feet in height, are relatively continuous and enclose a well-
defined trench, their sky-lines are often broken down by lateral notches.
Some of them belong to glaciated valleys, the bed-rock floors of which
lie below sealevel. Such inlets are more or less filled with deltas and
alluvial funs built out by rapid brooks and torrents. Still more numer-
ous are side-valleys that characteristically mouth at varying heights
above the fiord waters. From Kipsimarvik (the Hudson's Bay Post)
to the Xarrows, twenty-two well developed cirques or corries from
three to eight hundred yards in length were counted. The small
streams draining them leave the corries at altitudes varying from one
to two thousand feet above the sea. Identical in form and relations
with the amphitheatres lining the larger glaciated valleys of the Alps,
of Xorway, of Scotland, and of the Rocky Mountains, they may best be
explained as the result of very local but intense .erosion of small ice-
tributaries feeding the now vanished main glacier that once occupied
Nachvak Bay. (cf. Plate 5.) Other " hanging valleys " of much
greater dimensions are likewise found. The most important of these is
drained by the large stream furnishing most of the water in the pictur-
esque cascade " Korlortoaluk," two miles east of Kipsimarvik (Plate 6).
The main leap of this fall was found by barometric means to be three
hundred and seventy-five feet high ; the total height of cascading water
visible from the bay is five hundred and twenty-five feet, but the valley
floor really appears at a height of seven hundred and fifty feet above
the sea. The fiord is five hundred feet deep opposite the waterfall.
There is thus a total discordance of twelve hundred and fifty feet in the
altitude of the main and lateral valley floors. On an estimated gra-
dient of two hundred feet to the mile, the lateral valley turns north-

230 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

northeast and thus runs nearly parallel to the Bay. The valley
terminates, at a distance of eight or ten miles from the cascade, in a
number of deep corries. Unlike the fiord, the lateral valley flares
broadly and its floor is occupied by a half dozen lakes large and small.
It is itself provided with two branch-valleys which join it just above
the cascade. One of them ends in a glaciated col, the other in a mag-
nificent cirque a mile in diameter. The drainage of these three valleys
serves to sustain the waterfall all the year round.

It may also be noted that a line series of cirques and hanging valleys
look over the Kogarsuk. The most notable of these is the one whose
drainage, forming the Sennerkitte brook, cascades more than two hun-
dred feet into the Kogarsuk. It lies north of Mt. Ford, runs five
miles or more to the eastward, and is floored over with a number of
small lakes. (Plate 12.)

An early statement of an explanation for these apparently abnormal
valleys was given by Helland : " If a glacier fills a tributary valley, and is
thinner than that in the main valley, the depth to which it erodes its bed
must be less than the depth of the main valley. Hence many tributary
valleys must debouch high above the bottom of the main valley. In-
stances abound of tributary valleys debouching thousands of feet above
the beds of the main valleys along the steep sides of the fjords of western
Norway." 1 Helland does not seem to have recognized the full signifi-
cance of his idea in the general theory of glacial erosion ; perhaps be-
cause he was so fully persuaded at the time of the competency of glaciers
in this regard, that further emphasis was not necessary. The advantage
of living on the ground may here have aided perception just as the an-
cient belief of the Swiss peasants long antedated the theory of Char-
pentier and Agassiz that the Swiss glaciers were once much greater
than now. Davis, Gilbert, and Penck have recently and independently
developed the idea to something like its full importance. Reference may
be made to the detailed memoir of Davis for fuller information on this
subject.2

Soundings showed that Nachvak Bay is an excellent type of fiord.
Twenty-one casts of the lead sufficed to determine the submarine relief
in its main features. (Plate 12.) The greatest depth, of one hundred
and ten fathoms, was found at a station six miles from the Narrows,
where the water is nearly as deep. Two miles to seaward of the Nar-
rows, the bottom shoals to an interrupted rock-sill which forms a long

1 A. Holland, Quar. Jour. Geol. Soc, 1877, Vol. 33, p. 174.

2 W. M. Davis, Proc. Bos. Nat. Hist., PJ00, Vol. 29, p. 273.

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADui:. 231

line of breakers athwart the bay entrance.
Ships generally enter the bay by a broad chan-
nel running from Gulch Cape close against the
shore. The sill appears to be represented on
its bottom where a sounding gave twenty-five
fathoms of water. The sill thus rims about the
mouth of the bay in a gentle curve. To the
eastward the Admiralty charts indicate con-
siderably shallower water than that in the bay
itself. Up to a distance of twelve miles from
the Narrows, the average depth of the bay is
one hundred fathoms ; then the bottom rises
rapidly to a narrow bar running across the inlet.
Upon it the maximum depth obtained was
eighteen fathoms. The bar ends at either shore
in a projecting spur of bed-rock ; a fact that
seemed to indicate that we have here to do with
a rock-sill. Eight miles further west, a very
similar sill crosses the bay, with a maximum
depth of fifteen fathoms. To east and west of
this bar, eighty and sixty-eight fathoms respect-
ively were found. This coincidence in location
of sills with constrictions in the fiord seems to
be exceptional among the features of this type of
inlet. A longitudinal profile of the bottom is
given in Figure 2. The broadly U-shaped trans-
verse profile of the Tallek is probably of the
same general quality as the average cross-section
of the whole fiord.

The extremely rapid destruction of the fiord
walls, rendered so steep by glacial "over-deepen-
ing " has entailed the growth of abundant talus,
and thereby the declivity of the submerged slopes
has been diminished. This filling of the trench
is particularly noticeable at the base of the
numerous alluvial cones and fans which almost
invariably appear where strong lateral ravines
notch the cliffs. Creeping and leaping of the
glacial drift down the slopes is going on apace.
Large scallops or tongues of streaming clay,

Figure 2. — Longitudi-
nal section of Nachvak
Bay.

232 bulletin: museum of comparative zoology.

gravel, and small boulders abound. Mr. Ford states that, after a heavy
snowstorm, as many as twenty or thirty avalanches may, in the suc-
ceeding twenty-four hours be expected to fall within sight or sound of
the Post. These slides always bring a greater or less amount of loose
rock with them which, in winter, will find a temporary resting place on
the frozen surface of the fiord, and gradually build a fringe of debris
resting on the ice parallel to the shores. As might be expected, the
number of these falls is greater in the spring than at any other time of
the vear. The marks made by the bounding boulders where they strike
soft ground were found to be in the month of August, extremely fresh,
and must have been formed only a few days previously. Mr. Palmer
saw one boulder six feet in diameter fall from the wall of the Tallek, and,
after its last rebound, leap fully a thousand feet before it struck the sur-
face of the water. As in the Alps, there is a certain element of danger
in travelling on these slopes.

It is not to the credit of American geographical enterprise that the
Torngats are to-day unnamed, unmeasured, unknown in any scientific
sense ; yet they doubtless represent the highest land on the mainland of
the American Atlantic seaboard from Hudson's Strait to Cape Horn.
Lieber stated that " they are not less than 6,000 feet high, and some
peaks may be 10,000 feet high." Koch later remarked that " the highest
points of this range are opposite the island of Aulatsivik, and reach
elevations of from 8,000 to 9,000 feet."1

One short fortnight wTas quite insufficient to permit of any exhaustive
•work on the determination of altitudes, especially as there were other
and more important problems which engaged our attention while the
" Brave " lay anchored at Kipsimarvik. Partly for this reason, only a
few of the lower and nearer mountains were ascended. (Plate 12.)

Immediately back of the Hudson's Bay Post, on the northeast, a
rounded knob was found, with the aid of two standard compensated ane-
roids, to be twenty-eight hundred feet above high water. It was called
by our party " Mt. Elizabeth," after the young daughter of the Hudson's
Bay agent. This mountain is separated by a profound notch from " Mt.
Ford," named after the agent himself. It lies still farther to the north-
eastward ; its altitude, determined again by the barometers, is thirty-
nine hundred feet. From the summit a superb panorama is obtained
on all sides. Due north, two conspicuous peaks some four miles distant
across a deep east-west valley, cut off some of the view. The western-
most was ascended by Professor Delabarre and Mr. Adams and deter-

1 Quoted by Packard, " The Labrador Coast," p. 228.

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 233

mined at forty-four hundred feet in height. This peak is, so far as
known, the highest yet measured in Labrador. The other peak was not
climbed, but was estimated by these gentlemen to be fifty-three hundred
feet in height. Still higher summits dominate these on both sides of
Nachvak Bay. From Mt. Ford the writer saw two extremely sharp horns,
bearing N. 25° W. These must have been at least forty miles distant,
and in all probability are members of the group of the " Four Peaks"
mapped on the Admiralty chart as southwest of Eclipse Harbor. Judg-
ing from their conspicuous position in the sea of mountains, each of
them must be at least seven thousand five hundred feet high. South
of Nachvak Bay there are several fine peaks more than six thousand feet
in altitude.

The Geology. — Bell has already given us some notes on the nature
of the rocks in the bay. Additional observations made last summer
could evidently not form a very complete or systematic piece of work.
The Torngats, where examined along the long cross-section of the
bay, are seen to be essentially composed of contorted but generally
highly inclined crystalline schists, chiefly gneisses. The average strike
is N. 25° W. At Skynner's Cove (Plate 12) ledges of common gray
biotite gneiss and dark-colored hornblende gneiss striking N. 5° W. are
covered with many large boulders of an arkose-like rock enclosing
roundish, often large, grains of opalescent quartz. This rock was not
seen in place ; it may be related to a well-stratified, light colored I'ock
series that occurs on the opposite side of the bay. The latter series is
about five hundred feet in thickness, is manifestly sedimentary and lies
unconformably upon the gneisses. Unfortunately there was no opportun-
ity to make a landing and investigate the region closely. It may be that
the cover represents a continuation of the stratified rocks seen southwest
and west of Gulch Cape. The rocks of the Nachvak cliffs were sampled
by Bell, who reported to have found there a slaty breccia apparently
similar to the arkose of Skynner's Cove, and, as well, a " fine-grained
silicious schist." 1

Nine miles west of Skynner's Cove, on the north side of the bay, the
well-banded gneisses still strike N. 5° W. ; they are here cut by a re-
markable network of dikes exposed on a sheer fifteen hundred-foot cliff.
The parallelism of the dikes gave them at first sight the look of sills,
but closer examination showed that they are independent of the schis-
tosity of the gneisses.

Near the Hudson's Bay Post, the rocks are chiefly coarse, friable
1 Rep. Geol. Surv. Canada, 1882-3-4, Pt. DD, pp. 15, 16.

234 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

hornblende gneiss with abundant lenses of segregated hornblende and
biotite. The strike varies from X. 15° W. to N. 30° W., and averages
N. 25° W. From its trend, the Tallek was seen to be a strike-valley.
The same may be said of the valley of the Kogarsuk, which mouths
just west <>f Kipsimarvik. At the Post, thin sheets or dikes of eruptive
rock cut the gneiss. They are light purplish-gray in color, and recall
the anorthosites of the south. Gneiss similar to that at Kipsimarvik
composes the northern and western slopes of the Kutyautak.

At the head of the Tessyuyak, the variegated cliffs proved to be in
largest part made up of badly weathered ferruginous and silicious schists,
graphite gneiss, and graphite schist. With these was associated in con-
siderable development a peculiar breccia of angular quartz fragments em-
bedded in a black corneous matrix. The graphite occurs abundantly in
the form of disseminated flakes. No large mass of the pure mineral was
discovered in the short time at our disposal. That such a discovery is
possible appears from the fact that a rounded piece of pure graphite
measuring four by five inches has been found at the foot of the talus
near the great alluvial fan ; it is now in the possession of Mr. Ford.

The General Structure of the Coastal Belt.

From the foregoing account, it may be seen that, in the line of the long
coastal section from Belle Isle Strait to Nachvak, Labrador is underlain
by the crystalline complex. If, now, a review is made of the most
important structural element, strike-direction, one cannot but conclude
that the Archaean shield is rather definitely framed on this border, and
that the average direction of the coast-line is related to the tectonic
trend of the ancient mountain-system of which the Labrador plateau is
a diminished remnant. (Plate 11, Table I.). A similar parallelism of
structural trend and coast-line probably exists along the high mountain-
belt of eastern Baffin Land as far as Lancaster Sound.1 It should be
stated that the brief table and the map do not represent the only evi-
dence for this law as expressed in Labrador. Very often the general
attitude of the crystalline schists could be determined from the schooner
when under way, although no opportunity could then present itself for
accurate measurement of the strike. The impression thus gained is
sufficient to warrant our regarding the fidelity of the structure to a
general X. \V.-S. E. trend as of a higher order than is shown in the table
or in the sketch-map. On the other hand, it is evident that no record

i E. Suess, La Face de la Terre, 1900, vol. 2, p. 47.

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 235

TABLE I.
Strike-directions (true) in Newfoundland and Labrador.

Chabactee op Rocks.

Steike.

Great Brc'hat, Newfound-
land

Great Bre'hat, Newfound-
land

Fortune Bay, Newfound-
land

Kirpon Harbor, New-
foundland

Assizes Harbor, Labrador

St. Francis Harbor . .

Pottle's Cove, West Bay

Ice Tickle

Summer Cove, Aillik Bay
Head of Aillik Bay . .
Hopedale

Quirk Tickle, 25 miles S.

of Ford Harbor . . .

Ford Harbor

Island 2 miles N. of
Ford Harbor ....

Island 6 miles N. of
Ford Harbor ....

Cutthroat Tickle, 25
miles N. of Port Man-
vers

Mugford Tickle . . .

Mugford Tickle . . .

Hebron

Skynner's Cove, Nach-
vak Bay

Nine miles W. of Skyn-
ner's Cove, Nachvak
Bay, north side . . .

Great Cascade, Nachvak
Bay

Kipsimarvik

Five miles from head of
the Tessyuyak, Nach-
vak Bay, south side .

slates, sandstones
slates, sandstones
slates

slates and sandstones
gneiss

gneiss
gneiss

gneiss

quartzitic sandstone
quartzite ....
gneiss

gneiss
gneiss .

gneiss ,
gneiss .

gneiss

gneissic floor . . .
sedimentary cover .
gneiss, amphibolite .
gneisses

gneisses

gneiss .
gneisses

gneiss .

N. 65° W.
N. 35° E.

N. 25° E.

N. 20°-30° E.

N. 57°-82° E.

N. 45°-65° W.
N. 60° W.

N. 85° W.
N. 20°-60° W.
N. 35° W.
N. 55° W.

N. 45° W.
N. 5°E.

N. 90° E.
N. 5°W.

N. 30° E.
ca. N. 45° W.
ca. N. 45° W.
N. 45° W.
N. 5° W.

N. 5° W.

N. 30°-35° W.
N. 15°-30° W.

N. 35° W.

stratification

slaty cleavage

stratification

stratification
average

N. 70° E.
average

n. gu° w.

linear parallel
structure

stratification
stratification
gneiss con-
torted

gneiss con-
torted

gneiss con-
torted

gneiss con-
torted

strike gen-
eralized

average

N. 25° W.

vol. xxxviir. — no. 5

236 BULLETIN: MUSEUM OF COMPAKATIVE ZOOLOGY.

of the tectonic axis of the system can be made in the broad areas of
eruptive granites, diorites and gabbros so plentifully distributed along
the coast. In the northern part, from Cape Chidley to Cape Mugford,
the structural trend is represented in the present day topography of the
Torngats and, to a less degree, of the Kaumajets. Elsewhere, there are
only subordinate ranges of hills the directions of which lie parallel to
the strike. The peculiar attitude of the east-west crest-line of the Kig-
lapait cannot be explained until the constitution of that range is known.

It would be of great interest to determine the relation between this
Labrador trend and the structural axes of the Appalachian system. A
hint of that relation was suggested by the structures observed at Great
Brehat Harbor in Newfoundland (twenty miles south of Cape Bauld).
There a series of slates and feldspathic sandstones of unknown age but
very similar to the Cambrian (?) sediments of Kirpon, shows sti*atifica-
tion striking on the average N. 65° W. while a later and beautifully
developed slaty cleavage strikes N. 35° E. The cleavage has an Appa-
lachian trend ; the folds show what may be regarded as the Labrador
trend. If the future survey of northern Newfoundland proves that this
association of structures is general, it may not be too bold to consider
the region as at the nodal point of intersection of the two master struc-
ture lines of eastern North America.

These structural lines are likewise related respectively to the Labra-
dor and Appalachian continental shelves which are so typically devel-
oped. Where the two shelves meet, we have the Grand Bank of New-
foundland. Although not accentuated among the theories of the Bank so
far formulated, the possibility that the plateau on which it stands owes
its origin to tectonic movements at the intersection of the Labrador
and Appalachian structural axes, should not be overlooked. Thoulet
seems to have thrown final discredit on the older view of Maury that the
Bank has been built up from the deep sea by iceberg droppings ; but
Thoulet's replacing icebergs by coast-ice as the carrying agents still
leaves the question open as to whether the materials dredged up from
the Bank may not represent but a very shallow veneer coating the sur-
face of a submerged mountain-plateau.

Observations on the Surface Geology.

The most important problems in connection with the surface geology
of the coastal belt relate to glaciation and postglacial crustal movements.
These processes became most interesting, perhaps, when viewed in the

DALY: GEOLOGY. OF THE NORTHEAST COAST OF LABRADOR. 237

light of their effects on the existing topography. Among the incentives
which the writer had to join the expedition was the need of more detailed
observations than had yet been made on the general direction of ice-move-
ment in glacial times ; on the question of the nou-glaciation of the high
northern mountains ; and on the " lunoid furrows " that were long ago
found on the coast by Packard. The amount and character of post-
glacial uplift, the elevated beaches and other shore-forms associated with
that emergence likewise invited as close study as time and circum-
stances would permit.

Glacial Markings ; Direction op Ice-Movement.

Several visitors to the coast have repeatedly emphasized the difficulty
of discovering there the glacial striae and grooves which should normally
appear on ledges so strongly ice-worn as those of Labrador. For this
reason, it had been anticipated that our necessarily short halts at the dif-
ferent anchorages would not permit of our adding many localities to the
few where the course of ice-movement has been determined. It was there-
fore an agreeable surprise to find ice-markings at nearly every landing
place and often in great abundance. The intense power of the frost has,
in postglacial time, certainly caused some obliteration of such records ;
but the non-discovery of these is much more likely to have been occa-
sioned by the fact that search seems to have been hitherto largely con-
fined to the shore-zone recently emerged from the sea. The evidence
and amount of this emergence will be considered in the sequel. In the
southern part of the coastal belt, it is only the higher points which, dur-
ing the maximum postglacial depression of the land, remained above the
sea. As the land arose, wave-action, coast-ice and the exceptional power
of frost in the zone of flying spray, have tended to cause the destruction
of all glacial rnarks within the belt of land thus exposed.

Selected readings of the directions followed by glacial striation are
given in the following table (Table II.), and are plotted on the map.
(Plate 13.) It will be seen that at all elevations, both in the valleys
and on the hill-tops, the ice-movement was outward from the central
part of the peninsula. The result is to confirm the conclusion which
has been based on five recorded single observations made at Hamilton
Inlet, Indian Island, Davis Inlet, Nain, and Xachvak.1

Glacial Lunoid Furrows. — Chamberlin has shown, by the collection
of many examples, that glacial markings transverse to the course of the
1 R. Bell, Scot. Geog. Mag. 1895, vol. 11, map accompanying text.

23S

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

o
B

s»>

•a .

w

""a"

o 63 63

W W

H»s'a

o 3
.2 B

55°-60
N. 70°
N. 55°

N. 65°
N. 80°

w -g co §
55 ►SBfc

o o

cy ^

&

s

-a

o>

0) 01

CO CO

CD CD
CO 05

3'SJS

*i lo "3

r3 t3

"2 "co"0 '«

t- 93

a

S B

CO

if'i

M

os OS

OS 03

fc

#FH

.— ....

o

fa

CO

'co °co

CO CO

8

.m

S -a

0J

o

o

In

1 «

3

•d

5> S

o> •-

0>

©

,2

"O "S

graniti
arse gra
do.

•^ . -^ ■

o
g
1

■graine

do.

rse gn

grani

do

diori

do

H

CU 03

o

si

o

s o

u

5

o

o .

o

o o

oo oo

■§■8

lOO .

lO »o

o r~ m o

<M e«

CM CM

NH CM

«

1-1

3

3

CD

3

90° E.
70° E.

20° E.

82° E.
85° E.
85° W.

°-60° E.
70° E.
55° E.

So

of

• • • a •

SBoS
' o o 2°

£4
H
OQ

O

00o

• CO N 0' Tj(

. . . • o • •

• O

• CO •

Q

CD

55 CO COCOC055 «5^^h

• Z,<»

•55 55=*55

3

CO

.a
ft

£

55

55

B

§

•s?s

o

3 si?

O O MUMNM io2 .

oo

OOO o

o

i© 35

. t^ o o o

■fli ^ es

CM CM

>-l CMH

■J3 -g-fl

o

S « ho

"^-3

i

u

3

, Newfoundland
Island, Greenspond,

ay, near Cape Bauld,

s Harbor ....
Vest Bay ....
or, near Tub Harbor
pie Right, Rodney

do

Rodney Mundy Id.

Sloop Harbor (Brig

of 335' hill, Sloop

NW. of Conical Id.
iles W. of Pomiad-

do

Summer Cove, Aillik Bay

do. . .
E. side Aillik Bay ....

St. John's
Newell's ]

Newfou
Fortune B

Newfou
St. Franci
Head of \
New Harb
Bake Ap

Mundy

SW. side

W. shore
Harbor)

SW. flank
Harbor

Mainland

Ridge 3 m

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 239

>>

"x

W H

if

■^"^a.O-Oooo ° ° o oo o o •
i o ^ o cc o o Er "5 ° i ^ ° o o CO

o'^OO^cOMCO CO O 0' COO -<»< t~-

§^Z^^ fcZ^ ^^j £ Jz;

55 fc

s

„

s » ss

_ ._ ... 1) 1< M 1) 1/ _■ — — .- ._,._■

^nconnuoco " a M e ^ OT co en

u *"■* CO

c5 csj cj cs

£

c :i '-

o
a
as

'35 co « «

o

g

M

o>

o

55

o

B

D

c; =

_1

■o

fc, w '3

XI

^ co ■*«»£&► » ««

aico _ coco.^.j: 2 coco

~3 ■£ o O O O O ^ 'S « * 0) N -^ •£ "J

o

CD

u a e tuit* *» bo txbti

C ~ ^

«

E

A

c

<3

a

O o

a> .

0 1— . . . . O O O O O iO oo

12 ©

OCOCOOOOO oso oo . oo

.-S<2

© • -3 -C -5 T3 -3 -3 MNM CM CM oS CM t~-

o

■<

3

BBS ,H . o H W HBWH W H M

s

- —

OOn ' O 2J 0 O OOOO * O 0 o ""'.

O O ~a CO S-cO-O OMOU ifliOoiB

s

a

a

o

""* <N CO * "* • • • o ° ° ° ° ° ""* • 00 •* tt

9

Szjjzifc •& • • • ojziss ZZfcZ -fcizjaj

z

5

• . . jzj

o

o

- >S

iq iq * iQ * * ' o

a

a> o a

© £: E: tr o o o oo«o o © o o

O

3 es^

O CO CO . CM . . . o O CO OO o CO <-i O O

O.. . MMM MN M CO . CM t-

iff

i~~i cE ri cj • c8

«!-M

© ' ' "H ' '£ ' ' " S?

do ' • *£ • • " -g ' ' 'I""

_ • . .v. . . . .cS • . -BSjd

3 . . . o S .* *

^ a .'.'>*..; .'s j ; ;ii

H

•<
o
o

Head of Aillik Ba
Hopedale . .
do. . .
do. . .
do. . .
do. . .
do. . .
do. . .
Quirk Tickle, 25

Ford Harbor
Ford Harbor

do.
Island 2 Jo miles
Harbor . .
do. . .
Port Manvers .

do.
Cuttliroat Tickle,
of Port Manver
Mugford Tickle
Hebron . . .
Kipsimarvik, Nae
Great Cascade, N

240 BULLETIN: museum of comparative zoology.

ice are of frequent occurrence.1 In every case, the marking is typically
curved and of lunate pattern, the convex side of the curve heing
directed upstream in the one class of markings, downstream in the
other. There have yet lacked hoth criteria to distinguish the one divi-
sion from the other and information as to the relative frequency of
either kind in nature. Consequently, it has been hitherto impossible
to use these cross-markings extensively as indications of the direction
of ice-movement. With considerable interest and care the coastal belt
was examined for light on the question.

Packard has described and figured the "glacial lunoid furrows"
occurring at Indian Tickle just north of Hamilton Inlet. "These
crescent-shaped depressions, which run transversely to the course of the
bay, were from five to fourteen inches broad by three to nine inches
long, and about an inch deep vertically in the rock. Their inner or
concave edge pointed southwest, the bay running in a general S. W.
and N. E. direction. They were scattered irregularly over a surface
twenty feet square. When several followed in a line, two large ones
were often succeeded by a couple one quarter as large or vice versa." 2
We owe the name " lunoid furrows " to Be Laski, who gave the first
clear account of them as they appear on the ledges about Penobscot
Bay. The furrows are there from one inch to four or five feet in
breadth. " They are lunate in figure, . . . their steep walls invariably
looking towards the north, never directed south as stated in the " Reports
on the Scientific Survey" [of Maine] for 1862, p. 383." 3 His furrows
seem to be equivalent with the " crescentic gouges," " jumping gouges,"
ami "disrupted gouges" of Chamberlin.

It was not far from Indian Tickle that the present writer also found
the luni iid markings for the first time. At Bake Apple Bight on Rodney
.Munilv Island, excellent examples were discovered. The lunoid depres-
sii his, measuring from two to fifteen inches from horn to horn and from one
quarter to one and a half inches deep in the middle of the curve, pre-
sented an appearance essentially similar to that described by Packard.
The two slopes of the depression were always imequal. The longer one
made an angle of from two to twenty degrees with the general surface of
the roche moutonnee ; the other, limited above by the convex line bound-
ing the marking in plan, made angles of from fifty to ninety degrees with
tin' general ledge surface. The furrows or lunes occurred singly and

i T. C. Chamberlin, U. S. Geol Sur., 7th Ann. Rept, 1885-86, p. 218-223.

2 The Labrador Coast, p. 298.

8 J. De Laski, Amer. Jour. Sci., 1864, ser. 2, vol. 37, p. 337.

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 241

also aligned in series. In both cases the axis of symmetry and the
convex side of the curved pit were directed N. 55°— 60° E. What made
this locality of particular interest was the fact that one series of seven
furrows lay on the stoss side of a ledge covei'ed with undoubted glacial
stria? and grooves. These were seen to lie in the axis of the lunes.
The trend of the striae was likewise N. 55°-60° E. Such a relation of
parallelism between lune-axes and stria? was a strong witness to the
glacial origin of the lunes, and it was found to exist in the case of the
great majority of them at each locality on the coast. (See Table II.)

The best exposures of the luues discovered during the summer are at
Pomiadluk Point, at Aillik Bay and at Hopedale. (Plate 7.) At the head
of Aillik Bay near the highest of the elevated shore-lines, a fine group
of naked, highly polished roches moutonnees are marked with numerous
lunes from two to three feet in span. On the east side of the bay at the
shore near the prominent trap dikes opposite Summer Cove, the serial
arrangement of the lunes is exceptionally well developed. In one
instance twelve, and in another fifteen, of them were seen in line.

Single lunes, very rarely lunes grouped in series, are sometimes so
situated that their axes of symmetry are widely divergent from the
accompanying lines of striation. Such an attitude is, however, quite
exceptional, and in general, the perfection of the lunate form is greatest
when the axis of symmetry is orientated parallel to the stria? and
grooves. Like the latter, the lunes were found most commonly on the
stoss side of the glaciated ledge. As in Maine, the convex side of the
lune is always directed downstream with reference to the moving ice-
sheet that formed the grooves. This is true in the south and also at
Nachvak where the glaciation was but local.

There seems to be no reasonable interpretation of the lunes that does
not assign their location to glacial action. But it is difficult to imagine
how either clean ice or boulders caught in the ice and dragged over a
ledge could produce the actual furrow. From a somewhat prolonged
study of the typical examples at Hopedale, the writer was led to believe
that these lunes were only potential when the ice-sheet disappeared and
to extend the same idea to all such furrows. The tension or shearing
stress set up in the bed-rock by a boulder dragged along beneath the
ice, must oftentimes be enormous. This must be so if boulders are
really responsible for the deep striation and grooving of glaciated ledges.
Such shearing stress may be easily conceived to be here and there partly
relieved by the development of incipient cracks dipping gently forward
and, at the same time, sloping inward from each side toward the line

242 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

traced by the boulder. The formation of such a crack would, of course,
be facilitated by a pre-existing tendency of the rock to exfoliate follow-
ing surfaces parallel to the crack, but such a coincidence would be very
rare. The integrity of the rock-surface will henceforth be endangered
because frost can work upon these cracks in the same manner as it
works on joint-planes. The actual hollows would thus owe their existence
to the postglacial splitting action of frost, prying up and breaking off
prismatic fragments of the rock until these fragments had reached a
thickness appropriate to the steep inner face of the lune. The size of
the lune is limited by the distance measured along the crack through
which the rock has been rendered unsound.

The detailed features of the lunes seem to bear out this hypothesis.
In every perfect or only partially completed furrow, a thin crack forms
the prolongation of the gently sloping floor of the depression and is
extended into the yet undisturbed rock at the horns where, in a little
distance, the cracks disappear entirely. There is ofteu a very decided
difference in the freshness of the rock exposed on the steep aud gentle
slopes of the lune respectively. The latter may be distinctly weathered
and lichen-covered, the former almost perfectly fresh and evidently so
recently formed that no plant-life has had a chance to secure a foothold
on the unaltered surface. This contrast between the two slopes is
explicable by the hypothesis proposed ; it forms a difficulty in the way
of accepting any view which would derive a given lune in its present
form from the direct action of the ice or its graving-tool.

While it may be anticipated that most lunes will be developed in
parallel orientation with grooves, it will, by the shear-hypothesis, be no
less certainly expected that tensions will be set up by many boulders not
travelling in the direction of general ice-movement, and that lunes with
quite different orientations will result. As we have seen already, the
facts agree with this expectation. A further cause for the typical devel-
opment of furrows may be looked for in the structure of the rock on
which they occur. It may be for this reason that the lunes of Hopedale
are so well fashioned. There the glacial grooves cut across the schisto-
sity of the gneisses which will doubtless tear apart most easily in that
transverse direction. Where the schistosity is not at right angles to
the line joining lunes in sequence, the lunes are sometimes accordantly
unsymmetrical, as if there were a tendency to exfoliation in a direction
oblique to that line. Yet structure must play but a subordinate part in
the manufacture of lunes, for they are found on such widely different
rocks as the fine to medium-grained gneisses, coarse granitoid gneiss

DALY : GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 243

and coarse granite of Rodney Mundy Island, the granitites of Sloop Har-
bor and Pomiadluk Point, the diorite of Aillik Bay, the gneiss and trap
of Hnpedale and the vein quartz as exposed near Ford Harbor. Each
rock has evidently acted as a more or less homogeneous body with
reference to the deforming force.

Finally, it should be noted that postglacial frost is not of itself sufficient
to produce the systematic form and arrangement of the lunoid markings.
Examination of many ledges in Labrador on surfaces exposed by rifting
since the ice epoch, failed to disclose anything similar to the phenomena
described. To be sure, the familiar rough surfaces of ledges cleft by frost
acting on structural and less regular cracks, is sufficiently analogous to
the siu-face of a roche moutonnee covered with lunes, to suggest frost as
the common cause for the damage done, but the irregular depressions of
the first class lack the peculiar form and orientation of the second.

This explanation of the lunes naturally suggested experiment, but
little has been done in that direction. When glass is scratched with a
diamond point, a multitude of curved transverse cracks is generally pro-
duced. They are always short, highly inclined, and recall in a significant
way the "serrated striae " of Andrews,1 the "cross-fractures" of N. H.
Winchell,2 the "crescentic cross-fractures" and "crescentic cracks"
described by Chamberlin. They may be allied to " chatter-marks " as
well. Russell has described others in the Sierra Nevada.3 The convex-
ities of these curved cracks are always symmetrical to the glacial striye
and are directed upstream, just as the convexities of the cracks in the
glass of the experiment are symmetrical and directed in the sense
opposite to the movement of the diamond. The " crescentic cracks "
were seen in several localities in immediate association with lunes, but
with an invariably reversed attitude of the convexities. The former
class of markings, though commonly arranged in series, are much rarer
than the lunes, and seldom appear on any but the finer-grained and more
brittle rocks, quartzites, slates, and traps. They are usually under three
inches in length and thus much smaller than the lunes. Being simply
cracks, they do not form distinct hollows in the rock. Their steep
inclination doubtless explains the fact that the frost has not produced
such depressions as are likely to be developed where a crack has a low
dip into the rock-surface.

While thus this experiment does not throw direct light on the inter-
pretation of the lunoid furrows, it yet suggests an explanation for a related

1 E. Andrews, Amer. Jour. Sci. 1883, ser. 3, vol. 26, p. 101.

2 Geol. Nat. Hist. Sur. Minn., 1884, vol. 1, p. 548.

3 I. C.Russell, U. S. Geol., Sur. 8th Ann. Rept., 1889, p. 367.

244 bulletin: museum of comparative zoology.

group of markings, and affords reason to believe that both groups are
effects of the release of shearing stresses set up in the roche moutonnee
by a striating boulder.

Packard's early theory of the furrows has been amended to a form
which may be stated in his own words :

" The curved and crescentic or gouge shape of the mark appears to be clue to
the fact (1) that the glacier carried or pushed a more or less angular boulder
ov.r a granite nubble or spur, so that the pressure was greater than at other
points in the valley [the Aar Valley] ; (2) the more or less rounded boulder,
with its lower or under side perhaps somewhat flat, and so situated that the
iee rested only on the top, occasioned greater local pressure than where no
boulders were present ; (3) the boulder meeting with a slight obstacle suddenly
stopped, and the ice pushing it from behind caused it to slightly tip, so that an
immense pressure was brought to bear on the small surface, causing the forma-
tion of a gouge-like crescentic hollow, with the concavity towards the origin
of the motion, i.e., facing up the valley.

" In making my first explanation I wrongly inferred that there might be an
' advancing and receding motion of the glacier,' so as to cause the stone to
turn over.

" In some way, then, due both to the striking or pushing force of the glacier,
and to the local pressure resulting from the presence of a boulder between the
ice and the rock-surface, the boulder was not as with the rest of the ground-
moraine, pushed gradually and slowly onward, but hitched, thus causing it to
break off the lunoid fragment on a surface peculiarly liable, under great local
pressure, to exfoliate." 1

If the writer correctly understand Packard's view, both his hypothesis
and the shear hypothesis agree in removing the furrows from the category
of "chatter-marks," which imply that the gouging-tool must have been
lifted clear of the rock-surface in order to deliver its smiting blow.
That true lunoid furrows may be formed at intervals along an otherwise
continuous groove is shown in Figs. 32, 33, and 34 of Chamberlin's
classic paper.2 It is clear, in each case, that the boulder must have
pressed the ledge with great force as it traversed the inter-furrow space.
Packard makes the hollow as well as the tendency to exfoliate date from
the time of the actual passage of the boulder over the rock ; but he does
not show how localized downward pressure could produce a crescentiform
hollow with a steep scarp facing upstream.

The main conclusions from these considerations are : First, that cres-
centic cracks and lunoid furrows may be distinguished in the field ;
secondly, that while often very abundant, they form valuable criteria

i Amor. Geo]., Feb. 181)0, p. 105.

2 U. S. Geol. Surv.,7tli Ann. Rept.,1885-8G, p. 219-220.

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 245

for the direction of ice-motion in a region once glaciated, and still
strongly frost-bitten in the winter-season at least ; and, lastly, that both
kinds of marking may be regarded as directly or indirectly the product
of shearing stress set up in bed-rock at the gouging corner of a boulder
held in the advancing ice.

The Glacial Deposits.

Nothing is more striking in the glacial geology of the southern part
of the coastal belt than the almost complete absence of drift deposits.
Above the highest shore-line registering the limit of postglacial sub-
mergence, isolated boulders resting on bare rock, or small patches of till
a few inches or feet in thickness, represent the only glacial accumulations
seen at any of our landing-places save one of all those south of Nacbvak
Bay. North of Cape Porcupine and elsewhere, it is true, washed drift that
has been assorted and collected in the form of bay-plains and beaches
now elevated, were found ; but extensive deposits either directly ice-laid
or distributed by glacial streams, normally failed. The exception
referred to, applies to a small but well-defined frontal moraine situated
on the mainland opposite Copper Island near Seal Island Harbor.
Approaching the mainland through a narrow tickle, we were struck by the
sight of a flat-topped terrace, steeply scarped at the shore. This cliff,
averaging some fifty feet in height, was found to be composed of typical
till strongly charged with boulders of gnarled schists and gneisses, a
peculiar hornblende syenite, micaceous trap, pegmatite, aplite and vein
quartz. The beach facing the moraine had greater variety of composi-
tion than any other noted during the summer. Most of the boulders
were clearly erratic. The moraine is bilobate in plan, the curved ridge
trending roughly north and south. It is a curving wall one hundred
feet in maximum height and about a mile and a half in length. The
concave side of the curve looks toward the mouth of a strong east-west
valley opening toward the sea through a range of hills. The latter
approach one thousand feet in altitude and seem to have harbored at
least this one valley glacier in the closing stage of the ice period. On
account of the rarity of such deposits, and on account of its fine develop-
ment, it would have been of interest to study it in greater detail, but no
opportunity to do so was afforded.

The Glaciation of the Torngats.

In 1860, Lieber noted that on the mainland opposite Aulatsivik
Island, " wild volcanic-looking mountains form a water-shed in the

24b bulletin: museum of comparative zoology.

interior, whose craggy peaks have evidently never "been ground down
by land-ice into domes and rounded tops." Bell, in 1884, published his
observations on the Torngats. He stated: "The mountains around
Nachvak are steep, rough-sided, peaked, and serrated, and have no ap-
pearance of having been glaciated, excepting close to the sea-level. The
rocks are softened, eroded, and deeply decayed. . . . Throughout the
drift period, the top of the coast range of the Labrador stood above
the ice and was not glaciated, especially the high northern part." 1

But this serrate topography is not of itself sufficient to disprove that
the glaciatiou was general in the Torngats ; it implies with certainty
only that the glaciation of the range was weak at higher levels at the
time when the hummocky plateau of southern Labrador was remodelled
by the ice. Tarr has shown that, notwithstanding the augularity of its
summits, Mt. Ktaadn in Maine was completely covered by ice in the last
glacial epoch. The present form is understood in the light of the facts
that Tarr has emphasized : (1) The rapidity of frost attack is so great at
high altitudes that the shape of a glaciated knob may be significantly
changed even in postglacial time. (2) The summits have been longer
exposed to the weather than the valleys, since the latter would be occu-
pied by local glaciers during late glacial time. (3) Although the sum-
mit was ice-covered, it would presumably be above the zone of maximum
glaciation. (4) The lack of timber in the higher parts of the mountain
would give the weather permission to produce serrate topography which
would not be expected in the lower vegetated belt.2 In summarizing
the results of his studies of the Cornell glacier, Tarr notes further stric-
tures on the same criterion of non-glaciation.3 He states in effect that
we should expect less erosion by ice in the higher parts, because of the
slower motion of the ice in the upstream portion, and because of the
clean character of the ice above the debris zone. The valleys should
necessarily show more intense erosion than the mountain-tops, since
thick glacier ice would appear first in the valleys as the ice-cap waxes,
and still exists there after the waning ice-cap had deserted the summits.
He has, moreover, found erratics on Mt. Schurmann, a ragged nunatak
in the Cornell glacier. Lastly, he notes that serrateness may be only
apparent, the impression of raggedness being due to the common posi-
tion of the observer on the lee or down-stream side of a glaciated head-
land, as he approaches such a coast as that of West Greenland. "While

1 Rep. Geol. Sur. Canada, 1884, Part DD, pp. 14 and 37 ; cf. Report for 1885,
Part DD, p. 7, and Scottish Geog. Mag., Vol. 11, 1895, p. 340.

2 R. S. Tarr, Bull. Geol. Soc. Amer., 1900, Vol. 11, p. 437.
a Bull. Geol. Soc. Amer., 1897, Vol. 8, p. 252.

DALY : GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 247

it is true that the serrateness of the Torngats is real and extensively
developed, it will be well to present the evidence now in hand, which
shows that, above a certain relatively low line, the range was, during
the last advance of the great ice-sheet, not covered by a glacial mantle
differing in any important way from the snow-blanket which overlies the
country during the present winters.

It is important to note that the discussion of the question will refer
simply to the last advance of the ice. Until it has been determined
whether there was once an interglacial period in Labrador corresponding
to that recorded in the United States and southern Canada, we cannot
be sure that a glacial topography, developed during the first advance,
was not destroyed during interglacial time and during the time occupied
in the secoud advance. It seems probable that the interglacial interval
was longer near the zone of great terminal moraines than in Canada.
Still further north the ice may have lain on the country throughout the
whole glacial epoch. On the other hand, the Torngats may hold
exactly the same relation to the rest of the peninsula as that of the
Lofoden Islands to the Scandinavian mainland. An Alpine relief char-
acterizes the island group to-day, although the ice-cap of the first glacial
advance ran over the group far out into the Atlantic.

The first attempt to solve the problem was made on the slopes of Mt.
Ford. Ascending the mountain on the west side, typical roches mouton-
nees and undoubted erratics cease at the sixteen hundred-foot contour.
Above this, to the summit, the slope is one continuous Felsenmeer.
Although unremitting search was carried on during the ascent, not a
single erratic piece of rock was discovered above the contour mentioned.
The creeping, frost-driven and snow-driven, fragments of ferruginous
gneisses, trap, vein quartz, and syenite were found to be universally
sharp-cornered, never subangular, and always deeply weathered. The
few ledges protruding through the Felsenmeer were likewise profoundly
affected by frost and general weathering, and nowhere presented the
familiar smooth surface of a glaciated nubble of rock. While these
contrasts between the Felsenmeer fragments and decayed ledges above
and the fresh subangular erratics and roches moutonnees less than five
hundred feet below, were very marked, there yet remained the possibility
that the Felsenmeer really covered well glaciated bed-rock over which
the rock fragments have streamed in postglacial times. The rarity of
the ledges between sixteen hundred feet and the summit was such as
not to permit of a satisfactory conclusion based on the study of this one
mountain-slope. On the south side, where the descent of Mt. Ford was

248 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

made toward the cascade, more significant testimony was furnished by
the ledges. Two of these, each from one to two acres in extent, occur
on the flat top of spurs measured at thirty-two hundred and twenty-live
hundred feet respectively. In both cases the outcropping rock is a de-
composed, light-colored gneiss cut by numerous wide bands of black
trap. Each spur-top is separated from the main mountain by a col ;
rock-slides or snow-slides would not sweep over the spurs from the
higher slopes, and thus disturb from their original positions any erratics
that may have been dropped on the spurs by a glacier. Nor would
general creep have removed such erratics. Wherever ledges at all
heights up to twelve hundred feet and above the highest shore-line of
postglacial submergence, were examined elsewhere in Labrador, glacial
boulders were very plentiful. Thus, under essentially similar climatic
conditions, the hills about Hebron have preserved their coating of glacial
drift even on slopes much steeper than on these spurs of Mt. Ford.
It was, therefore, a matter of no small moment to find, after prolonged
search, an utter failure of erratic material on both spurs. Here, too,
the bed-rock shows no trace of glacial smoothing.

At sixteen hundred feet, in the notch between Mount Elizabeth and
Mount Ford, the Felsenmeer ceases in a steep talus-like slope. Below
that level the valley-floor is composed of typical, well-polished and
striated roches moutonnees covered with fresh subangular boulders, which,
from their position, being often delicately poised on the tops of nubbles
sloping outward on all sides, were evidently not brought thither from up
the valley by waste-streams or snow-creep of present day dimensions.
The erratic nature of these boulders of diorite and fine grained ferrugin-
ous schist resting on coarse hornblende gneiss, was clear in the field.
The ledges, though often as steep as others occurring above the sixteen
hundred-foot level, and though similarly devoid of vegetation, are yet
hardly at all riven by the frost. The stria? and grooves are well devel-
oped and are invariably directed down the valley. At the cascade,
" Korlortoaluk," they are confluent with grooves transverse to them-
selves and running seaward parallel to Nachvak Bay. Both sets of
grooves agree with corresponding sets of lunoid furrows in indicating
that the Nachvak trunk glacier flowed eastward and received at the
cascade a south-flowing tributary. (Plate 12.)

Thus, below sixteen hundred feet on the west and south slopes of
Mt. Ford, there is a well glaciated zone. Above that level, steep
slopes of streaming Felsenmeer have deeply covered the ledges. At
twenty-one hundred feet and above, the ledges are in the utmost con-

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 249

trast with those lower down. This contrast is such as to enforce the
belief that glacier ice did not reach higher than about twenty-one hun-
dred feet above the present sealevel.

Similar evidence was found on Mt. Elizabeth, and the limit of glacia-
tion was put at the same level as on Mt. Ford.

Although of no special importance in connection with the problem
now under discussion, brief reference may be made to a curious novelty
that was found in the search for the glacial limit on the twenty-five hun-
dred-foot ridge east of the cascade, Korlortoaluk. About three-eighths
of a mile from the waterfall at the measured altitude of sixteen hundred
and twenty feet, there occui's a shallow col drained south ward into the
Bay and northward into the main hanging valley. This col is floored
over completely with glacial boulders packed closely together so as to
form two smoothly graded slopes gently declining in the directions of
the drainage-lines mentioned. The total length of the doubly graded
pavement is about two hundred and fifty yards. At either end it
terminates in much steeper slopes of bed-rock. The width varies from
thirty to fifty yards. The pavement is bounded laterally by glaciated
ledges covei'ed with erratics ; the latter are plainly in a much fresher
condition than the materials of the general Felsenmeer. That the de-
posit is not of the nature of a barrier-beach is clear from the subangular
form of the boulders that show no sign of having been water-worn.
There is, moreover, unequivocal proof that the land during postglacial
time was not submerged more than about two hundred and fifty feet at
Nachvak Bay. Similar doubly graded slopes of rock-fragments wTere
discovered on Mt. Elizabeth at altitudes of twenty-two hundred and
twenty-three hundred and fifty feet, but these seemed to be simply parts
of the ordinary streaming Felsenmeer, as the rock-fragments were quite
angular and deeply weathered. The two types may be analogous in
origin, but it is difficult to imagine how an ice-sheet could veneer a col
deeply and completely with typically ice-worn boulders. Without even
a woi-king hypothesis to go upon, the question of origin must here be
left unanswered.

But the most satisfactory locality yet studied occurs in the Kogarsuk
Valley. About two and a half miles from the delta of the brook and on
the eastern slope of Kaputyat Mountain there is a series of ten lateral
moraines at elevations of from seven hundred and fifty to seventeen hun-
dred feet above sealevel (Fig. 3). The deposits are composed of large,
relatively fresh, subangular boulders with a small intermixture of clay.
The form of each moraine is that of a steeply scarped bench or of a distinct

250 bulletin: museum of comparative zoology.

wall. One of the best defined ridges, at an elevation of twelve hundred
feet, is from twenty to fifty feet high and some four hundred yards in
length. The longitudinal and transverse profiles are in every way similar
to those of a Swiss lateral moraine. Though generally arranged parallel
to the valley-wall, in some instances the moraine is winged out from it
and is then characterized by many shallow kettles and small ponds. The
general slope of the moraine-crests is to the southward toward Nachvak
Bay, and is sympathetic with the inclination of three well-marked rock-
benches on the east side of the Kogarsuk Valley. Combining the evi-
dence of groovings with these facts, it was evident that the glacier that
was responsible for the moraines and benches flowed southward toward

Figure 3. — Cross-section of the Kogarsuk Valley showing its bed-rock floor overlain
by lateral moraines (solid black) and trenched ground moraine (dotted lines). Kock-
benches on the right. Lookiug north.

the great fiord. At twelve hundred and fifty feet on the flat top of the
long spur which lies in the angle between the fiord and the brook, a
crescent-shaped frontal moraine nearly four hundred yards in length
lies, as it were, stranded there, as the Kogarsuk glacier retreated to its
narrower channel on the east.

The belt of lateral moraines, here and there broken down in the
paths of entering side streams, extends five miles to the northward from
its southern extremity. A little more than three miles from the
mouth of the brook, the valley widens out in a notably flat stretch
estimated to be at least eight miles in length and from one to three miles
in width. The bed-rock is here deeply buried in a till-deposit which
seems to be transitional into the terraced lateral moraines on the we&t

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 251

much after the manner of similar deposits in the High Sierras.1 It
abuts directly agaiust the bed-rock on the east where the lateral
moraines are wanting. This great deposit, so unlike in composition
and form any other considerable glacial product seen in northern
Labrador, was interpreted in the field as a ground moraine. It is
trenched to a depth of more than five hundred feet by the Kogarsuk,
toward which the rough plain slopes on either side.

From seventeen hundred to nineteen hundred feet on Kaputyat moun-
tain, fresh subangular erratics rest on the ledges. This zone is as recog-
nizably glaciated as that of the morainal ridges below. But from nineteen
hundred to two thousand and fifty feet there occurred an abrupt transi-
tion into the region of an interrupted, typical Felsenmeer in no way
markedly different from that on Mt. Foi'd. The rock-fragments are
sharp-angled, rusted to the deep brown color of the adjacent ledges and
strikingly different from the gray tints of the fresher ledges and erratics
below. The transition zone is represented, too, not only in the color, but
also in the form, of the ledges. Though rapid, the change can be dis-
tinctly observed from well smoothed profiles to those which are extremely
ragged. These various phases of transition would be expected if the
glaciation of the valley had been purely local ; its existence in no way
obscures the essential contrast of the two limiting zones.

No "erratics could be found above two thousand and fifty feet. Above
and below the transition zone, the general declivity is constant in
amount, and we cannot ascribe their absence above the zone to a more
rapid rate of creeping. The x-esidence of erratics on the tops of isolated
spurs in the lowest zone forbids the idea that they could have crept
thither from the higher zones.

The limit of glaciation in the Kogarsuk Valley is seen to be very
nearly of the same altitude as on the flanks of Mt. Ford and Mt. Eliza-
beth, namely at twenty-one hundred feet above the sea.

Taking the Nachvak region as a sample of the whole of the higher
Torngats, the general conclusion is that these mountains have not
suffered, during the last advance of the ice-cap, even the limited amount
of glacial erosion that may be discerned on the summits of Mt. Ktaadn,
the Presidential range of New Hampshire, Ben Nevis and the neigh-
boring peaked mountains of the western Scottish highlands, or the rag-
ged outliers of the Scandinavian plateau. It is probable that the higher
parts of the Kiglapait and the Kaumajet massifs similarly formed nuna-
taks overlooking the late Pleistocene ice-sheet.

i I. C. Russell, U. S. Geol. Sur., 8th Ann. Rep., 1889, p. 360.

VOL. XXXVIII. — NO. 5 4

252 BULLETIN: museum of comparative zoology.

The Zone of Postglacial Emergence in Northeastern Labrador
and in Newfoundland.

The path of the " Brave " was such as always to keep us in view of
memorials of recent uplift of the land. These were found to he so
numerous and so striking that one's note-books rapidly filled with the
data of their nature and occurrence. If all other sources of geological
interest failed on the coast, the singularly fresh records of emergence
are yet sufficient to refute Lieber's statement that " the geology of
Labrador is of extraordinarily little interest." The account of the
cruise would be incomplete without reference being made to the obser-
vations made in connection with this subject.

Packard's summary in " The Labrador Coast " makes it unnecessary
to review the earlier studies. Still more recently, Low has published
the elevations of raised beaches in Ungava Bay, Hudson's Bay, James
Bay, and in the interior of the peninsula.1 So far as known, all the
beaches described by both authors are of postglacial date and are
related to the elevated shore-lines early described in the Gulf of St.
Lawrence. On account, however, of the scant information we possess
concerning the actual heights of the Atlantic coast beaches, no very
definite correlation of these with one another or with the beaches
farther west has yet been made. Thus an accurate idea of the amount
or kind of elevatory movement that the Labrador peninsula has suf-
fered in postglacial time, is yet lacking. "What light last summer's
coastal studies throw on the question will first engage our attention.

The bai-ometric method was used in fixing elevations. Four standard
compensated aneroids were employed. The accessibility of the sea-
level and of the British Admiralty's triangulation stations as check-
points, of course, give the aneroid an exceptionally large share of
advantage as compared with that enjoyed by the same instrument
working inland. It is believed that the error seldom exceeded five
feet. At any rate, the accuracy obtained was sufficient for the main
purposes of the study.

It was found to be impossible to group the beaches with reference
to distinct levels, the deformation of which would indicate the type or
types of crustal movement in postglacial time. At only one anchorage
was there discovered a correspondence among the different beaches
which showed that there were special beach-building periods in the
whole time since uplift began. This was at Kirpon Harbor near Cape

1 A. P. Low, Rep. Geol. Surv. Can. 1895, p. 308, and 1899, Part L.

DALY : GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 253

Bauld, Newfoundland. Wave-built terraces were found at six, twenty-
two, thirty-five, and seventy feet above high-water on the hills over-
looking the harbor itself. Others measured six, twenty-two, thirty-five
and one hundred and twenty feet in Mauve (Noddy) Bay. A third
set on Kirpon Island measured six, twenty-two, thirty-five, seventy,
and one hundred and twenty feet, and were particularly well exposed
in Grand Galets Bay. The one hundred and twenty-foot beach-terraces
match well with the very strong rock-benches on the headlands com-
posed of the relatively soft slates about the entrance to Mauve Bay.
At Grand Galets, where the indentation is deep both vertically and
horizontally, terraces at all the levels may be seen. We hoped to find
similar correspondences among the beaches across the Straits, but
failed to do so. In Labrador nothing comparable to the beautiful
system underlying the occurrence of the warped ancient shore-lines
of the Great Lakes could be determined. It is impossible to deny
that the uplift has been on the Labrador coast, as so often illustrated
elsewhere, spasmodic, but it does not follow that this intermittent
character will be reflected in the raised beaches of the present day.
In most cases an ancient beach has been composed and located where
it is because of special local conditions of formation and not because
there occurred a halt in the uplifting process. Deposition takes place
wherever the required protection against under-tow and shore-ice in
the presence of appropriate off-shore depth, is afforded. (Plate 10.)
If the hardness of the rocks, the off-shore depth, or the fetch and direc-
tion of the more effective waves were to change at a given point, the
balance of conditions leading to beach-formation would be destroyed.
Thus a beach growing under the former circumstances, might grow
faster under the new; or, on the other hand, be demolished, its debris
forming a new beach elsewhere or helping to raise the sea-floor whither
it was dragged. Probably no single factor in producing such changes
on the Labrador coast has been so important as vertical movements of
the land.

Not only, thus, will those beaches that were really formed during
halts in the elevatory process, be difficult to distinguish from those
developed in protected bays as uplift takes place ; the record of halts
will be further masked by the appearance of beach-like deposits laid
down on steep-to shores in several fathoms of water. An interesting
example has been described by Packard as "a truly noble beach."
It occurs on the south side of Sloop Harbor. It is about two hundred
yards long and roughly graded from the one hundred and fifty -five-foot

254 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

crest of the ridge enclosing the harbor on this side to the shore two
hundred and fifty yards away. Most impressive was the view along
the rugged accumulations of boulders, rounded or subangular and
varying in diameter from six to ten feet. From the position of the
" beach," it was clear that the enormous energy required to round off
such boulders was derived from waves of great fetch and moving in
water of some depth from the southward ; that is, it looks as if the
cyclopean Felsenmeer had been built up in the lee of the ridge at a
time when the ridge was here entirely submerged. The heavy Atlantic
breakers of that time rifted the masses from the bed-rock and threw
them over the col into the protected place where we now see them.

When we also consider the fact that long stretches of the coast are
too exposed to permit of the growth of beaches at all, it is clear that
the discovery of general levels may be made only after several seasons
of careful field-work. Even after such effort be expended, the quest
may prove to be fruitless. The work that was done in this one
summer certainly led to negative results.

Location of the Highest Elevated Shore-Line. — On the other
hand, a considerable degree of success was attained in the attempt to
fix the highest shore-line, i. e., the most ancient of postglacial levels
now warped into a form which is a resultant of all positive and negative
movements of the land since uplift began. Along the eleven hundred
miles of coast from St. John's to Cape Chidley, no trustworthy estimate
had been made as to the position of this old level. The desirability of
filling in this gap in our knowledge is evident. De Geer has attempted
to construct a map of northeastern North America similar to his classic
one of the Scandinavian Peninsula, showing the character of postglacial
uplift.1 His " isobases " join all points of equal uplift. The map was
left incomplete, since, for lack of information, the isobasic curves could
not be produced into eastern Labrador. Yet for the purposes of geo-
logical theory, it is of the highest importance that this edge of the
glaciated tract should be similarly treated.

The principle used in the determination of the highest shore-line
seems to have been in the minds of Packard and Hind during their
early visits to the coast, but was applied by them only very locally.
Shaler, and later Stone, used it on the coast of Maine, and it has been
employed in Baffin's Land by Watson and Tarr, and still more exten-
sively in Norway, Sweden, and Finland by De Geer and other Scandi-
navian geologists. The criterion is, in a word, that, on appropriate

1 G. De Geer, Proc. Bost. Soc. Nat. Hist., 1892, vol. 25, p. 454.

DALY : GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 255

headlands and hills where the surf from the open ocean can be felt on a
glaciated shore during submergence, the present lower limit of undis-
turbed glacial erratics marks with but subordinate error the highest
shore-line of postglacial submergence.

For the first time during the cruise, conclusive evidence of the value
of the criterion on the Labrador shore was obtained at Ice Tickle Harbor.
It is of importance to review the conditions under which it was there
employed. These conditions are essentially the same as those found at
the numerous other stations occupied up and down the coast.

The situation, topography, and bed-rock geology of the islands on
either side of the tickle have already been described (p. 213). The dark-
hued trap-ridges are dotted over with light gray gneissic boulders offering
strong contrast in color and composition with the ledges beneath. These
boulders are subangular, not water-worn, often greatly decomposed, and
have evidently lain long in their present positions. They are associated
with a small proportion of boulder-clay which is, in places, actively
creeping down the slopes. Occasionally the boulders are pei'ched and
may be easily rocked with the hand. All these boulder-covered ridges
are over 265 feet high. Those of less height are devoid of boulders;
those of greater height may be divided into a boulder-covered and a
boulderless zone separated from each other by the 265-foot contour.
(Plate 8.)

That the boulders are truly glacial erratics, reposing practically where
the great Labrador ice-sheet left them, can hardly be doubted. The
only alternative origins which have suggested themselves are two in
number. Either the boulders, originally deposited elsewhere by the
land-ice, were thrown up by strong storm-waves, or they were brought
thither by floe- or shore-ice. In spite of the unlikelihood of these
hypotheses, it was held that evidence should be obtained that would
thoroughly test them.' The long ridge on the southern shore of Ice
Tickle Island threw ample light on the question, and serves as an
extremely good type locality for demonstrating the criterion.

This ridge, over a mile in length, and generally about three hundred
feet in height (325 feet, measured barometrically at the highest point)
is a residual hill situated where a thick trap dike projects above the
softer gneisses. Its axial trend is roughly east and west. Its sides are
very steep, running together at a sharp edge, whence one looks directly
out over Hamilton Inlet on the south and over a deep valley and beyond
over the open sea north of the island. There seemed to lie no possibility
that waves breaking on the ridge, submerged to the 300-foot, or any higher

256 bulletin: museum of comparative zoology.

contour, could under normal conditions, transport large boulders from
the adjacent depths and lodge them, delicately poised, on the summit of
the ridge. The explanation by floating ice would leave out of account
the great rarity of trap boulders, though these should be the commonest,
since, during submergence, the only shoals for many miles around
whence floating ice might derive rock-fragments other than glacial
erratics, would be underlain by the trap of the present hills. Still less
would the sudden cessation of this sort of deposition at the 265-foot
level as the land arose, be explained. Finally, both hypotheses suffer
from the difficulty that boulders, perched on the ridge-summit as we now
see them, could not be left in situ if the rate of uplift were anything
else than catastrophic. A single heavy storm from the southward would
doubtless suffice to sweep all loose material from their precarious posi-
tion. The conclusion that the boulder-covered zone has never been
submerged since the general ice-sheet retreated from the country, can-
not be escaped. On the other hand, the boulderless zone is a wave-
swept zone.

Other evidences for former submergence of the lower zone are clear
and convincing. The smooth, unbroken surfaces of the roches moitton-
nees above 205 feet contrast strongly with the jagged and riven ledges
below that level. At the upper limit of the boulderless zone, low, but
steep and rugged cliffs look northward over frequent pockets and
graded slopes of well-rounded pebbles. Similar deposits were observed
at 145, 155, and 175 feet, while an unusually fine boulder-beach fifty
yards long and twenty in width occurs at 205 feet. The boulders,
averaging six inches in diameter, are not covered with vegetation, and
bear remarkable resemblance to the raised beaches of Hogland in the
Gulf of Finland. Such beaches are rare on both Ice Tickle Island and
on Rodney Mundy Island. That they were here not discovered more
plentifully is due to the general steepness of the ridges and the conse-
quent lack of places of lodgment for loose material ; to the hardness of
the rocks hereabouts, coupled with the shortness of the time allowed for
beach-building; and to the abundance of moss and other peat-formers
that develop thick vegetable mantles over most graded slopes. An indi-
cation of the great variety of form assumed by elevated sea-chasms,
benches, and boulder-deposits in the wave-swept zone at other localities,
will appear in the following pages.

Occasionally it was found that single boulders in exposed situations
occurred below the accepted level of the highest shore-line. These were
either too large to be moved by the waves, or had rolled down from the

DALY: GEOLOGY OF THE NOETHEAST COAST OF LABRADOR. 257

upper zone since emergence had taken place. Greater care had to be
exercised with slopes covered with streaming drift. In such cases,
masses of clay, small rock-fragments, and boulders compose small tongues
or scalloped terraces, the forms assumed by these materials as they are
washed each year farther and farther down the hill-side. Similar terraces
were photographed in Alaska by members of the Harriman Expedition.

Between landing-places, a tolerably good idea of the altitude of the
highest shore-line could sometimes be attained. The treeless nature of
the headlands caused the boulders of the upper zone to stand out with
great clearness in the different profiles of the hills. As our schooner
hugged the shore pretty closely on the northward, or "downward"
journey, the line could often be located within an error often or twenty
feet. If by good fortune, the cairn of a triangulation station were also
in the view, it was often possible to secure quite useful information con-
cerning the line. A difficulty in using the cairns was, however, found
along the northern half of the coast. Not only are the charts of that
section very incomplete and inaccurate ; it was often not possible to
distinguish the Admiralty cairns from those erected in great numbers
•by the Eskimo on prominent hills, — the " American men " of the
fishermen.

Finally, it is worth noting that the boulder-limit does not exactly
represent the actual former level of the sea, which will be a few feet be-
low the limit. The ancient waves would have an effective reach for
some distance above high-water mark.

Table III. summarizes the results of the observations on the highest
shore-line. At most of the landing-places the land was high enough to
show the line. At others, all the hills ascended were found to be clean
swept. (Plate 8.) For each of the latter the elevation of the highest
hill is given in the table. There also appear the estimated heights of
the line determined from the schooner on islands and headlands sur-
mounted by triangulation cairns. The table shows that the uplift on the
Labrador coast has been greatest near Hopedale. Hamilton Inlet owes
in part its depth, and, indeed, its very existence as an inlet (it is but
10 fathoms deep at the Narrows), to the fact that the part of the
plateau on which it lies has not been elevated as much as the land to
north and to south. The line rapidly rises as it crosses the Strait of
Belle Isle, and seems to be about 500 feet in height along the whole
eastern shore of Newfoundland. It was last observed at St. John's.
Signal Hill (508 feet) is clean swept. The ridge on the south side of
the NaiTows is boulder-covered and the line was estimated at the dis-

258

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

TABLE III.

Observed Heights of the Highest Postglacial Shore-Line in
Newfoundland and Labrador.

Note. Numbers enclosed in brackets refer to altitudes estimated at a distance by the

use of the boulder limit and the triangulated points of the Admiralty charts. Numbers
followed bv the plus sign give the heights of clean-swept hills too low to show the highest
shore-line.

Average Altittjde

OBTAINED FOR HIGH-

Altitudes of Beaches in the

Locality.

EST Shore-Line, in
Feet.

Wave-swept Zone.

St. John's, Newfoundland .

508 +

washed gravels at ca. 400

[575]

feet on Signal Hill

Cape Rouge Harbor, New-

505

Kirpon Itland, Newfound-

450 +

beaches at 6, 22, 35, 70, 120

feet

St. Francis Harbor . . .

365

rolled boulders at 260 feet

Venison Tickle ....

[325-340]

Near Domino Harbor . .

[ca. 300]

[260]

Ice Tickle

265

washed gravels at 145, 155,
175, 250, 255 feet ; beaches
at 140, 205 feet

265

beaches at 115, 140, 160,215
feet

Mainland N. W. of Conical

290 +

Pomiadluk Point ....

345

gravels and beaches at 55,
65, 245, 250, 255, 260, 275,
315, 320, 335 feet

Cape Strawberry . . .

[350]

Aillik Bay

355

beaches (many below 100
feet) ; rolled boulders at
255 feet

390

beaches at 125, 270, 370, 385
feet

Quirk Tickle, 25 miles S.

of Ford Harbor . . .

340

rolled boulders at 250 feet ;
beaches at various levels

290

beaches at 145, 200, 245 feet;
many at lower levels

Black Island Harbor

(Newark Island) . . .

290

several barrier-beaches at 60

feet

285

beaches and bars at 210 feet
and lower

Cutthroat Tickle, 25 miles

N. of Port Manvers . .

270

Mugford Tickle ....

265

260

several beaches up to 225 feet

Kipsimarvik, Nachvak Bay

250

man}' small pockets of beach-
gravels up to 185 feet

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 259

tance of the width of the
Narrows, to lie at the
575-foot contour. No op-
portunity presented itself
for the ascent of this
southern ridge ; hence the
line at St. John's could not
be placed more accurately.

The observations at the
Hudson's Bay Post in
Nachvak Bay gave as in-
disputable evidence of the
relatively low position of
the highest shore-line as
was obtained at the other
coast stations. The po-
sition of the boulder-limit
and the entire absence of
elevated beaches at heights
greater than 250 feet mean
that that elevation is very
close to the level of greatest
postglacial submergence.

If the record of the
table be put in graphic form
(Figure 4) it becomes still
more evident that along
the eleven hundred miles
of coast the elevation has
been differential. The pro-
nounced warping of the
highest shore line is incom-
patible with the view that
changes in the position of
the level of the sea over
great stretches of the
earth's surface, are pro-
duced solely by independ-
ent vertical movements of
the surface of the ocean.

.. ST. JOHNS 575'

L_. C.ROUGE 505'

BELLE ISLE

ST. FRANCIS HARBOR 365'

. GREADY 260'

HAMILTON INLET
■-ICE TICKLE 265'

. POMIADLUK POINT 345'
.HOPEDALE 390'

.QUIRK TICKLE 340'
. FORD HARBOR 290'
PORT MANVERS 285'
CUTTHROAT TICKLE 270'
MUGFORD TICKLE 265'

HEBRON 260'

I NACHVAK 250'

Figure 4. — Curve showing the present warped
condition of the highest postglacial shore-line between
St. John's and Nachvak Bav.

260 bulletin: museum of comparative zoology.

A]on<>- the line on which our observations were made, there has been
unequal positive uplift of the earth's crust. The force responsible for
great piece of work has been applied locally and in varying degree.
The result is that to-day the actual distance from the centre of the earth
of every point on that line is greater than it was at the close of the
glacial period.

The coastal belt shows a degree of elevation considerably less than
that demonstrated for the region east and southeast of James Bay. Sup-
plementing the data of Low's and De Geer's maps with the writer's
observations on the northeastern coast, oue is led to the conclusion that
the uplift of the glaciated tract of this part of America has been greatest
near the region of the central neve.1 The result is to strengthen De
Geer's parallel between the postglacial behavior of the earth's crust in
northeastern North America and in northwestern Europe. The bearing
of this conclusion on the theory of isostasy is obvious. Perhaps the re-
latively great uplift of Newfoundland is connected with the local charac-
ter of its glaciation, which, according to Chamberlain, was not due to an
extensiou of the ice-fields of the mainland.2

Boulder Barricades. — There is not wanting an indication that on
the Labrador coast, at least, the land is higher to-day than in any other
part of postglacial time. Wherever the shore slopes are not too steep, the
coast is belted with lines of innumerable large boulders visible between
low and half tides. These accumulations may be called " barricades."
The name is recognized as appropriate by any one who attempts to land at
low water by forcing a small boat through a gap in the nearly submerged
wall of boulders. The barricade is situated twenty to one hundred or
more feet from the shore according to the slope of the foreshore ; the
distance is greater as the slope is the more gentle. Plate 9 gives a
typical view of oue seen at Ford Harbor. Lyell long since figured an
example in the " Principles of Geology (ed. 11, 1892, Vol. I. p. 381).

Of particular significance is the fact that, in practically every case
where one of these accumulations was examined, it proved to be com-
posed essentially of large glacial erratics. These are believed, in most
cases, to have been derived from the wave-swept zone immediately above.
As the land emerged, the boulders were dragged down in the undertow
and lodged just below the level where the surf could move them. The
relative absence of boulders between the shore and the barricade is also
explained in part by the action of coast-ice, which floats off such boul-
ders when the ice-foot breaks up. If the boulders happened to be

1 Proc. Bost. Soc. Nat. Hist., 1892, vol. 25, p. 464.

2 Bull. Geol. Soc. Amer.,1894, vol. G, p. 407.

DALY : GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 261

dropped again just at the limit where, on account of the depth, the surf
is no longer effective in moving boulders resting on the bottom, the
erratics would further build up the wall constructed by the undertow.
Very close to that limit of effective wave-wash, the winter-ice, because
of the depth at which the boulders lie (submerged at high tide), would
not be able to buoy up the heavy masses and float them away.

If the land had, in postglacial time, ever been higher than to-day,
erratics won from the wave-swept zone would have moved seaward
beyond their present position through the operation of the causes just
described. During a subsequent uplift of the land, the boulders could
not, in any large number, be recalled to the new shore-line. The actual
magnitude of the average barricade is certainly too great to warrant the
belief that the waves would, under this condition, undo what they had,
during the progress of elevation, accomplished. The character of the
material making up the barricade and its organic relation to the wave-
swept zone forbid, thus, the assumption of a secondary uplift following
a former greater postglacial depression of the land than we now see.

One is forced to reject the idea that the barricades have been princi-
pally formed by 'longshore transport of boulders. The walls are devel-
oped very uniformly at the mouth and head, and along the sides of long
bays. If they were the result of 'longshoi-e deposition by floating ice,
we should expect the accumulation of dropped boulders to be quite
uneven, most pronouueed where floes and pans most frequently strand,
and, at other points, scarcely begun. It cannot be denied that coast-ice
does carry boulders in this way, but one may justly question its ability
to have performed so great an amount of work as that demanded in the
construction of the barricades. For this hypothesis implies that the sea-
floor along the broad track of the annual field-ice is peppered over with
glacial erratics far greater in number than could have been furnished by
the wave-swept zone, if that zone had had anything like the average
proportion of the drift now seen above the highest shore-line.

Continuance of Elevation. — We cannot doubt that the elevatory
process continues in both Labrador and Newfoundland . The almost
universal belief of the old settlers on these shores is that in no other
way can the changes in depth at familiar localities be explained. With
no theory to support or refute, many reputable observers among the
fishing population state that they have time and again noted, during
periods of from thirty to sixty years, cases where rock-ledges have come
perceptibly nearer the sea-surface, where new channels have had to be
sought among the shoals for the passage of their fishing-boats, and where

2G2 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

the stages must be again and again lengthened over their bed-rock foun-
dations in order to secure a depth of water sufficient to float the small
craft. Mr. Mark Gibbons, of St. John's, has made a study of the question
for forty years, and has come to the conclusion that elevation is in progress
along the whole coast. He believes that the rate of uplift is about twice
as rapid in northern Labrador as in Newfoundland. He has found among
the older settlements of the island some where the inhabitants are
in a very unfavorable position for plying their industry on account of
the rim of just submerged rock-ledges that obstruct the harbors. He
has asked the older men why they chose such locations for settlement.
The reply was that they or their fathers had made these harbors when
the conditions were very different from the present, namely when the
harbors were deeper. Such qualitative evidence, however great in
amount, must yield in value to the testimony of even a few bench-marks
carefully distributed along the coast. It is hoped that, with the co-
operation of Bishop Marten of the Moravian church, of the missionaries
under his charge at the various stations, of Dr. Grenfell and of Mr.
Ford at Nachvak, a description may be published during the coming
year of eight bench-marks fixed by these gentlemen. One of the stations
is planned for northern Newfoundland.

The Scenery of the Emerged Zone. — The coastal landscapes exhibit
many details which are those expected after the sea-bottom has been
exposed by elevation. They figure among the most striking proofs of
crustal movement.

Within the labyrinth of islands, wTave-cutting has done little toward
modifying the condition of the glaciated ledges. Outlying islands and
headlands, as, for example, at Sloop (Brig) Harbor, Hopedale, Cape
Harrison and Cape Strawberry, are usually very ragged and wave-worn
throughout the wave-swept zone. The same contrast holds in the case
of many individual islands which are uninjured on the landward side but
strongly fretted on the seaward face. On the wooded hills of Great
Bre'hat and Cape Eouge in Newfoundland, where the slates and other
sediments present relatively small resistance to sea-attack, the highest
shore-line was the more easily determined because of the dissimilarity of
the ragged zone of emergence and the smoother, erratic-covered zone.
Benches and strong wave-cut cliffs, at all elevations up to the level of the
highest shore-line, appear at almost every exposed point on the Labrador.
An exceedingly picturesque sea-cliff occurs just above a 205-foot beach
on Pomiadluk Point. Others were particularly noted on Cape Harrison,
Cape Strawberry, at Sloop Harbor, at Ice Tickle, and at Aillik Bay.

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 263

Very rarely are the benches of great breadth or length ; nor are they
usually horizontal. They have been developed by rifting on gently
curving master-joints. They are not the continuous, horizontal terraces
of our text-book diagrams. The benches cannot be taken to mean so
many halts during the process of elevation ; but are rather determined
in size, form, and position by the controlling planes of jointing.

Elevated sea-chasms, often located on trap-dikes, indent the cliffs in
large numbers. The great length of some of these is quite extraordinary.
A half mile northwest of the Mission House at Hopedale and at an alti-
tude of 325 feet, a chasm three hundred yards in length has been worn
out during submergence by waves that followed the trend of a trap dike.
The vast majority of the fossil chasms are similarly located on trap
dikes. The latter, with respect to subaerial destruction, may be hard
in comparison to the country-rock and project above the highest shore-
line as ridges ; below that line the same dikes may prove less resistant
to the attack of the waves than the country -rock and have thus served
to locate chasms. Examples occur on the shores of Aillik Bay. The
position of the highest shore-line is beautifully shown at the upper ter-
mination of a dike-chasm, still floored with rounded boulders, that forms
a conspicuous landmark above the anchorage at Ford Harbor. Above
the line, the surface of the dike is flush with the glaciated gneiss. The
most picturesque chasm seen during the summer is one 250 yards long,
75 feet deep, and 20 feet wide that was found on Long Island, at
American Tickle. The waves still reach nearly to the head of this chasm,
which is still being slowly deepened.

The waste of the fossil cliffs is seen at the present day in the raised
beaches and other boulder accumulations that also represent, in part, the
rearranged drift that once lay scattered over the emerged zone. In
Newfoundland and southern Labrador, these deposits are as a rule
covered deeply under peat, which seems to prefer such graded slopes as
a place for rapid growth. The form and location of the slope was, in
such cases, used in connection with the natural sections of brook-beds
to indicate the nature of the deposit. North of Hamilton Inlet, the
lack of a vegetable cap has rendered the exposures extremely good and
the study easy and rapid. (Plate 10, a and b.) Perhaps the finest ex-
hibitions of the beaches were seen at Sloop Harbor (altitudes measured
115, 140, 160, and 215 feet), Aillik Bay, Hopedale, Pomiadluk Point
(at 55, 65, 230, 250, 260, 315, 320, and 335 feet), and Port Manvers.

The development of the beaches was naturally found to be a function
of four conditions, the relative amount of drift on the shore, the

264 bulletin: museum of comparative zoology.

amount of resistance offered by the bed-rock to wave-erosion, the degree
of protection afforded by outlying islands or shoals against the ocean-
swell, and lastly, the height of the beach above the sea.

Near Cape Porcupine and north and south of Paul's Island, two
typical regions are seen where the drift seems to have been left by tbe
ice-sheet in much greater amount than elsewhere on the coast. Con-
sequently, the "beach" of other parts becomes a coastal plain at the
Cape, or a huge sand and gravel spit in the lee of many an island near
Ford Harbor.

In Labrador, the fossil shore-lines have repeated the process exempli-
fied on the existing shore of eastern Newfoundland. The porphyritic
granite of Greenspond is suffering but very slow attack by the waves,
and the coast is there for many miles devoid of beaches. Approaching
Change Island from the south, one is struck by the more advanced
development of the shore-line as illustrated in the abundant pocket-
beaches and barrier-beaches wrought out of the fissile slates, quartzites,
and schists.

Lying well within the island-belt, the hills about Rogue's Roost, Quirk
Tickle and Cutthroat Tickle possess comparatively few beaches. In
the time of submergence, the outer islands bore the brunt of erosion by
the waves ; the enfeebled Atlantic waves were unable to damage seri-
ously the bed-rock so that it should furnish an essential part of the
beach material. Where beaches do occur in this situation, they are
almost entirely composed of glacial erratics, — a constitution that goes
far to explain the usual contrast of the higher and lower beaches. The
latter are the better developed both as to size and to the degree of round-
ing which the erratics have suffered through wave wear. This contrast
is natural, for the lower deposits are mainly made up of boulders and
pebbles that have been derived from the upper part of the emerged
zone, and thus have been longer in the mill of the surf than the ma-
terials earlier lodged farther up the slope. At the same time, the loss
of bulk by attrition has not been sufficient to counterbalance the gain
to the lower beaches produced by the winning of erratics from above.

On some of the sand beaches, dunes apparently fossil and dating from
a lower stand of the land, were discovered at varying heights above the
sea. These are especially large and numerous in West Bay just south
of Hamilton Inlet. At Pottle's Cove (West Bay) a fifteen-foot pocket
beach two hundred yards long has resting upon it a number of typical
dunes ; the top of one of these is forty-five feet above high water mark.
A much larger area of dunes fifteen to twenty feet high, covers an

DALY : GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 265

extensive thirty-five-foot terrace on the south side of West Bay. The
antiquity of the dunes is indicated by the thick vegetable cap that has
long kept them stationary and in essentially their present form.

The settlements on the Labrador are largely confined to the graded
terraces close to sealevel. Their accessibility, relatively smooth surface,
and naturally sheltered positions, have determined the location of houses,
fishing-stages, and the long lines of bawns on which the codfish are
dried. The inevitable graveyard is always situated on a raised beach,
for only there is there sufficient depth of loose material on the bed-
rock.

True pocket-beaches could not be sharply separated from boulder and
pebble deposits that have been formed on steep-to shores or in the lee
of submerged rock-knobs in several fathoms of water. With the latter
are associated the very common aggi*egations of boulders flooring the
little rock-basins situated among the mammillated ledges of the emerged
zone. In this record therefore, the , term "beach" may, in certain
instances, be arbitrarily used.

One of the commonest forms assumed by the shore-detritus, is that
of the barrier-beach or of its relative, the tombolo or tying-on bar.
Fine examples of these occur at Port Manvers and at the eastern ex-
tremity of Paul's Island. The settlers on the coast recognize that
certain peninsulas have been formed by the tying on of rocky islands to
the mainland ; such islands of an earlier time are locally called " barred
islands." The recency of the shore-uplift is well shown in the numer-
ous fresh-water ponds lying back of raised barrier beaches. Their
basins represent either true coastal lagoons or the depressions located
on the landward side of submarine bars. Often a series of bars form,
with the tied-on rock-islands, the rim of a pond. Examples can be
found at Mauve Bay, Newfoundland, at Pottle's Cove, West Bay, at
Ford Harbor, and at Black Island Harbor. In every case, the pond
exhibits extremely little infilling with wash from the adjacent hills, nor
has the growth of peat significantly diminished the size of the basin.
The freshness of form corroborates in striking degree the other evidence
that goes to show how lately the land has emerged from beneath the
sea. Occasionally the bar is cut through by a stream that has thus
destroyed the integrity of the basin lying back of the bar.

Not the least conspicuous features of the views obtained along the
coast, are the fossil spits such as those at Jigger Island, Sandy Island,
Ford Harbor, John's Harbor and in the vicinity of Nain. Tailing off
with beautifully even slopes from a few hundred feet to more than a mile

266 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

in length, the spits invariably lie on the lee side of the islands to which
they are attached. In most cases they appear to be still growing on
their points although the flanks may be strongly cliffed by the waves ;
sometimes a spit will form a continuous bar from one rock-island to
another.

Finally, attention should be called to the largest single deposit occur-
ring in the zone of emergence and the only relatively large example of its
class of geographic form on this coast : the coastal plain north of Cape
Porcupine. From the cape it stretches fifteen miles northwestward to
Tub Harbor at the North of Hamilton Inlet. The plain averages nearly
four miles in breadth. It is covered with a thick growth of scrub tim-
ber which does not conceal its well graded character. The upper limit
of the plain surface was estimated from a distance to be about two hun-
dred and fifty feet above the sea ; thence the smooth slope descends to
the straight cliffs now being driven back by the actively encroaching sea.
The plain has apparently lost rather more than a mile of its breadth in
this way. There was a comparatively long halt in the process of eleva-
tion when the sealevel was about thirty-five feet above its present
position ; at that time there was developed a distinct bench that is
visible in West Bay. The plain is underlain by stratified sands, and
clays in which there are embedded a great number of large boulders,
including anorthosite from the interior. The bulk of these materials
may doubtless be referred to the drift as their original source. Many
small consequent streams have been extended down the slope of the
plain and are now deeply entrenched beneath its surface. The finest
sand-beach on the Labrador sweeps in a great curve along the present
shore.

The clays of the coastal plain seemed to promise that in them, if
anywhere on the coast, fossiliferous beds might be discovered ; but,
even after prolonged search, the hope was destined to disappointment.
Nor was better success to be had when other deposits of the coast were
examined. It may be that organic remains are truly rare in them, but
the short time permitted for the investigation of the beaches could not
at all warrant this as the final conclusion. The rich finds of Packard
at Hopedale and on the coast to the southward, certainly point to the
expectation that the Labrador Quaternary may some day afford data
sufficient for a fruitful comparison with beds of the same age all about
the north Atlantic.

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 267

The Need of Further Exploration.

One cannot leave the consideration of this huge field of research, the
northeast coast of Labrador, without indicating some of the directions in
which repaying investigation might be carried on. Packard said in 1891
that " the Labrador Peninsula is less known than the interior of Africa
or the wastes of Siberia." : Since that time a harvest of new facts has
been reaped by Low in connection with the Canadian Geological Sur-
vey, and the secrets have to some extent been told. Packard has him-
self done much to remove this stain on the banner of American geological
and geographical enterprise, in the publication of his book, embodying
as it does many original observations. Yet, in his account of the coast,
he was forced to sketch in but the briefest fashion, that part of it which
is by long odds the most interesting, the region north of Hopedale. It
is probable that for many years it will be impossible for government
surveyors to be called away from economically more important fields to
make thorough exploration of the coast. That work seems marked out
by nature for private ventures.

Escaping from the heat of an American or Canadian summer, the
explorer of northern Labrador will find a bracing, health-giving climate
calling forth strenuous and welcome exercise of body and mind. If he
be particularly interested in geological structures and processes, he will
find, in the lack of soil and forest-cover, most fortunate conditions for
rapid observation. Using a steamer sheathed for ice-navigation, an
exploring party might be on the ground in early July and, from one end
of the coast to the other, rarely fail to find a snug harbor close to the
important points of attack.

The Kiglapait is unmeasured, unmapped and absolutely unknown as
to composition. The Kaumajet sediments, covering several hundreds of
square miles, present important structural and stratigraphic problems.
Their age is quite undetermined, like that of the stratified rocks at Aillik
Bay, at Pomiadluk Point and at Ramah. The Torngats afford a field of
operations which it will take many seasons even to reconnoitre. In the
last mentioned range are the highest mountains on the Atlantic sea-
board of America, unmeasured and almost entirely unnamed. It would
be of much importance to fix the elevation of the highest postglacial
shore-line in the interior of the peninsula as well as on the south coast,
in Newfoundland, Cape Breton and across Hudson's Strait, where isolated

1 The Labrador Coast, Preface, p. 5.
vol. xxxvm. — no. 5. 5

268 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

observations have already been made. The gross imperfection of the
o-overnment charts for the region north of Cape Harrigau is such that
no just idea is given of the splendid fiords that indent the plateau.
These should be sounded and mapped if the great fishing fleet is to find
appreciable help in their arduous calling from those who can afford the
leisure to do this necessary work. The Labrador should be mapped at
least as carefully as the coast of northern Norway. With the mapping,
detailed observations of value on the hydrography of the coastal waters
could be carried on. The remarkable tides of Ungava Bay, the marine
zoology of the coast, particularly the study of the jelly-fishes, the fixing
of bench marks to show the rate of elevation on the coast, the study of
the fossiliferous beds of the raised beaches, — these and other subjects
of research await the explorers of the future.

DALY: GEOLOGY OF THE NORTHEAST COAST OF LABRADOR. 2G9

EXPLANATION OF PLATES.

PLATE 1.

The basalts and underlying sediments of the Mugford Series as exposed on the
northwest side of Mugford Tickle. The basement of crystalline rocks is seen
on the left. It disappears near the middle of the view in consequence of the
strong flexure revealed in the attitude of the stratified rocks above. The cliff
is about 1,800 feet in height where it rises above the exposed basement.

PLATE 2.

The unconformity of the Mugford Series and Crystalline Complex, at Cape Mug-
ford ; cliff here from 1,800 to 2,000 feet in altitude.

PLATE 3.

The Tallek, or southern arm of Nachvak Bay ; view taken from the height of
about 1,500 feet on the southern slope of Mt. Elizabeth. Mt. Idyutak on the
left.

PLATE 4.

The highest cliff in the Tallek ; the cloud-capped peak on the left is the summit
of Mt. Idyutak, 3,400 feet in height; view taken at a distance of about one
mile from the shore.

PLATE 5.

Glacial amphitheatre near Mt. Razorback, north of entrance to Nachvak Bay ;
looking north.

PLATE 6.

Hanging Valley of the Korlortoaluk, Nachvak Bay (see Plate 12) ; Mt. Elizabeth
on the left, separated from Mt. Ford on the right by a glaciated col ; this col is
drained into the main hanging valley, which comes in from the right ; looking
north.

270 BULLETIN : MUSEUM OF COMPAKATIVE ZOOLOGY.

PLATE 7.

Glacial lunoid furrow, one foot from horn to horn, and an inch and a quarter in
depth ; found with many others on hornhlende granite at Hopedale.

PLATE 8.

Wave-swept zone as represented in Bear Island, a granite knob cut by trap dikes ;
about 100 feet in height.

PLATE 9.

Boulder barricade at Ford Harbor ; at half tide; looking east.

PLATE 10.

a. Elevated gravel beach and sand dune at West Bay, south side of entrance to
Hamilton Inlet ; beach 35 feet above high water.

b. Near view of boulder beach 100 feet above high water ; west (lee) side of Brig
Harbor Island.

PLATE 11.

Sketch map of the northeast coast of Labrador, showing data concerning bed-rock
geology. Heavy straight lines : observed strikes. Sinuous lines : average strike
of contorted schists. Horizontal lining: Ramah series. Vertical lining : Mug-
ford series. Solid black : Nain gabbros.

PLATE 12.
Sketch map of Nachvak Bay. Heights in feet ; depths in fathoms.

PLATE 13.

Sketch map of Newfoundland and part of the Labrador peninsula, showing data
concerning glacial geology and postglacial elevation. Arrows indicate observed
directions of glacial striae ; numbers show local elevations of highest postgla-
cial shore-line about present high water. (See Table III.)

! l:i

l I

1:

.\i

i

■•mi,;. ! I

f. t 2

?S&5*

lllS§fc

Dalv. Labrador.

Plate 10.

6LIOTYPE CO., BOSTON.

RAISED BEACHES. A. WEST BAY. B. BRIG HARBOR ISLAND.

Daly. Labrador.

Plate ii.

CAPE CHIDLEY

NANUKTUT I .
^APE MUCFORD

CUTTHROAT TICKLE

MGUAPA1T-

CAPE HARRISON

CE TICKLE
HAMILTON INLET
WEST BAY

ST FRANCIS HR

y^?

BED-ROCK GEOLOGY.
SCALE : too miles to one inch.

Daly. Labrador.

Plate 12.

SKETCH MAP OF NACHVAK BAY.

SCALE : 6 miles to one inch.

Contour Interval : 500 feet.

Daly. Labrador.

Plate 13.

ys CAPE

CHIDLEY

UNGAVA

■< /k

BAY (

^

. .250 NACHVAK BAY
^Z/;...260 HEBRON

<

**^"ftcrr^ , 265 CAPE MUGFORD
*^Rj. 270 CUTTHROAT TICKLE

y

fc^T...28S PORT MANVERS

<p

J^Ea^< . . . 290 FORD HARBOR

\$»C— -340 QUIRK TICKLE

*fj \C . 390^«H0PEBAIE

]/^-rJ*C/J 555 *1LLIK BAY

US' »^.545 PONIADUJK POINT

/ Z^jtK.. - . .285 ICE TICKLE

— -J^~~T! »WIST BAY
■O C^J JSA* <26<" OREABY

ySjl XT ^^»J-- '300) DOMINO HARBOR

^ «« . . . 1 325 -40J VENISON TICKLE

j£ 1 MS ST FRANCIS HARBOR

■P ^

S 6

y /\S* 450+ KIRPON HARBOR

— ~^. l'A FORTUNE £AY

f J -</«... 505 C.ROUGE HARBOR

^

w

NEWFOUNDLAND

.(575) ST JOHN'S

GLACIAL STRIATION : POSTGLACIAL ELEVATION.
SCALE: 170 miles to one inch.

The following Publications of the Museum of Comparative Zoology
are in preparation: —

Reports on the Results of Dredging Operations in 1877, 1878, 1879, and 1880, in charge of Alex-
ander Agassiz, by the U. S. Coast Survey Steamer " Blake," as follows : —

E. EHLERS. The .Annelids.

C. HARTLAUB. The Comatulse, with 15 Plates.

H. LLTDWIG. The Genus Pentacrinus.

A. MILNE EDWARDS and E. L. BOUVTER. The Crustacea.

A. E. VERRILL. The Alcyonaria.

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in charge of Alex-
ander Agassiz, on the U. S. Fish Commission Steamer " Albatross," from August, 1899,
to .March, 1900, Commander Jefferson F. Moser, U. S. N-, Commanding.

Illustrations of North American MARINE INVERTEBRATES, from Drawings by Br/BK-
hardt, Sonrel, and A. Agassiz, prepared under the Direction of L. Agassiz.

LOUIS CABOT. Immature State of the Odonata, Part IV.
E L. MARK. Studies on Lepidosteus, continued.

" On Arachnactis.

R. T. HILL. On the Geology of the Windward Islands.
W. McM. WOOD WORTH. On the Bololo and Palolo of Fiji and Samoa.
A. AGASSIZ and A. G. MAYER. The Acalephs of the East Coast of the United States.
AGASSIZ and WHITMAN. Pelagic Fishes. Part II., with 14 Plates.
J. C. BRANNER. The Coral Reefs of Brazil.

Reports on the Results of the Expedition of 1891 of the TJ. S. Fish Commission Steamer
" Albatross," Lieutenant Commander Z. L. Tanner, U. S. N., Commanding, in charge of
Alexander Agassiz, as follows : —

A AGASSIZ. The Pelagic Fauna.

" The Echini.

" The Panamic Deep-Sea Fauna.

K. BRANDT. The Sagittse.

" The Thalassicolae.

C. CHUN. The Siphonophores.

" The Eyes of Deep-Sea Crustacea.

W. H. DALL. The Mollusks.
HE. J. HANSEN. The Cirripeds.
W. A. HERDMAN. The Ascidians.
S. J. HICKSON. The Antipathids.
W. E. HOYLE. The Cephalopods.
G. VON KOCH. The Deep Sea Corals.
C. A. KOFOID. Solenogaster.
R. VOX LENDENFELD. The Phosphor-
escent Organs of Fishes.

H. LUDWIG. The Starfishes.

J. P. McMURRICH. The Actinarians.

E. L. MARK. Brachiocerianthus.

JOHN MURRAY. The Bottom Specimens.

P. SCHIEMENZ. Pteropods and Hetero-

pods.
THEO. STUDER. The Alcyonarians.
M. P. A. TRAUTSTEDT. The Salpidae and

Doliolid*.
E. P. VAN DUZEE. The Halobatidae.
H. B. WARD. The Sipunculids.
H. V. WILSON. The Sponges.
W. McM. WOODWORTH. The Nemerteans.
" The Annelids.

PUBLICATIONS

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletin Vols. I. to XXXVII. ;
of the Memoirs, Vols. I. to XXIV.

Vols. XXXVIII., XXXIX., and XL. of the Bulletin, and Vols.
XXV., XXVI., and XXVII. of the Memoirs, are now in course
of publication.

The Bulletin and Memoius are devoted to the publication of
original work by the Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural History, and of work by specialists based
upon the Museum Collections and Explorations.

The following publications are in preparation : —

Reports on the Results of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander C. D. Sigsbee, U. S. N., and Commander J. R. Bartlett, U. S. N.,
Commanding.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission
Steamer "Albatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in
charge of Alexander Agassiz, on the U. S. Fish Commission Steamer
"Albatross," from August, 1899, to March, 1900, Commander Jefferson F.
Moser, U. S. N., Commanding.

Contributions from the Zoological Laboratory, Professor F. L. Mark, Director.

Contributions from the Geological Laboratory, in charge of Professor N. S.
S baler.

These publications are issued in numbers at irregular inter-
vals ; one volume of the Bulletin (8vo) and half a volume of the
Memoirs (4 to) usually appear annually. Each number of the
Bulletin and of the Memoirs is sold separately. A price list
of the publications of the Museum will be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 6.

LEUCITE-TINGUAITE FROM BEEMERVILLE, NEW JERSEY.

By John E. Wolff.

J CAMBRIDGE, MASS., U.S.A.:
PRINTED FOR THE MUSEUM.
February, 1902.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 6.

LEUCITE-TINGUAITE FROM BEEMERVILLE, NEW JERSEY.

By John E. Wolff.

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

PRINTED FOR THE MUSEUM.

February, 1902.

FfP ] 3 1902

No. 6. — Contributions from the Harvard Mineralogical
Museum, VI.

Leucite-Tinguaite from Beemerville, New Jersey. By John E. Wolff.1

Introductory.

The post-Cambrian eruptive rocks of Sussex County, New Jersey, have
been the subject of several interesting papers,2 which have described the
elaeolite-syenite and some of the associated rock masses (ouachitite), and
several of the dikes, but not the rocks of the whole complex. In 1896
and other years the writer was engaged with the areal geology of this
region and made rock collections from the old localities and from several
new ones discovered in going over the ground ; a list of which with map
has been published.3 He hopes to complete the description of this
material, of which the present paper forms the first instalment.

Leucite-Tinguaite.

This rock occurs as a dike in the main mass of the elaeolite-syenite
two miles northwest of Beemerville, and near the southwest end of the
syenite. It was found in a field just under the 1300-foot contour of
the topographic map and a short distance east of the road over the
mountain. The dike is fifteen inches wide, cutting coarse syenite, with
a strike N. 35 "W. and dipping 45 northeast ; it is exposed for but a
few feet.

The rock has a dull greasy green color, is dense grained and contains
distinct phenocrysts of nepheline in hexagonal prisms and large pheno-
crysts of pseudo-leucite, which weather white on the rock surface and
are greenish white in the fresh rock ; they have a diameter of fifteen
centimeters at the maximum and occur in bands parallel to the walls
of the dike. Their form is that of the leucite eikositetrahedron, the
crystal forms rather perfect, and the outlines linear and sharp in cross

1 Published by permission of the Director of the U. S. Geological Survey.

2 B. K. Emerson, Amer. Journ. Sci., 23, p. 302 and 376, 1882. J. F. Kemp,
Amer. Journ. Sci., 38, 1889, p. 130; the same, Trans. N. Y. Acad. Sci., 11, p. 60,
1892 ; Amer. Journ. Sci. 45, 1893, p. 298 ; Amer. Journ. Sci., 47, 1894, p. 339.

» Geol. Survey N. J. Annual report, 1896, p. 91.

VOL. XXXVIII. NO. 6

274 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

section. The nepheline and ldTicite phenocrysts are evidently somewhat
vicarious, for zones of the rock rich in one are poor in the other.

Microscopically the rock is composed of the following minerals :

Phenocrysts of nepheline and pseudo-leucite ; few crystals of augite
and occasional plates of biotite.

Ground mass : aegirine-augite needles, orthoclase, nepheline as essen-
tials, with a little analcite, biotite, calcite, etc.

The nepheline phenocrysts are glassy and fresh, bounded by base and
prism. The pseudo-leucites are composed of an aggregate of oi'thoclase,
partly in irregular grains, more often in radiating prisms, forming even
imperfect sphero-crystals, several of which may occupy one original
leucite. Between the orthoclase there is a comparatively small amount
of nepheline, usually in irregular patches, rarely in prisms, and this
nepheline lies between the orthoclase rays. The space in the centre of
the original leucite, especially that between the curving boundaries of
two opposite spheroids of orthoclase is filled with a clear isotropic sub-
stance with refractive index lower than the feldspar, traces of cubic
cleavage, gelatinizing with HC1, and determined as analcite. Into this
the square ends of the orthoclase rays sometimes project with a struc-
ture curiously analogous to that described by Iddings (" Obsidian
Cliff") in the obsidian spherulites, where rays of orthoclase project
into a quartz or tridymite paste. Associated with the analcite there
occurs an undetermined zeolite with equally low index, good double
cleavage and distinct though feeble birefringence. The very early period
of formation of the leucites is shown by their penetration by tongues of
the rock carrying nepheline and augite phenocrysts.

The pyroxene phenocrysts are a deep green, slightly pleochroic, aegir-
ine-augite, with the direction of negative extinction about 38° to c'.
They sometimes have a centre of colorless or reddish augite and an
aegirine border. Small masses of purple fluorite are deposited within
or around them and they are frequently filled with secondary biotite.
Associated with them are a few melanite crystals, titanite masses, areas
of colorless fluorite and astrophyllite(l).

The ground mass comprises a fine net work of slender pyroxene crystals
in a background of colorless minerals. These acicular pyroxenes are
pale to deep grass green in color, slightly pleochroic, parallel c' grass
green perpendicular to c' olive green; most of them extinguish nearly
parallel to c', others as high as 19° ; in either case this is the negative
optical direction and they would pass for aegirine or aegirine-augite.
The colorless background is resolved in polarized light into irregular

WOLFF: LEUCITE-TINGUAITE FROM BEEMERVILLE, N.J. 275

areas in which three elements occur. The most abundant is orthoclase,
identified by its index of refraction lower than balsam, cleavages, and
biaxial character; the analysis shows it can contain little sodium.
Nepheline comes next, identified by its higher index, and lastly there
are some isotropic areas of very low index with irregular cracking which
are probably analcite.

A little biotite, calcite, fluorite, quartz, brilliantly polarizing can-
crinite(l) and small areas of indeterminate zeolites are in insignificant
amount. Grains of pyrite are scattered through the rock.

Analysis.

The analysis was made essentially according to the methods used by

Hillebrand.1 While it shows the general chemical characters of allied

tinguaites, yet a striking difference is shown in the predominance of

ferrous over ferric iron, combined with the high and nearly equal amount

of the alkalies. The ferrous iron determination was made in duplicate

by the hydrofluoric acid method, using a doubly tubulated bell-jar and

sand bath instead of the water bath of Cooke (Pebal-Doelter method,

see Jannasch, " Leitfaden der Gewichts- Analyse," p. 269), and the two

determinations differed by only 0.02 per cent. The only mineral in the

rock containing iron (except pyrite, which is separated in the tabulation,

and a little biotite, etc.) is the pyroxene, and this is almost all in the

ground mass where it has the optical characters of aegirine or aegirine-

augite. A glance at the molecular proportions will show that there is an

excess of the alkalies over the (AlFe)203 of 21 molecules and as there is

no sodalite present it seems necessary to suppose this excess in sodium

present in the pyroxene, where it must be combined in some way with

ferrous and not with ferric iron, as would be the case were it aegirine.

To better illustrate the difficulty, if we use all the ferric iron, with one

sixth ferrous, and the proper soda for aegirine proper, distribute part

of the remaining soda and the potash with the proper alumina and

silica between nepheline and orthoclase, part of the ferrous iron with

the lime and magnesia in the hedenbergite molecule, we will find with

3 per cent in the rock of aegirine, 13 per cent hedenbergite and 6 per

ii
cent of a compound Na2Fe(Si03)2. The investigations of Merian a

and Mann 8 indicated that aegirine-augite contained sodium in some

1 Bull. U. S. G. S., No. 176. » N. J. Min. B. B., III. 1884, p. 252.

8 N. J. Min. B. B., II. 1884, p. 172.

276

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

A

B

C

D

E

F

G

SiO., . . .

50.00

50.88

848

53.56

50.36

52.91

51.93

Al.,03 • •

. 20.03

20.38

200

24.43

19.34

19.49

20.29

Fe,03 • •

0.98

1.00

6

2.19

6.94

4.78

3.59

FeO . . .

3.98

4.05

56

1.22

2.05

1.20

MgO . . .

0.69

.70

17

.31

0.29

.22

CaO . . .

3.41

3.47

62

1.24

3.43

2.47

1.65

Na20. . .

8.28

8.43

136

6.48

7.64

7.13

8.49

K20 . . .

8.44

8.59

91

9.50

7.17

7.88

9.81

HoO + 110
HoO-110

s 1.50
3 0.10

M

loss.
3.51

M

0.99
0.10

co2 . . .

0.22

0.22

6

abs.

.25

Ti02 . . .

0.99

1.01

12

abs.

.20

P205 . . .

0.21

0.21

1

. . .

trace.

.06

FeS2 . . .

0.54

0.55

4

S03

= 0.52

= .67

Fl

. not det.

0.27

CI

trace.

0.53

0.70

MnO . . .

0.50

0.51

7

0.10

0.41

0.44

trace.

BaO . . .

abs.

X. 0.48

0.09

SrO ...

0.09

0.07

99.87

100.00

1446

99.96

98.80

100.25

100.58
0 = .27

100.31

Specific gravity, 2.701

at 20° C.

A. Leucite-tinguaite, Beemerville, by J. E. Wolff.

B. The same calculated to 100 water free.

C. Molecular ratios.

D. Elaeolite-syenite, Beemerville, by L. G. Eakins, Bull. U. S. G. S., No. 150, p. 211.

E. " " " by F. W. Love (Kemp, Trans. N. Y. Acad. Sc.,
Vol. XI. p. 60, 1892).

F. Leucite-tinguaite, Magnet Cove, by J. F. Williams, Arkansas Geol. Survey,
Vol. III. p. 287.

G. Leucite-tinguaite, Bearpaw Mts., Montana, by H. N. Stokes (Weed and
Pirsson, Am. Journ. Sc, II. p. 196, 1896).

WOLFF : LEUCITE-TINGUAITE FROM BEEMERVILLE, N. J. 277

other combination with ferric iron than as aegirine ; the present case
indicates some combination with ferrous iron as yet unknown, and of
course purely hypothetical.

A rough quantitative mineral calculation of the rock gives :

Pyroxene, 22 per cent.
Nepheline, 36 per cent.
Orthoclase, 38 per cent.
Titanite, Apatite, etc., 4 per cent. ;

the analcite actually present is excluded but would of course lower the
percent of nepheline. Analyses D and E are introduced as the only
published ones of the elaeolite syenite of Beemerville ; the incomplete
analysis E shows much similarity with the tinguaite. Analyses F and
G of leucite-tinguaites differ mineralogically from the Beemerville tin-
guaite by containing sodalite.

It should be mentioned that J. F. Kemp {loc. cit.) has described two
camptonitic dikes — six, and nine and a half miles respectively distant
from the Beemerville dike — in which certain curious spheroids without
definite crystal outlines and composed mainly of analcite are referred to
original leucite, of which some traces are stated to remain.

Mineralogical Laboratory,

Harvard University, Jan., 1902.

The following Publications of the Museum of Comparative Zoology
are in preparation : —

Reports on the Results of Dredging Operations in 1877, 1878, 1879, and 188<>, in charge of Alex-
ander Agassiz, by the U. S. Coast Survey Steamer " Blake," as follows: —

E. EHLERS. The Annelids of the " Blake."

C. HARTLAUB. The Comatulse of the " Blake," with 15 Plates.

H. LUDWIG. The Genus Pentacrinus.

A. MILNE EDWARDS and E. L. BOUVIER. The Crustacea of the "Blake."

A. E. VERRILL. The Alcyonaria of the " Blake."

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in charge of
Alexander Agassiz, ou the U. S. Fish Commission Steamer "Albatross," from August,
1899, to March, 1900, Commander Jefferson F. Moser, U. S. N., Commanding.

Illustrations of North American MARINE INVERTEBRATES, from Drawings by Bork-
hardt, Sonrul, and A. Agassiz,. prepared under the direction of L. Agassiz.

LOUIS CABOT. Immature State of the Odonata, Part IV.
E. L. MARK. Studies on Lepidosteus, continued.

" On Arachnactis.

R. T. HILL. On the Geology of the Windward Islands.

W. McM WOODWORTH. On the Bololo or Palolo of Fiji and Samoa. *

A. AGASSIZ and A. G. MAYER. The Acalephs of the East Coast of the United States.
AGASSIZ and WHITMAN. Pelagic Fishes. Part II., with II Plates.
J C. BRANNER. The Coral Reefs of Brazil.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission Steamer
"Albatross," Lieutenant Commander Z. L. Tanner, U. S. N., Commanding, in charge of
Alexander Agassiz, as follows: —

A. AGASSIZ. The Pelagic Fauna.

" The Echini.

" The Panamic Deep-Sea Fauna.

K. BRANDT. The Sagittse.

" The Thalassicolie.

C. CHUN. The Siphonophores.

'•' The Eyes of Deep-Sea Crustacea.

W. H. DALL. The Molluslcs.
H. J. HANSEN. The Cirripeds.
W. A. HERDMAN. The Ascidians.
S. J. H1CKSON. The Antipathids.
W E. HOYLE. The Cephalopods.
G. VON KOCH. The Deep-Sea Corals.
C. A. KOFOID. Solenogaster.
R. VON LENDENFELD. The Phospho-
rescent Organs of Fishes.

H. LUDWIG. The Starfishes.
J. P. McMURRlCH. The Actinarians.
E. L. MARK. Branchiocerianthus.
JOHN MURRAY. The Bottom Specimens.
P. SCHIEMENZ. The Pteropods and Hete-

ropods.
THEO. STUDER. The Alcyonarians.
M. P. A. TRAUTSTEDT. The Salpidas and

Doliolidse.
E. P. VAN DUZKE The Halobatidae.
H. B. WARD. The Sipunculids.
H. V. WILSON. The Sponges.

W. MoM. WOODWORTH.

The Nemerteans.
The Annelids.

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletin Vols. I. to XXXVII. ;
of the Memoirs, Vols. I. to XXIV.

Vols. XXXVIII., XXXIX., and XL. of the Bulletin, and Vols.
XXV., XXVI., and XXVII. of the Memoirs, are now in course
of publication.

The Bulletin and Memoirs are devoted to the publication of
original work Iry the Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural History, and of work by specialists based
upon the Museum Collections and Explorations.

The following publications are in preparation : —

Reports on the Results of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander C. 1). Sigshee, U. S. N., and Commander J. R. Bartlett, U. S. N.,
Commanding.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission
Steamer " Albatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in
charge of Alexander Agassiz, on the U. S. Fish Commission Steamer
" Albatross," from August, 1899, to March, 1900, Commander Jefferson F.
Moser, U. S. N., Commanding.

Contributions from the Zoological Laboratory, Professor E. L. IMark, Director.

Contributions from the Geological Laboratory, in charge of Professor N. S.
Shaler.

These publications are issued in numbers at irregular inter-
vals ; one volume of the Bulletin (8vo) and half a volume of the
Memoirs (4 to) usually appear annually. Each number of the
Bulletin and of the Memoirs is sold separately. A price list
of the publications of the Museum Avill be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 7.

RIVER TERRACES IN NEW ENGLAND.

By W. M. Davis.

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

PRINTED FOR THE MUSEUM.

October, 1902.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 7.

RIVER TERRACES IX NEW ENGLAND.

By W. M. Davis.

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

PRINTED FOR THE MUSEUM.

October, 1902.

OCT 31 19P2

No. 7. — River Terraces in New England. By W. M. Davis.

TABLE OF CONTENTS.

I. General Statement ....
Theories of River Terraces

II. Preliminary Inquiry . . .

Various Kinds of Terraces
Terrace Patterns ....
Terraces carved by Streams

of Diminishing Volume .
Terraces carved by Streams

of Increasing Slope . .
Terraces carved by Streams

of Diminishing Load . .
Preservation of Terraces

by Rock Ledges . . .
Origin of Terraces in New

England

III. The Theory of River Ter-

races

Plan of Statement . . .
Behavior of a Wandering

River

Terminology of Wandering

Rivers

Ideal Terrace Patterns

Early Stage ....

Ideal Terrace Patterns

' Middle Stage . . .

Ideal Terrace Patterns

Late Stage ....
Defended Terrace Cusps

Early Stage

Slipping Meanders and

Blunt Cusps

Compressed Meanders and

Sharp Cusps . . . .
Terrace Fronts near De-
fended Cusps . . . .
VOL. xxxviii. — no. 7

PAGE

282
282
284
2?4

285

IV.

288

290

293

294

296

297
297

298

302

303

3)6

309

311

313

314

315| V.
1

Defended Cusps : Later

Stage 316

Diminished Swinging of the

Meander Belt .... 318
Distribution of High-scarp

and Low-scarp Terraces 319
Effect of Rock Barriers . . 320
Relation of Terrace Pat-
terns on the two Sides of

a Valley 322

Ratio of Sweeping, Swing-
ing, and Degrading . . 322
Relation of the Preceding
Deductions to the Obser-
vations described in the
Following Sections . . 324
Observations of River Ter-
races in New England . 327
Valley of the Westfield

River, Mass 327

Eastern Section . . . 327
Western Section . . 331
Little River .... 333
Valley of Saxton's River, Vt. 334
Western Section • . . 334
Eastern Section . . . 337
The Connecticut below

Bellows Falls, Vt. . . 338
The Connecticut below

Turners Falls, Mass. . 342
The Connecticut near

Springfield, Mass. . . 343
The Merrimac between
Concord and Manches-
ter, N. H 344

Bibliography 346

28!

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

I. General Statement.

Theories of Biver Terraces. The terraces carved by streams in the
washed drift of our valleys have been frequently studied and described
since the beginnings of geological investigation in New England. In
nearly all cases more attention has been given to terrace pattern as
seen in vertical cross-section than as presented ir. horizontal plan. The

Fig. 1.

cross-section is usually repi'esented as in Figure 1, in which the
depth of the rock-floored valley is made greater than that of the new
valley carved in the drift filling. A notable feature of such terrace
sections is that the open space measiu*ed across the valley between the
scarps of the low-level terraces is narrower than that between the scarps
of the high-level terraces ; and this fact has frequently given rise to
the supposition that the volume of our streams to-day is less than that

Fig. 2.

of the streams by which the high-level terraces were carved. It will
however be shown from what follows that the characteristic cross-section
of a terraced valley in which the river has not yet reached rock bottom
exhibits few stepping terraces, and is fairly represented by Figure 2 ;
while if stepping terraces are present a characteristic section, on which
the most significant points for a mile or more up and down the valley
are projected, would show that the base of many of the terraces is
determined for short distances (ten to fifty feet) by a rock ledge, as in

DAVIS : RIVER TERRACES IN NEW ENGLAND.

283

Figure 3, or better in Figure 4. This factor in the development of
terraces was first recognized, as far as my readiug has gone, by Hugh
Miller (the younger), whose view will be presented in abstract on a
later page.

Fig. 3.

Although the controlling ledges occupy a very small fraction of the
terrace length, they are of dominant importance ; and there can be little
doubt that the finest flights of stepping terraces in New England are to

Fig. 4.

be thus explained. Terraces of the kind shown in Figure 5 are different
from those here studied, as will be more fully stated in the next
section.

Fig. 6.

When the terrace pattern is considered in plan as well as in cross-
section, it appears that our terraces may be accounted for, first, by the

284 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

behavior of a meandering and swinging stream, slowly degrading a pre-
viously aggraded valley without change in volume ; and, second, by the
control exerted here and there over the lateral swinging of the stream
through the discovery of rock ledges, as suggested by Miller. The
following pages are devoted to a fuller consideration of this conclusion.

II. Preliminary Inquiry.

Various Kinds of Terraces. For the sake of clearness it is desirable
to exclude at the outset all kinds of terraces other than those here
studied. The terraces that occupy so many of our valleys are known as
river terraces, drift terraces, or alluvial terraces. They have as to
origin nothing in common with the terraces of sea-shores, such as occur
on the coastal slopes of Cuba ; or with the lake-shore terraces so well
developed in the basins of Bonneville and Lahontan. The}T bear little
resemblance to structural rock benches, such as break the slopes of
valley sides in dissected plateaus, as in West Virginia or on a still
larger scale in the Colorado canyon. They have little likeness to the
silt and gravel-covered rock terraces formed when a graded river,
revived by uplift, cuts a new valley in its former valley floor, as along
the gorge of the Rhine on its way through the Schiefergebirge of
western Germany.

Our New England drift terraces have a flat and nearly level upper
surface or plain, limited backwards by rising ground and forwards by
falling ground, and to that extent they resemble the terraces of all the
classes above mentioned ; but they have certain well-marked features of
their own. They are evidently the river-carved remnants of a body
of stratified clays, sands, or gravels that once occupied in larger volume
than to-day the rock-floored valleys of still earlier origin. Their upper
surface, the terrace plain or floor, slopes with the fall of the stream by
which their scarped face or front has been eroded; and in this they'
differ from sea and lake shore terraces and from structural rock benches,
none of which have any particular relation to the slope of neighboring
streams. They consist of unconsolidated, stratified drift ; if a ledge
appears in any part of a drift terrace it is manifestly an accidental
element, although as will be shown it may exert a controlling influence
on the pattern of the terrace front ; in this our drift terraces differ from
the structural rock benches of valley sides in dissected plateaus, and
from the rock terraces that represent the former valley floors of
revived rivers, both of which consist essentially of rock, even though the

DAVIS: RIVER TERRACES IN NEW ENGLAND. 285

latter may bear a veneer of river drift on their surface, as in Figure 5.
Moreover, drift terraces are in nearly all cases developed with much
more irregularity of pattern than is the case with the terraces of other
kinds. A single drift terrace — unless it be the highest one of a series
— is seldom traceable many miles along a valley side ; its length may
be only a few hundred yards. Terraces of other kinds are usually much
more persistent.

Our drift terraces differ, furthermore, from other terraces in the place
that they occupy in the geographical cycle. They are not products of
normal erosion during an undisturbed still-stand of a land mass, but are
the consequence of some relatively short-lived episode during which a
greater or less departure is made from the normal progress of a cycle.
The terraces of New England occupy well-opened rock-floored valleys of
earlier origin, and thus imply the previous attainment of maturity in the
cycle which witnessed the development of our hills and valleys. The
glacial period witnessed certain modifications of the preglacial valleys
and closed with the accumulation of abundant drift in them, as well as
with certain changes of level by which the rivers were prompted to wash
the valley drift away. Postglacial time has allowed the rivers to enter
well upon this task : yet, even when the task has been completed, the
normal cycle of erosion in New England will not have advanced far
beyond its preglacial phase ; so brief are the glacial and terrace episodes
compared to the time required for baseleveling a region of resistant
rocks.

Systematically considered, river terraces may be best associated with
the forms assumed by the waste of the land on the way to the sea.
Flood plains and alluvial fans are representative examples of the form
assumed by land waste while it is stopping on its way down a valley.
Terraces are examples of the forms assumed by waste that still remains in
its stopping-place after part of its volume has been swept forward again.

Terrace Patterns. Before entering upon the consideration of the
process of terracing it will be well to examine briefly the more character-
istic elements of terrace pattern, especially as seen in horizontal plan.
The plain or floor of a drift terrace frequently presents a rapid variation
in width, usually terminating in points at its up and down stream ends,
as in Figure 6. The borders are prevailingly formed of curves of greater
or less length, but of tolerably uniform radius, concave to the stream
and frequently uniting in cusps. When several cusps are grouped, one
back of the other, so as to form a strong salient, they may be called a
terrace spur. Convex borders fronting the stream occur but rarely.

236

bulletin: museum of comparative zoology.

The highest plain of a flight of terraces backs against the ascending
slopes of the older valley side and accepts their outline as its border, as
in Figure 6 ; while each lower terrace, as well as the existing flood plain
— the " intervale " or " interval " of New Englanders — backs against
the scarp of the next higher terrace ; thus the intermediate members of
a flight of terrace steps possess similar but not necessarily parallel out-
lines, front and back ; the cusps between the curves all point towards
the stream. The back border of a terrace i3 frequently followed by a
marshy channel from which the terracing stream has been withdrawn
by a short-cut or cut-off (as is more fully considered below) before the

Fig. 6.

channel was filled ; terrace plains thus characterized may slope gently
away from the axis of the valley towards their back border, and if they
are of moderate breadth the backward slope may be a relatively con-
spicuous feature, as in the lower terrace in the middle of Figure 6.
Terrace of this kind were called "glacis terraces" by Hitchcock (58). *
They are of very common occurrence, and serve to show that the sudden
withdrawal of the terracing stream from a roundabout channel to a more
direct course has not been unusual.

The scarp of a terrace connects the front border of the plain above
with the back border of the plain below. Its sloping surface therefore

1 Numbers in parentheses after an author's name are page references to his
writings, cited in the Bibliography.

DAVIS: RIVER TERRACES IN NEW ENGLAND. 287

presents a succession of curved re-entrants, separated by salients of
greater or less acuteness. It is well known that the curved terrace
fronts have been carved by the successive encroachments of a curved
stream which once swung against their base, and that the stream has
swung laterally at least as many times as there are terraces ; but the
behavior of the swinging stream has seldom been traced in detail.

Although the plain and the descending scarp at its front are usually
taken together as bounding a terrace, these two surfaces are not ge-
netically connected in river terraces as they are in constructional lake-
shore or delta terraces. River terraces being of destructional origin, it
is the ascending scarp at the back of a terrace that should be asso-
ciated with the plain beneath and in front of it. The line along the
re-entrant edge between the plain and the ascending slope at its back is
the most significant of all terrace lines. The front line of a terrace
plain is of less significance, for it is determined merely by the slipping of
the sands and clays down to the line of the undercut scarp at the back
of the next lower terrace ; the front line of a terrace plaiu is therefore
of value only in so far as it represents the back line of the next terrace
beneath.

Terrace scarps are steepest where the cutting stream has most re-
cently swung against their base. In a series of stepping terraces, the
youngest and steepest scarps are at the bottom of the flight ; but when
all the terraces of intermediate levels are destroyed by a chance lateral
swing of the stream so that it undercuts even the highest terrace plain,
then the whole descent from highest to lowest level may be fresh-cut
with sharp edges at top and bottom. In older terraces, the scarps
weather to a gentler slope, and the edges are rounded off. A convex
slope of erosion is thus formed above and a concave slope of deposition
below. The older the terrace, the greater the part of its front is occupied
by rounded slopes and the gentler is the slope of the shortened tangent
between them. At the same time, the salients or cusps between the re-
entrants of the scarps as seen in plan lose their original sharpness of
definition and become blunt and dulled. There has been no attempt to
show details of this kind in the accompanying diagrams.

Gulches are often worn in terrace fronts by wet-weather streams, and
fans are spread on the terrace plain below. The abandoned stream
channels at the back border of a plain are usually taken as guides for
surface drainage, whose gathered waters dissect the plain where it is
cut off by the next lower terrace. A rather systematic drainage pat-
tern is thus developed, as in Figure 6.

288 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

The several theories by which terraces have been explained may
now be reviewed.

Terraces carved by Streams of Diminishing Volume. The primi-
tive explanation of terraces is that the whole space between the upper
terrace scarps represents the channel of a huge river by which the valley
was once drained, and that successive diminutions of volume to that of
the present river are indicated by the decrease of breadth between the
terraces in descending order. Although this view has sometimes re-
ceived distinguished advocacy, it has never gained general acceptance
among geologists or geographers. It has, however, been very generally
supposed that the present rivers are much smaller than they were
when they began the work of terracing ; hence it is desirable to con-
sider the special features that should appear if a decrease in stream vol-
ume had actually taken place.

The best indication of the volume of the stream by which a terrace
has been carved is afforded by the curvature of its frontal scarp. If the
scarps of the low-level terraces have a radius and an arc of curvature
similar to these elements in the existing river meanders, and signifi-
cantly smaller than in the high-level scarps, while curves at interme-
diate levels show intermediate values, a diminution of stream volume
may be fairly inferred. If the radius and arc of curvature are of about
the same measure in the three cases, no change in stream volume is
indicated.

On the other hand, if a stream were charged with abundant and
coarse load in the last stages of its aggrading action, as seems to have
been frequently the case in New England, its slope must have been
relatively strong ; and a graded river with a heavy load on a strong
slope does not develop curves of as small radius as it would when sub-
sequently flowiug with the same volume but with a finer load on a gen-
tler slope ; hence a large radius of curvature in the uppermost terraces
should not alone be taken as an indication of large volume ; large arc of
curvature should also be found before large volume is inferred. It is
for this reason that some of the uppermost terraces of the Connecticut
and the Westfield rivers, whose scarps seem to sweep in curves of
greater radius than do those of the low-level terraces, cannot alone give
assurance of a former greater volume for their rivers.

It is, however, rendered very probable by what is known of the later
stages of the glacial period that many of our streams had greater volume
then than now. The most effective cause for greater volume was the
constraint of the ice sheet, whereby the drainage from the basins of

DAVIS: RIVER TERRACES IN NEW ENGLAND. 289

north-flowing rivers was turned over the divides into the valleys of
south-flowing rivers. This may have heen the case while the ice still
covered the northern basins, their waters (as far as they had any) then
running as subglacial streams, which may have been forced to ascend
slopes and cross divides. Effective constraint may also have been pro-
vided after the ice had at least in part withdrawn from the northern
basins, but when it remained in sufficient force to obstruct their normal
outlets, thus forming lakes whose overflow ran across a pass in the
divide to some southern valley. Another cause for increased volume of
our south-flowing rivers was the importation into their basins of a con-
siderable snowfall that was received on the ice sheet over some northern
basin. A fourth cause for increased volume of our rivers lies in a pos-
sibly greater precipitation during the later stages of the glacial period
than at present. A fifth cause lies in a relatively rapid meltiug of the
retreating ice sheet. It is eminently possible that these various causes
may have contributed effectively to an increase in river volume while
the New England valleys were aggrading with drift ; but it does not
follow that volumes decidedly larger than those of to-day were continued
into the period of terracing.

Except where direct evidence is given by curvature and arc of high-
level terrace scarps, a formerly greater volume of the terracing streams
should be regarded only as a possible, not as an actual occurrence. It is
especially desirable that large bulk and coarse texture of terrace deposits
should not be too readily accepted as evidence of former greater volume
of streams ; for bulk of deposits is a function of time as well as of
rate of action, and texture is a function of slope as well as of
stream volume and velocity. Hence until time and slope are shown to
have been insufficient to account for bulk and texture of deposits, it is
not compulsory to account for them by greater stream volume.

Even if decrease of volume has been of general occurrence during the
period of terracing, it has nevertheless not been in control of terrace
development ; for if it had been, stepping terraces should be much more
abundant than they are to-day. As a matter of fact, the diagrams by
which terraced valleys are ordinarily represented give an exaggerated
idea of the prevalence and perfection of these graceful forms. It is
rare to find a long flight of stepping terraces on both sides of a valley ;
it is rare to find a flight of terraces continued for any long distance
along a valley side ; when more than three or four low steps are to be
counted, it is usually only for a moderate fraction of a mile that they
persist. A large part of the length of our terraced valleys is bordered

290 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

by a few terraces of strong scarps, or by a high terrace with one or two
lower ones beneath it ; and it is not uncommon to find at least one side
of a valley enclosed by a single scarp in which the whole descent is made
at once from the highest terrace plain to the lowest. If terracing had been
due to a general decrease in the volume of our rivers, stepping terraces
should be much more prevalent, and broad flood plains between the high
scarps of a single terrace on each side of the valley should be much more
rare than they are ; and when the whole descent from high terrace to flood
plain is made in a single scarp on one side of the valley, stepping ter-
races with broad treads should be well developed on the opposite side ;
but no such arrangement of terrace form can be said to prevail. De-
crease of river volume must therefore be at most a subordinate cause of
terracing, if, indeed, it is not as a rule a negligible factor in their
production.

This conclusion seems to have been clearly in the mind of Adams,
state geologist of Vermont, who in 1846 wrote as follows: "The first
stage in the process in which the terraces originated, the deposition of
the materials, we have before referred to the older pleistocene. The
process of denudation must have next followed, when the rivers, cutting
down their channels through the drift barriers, lowered them gradually
above the barriers. Flowing through the level deposits of sand, they
must have formed serpentine channels, as rivers do now in alluvial
plains ; consequently by increasing the convexity of the bends, and then
cutting them off or wearing away their headlands and shifting their
beds, they would be meanwhile removing the greater part of the ma-
terials thus disturbed. By this process the greater portion of the orig-
inal plain must have been carried off, and it is not necessary to suppose
that the distance between opposite terraces is any indication of greater
magnitude of the river, but only of its shifting its channel" (1-45,
146).

Terraces carved by Streams of Increasing Slope. When the
basin of an aggrading river system is slightly tilted it may be expected
that those streams whose slopes are decreasing will aggrade their valleys
more rapidly than before (unless their point of junction with a degrading
stream may be lowered more than their headwaters are depressed by
tilting) ; while those whose slopes are increasing will change their action
from aggrading to degrading. It is well known that New England has
suffered a differential elevation in postglacial time. The postglacial
clays of Lake Champlain and of southern Maine were deposited when the
sea stood three hundred feet or more above its present level. The clays

DAVIS: RIVER TERRACES IN NEW ENGLAND. 291

of the Connecticut valley in Massachusetts were, according to Emerson,
deposited in lakes or bodies of slack water at or very close to the sea-
level of their time, but the clays now reach elevations approaching two
hundred feet. No postglacial changes of level of such amounts are
known to have taken place along the southern New England coast. Our
south-flowing rivers have therefore been accelerated, while those flowing
northward have been retarded ; and to this differential tilting Shaler
has ascribed the weak terracing by streams of the latter class in contrast
to the active terracing by those of the former.

While it is thus made very probable that the erosion of valley drift
was determined by the unequal elevation of New England in postglacial
time, it does not follow that individual terraces are in any close way
related to this movement. Several cases must be here distinguished.

The northern uplift may have been accomplished in a single movement
and so rapidly as to have revived the streams to an unusual activity of
erosion, whereby they deepened their valleys quickly for a time, and did
not begin to swing laterally, in the manner essential to terracing, until
they had developed new grades of gentle declivity after the rapid uplift
had ceased. In this case only a single high-level terrace and no inter-
mediate terraces would be formed, and there would be but few low-level
terraces.

A second supposition includes cases of repeated rapid uplifts separated
by deliberate pauses, each of which would produce a result similar to
that of the previous case. Here we should expect the river to have
swung laterally at as many different levels as there had been pauses
during the total uplift ; and the flood plain formed during each pause
would be of relatively persistent occurrence down the valley. But in
order to protect the terrace remnants of the successive flood plains from
being consumed by the river when it swings from side to side at lower
levels, it is necessary to postulate that the movements of uplift should
succeed each other at shorter and shorter intervals, so that the later-carved
flood plains should be narrower than the earlier ones. The chief objection
to this supposition is not so well directed against the postulate just men-
tioned as against the requirement of correlated levels in the terrace on
the two sides of a valley. Such correlation is occasionally found, but it
is by no means characteristic of our terraced valleys in general. The
terrace levels are usually so discordant on the opposite sides of a valley
that they cannot be considered the records of still-stands of the land
between times of rapid uplift.

A third supposition considers an uplift so slow that the south-flowing

292 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

rivers were never much accelerated ; for during slow uplift the larger
rivers might continue to swing actively from side to side, while all the
time degrading the valley floor. In this case terraces might be cut at
many different levels on opposite sides of the valley, according to the
habit of the river in its lateral swinging.

The third supposition seems most appropriate to Xew England, for all
of our valleys in which terraces are well developed exhibit flood-plain
remnants at many levels, high and low, and neither so few in number
nor so accordant in relative altitude above the river as to imply that
lateral swinging had occurred only during the quiet intervals between
rapid uplifts. But it should be noted that this conclusion applies better
to the valleys of good-sized streams or rivers than to those of small
brooks ; for the latter frequently show only faint terraces or no terraces
at all, even though they are branches of rivers whose valleys are well
terraced. This seems to mean that an uplift which was so slow that a
good-sized river could easily keep pace with it by down-cutting, may
have been too fast for such a result in the case of a small stream. While
the able-bodied rivers may thus have been always effectively at grade,
leisurely swinging from side to side and at the same time slowly wearing
down their valley floors, the small streams may have been for much or
all of this time above grade, and therefore unable to widen their little
valleys, although actively engaged in deepening them. On the other
hand, even the largest rivers have not been able to maintain a graded
channel in the rock ledges upon which they have been here and there
superposed by the drift cover. They are still actively cutting down such
ledges, but they are not yet able to widen the rock-notch that they are
cutting ; thus imitating the condition of their smallest branches, which
have not yet been able to widen their little valleys even in clays and
sands. Boulder clay or till is of a resistance between the feebleness of
stratified drift and the strength of rock ledges. If a mass of till is
discovered the stream may be successful in cutting down its channel to
grade, and yet unsuccessful in opening a valley floor ; and thus a boul-
der-clay " shut in " may be produced between open valley floors or
" intervals " that have been eroded in weak stratified drift farther up
and down stream. Little river, a mile southwest of Westfield, Mass.,
offers examples of this kind (page 333).

The small changes made in rock ledges during the development of an
extended series of river terraces serves to indicate how short is the
duration of the episode in which the alluvial filling of a valley is terraced,
in comparison with the time needed for the erosion of the rock-bound

DAVIS: RIVER TERRACES IX NEW ENGLAND. 293

valley itself, or still more with a whole cycle of erosion, in which a
mountain mass is reduced to a plain of degradation.

While slow uplift is thus seen to be consistent with the production
of many terraces, it is not consistent with their preservation, for
it does not explain the diminution of the interscarp space from the
higher to the lower levels. Indeed, the present rivers might tend to
develop broader flood plains by strong lateral swinging at the faint
grades now assumed than they had developed at the stronger grades
during the earlier stages of possibly more active uplift and heavier
load ; and the broad low flood plains would necessitate the undercutting
of all or nearly all the earlier high-level terraces by the pi'esent stream,
and the concentration of nearly all the separate scarps in a single
high-terrace front, as in Figure 2. Examples in which this condition
has been actually attained are to be found in the valleys of various
rivers, as will be more fully set forth on pages 328, 342, and 344.
Single high-scarped terraces are indeed so common as to warrant the
conclusion that high-level and intermediate terraces would nearly
always be destroyed by the swinging of the river at a lower level,
but for the occurrence of some special conditions by which they are
preserved.

Terraces carved by Streams of Diminishing Load. A graded
river may be caused to degrade as well by diminishing its load as by
increasing its slope, volume remaining constant. A diminution of
load since the stage of glacial retreat is highly probable, for not only
the streams that issued from the ice sheet but those also which
washed the freshly-exposed drift-covered land surface were in all
probability highly charged with detritus in late glacial and early
postglacial time. Indeed, increase of load may have been almost as
potent a cause of filling the valleys with washed drift as was the
depressed attitude of the land in the north and the consequent en-
feebled slope of the south-flowing rivers. As the ice disappeared and
as the land surface was more or less covered wTith vegetation, the
load of the streams should have been lessened, and they must there-
upon have set to work to degrade the valleys that they had just
before been aggrading, even if no change of slope had taken place.
This process, if working alone, must have been very gradual, and
might therefore have allowed plenty of time for lateral swinging and
terrace carving. But, as before, no explanation is here found for the
production of stepping terraces. On the contrary, when the diminu-
tion of load was further advanced the rivers would degrade their valley

294 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

floors more arid more slowly, aud the tendency would then be to destroy
all the earlier terraces by broadening the flood plain to a maximum.

That the rate of degradation by our rivers was really slow is
proved by the flights of stepping terraces here and there in different
valleys ; and that the normal tendency of the larger rivers is to de-
stroy nearly all the earlier-made terraces by opening broad flood plains
at low levels is proved by the frequent occurrence of high scarps
descending from the highest terrace plain nearly or quite to the
lowest. Hence it is for the preservation of high-level and intermedi-
ate terraces rather than for their production that a more efficient
cause than any yet discussed is to be sought.

Preservation of Terraces by Rock Ledges. What is more natural
than that a river, swinging from side to side as it slowly degrades its
valley floor, shall here and there be restrained on coming against a ledge
projecting from the sloping valley wall ; and that the deeper the valley
is excavated the less breadth of free swinging can remain ! This idea
was first given explicit statement in Miller's paper on "River Terracing;
its Methods and Their Results," as illustrated by observations in Scot-
land. After a review of earlier writings this author says :

" The modern rivers . . . have struck rock at very variable depths.
In hundreds of cases, after winding freely about, encountering only soft
clays and the like, and constructing terraces of various kinds, they have
here and there become rock-bound, and prevented from pursuing their
work of terrace-building after their former manner, as well as from de-
stroying the terraces they had already made " (298). "When . . . the
rivers commenced to work upon shallow, wide-bottomed valleys, soft and
yielding in their nature, except where crossed by bars of rock . . . , they
proceeded to plane far and wide, travelling from breadth to breadth to
an extent uever now equalled. With banks nowadays eight or ten
times as high, and rock-bound at perhaps ten times as many points, it
is no wonder that the modern rivers should seem to have ' run in ' '
(299, 300).

Rock ledges, however, are not here given the importance they deserve.
The reader will not surely gain from Miller's article a full measure of the
value of ledges in determining the pattern of terraces, and of stepping
terraces in particular. Hence a more detailed statement of the relation
of ledges to terrace pattern and to terrace development seems desirable,
especially with reference to the valleys of New England, where this ex-
planation of terraces has not previously been applied.

It should be further noted that certain postulates of Miller's essay do

DAVIS: RIVER TERRACES IN NEW ENGLAND. 295

not command entire assent. He states that " it is not allowable to have
recourse to coast elevation, or climatic changes, or periodicity of any kind,
without first proving that the terraces range in opposite pairs " (30-t,
305). This seems to be an unnecessary limitation of possibilities ; for,
us is here explained on page 305, a river that is impelled to gradual '
degradation by a slow rising or tilting of the hind may produce unpaired
terraces as it wanders to and fro across its valley floor. On another
page Miller concludes that rivers " cannot but concentrate their channels
as they excavate them, unless the amount of planation is out of all
proportion to the rate of deepening " (300), and seems to imply in this
statement that a large ratio of lateral erosion to degradation, such as is
here assumed for our New England rivers, and further consid-
ered in later sections, is an improbable ratio. To this it may be an-
swered that the occurrence of stepping terraces at one and another point
in our larger valleys certainly justifies the assumed ratio by showing
that lateral swinging should be measured in hundreds or thousands of
feet at many successive stages of degradation, while the total degradation
is usually to be measured in tens of feet and seldom exceeds one or two
hundred feet. A possible reason for the diffei'ence of values given to
this ratio may be that Miller's studies were directed to the moderate-sized
rivers of Scotland, while the best terraced valleys in New England are
those of large rivers like the Connecticut and the Merrimac and
their stronger branches ; and, as has been already pointed out, a large
river may swing actively during an uplift that gives a small river no
time for anything but down-cutting. These two items are, however, of
secondary importance in Miller's theory compared to rock ledges.

In reviewing various other essays of earlier dates several suggestive
passages have been found, hinting at the importance of rock ledges.
Adams makes the following statement : " If a terrace has been formed
before the complete removal of the obstructions in the channel [the con-
text shows that these obstructions are ' solid rock '], the same process
must have been repeated within the new and narrower level of interval.
We should thus have a second terrace. Repetitions of the process in
cases where the obstructions were not entirely removed would occasion
agreater number of terraces " (146). Something more explicit is found in
Edward Hitchcock's "Surface Geology." In describing a middle section
of the Connecticut valley, where the terraces became famous from the
writings of this author, it is said that " the rock often projects through
the terraces" (18), but the service of the rock in protecting the over-
lying terrace from being cut back is not announced. Farther on a

VOL. XXXVIII. — NO. 7. 2

296 bulletin: museum of comparative zoology.

description is given of the basin of the Westfield river, where the effect
of ledges in determining the number and pattern of the terraces is very
striking (see page 331) ; here it is briefly stated that "the materials of
which all these terraces are formed are clay, sand, and gravel, though
the red sandstone shows itself occasionally near the river " (20). The
secret is told in an account of the terraces of the Deerfield : " The river
would encroach still further upon this hill, had it not struck a ledge of
red sandstone, which will at least retard its lateral erosion" (19) ; and
again, " the reason why those [terraces] on Pine hill remain, I find to
be that they rest on a protuberant mass of red sandstone. On the west
side of the hill ... is an ancient bed of Deerfield river . . . which was
prevented from making any further lateral encroachments by the under-
lying rock" (20). Yet in spite of the understanding thus shown of the
importance of ledges in these particular instances, the generality of the
relation of ledges and terraces is not brought forward ; and the above
instances of the local restraint exercised by ledges have never been
quoted, so far as I can find, by any of the many readers of Hitchcock's
well-known essay.

A fuller recognition of the part played by defending ledges is to be
found in Emerson's " Geology of Old Hampshire County, Massachusetts."
It is here said that the Connecticut river in the neighborhood of Holyoke
"has now cut its bed deep in the sandstones and is thus prevented from
oscillating" (730). A little farther down the valley "the river early
became entangled in rock and has cut only vertically " (733). In the
northern part of the state " the river everywhere cut down rapidly to
rock and has not swung widely to east and west, but has been condemned
from the beginning to rock-cutting" (733). "Across Chicopee there is
a fine, low terrace bounded on the east by a high scarp of the high ter-
race, which everywhere shows till in great force beneath the sands of
the old lake " (730 ; see also 627, G32). It should be noted, however,
that these passages are chiefly concerned with the occurrence of trenches,
floored with rock and lined on both banks with ledges. The part played
by an isolated ledge of rock or by a bank of till in preventing the
further swinging of the river and thus defending the terraces above it is
not brought forward.

Origin of Terraces in New England. Three conclusions may
now be stated in order that the reader may have in mind the end to
which the preceding and following pages lead. First, a diminution of
stream volume may have taken place during the terracing of our New
England valleys, but it has not been essential to the production of the

DAVIS : RIVER TERRACES IN NEW ENGLAND. 297

observed terraces. Second, the terracing rivers have slowly degraded
their aggraded valleys while actively swinging from side to side : deg-
radation probably being the result of the combined action of a slow
northern uplift and a gradual decrease of load, and in spite of a probable
decrease of volume. Third, the chance discovery of rock ledges by the
swinging river is the chief cause of the systematic diminution of inter-
scarp space and of the preservation of terraces, as seen in a typical
section of stepping terraces.

A full statement of the process of terracing therefore involves a con-
sideration, first, of the behavior of a free-swinging, slowly degrading
river; and, second, of the constraint that may be imposed on such a
river by the accidental encounter with previously buried ledges at
various points in its course and at various stages in its history. This
anticipatory statement may aid in the understanding of the following
pages.

III. The Theory of River Terraces.

Plan of Statement. The previous paragraphs have given a gen-
eral consideration of several theories of river terraces, with the result of
deciding that one of them offers a much better explanation of observed
facts than the others. It is now proposed to examine the successful
theory with more care, first, by making a somewhat detailed study of
such processes of river action as are involved in the theory ; second, by
deducing with some minuteness the various patterns of terraces that can
be formed by these river processes. It will then be possible to make in
Part IV. a thorough test of the verity of the theory by confronting its
deduced consequences with the facts determined by observation.

It should be understood that the deductive character of the succeed-
ing paragraphs is more apparent than real. Many features of river
work here presented as deductions were discovered by observation. It
is true that an expectation of certain occurrences had been aroused by the
deductive consideration of certain processes, but there has been so con-
tinual an interweaving of observation and theory during the growth of
these pages that it is now rather difficult to determine the order in
which the various items here recorded came to mind. It is therefore
chiefly for the sake of a continuous presentation of the theory of river
terraces that a largely deductive treatment is here adopted. When
the whole theory has been apprehended, it is relatively easy to test its
verity by observations pertinent to its different parts.

298 bulletin: museum of comparative zoology.

The other theories of terracing, regarded as unsuccessful in the pre-
liminary inquiry of Part II., should be more thoroughly considered before
they are discarded ; but, inasmuch as the further they are examined
the less competent they seem to explain the facts observable in the ter-
raced valleys of New England, it does not seem worth while to give
them any more explicit consideration here.

Behavior of a Wandering River. The diagrams introduced in
the following sections represent several successive stages in the process
of slow degradation by a wandering river. The postulates as to river
behavior on which the diagrams are based are: (1) The degrading
stream continually maintains an essentially graded condition. (2)
The lateral swinging of the meandering channel is very much faster
(a hundred-fold, for example) than the degradation of the valley floor.
(3) The breadth over which a free river (not constrained by ledges)
tends to swing laterally is greater than the breadth of the meander belt
(the belt included by tangents to the meandering channel). (4) An
individual meander tends to enlarge its radius and to work its way
down the valley until it may be abandoned at season of high water for a
short cut across a flood-plain lobe, or at any season (but usually at high
water) for a cut-off through the narrowing neck of a lobe. It is believed
that abundant justification may be found for all these postulates, either
in the observed behavior of a graded river, or in the success with which
they lead to an understanding of the peculiar patterns of our terraces.
The several postulates may now be reviewed.

(1) If the change in the ratio of load to carrying power (volume and
slope) proceed very slowly, a river may remain in an essentially graded
condition all through the process of aggrading or of degrading its valley,
and through the change from aggrading to degrading. It is true that
the graded condition depends on a balance between load and carrying
power, and it would at first sight appear that any change in either
quantity would destroy the balance and throw the river out of grade. But
if the change is only by a quantity of the second order — that is, if
either load, volume, or slope is changed only by a differential of its
value in a unit of time — adjustment to the new condition will follow
so immediately that no failure of adjustment will be noticeable. A
similar maintenance of an essentially graded condition obtains in the
degradation of graded (waste-covered) hillsides as they pass from ma-
turity to old age. It is, however, not likely that the very slow degrada-
tion of a valley floor which accompanies the advance from maturity to
old age in a normal cycle of rock erosion will result in terraces, because

DAVIS : RIVER TERRACES IN NEW ENGLAND. 299

the general processes of weathering may lower the flood plain about as
fast as the river sinks to a fainter and fainter slope. The rate of degra-
dation of a terracing river in a drift-filled valley, resulting from a
climatic change or from a land movement, may therefore be allowed a
decidedly higher value than that of an ageing river in a normal, undis-
turbed cycle.

(2) It is well understood that a graded stream may work actively in
wearing or building its banks laterally, however slowly it aggrades or
degrades its valley floor.

(3) There are abundant examples of rivers whose lateral oscillation
or swinging carries them from side to side of a flood plain that is
much wider than their meander belt. The Mississippi is a noted case
of this kind. Its meander belt is six or eight miles wide in a flood
plain whose enclosing bluffs are from twenty to sixty miles apart. A
similar relation may be seen in many meadow flood plains, drained by
small brooks.

(4) The fourth postulate involves a principle of river action which
may be familiar to hydrographers, but which is nevertheless not com-
monly stated. Imagine a stream in a broad flood plain passing from a
straight stretch or tangent to a well-defined curve. On the tangent A B,
Figure 7, the thread of fastest current might, as far as local control is con-
cerned, lie along the middle of the channel, or indifferently on one side
or other of the middle line. On entering the curve the fastest current
is gradually shifted, B C, towards the outer bank of the channel, and
there flowing steadily all around the curve C D it determines the line of
greatest depth. On passing from the curve the thread of fastest cur-
rent is necessarily delivered to the next down-stream tangent, D E, on
the down-valley side of the channel, and only after flowing for a signifi-
cant distance will the fastest current gain a path near mid-channel. If
the curves or meanders are close set, so that one curve passes directly
into the next one with no intervening tangent, then the thread of
fastest current must, on passing the point of inflexion, enter the up-
stream end of the next curve near its inner or convex bank, and only
gradually be displaced towards the outer bank as inertia has time to
bring about its usual effect.

As a result of this systematic displacement of the fastest current
from the mid-channel line the bank that it approaches will be worn
away, while deposition will take place along the bank from which the
fast current is withdrawn. The stream will therefore tend to wear
away the bank on the outer side of its curves (but perhaps failing to

500

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

begin this action just at the beginning of the curves), and on the down-
valley side of short tangents. The curves will thus increase in radius
and arc and the meander belt will widen, while each meander will tend
to move slowly down the valley. The flood plain must be scoured out
for a certain stretch, N N1, around the concave banks and along the up-
valley side of every lobe; while a scroll of new flood plain, M M, is
added around the end and on the down-valley side of the lobe.

Fig. 7.

All this may be easily recognized in the meanders of a meadow brook ;
and that it occurs even in the Mississippi is abundantly illustrated in
the large-scale maps (three inches to a mile) published by the Missis-
sippi River Commission (see specially Charts 36, 38, 39), where the
lines of flood-plain growth pass around the end and along the down-
valley side of each lobe, while they are cut across by the encroaching
river on the up-valley side of the lobe, as indicated in Figure 7. Still

DAVIS: KIVER TERRACES IN NEW ENGLAND.

301

more definite proof of this feature in the behavior of a meandering river
is found in the new edition (1900) of the Preliminary Map of the Mis-
sissippi (one inch to a mile), in which a red overprint indicates the new
position of the channel, as determined some fifteen years after the
previous survey (see especially sheets 14, 16, and 18).

The same down-valley shifting of the meanders is seen in the enclosed
meanders of many rivers, typified in North branch of the Susquehanna,
Figure 8. This fine river has in-
cised its course beneath the up-
lands of northern Pennsylvania.
The upland spurs that enter the
river curves have been subjected
on their up-valley sides to a per-
sistent sweeping that is but little
less effective than that by which
the curved re-entrants between
the spurs have been scoured out.
The up-valley side of the spurs
have strong bluffs, as different
from their gentle down - valley
slopes as are the high lateral
bluffs that enclose the curves
from the gentle terminal slopes of
the spurs. It is noteworthy that in this case of a rock-walled valley, the
down-valley shifting of the curves does not seem to have been more
than fifteen or twenty times greater than the degrading of the river
channel in the latest period of valley trenching : while in the case of
terracing streams in drift-filled valleys, the first of these changes ex-
ceeds the second in a much higher degree. It may be further noted
that the North branch of the Susquehanna above Wilkes Barre seems
for some time past to have ceased deepening its valley, for narrow
double-curved scrolls of flood plain are now systematically added to the
outer end and the down-valley side of the spurs, as may be especially
well seen just above Tunkhannock, and as is indicated by the dotted
areas M M', opposite the under-cut bluffs, N N', Figure 8.

Other examples of meandering valleys, exhibiting the systematic lateral
growth and down-valley shifting of the meander curves, might be
instanced ; a few of them are mentioned in my paper on the Drainage of
Cuestas (89). It is my hope to give at another time a fuller account
of this phase of river development, and then to show how satisfactorily

Fig. 8.

302 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

the stage of development may be stated in terms of the flood-plain
pattern.

The four postulates above announced concerning river action may
therefore be taken as well supported.

A natural limit is set to the dimensions of a growing meander curve
on a flood plain by the formation of short-cuts across flood-plain lobes at
time of high water, or of cut-offs when the narrowing neck of a lobe is
finally worn through, a roundabout course being in both cases abandoned
for a more direct one. It may therefore be expected that the abandoned
channels, such as are preserved in ox-bow lakes on existing flood plains,
and in swampy half-filled channels at the back border of many terraces,
will on the average show a larger radius of curvature than the curves
of the existing river; and the maps of the Mississippi give some support
to this expectation. Emerson has pointed out (735) the tendency of our
New England rivers and streams to form loops or ox-bows on the right
of their general course, from which they return to a nearly direct course
by short-cuts or cuts-offs, only to begin again the work of right-handed
loop-cutting. He states that the Connecticut near Northampton, Mass.,
has seven deserted loops on the right (west) and none on the left ; some
of its tributaries have sharp bends and ox-bows thirty times as numerous
on the right as on the left of their course. It is naturally suggested
that this asymmetry is the result of the deflective force arising from the
earth's rotation.

Terminology of Wandering Rivers. The terms already introduced
regarding rivers that wander about their flood plains may now be sum-
marized and somewhat extended. The space enclosed between tangents
drawn outside of the curves or meanders of the stream is the
meander belt. This belt will widen while the meanders are normally
wearing their outer bank ; but on the occurrence of a short-cut across a
flood-plain lobe or of a cut-off through the narrowing neck of a lobe, the
belt will locally collapse. Here the river course becomes relatively
direct for a time, only to develop serpentines again as new meanders are
established. The progressive movement of the meanders down the
valley will be called sweeping. Up-stream and down-stream will be
used in their ordinary sense, but up-valley and down-valley will be
substituted when it is desired to indicate a more general direction than
that of the circuitous channel.

The lateral movement of the meander belt from one side of the valley
floor to the other will be referred to as swinging. It is not always
possible to distinguish between the true lateral swinging of the meander

DAVIS: EIVER TERRACES IN NEW ENGLAND. 303

belt as a whole and the more local shifting of an irregularly sweeping
meander. The compound movement of sweeping meanders in a swinging
meander belt will be called wandering, this term being fully justified
when it is noted that many unsystematic irregularities must be devel-
oped in a stream channel, whereby it will depart significantly from the
simple and regular movements here considered. The whole breadth of
the valley floor that may be worn down by the stream will be called the
belt of wandering; this corresponds to many of our flood plains or
" intervals."

In an ideal case, a regularly growing pattern of meander curves might
be imagined slowly sweeping down a valley, the meander belt collapsing
here and there, now and then, but growing again to its ordinary breadth
as new curves are developed in the place of the old ones. At any point
in the valley, an endless procession of meanders would sweep past.

If it be now supposed that the wandering stream is slowly degrading
its valley floor, each meander will sweep past a given point at a slightly
lower level than that of its predecessor ; and each time the meander
belt swings across the valley from one side to the other and back again,
it will return at a distinctly lower level than that at which it left. The
flood plains formed at different stages of this leisurely pi'ocess will differ
in altitude, and all of them will be inclined gently down the valley. It
is the remnants of these flood plains that form our terrace plains.

Ideal Terrace Patterns : Early Stage. Soon after the stage of
degradation has been definitely established and the meandering stream
begins to swing across the valley at a little lower level than before, a
condition represented in Figure 9 may be reached. In this figure, as in
a number that follow, it is supposed that the view is taken from a
considerable height, looking northwest across the valley of a south-flow-
ing river. The terrace plains are left blank in most of the diagrams.
The western meander in the foreground of Figure 9 is now scouring out
a curve in a low concave terrace scarp, B, the ninth of its kind within the
limits of the diagram. A small portion of a terrace, A, of slightly less
height, is shown in the immediate foreground ; it may represent the work
of the preceding westward meander, while the next following westward
meander is cutting out a deeper terrace, C, in the background. Terrace
A may be taken as one of the first marks made by the degrading stream.
Terrace B is of greater height than A, because A has been under-cut and
consumed in the production of B, except in the immediate foreground.
Terrace C is as yet independent of B, and therefore shows a height to be
measured only by the few inches or feet of depth to which one sweeping

304

bulletin: museum of comparative zoology.

meander cuts below the plain of its predecessor. The curves and cusps
of Terrace B result from a vacillation of the meander during its

Fig. 9.

down-valley progress ; a rather sharp cusp being left between B 2

and B 3, while the curves from B 3 to B 9 have arcs so small as to

join in a nearly straight terrace front.

It may be noted that the small curves
by which a nearly straight terrace front
is usually formed are as a rule to be ex-
pected only towards the side of a valley,
and less commonly on a terrace spur that
advances into the valley. For example,
in Figure 10 the little curves A, C, D, E,
record so many positions of the meander
apex, and will not now be destroyed until
a later meander under-cuts them. Scarps
of this kind may be called one-sweep scarps ;
and their cusps, one-sweep cusps. Several
similar cusps are shown in Figure 9. The
longer curve, E E', cut by the advancing
front of the meandering river, may be
abandoned if the river is diverted by a
Fig. 10. short-cut to a new course, and if so the

advancing terrace spur will be smooth trimmed.

DAVIS: RIVER TERRACES IN NEW ENGLAND. 305

A terrace whose scarp has been almost evenly trimmed by the small
vacillations of a down-sweeping meander will face the axis of the valley.
A terrace whose scarp has been under-cut by the forward half of a down-
sweeping meander will face obliquely up the valley, as in the foreground
of Figure 10. Inasmuch as the normal progress of meander-sweeping is
down the valley, it would seem, at first thought, that no terrace scarps
could be carved so as to face in that direction ; but on second thought it
will be seen that the lateral growth of a meander may cause part of the
curve to grow up-valley faster than the meander is carried in the other
direction by the normal down-valley sweeping of the meander system ;
and in this case a terrace scarp facing obliquely down-valley will be
carved. An example of this kind is shown near the foreground of Figure
11. It is manifest that the development of terraces facing down- valley
will be favored wherever the down-valley sweeping of a group of mean-
ders is for any reason checked while the enlargement of their curves is
continued.

There can be little doubt that the height of terraces produced by the
action of successive meanders would be very small, hardly measuring as
many inches or quarter inches as actual terraces measure in feet. Let
it therefore be now supposed that after a series of one-sweep scarps has
been carved, the river swings away from the western side of its valley
and for a time occupies itself in carving scarps on the eastern side.
Many meanders will have swept down the valley during the eastward
swing of the meander belt, each meander leaving its faint scar on the
valley floor. When the river swings westward again, it will be working
at a lower level than before, and as it then once more undercuts the
high plain, a distinct terrace with a scarp of ten or twenty feet will be
formed. Terraces of distinctly different levels may therefore usually
be taken to represent different swings of the meander belt ; terraces
that represent only the sweeps of successive meanders while the belt
remains almost stationary must be so faint as to be hardly noticeable.

A cusp which results from the slightly vacillating forward sweep of a
single meander, as B 6, B 7, etc., Figure 9, has already been called a
one-sweep cusp. The terrace plain extending forward from the base
of such a cusp will be a smoothly continuous surface on both sides of
the apex. When the two parts of a terrace plain on either side of a
cusp differ in height by a foot or two, they are probably the product of
different (but not necessarily successive) meauders ; and such a cusp
may be called a two-sweep cusp, because the two levels probably repre-
sent different sweeps in the meander belt. When the difference in

306

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

height is a number of feet, the cusp is probably a two-swing cusp,
because the two levels are then best explained by different (but not
necessarily successive) swings of the meander belt. Examples of these
forms are often met with.

Ideal Terrace Pattern : Middle Stage. If the sweeping and swing-
ing of the river continue until, as in Figure 11, a fifth return of the
meander belt to the western side of the valley is accomplished, a ter-
race pattern of some complication may result. Few and small remnants
of the higher terrace plains are to be expected at this stage, for they
have been repeatedly undercut and destroyed. Larger and more nu-
merous remnants of the lower plains may be still preserved, for they have

Fig. 11.

been less frequently attacked. A triangular portion of the third-swing
plain is shown in the middle of the figure ; on the right appears a still
larger piece of the fourth-swing plain, from which the river was with-
drawn by a short cut across the flood plain (not shown in the figure),
leaving an unfilled channel which now guides a small brook. The fifth
westward swing of the river has, during the down-valley sweep of a
single meander over the length of the diagram, undercut and destroyed
part of the fourth-swing plain on the right ; it has destroyed all of the
fourth-swing and part of the third-swing plain in the middle and all of
the earlier plains toward the left, where the meander undercuts the high
scarp and recent landslips have occurred. The greater the number of
swings, the smaller and rarer will be the remnants of the higher ter-
race plaius, unless some special control is present to preserve them.
A special interest attaches to the form and arrangement of the cusps

DAVIS : RIVER TERRACES IN NEW ENGLAND.

307

that are produced by the chance intersection of the curved terrace
fronts. In the two-sweep or two-swing cusps, two patterns may be
produced, according as the higher scarp is on the up-valley or down-
valley side of the cusp, as in Figures 12 and 13. The higher scarp is
evidently the younger of the two that unite in the cusp ; it is con-
tinued in direction, but with less height beyond the point of intersection.
From the appearance of a letter Y in the shaded scarps of the diagrams
forms of this kind may be called two-swing (or two-sweep) cusps with
an up-stream or a down-stream Y-stem, as the case may be. All the
elements of such cusps are variable. The location of the cusp is at the
chance intersection of two lines that have no particular relation to each
other. The angle of the Y may vary through a large range. The
heights of the scarps have no definite relation, except that the higher
single scarp must be the sum of the other two.

Fig. 12

A series of one-sweep cusps may occur with some regularity along a
valley side ; but two-sweep and two-swing cusps cannot be expected to
show so definite an arrangement, unless under the control of something
more systematic than the action of the wandering stream. In Figure
11, for example, two series of one-sweep cusps in the middle of the dia-
gram stand in normal down-the-valley order; but all the two-swing
cusps are located indifferently to one another. So on the left side of
Figure G ; but on the right side of that figure, three one-sweep cusps on
successive terraces are placed in line, one forward of the other, as if in
some way systematically related or subject to some common control.

In view of these various ideal combinations, it is evidently desirable
to analyze the special configurations that may be due to river action
alone, in order to detect more surely the patterns that must be referred
to some other cause.

A later-made one-sweep cusp may occasionally chance to stand in
front of an earlier-made one-sweep cusp, as in the further part of Fig-

308

BULLETIN* MUSEUM OF COMPARATIVE ZOOLOGY.

ure 14; but it is evidently out of the question that four cusps should
gain a systematic position of this kind, as in the center of Figure 15,

Fig. 14.

without some common control. Whenever such an arrangement is
found, special examination should be made of it. Again, while a two-
swing cusp is of common occurrence, a three-swing cusp, Figure 16,

Fig. 16,

must be rare, for it involves the intersection of three unrelated lines in
a systematic manner; and a four-swing cusp, Figure 17, is practically
an impossible occurrence. True, a three-swing cusp must be produced

Fig. 18.

at a certain stage of the change shown in Figure 18, where a sweeping
meander is undercutting its scarp and thus pushing one two-swing cusp,
A, towards another two-swing cusp, B ; for at a certain stage in the

DAVIS : RIVER TERRACES IN NEW ENGLAND.

309

undercutting, A will be pushed into coincidence with B, thus forming a
three-swing cusp. But it is very improbable that this temporary stage
will be preserved. The undercutting will continue and the tempo-
rary three-swing cusp will then be divided into two two-swing cusps, C
and D. The three-swing cusp can be preserved only when, just at the
moment of its formation, the stream is withdrawn by a short-cut or a
cut-off, and such a coincidence must be of very rare occurrence. With-
drawals of the stream may, however, happen likely enough before or after
the momentary stage of the three-swing cusp ; and the various patterns
thus producible are indicated by the full and broken lines in Figures 18
to 21. The eight possible cases of this kind result from different com-
binations of the up-valley or down-valley half of a meander with two-
swing cusps of up-stream or down-stream Y-stems. Evidently, then,
no combination of unguided sweeping and swinging meanders will pro-
duce an orderly grouping of cusps such as is shown in Figure 15.

Fig. 20,

Ideal Terrace Patterns: Late Stage. When the causes that de-
termine the degradation of a valley floor weaken and disappear, the
stream will repeatedly swing to and fro on about the same plane. Even
the basal terraces of a series may then be almost completely swept
away by the wandering river, as in Figure 22, and the whole descent
from the highdevel terrace to the existing floodplain may be, for con-
siderable distances along the valley side, united in a single strong escarp-
ment. The conditions under which this result may be brought about
are : first, the attainment of nearly fixed values of volume and load,
such as might be reached when a glacial climate had given way to a
milder climate and the latter had become well established ; second, the
cessation of any slow uplift by which degradation had been initiated
or aided ; third, superposition of the stream on a strong rock sill on which
corrasion is very slow. Under these conditions, a stream would almost
cease to degrade its channel, and would then devote practically all its

310

BULLETIN: museum of comparative zoology.

energy to lateral cutting. The stream would wander back and forth
across its valley floor, unimpeded save by its flood-plain bank, until it
came against a lateral terrace ; and the terrace would be worn back
to the limit of the belt of wandering. Sooner or later the stream
would consume all the high-level and intermediate terraces, pushing
back their united scarps into a single scarp by which the highest
terrace plain would then descend at once to the flood plain. Stepping
terraces would no longer characterize this stage of valley development;
but a few low terraces might remain not yet consumed here and
there, as in Figure 2.

Fig. 22.

Miller seems to suggest that an increase in the height of terrace scarps
is in itself a cause for a decrease in interscarp space. " The restraint
exercised by rock upon the modern rivers strengthens their natural
tendency, of which sufficient account has not been made, to occupy nar-
rower portions of valleys the more they deepen them. It has been too
hastily concluded, because rivers now occupy narrowed valleys flanked
by terraces comparatively broad, that therefore they have vastly shrunk,
— from dimensions, in fact, proportional to the greater breadth " (299).
He then goes on with the statement already quoted regarding the
" banks nowadays eight or ten times as high" as formerly. It thus
seems to be implied that it is " natural " for the interscarp space of a
stream of constant volume to lose width ; and that not only rock ledges
restrain the breadth of the belt of wandering, but that the increase in
height of the enclosing scarps also has a share in determining this feature
of valley form. A similar opinion is expressed by Gilbert (133). This

DAVIS : RIVER TERRACES IN NEW ENGLAND. 311

seems to me an unnecessary conclusion, unless rapid elevation up to a
recent date be postulated also, as is perhaps implied by Miller in a later
sentence. Certainly, so far as increase of scarp height in New England
valleys is concerned, it has not sufficed to prevent the broadening of the
valley floor and the consumption of terraces at higher levels, so long as
the terraces consist only of clay, sand, and gravel.

It may be noted that if there had been a diminution of volume during
the deepening of a drift-filled valle}-, the obliteration of stepping terraces
would be delayed, but not prevented. It has already been explained
that some diminution in stream volume is certainly probable. It may
now be added that many valleys have, in spite of this very probable
decrease of stream volume, already reached in one or another part of
their length the late stage of terracing just described, in which all the
descent from the highest terrace to the flood plain is concentrated in a
single scarp, and that in many other parts of these valleys only a few
basal terraces remain beneath the strong scarp of the high terrace. It
thus becomes all the more probable that diminution of volume is not an
important cause of the decrease in the breadth of the interscarp space,
and that where stepping terraces occur, they must be in large part re-
ferred to some special and local cause. Such a cause is found in the
presence of rock ledges, as suggested by Miller ; and to that element of
the problem we may now turn.

Defended Terrace Cusps : Early Stage. It has thus far been
tacitly postulated that no buried ledges should be discovered by the
wandering river. Such, indeed, is the condition usually assumed in the
cross-section of a series of typical terraces, as in Figure 1. Let a new
series of terraces now be developed, in which ledges shall here and there
be discovered as the river degrades its valley floor to greater and greater
depths. It is evident that the number of such ledges may vary greatly.
They might be numerous and frequently encountered by a terracing
river in a narrow valley with rugged rock, walls and bottom ; they might
be almost absent and hardly ever discovered in a broad valley that had
been heavily aggraded. In all cases it is important to note that the
slope of a ledge face will seldom be as steep as the average slope of a
terrace front, which may be as much as 30° in freshly cut scarps.

As before, the river wanders about freely so long as it is working
on unconsolidated sands and clays ; and thus several low terraces
may be formed in the manner already described. But when a ledge
is encountered in the river bank, as at the left forward edge of Fig-
ure 23, the rock is practically indestructible. The stream will in a
vol. xxxviii. — no. 7 3

312

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

comparatively short time swing away, after having altered its course
more or less in a fruitless effort to wear back the obstacle. The
ledge thus comes to determine a cusp in the terrace front. A salient
of this kind may be called a defended cusp, in distinction from the
accidental or free cusps described in the previous sections ; the ter-
race behind it cannot be destroyed by the stream.

Following the colloquial style often adopted for field descriptions,
a terrace of this kind is sometimes entered in my notes as a"can't-be,"
in contrast to the low " not-yet " terraces of Figure 22.

It should be noted that the ledge here considered does not deter-
mine the depth to which the river may work ; the rock is exposed

Fig. 23.

only in the river bank and enters but a little distance into the
channel. The slope of the river and the depth to which it has cut at
this or any other point in its course are assumed to be determined
in all cases thus far detailed by the maintenance of an essentially
graded channel with respect to some controlling baselevel further
down stream ; the sea at the river mouth, a larger river into which
the smaller stream enters, or a broad sill of rock that stretches all
across the channel, somewhere further down the valley.

When the withdrawing stream swings back again at a lower level,
as in Figure 24, it cannot often under-cut and destroy all of the
terrace on whose back border the first ledge rises, because, as has
been noted, the slope of the ledge is seldom so steep as that of the
terrace scarp. A second encounter with the ledge will usually be

DAVIS: RIVER TERRACES IN NEW ENGLAND.

313

made before the swinging stream has entirely consumed the terrace
of the previous swing. Every return of the swinging river against a
sloping reef of rocks will thus be recorded by a little strip of terrace
behind the defending ledges, and a flight of stepping terraces will
necessarily be produced wherever a large group of ledges slopes under
the valley drift into the belt of river action. The length of the
ledge exposed at the back of each terrace may be but a few feet,
but its effect will be prolonged by a trailing terrace, as it may be
called, stretching hundreds or thousands of feet along the valley side.
It is in this way that the best flights of stepping terraces are produced.

Fig. 24.

The special features of terraces associated with defending ledges are
next to be examined.

Slipping Meanders and Blunt Cusps. It has not yet been possible
to discover by observation how a down-sweeping meander will make its
passage past a ledge ; but it may be inferred from the forms of terraces
found in various valleys that a stream has two methods of procedure in
such an exigency. The first is considered in this section, and is illus-
trated in Figures 25 to 28 ; the second is taken up in the next section,
with Figures 29 to 31.

Let it be supposed that the river in Figure 25 is now making irs fourth
swing against the western side of its valley, and that a buried ledge lies
a few hundred feet back of the group of free cusps in the middle of the
diagram. The ledge is discovered in Figure 2G ; it lies somewhat below
the apex of a down-sweeping meander. Assuming that the meander

314

bulletin: museum of comparative zoology.

curve is to remain practically unchanged, it can pass the ledge only by
withdrawing somewhat towards the axis of the valley ; it may thus, as
it were, slip by the obstacle that stands immovable in its way. The
stream is represented as having just slipped past the ledge in Figure 27,
and as having swept somewhat farther down the valley in Figure 28.

Fig. 25.

Fig. 26.

All records of the first and third swing of the river are now destroyed,
so far as this part of the valley is concerned ; the terrace front shows a
high, defended, one-sweep cusp, a free two-sweep cusp with an up-stream
Y-stem, and a free one-sweep cusp.

The ledge at the base of the defended cusp may come to be more or
less concealed by the sands that are washed down from the weathering

Fig. 27.

Fig. 28.

scarp. In time it may be entirely covered, and its presence will then
be known only if a road cut or a boring reveals it. It is therefore quite
possible that some apparently free cusps are really defended cusps, with
their defending ledge ambushed beneath a thin cover of soil.

Compressed Meanders and Sharp Cusps. Let it now be supposed
that the river in Figure 2G is unable to slip past the ledge. The front

DAVIS : RIVER TERRACES IN NEW ENGLAND.

315

of the curve is held fast ; the apex of the curve bends outward and cuts
a curved re-entrant in the terrace front next up-stream from the ledge,

Fig. 29.

Fig. 30.

Fig. 31.

as in Figure 29. The meanders further up-stream continue their advance
and the meander uext to the ledge is therefore compressed to a relatively
strong curvature, as in Figure 30.
The defended cusp is now sharp-
ened. It may come to point
somewhat up the valley. The
compressed meander cannot slip
by the ledge ; there is no escape
for the stream save by a short-
cut across the narrowing flood-
plain lobe at time of high water ;
thus the condition of Figure 31
will in due time be developed.
A rather sharp cusp, one of whose sides faces up the valley, will be
produced, and the great concave scarp adjacent to it will have an
abandoned channel at its base.

Terrace Fronts near Defended Cusps. The difference between
the behavior of slipping and of compressed meanders may be inferred to
depend on the position of the obstructing ledge with regard to the apex
of the meander. If the ledge is discovered near the apex of the meander,
the stream may slip past the obstacle, as in the example first given.
If the ledge is encountered near the point of river inflexion — that is,
on the tangent between two meanders, — compression of the meander
up-stream from the obstacle is likely to result.

In both these cases the defended cusp is likely to be associated with
several short curves and free blunt cusps for a little distance down-
valley ; while a rather long re-entrant curve usually joins its up-valley

316

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

side. The short curves and blunt cusps may be blended so as to pro-
duce a convex terrace front for a little distance down-valley from the
ledge ; and this convex front will be separated from the long concave
re-entrant by a more or less pronounced angle at the defended cusp.
The reason of this may be easily understood from Figure 32. The ledge
was here first discovered by a down-sweeping meander, of whose first
work only the short scarp A remains. The meander was then com-
pressed so as to scour out a large concave re-entrant, and was with-
drawn from the channel, B, at its base by a short-cut (outside of the
diagram). A meander on the new course of the stream", swinging

Fig. 32.

westward, trimmed off the terrace front in successive lines, C, D, E ;
but as this meander had no definite relation to the ledge, and as the
general sweeping of its curve was down-valley, it did not trim the ter-
race front with any special regard to the down-valley side of the de-
fending ledge. In brief, the tendency of a stream to sweep its meanders
down-valley commonly l'esults in trimming off the terrace front close to
the up-valley side of a defending ledge, but it is only by chance that the
terrace is worn away close on the down-valley side of the ledge. The
pattern hei-e deduced may be matched in many actual examples.

Defended Cusps : Later Stage. After a ledge has once been dis-
covered by the swinging river, there is much probability that the for-
ward reacli of its under slope will present an obstruction to the stream
every time it swings again towards the valley side. For example, let

DAVIS : RIVER TERRACES IN NEW ENGLAND.

317

the stream of Figure 31 be supposed to have returned from a swing to
the eastern side of its valley floor. It will now be at a lower level than
before and will therefore be halted somewhat in advance of the defended
cusp previously developed. Figure 33 shows the blunt cusp of a slip-
ping meander thus determined. Another swing out and back having
been accomplished, Figure 34 shows the work of a compressed meander,

Fig. 33.

Fig. 34.

A, which for a time was held up-valley from a third exposure of the
long-sloping ledge ; but the stream was then withdrawn from its round-
about course by a short-cut, after which a sweeping meander wore out
the three short curves, B, down-valley from the ledge ; and still later
the next down-sweeping meander, C, trimmed off the terrace front close
to the ledge preparatory to slipping past the obstacle and pushing still
further back the down-valley side
of the terrace front. Another
stage is shown in Figure 35. Here
the eighth westward swing of the
stream is recorded (compare Fig-
ure 25). Part of the plain formed
by the second swing happens to
be still preserved, but not of the
first and third. All the later
swings, fourth to eighth, are well
indicated. A strongly-compressed meander of the eighth swing has
trimmed off all the terraces a little distance up-valley from their ledges,
and would have trimmed them still closer but for being withdrawn by
a short-cut. A later meander of the same swing is less successfully
wearing away the down-valley extension of the terraces.

It is evident that an infinite variety of terrace patterns may be ex-

Fig. 35.

318 BULLETIN': MUSEUM OF COMPARATIVE ZOOLOGY.

pectcd in association with defending ledges ; yet they must all conform
to certain general laws of development.

If the ledge lies at a low level, the greater part of the terraces that
have been cut at swings of higher level will have been destroyed before
the ledge lias a chance to defend them. If the ledge is of large size,
rising nearly to the highest terrace level and standing forward in such a
position that the stream may frequently swing against its buried slope,
a whole flight of stepping terraces may be not only formed, but pre-
served by it. Here the early terraces, unlike those of free-swinging
rivers, are defended by ledges, and cannot here be attacked by
the later swings and sweeps of the stream. They are subject to destruc-
tion only by general weathering and washing of the valley sides. It is
evidently, then, to the largest and highest and most outstanding ledges
that one must go in order to find the fullest record of the number of
swings that a river has executed during the excavation of its valley, for
only on such ledges are the records of river terracing well preserved.
Elsewhere they are for the most part swept away. Even here some
swings may not be recorded. In short, the maximum number of ter-
races shows only the minimum number of river swings.

Diminished Swinging of the Meander Belt. The greater the
depth to which the valley floor is degraded, the more frequently may
ledges be found, and, as a rule, the nearer will they stand to the axis of
the valley. The number of defended cusps will therefore tend to in-
crease as the valley deepens. The breadth of free swinging will at the
same time decrease, and the space between the scarps of the lower ter-
races will necessarily be less than the space between the higher terraces.
This principle, first stated by Miller, seems to be essential in explaining
the stepping terraces of New England.

It must frequently happen that ledges approach the axis of a valley
more closely at one point than at another. The valley may be well
beset with buried reefs for a fraction of a mile or more, and then may
be relatively free from ledges for several miles up and down stream.
Where the ledges are numerous, the valley will be narrowed, and the
terraces will be preserved in good number; but in the stretches that are
comparatively free from ledges, or in which ledges are found only at
low levels, the valley floor may be broadly opened, and but few of the
many Hood plains that the river there formed at various levels will
be preserved. These open basins, often bordered by a single high-
scarped terrace, have attracted less attention than they deserve in
the discussion of terracing ; and well-developed flights of terraces

DAVIS: RIVER TERRACES IN NEW ENGLAND. 319

have been given an importance that their restricted occurrence hardly
warrants.

Distribution of High-Scarp and Low-Scarp Terraces. Ledges
may be gradually disclosed at various points up and down the valley,
each one having an effect in cusp-making and terrace-keeping appro-
priate to its position with respect to the valley axis. The frequent
swinging of the meander belt from side to side during the slow degrada-
tion of the valley floor requires that the discovery of every ledge lying
well within the belt of wandering should be made soon after the stream
has degraded the valley floor to the level of the ledge top. If the valley
floor is deepened ten feet during a complete swing of the meander belt
to and fro, it would be very unlikely that a ledge within the breadth of
swinging should escape discovery until the valley floor was worn down
twenty feet below the ledge summit. It would be impossible for the
discovery to be postponed so long that the stream should first encounter
the ledge fifty feet below its top, unless the ledge were situated rather
far to one side of the valley axis, where it might not be encountered by
the down-sweeping meanders every time the meander belt swung across
the valley. The more likely case is that an actively swinging, slowly
degrading stream will discover the upper part of every ledge lying
well within the belt of wandering, and that thereafter the stream will
frequently swing against the slope of the ledge at lower and lower levels
as the valley floor is deepened ; unless, indeed, another ledge, nearer the
axis of the valley, and with a lower summit than the ledge already dis-
covered in its neighborhood, prevents the river from swinging laterally
so far as it had at higher levels.

This specialized conception of the terracing process leads to some
reasonable deductions as to the distribution of high-scarp and low-
scarp terraces. They are summarized in Figure 36. A low-scarped
terrace is formed near the western border of the belt of wandering
shortly after degradation has begun (see further side of figure). After
six swings the river discovers a ledge (also on the further side of the
diagram) somewhat within the belt of wandering. Then all the terraces
of earlier swings lying back of this ledge will be preserved. On every
later westward swing the river is halted nearer and nearer to the
axis of the valley. Thus a flight of stepping terraces is formed in con-
nection with a series of defended cusps ; but on account of the absence
of ledges on the near side of the diagram and of the increased breadth of
wandering as the later stage of terracing is approached, the river de-
stroys all traces of the earlier terraces in the foreground, where the

320

bulletin: museum of comparative zoology.

tenth swing produces a single scarp by which the highest plain de-
Bcends to the river level. Then the eleventh and twelfth swings are
held off from the high scarp by a lower ledge on whose slope two
low-scarped terraces are carved. It may therefore be concluded that low
undefended high-level terraces of early swings are most likely to be pre-
served back of defended cusps of later swings ; that the undefended
terraces of early swings will probably be swept away in the production
of a single high-scarped terrace wherever broad swinging at low levels
is not prevented ; and that when high scarps occur in a flight of step-
ping terraces they are more likely to be found at or near the top than
near the bottom of the flight.

Effect of Rock Barriers. Superposition upon strong rock barriers
has already been considered on pages 292 and 309, in so far as it deter-
mines the separation of a valley into several compartments, in each of
which the flood plain is thenceforward graded with respect to the
next down-valley barrier. This is a very familiar condition in New
England, as the water-power in the falls or rapids on the down-valley
side of the barriers has repeatedly determined the location of our manu-
facturing villages and cities. In the present section, brief consideration
is given to the effect of rock barriers in producing a fixed node, as
Emerson has called it (736), in a stream that elsewhere vibrates freely
as it meanders and swings on its flood plain. It is, however, not yet

DAVIS: RIVEK TERRACES IN NEW ENGLAND.

321

clear how a wandering stream will behave up and down valley from a
fixed node. Several suppositions may be made.

First, the meanders will sweep down the meander belt, and the
meander belt will swing to and fro across the valley, but the ampli-
tude of both movements will be decreased as the node is approached,
and extinguished as it is reached. So far as my observations go, this
condition is more appropriate dowu-valley than up-valley from a fixed
node. Below the node, slight curves may be formed ; these may develop
into normal meanders, Figure 37 (the river flows to the right), as they
sweep away from the sill; but such development will probably be grad-
ual, and hence the valley floor will widen gradually in that direction.

""/IliiV"'-- , ■■"Hit , — ■"„„,.' „^,\ll'"ll"l»//ll1"1,",4

Fig. 37.

Second, the meanders may continue almost in full force as they
approach the node from up the valley, merely changing in the lowest
part of their course that leads directly to the sill. This might involve
the introduction of a " kink " into the meander system, at the point
where a change is made from the normal down-sweeping curve to the
constrained course that leads to the ledge. That such a sharp bend is
possible seems to be shown by certain peculiar forms in the meanders of
the Theiss on the plain of Hungary ; it being probable that bends of this
kind result from the faster down-sweeping of some meanders than of
others. The considerable breadth of flood plain often observed next up-
stream from a node supports this supposition.

Third, the fixed node may perhaps induce the formation of free nodes,
evenly spaced from the ledge of superposition ; then between the fixed
node and the free nodes, the stream might vibrate as a stretched string
vibrates when it is lightly " stopped " at a third or a quarter of its
length. Symmetrical free terrace cusps would result from this process.
So systematic a movement would seem to be possible only in rare

322 bulletin: museum of comparative zoology.

cases, if at all ; but the terraces of Chicopee river above Beecham Falls,
a few miles northeast of Springfield, Mass., give some support to this
possibilitv. Further observation is needed in this direction.

When two barriers occur near together, leaving a free space of half a
mile or so between them, the river is fixed at two nodes, but may
vibrate between them. A remarkable case of this kind is found in the
valley of Saxtons river at Bellows Falls, Vermont, as described on
page 337.

Relation of Terrace Patterns on the Two Sides of a Valley. It
has been already pointed out as a matter generally accepted by many
observers, that the terraces on the two sides of a valley need not neces-
sarily agree in number or in height. The relations of terrace patterns as
seen in plan on the two sides of a valley have been less considered. It
is desired here to indicate certain relations that seem to obtain in special
cases.

When a group of defended cusps occurs in a valley of moderate
breadth, the stream must have been repeatedly deflected across the
valley by the defending ledges, so as often to impinge upon the opposite
side of the valley in about the same place. Hence re-entrants of more
than usual size may there be worn out, next up-valley from which a
group of free cusps may thus come to stand about opposite the defended
cusps. If the meander next above the ledge is somewhat compressed,
the stream may strike more squarely across the valley and under-cut the
down-valley side of a terrace with somewhat greater vigor than usual.
The valley of the Westfield river, a mile or so up-stream from West-
field, Mass., offers some remarkable examples of this kind (page 333).

When a side stream enters the valley of a degrading main stream, it
tends to push the main stream away, and thus causes it to wear out
re-entrants opposite to the entrance of the side stream. Reacting from
such re-entrants, the main stream will strike across the valley and scour
out another group of re-entrants belowT the mouth of the side stream.
When this transverse deflection of the main stream is confirmed by the
occurrence of a guiding ledge, the re-entrants will be all the more per-
sistently and repeatedly carved out. The Connecticut seems to show
an example of double control by the Westfield and by ledges in the
southern part of Springfield, Mass. (page 344) ; and the Westfield itself,
two miles east of Westfield village, offers a similar example of double
control (page 330).

Ratio of Sweeping, Swinging, and Degrading. The foregoing analyses
of the process by which a graded and wandering river may degrade and

DAVIS: RIVER TERRACES IN NEW ENGLAND. 323

terrace its valley suggests a method by which a quantitative determina-
tion may be made of the ratios of sweeping, swinging, and degrading.
If numerous measures are taken of the difference of level between
adjacent terraces in a certain section of a valley, it maybe expected that
two groups of minimum differences should be found ; the group of
smaller values representing the deepening of the valley floor in the inter-
val between the down-valley sweeping of two successive meanders ; the
group of larger values representing the deepening between two succes-
sive swings of the meander belt.

If measures of this kind were taken in different sections of a valley
system, it might be possible to determine from their variations whether
an even regional uplift or a tilting were chiefly responsible for the
activity of the river in carving its terraces, as the following considera-
tions will show.

In the case of uniform uplift over a large area, let it be assumed that
the movement was rather quickly initiated, and then steadily continued
until it rather rapidly weakened at its close. We should then expect
that the terrace scarp marking the interval between two lateral swings
of the meander belt would be of greatest measure and relatively constant
in the lower course of the main river ; but the slow initiation of the up-
lift might possibly be recorded by a few low terraces at the top of the
series ; the slow close of the uplift and the very slow degradation of the
valley floor in later time might be recorded by a few terraces of lower
and lower scarps at the base of the series. If any terrace in the lower
course of the river could be followed up the valley, it would assume
a relatively higher and higher position in the series; for when the river
had in its lower course worn down its valley floor to a low grade at the
close of the period of uplift, there wTould still be a considerable amount
of degradation permitted to the middle and upper part of the river. As
a result, the very low terraces at the base of the series near the river
mouth might gain a higher rank and a greater scarp height further up-
stream. If the terraces could be followed up to the headwaters of the
river system, they would become narrower and finally disappear in
single-scarped V-shaped valleys. So far as they could be recognized,
the upper members of a headwater series might correspond in date
to the basal members near the river mouth ; while the basal members
of a headwater series would decrease in scarp-height as they were
traced down-valley, until they at last merged iu the even flood plain
of the middle or lower river course. So many are the irregularities
of drift terraces that there has not to my knowledge been any sys-

324 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

tematic attempt to discover the facts by which these deductions might
be con fin noil.

In tbe case of a tilting, with fulcrum at the river mouth and at right
3 to the general river course, the maximum height of two-swing
terrace scarps would be found somewhere near the middle course of the
river; and the scarp height would thence decrease down-valley and up-
valley. It seems that, in such a terrace system as that of the Connecti-
cut, it may be possible to apply this test and thus gain from the
dimensions of the terraces a direct proof as to the kind of movement by
which the work of terracing was initiated, as well as a confirmation
of the evidence already in hand regarding the nature of postglacial
movement in the New England province.

Relation of the Preceding Deductions to the Observations
described in the Following Sections. The facts presented in the
following sections are chiefly details of structure and form, directly
observable ; changes of form are occasionally noted, but these are of
relatively small measure. All these details are but the present members
of a long series of facts, every one of which might have been recorded,
had observers been living to witness them ; and then the origin of ter-
races would be fully understood. Bat the earlier members of the series
are hopelessly lost to observation from being prehistoric. In their
unavoidable absence, theory attempts to supply a series of conditions,
pictured by the reasonably guided imagination, which shall imitate the
series of past facts, and thus, as it were, call them to life, bring them
into the field of vision. The success of the theory is not to be measured
so much by the apparent reasonableness of its fundamental suppositions,
or by the definiteness with which various imaginary consequences may
be deduced from it, as by the accuracy with which the observable
members of the deduced consequences imitate the facts of actual occur-
rence. The greater the number of peculiar categories of observed facts,
the greater the probable correctness of a theory whose deduced conse-
quences can match all of them. Hence the importance of minute
observation and careful generalization on the one hand, and of accurate
and detailed deduction on the other. Hence also the importance of
carefully distinguishing these unlike processes in order that their results
may be systematically confronted in an unprejudiced comparison. The
elaboration of the deductions in the preceding sections therefore seems
to be as necessary a part of the study of terraces as is the accumulation
of observations for presentation in the following sections.

The theory of terracing has here been presented before the observa-

DAVIS : RIVER TERRACES IN NEW ENGLAND. 325

tions of terraces are detailed, because it is the theory with its deduced
consequences and not the facts that are on trial. Furthermore, it is only
after the presentation of the theory that the pertinent facts can be con-
veniently selected from among many others and that their bearing can
be clearly appreciated. True, the attempt might be made independent
of any theory to observe all facts thoroughly and to record them minutely,
in the hope of including every item that could be asked for in the test-
ing of whatever theory should afterwards be invented ; but under this
method of work, items of minor importance are confused with those of
major importance, and their recital becomes so long that the beginning
is forgotten before the end is reached. As a matter of fact, observational
study of this kind is notoriously incomplete. Indeed, the terrace prob-
lem, like many others, gives striking illustration of the difficulty if not
the impossibility of really seeing all the essential facts when only the
eyes of the observer are trusted ; and it illustrates at the same time
the critical power that is given to observation when it is directed towards
significant points, instead of being allowed to wander in the vain hope
of finding all the facts before theorizing is begun. For example, if it is
not already manifest from the deductions of the preceding paragraphs
that the terrace spurs formed of grouped cusps and the outcropping
ledges that are associated with them are of particular significance, no
doubt will remain on this point when the observations detailed in the
following paragraphs are reviewed ; yet in all that has thus far been
written on this subject in New England, no description of grouped cusps
is to be found, and no recognition of the significance and the generality
of the relation between ledges and cusps is recorded. It is as if it had
been thought that all parts of a terrace are equally significant ; that
when ledges appear at the terrace base they are of no particular impor-
tance. Even the citations made above from the writings of Edward
Hitchcock do not show that that careful observer thought the ledges he
described were of any more than local importance ; and certainly no later
observer has been led by Hitchcock's essay to understand the control
that ledges exercise in determining terrace pattern and terrace preserva-
tion. Yet after apprehending this control and discovering the suggestive
relation that must obtain between ledges and cusps, the observer no
longer strays over his field ; he directs his steps and secures in the least
possible time the greatest possible results.

Largely deductive as the preceding portion of this essay is in its
present form, the reader should not suppose that it was prepared inde-
pendent of observation. The actual progress through the problem has

326 bulletin: museum of comparative zoology.

involved repeated alternations of external and internal work; the col-
lection of observations and the induction of generalizations on the one
hand ; and on the other hand the invention of hypotheses, the deduc-
tion of their consequences, the confrontation of deductions with general-
izations, the evaluation of agreements, and the repeated revision of the
whole process. It is not profitable to expose the personal history of a
study all through these stages, for the convenience of the reader is best
served by a careful separation of its two phases ; and to the second of
these we may now turn in Part IV. with no more delay than is required
for the citation of the following pertinent extract from Playfair's Illustra-
tions of the Huttonian Theory of the Earth. After pointing out that to
wait for the completion of discoveries in other sciences before theoi-izing
in geology " would not be caution, but timidity, and an excess of pru-
dence fatal to all philosophical inquiry," this lucid writer of a century
ago proceeds as follows : —

" The truth, indeed, is, that in physical inquiries, the work of theory
and observation must go hand in hand, and ought to be carried on at the
same time, more especially if the matter is very complicated, for there
the clue of theory is necessary to direct the observer. Though a man
may begin to observe without any hypothesis, he cannot continue long
without seeing some general conclusion arise ; and to this nascent
theory it is his business to attend, because, by seeking either to verify
or to disprove it, he is led to new experiments, or new observations.
He is also led to the very experiments and observations that are of
the greatest importance, namely, to those instantice crucis, which are
the criteria that naturally present themselves for the trial of every
hypothesis. He is conducted to the places where the transitions of
nature are most perceptible, and where the absence of former, or the
presence of new circumstances, excludes the action of imaginary causes.
By this correction of his first opinion, a new approximation is made
to the truth ; and by the repetition of the same process, certainty is
finally obtained. Thus theory and observation mutually assist one
another; and the spirit of system, against which there are so many and
such just complaints, appears, nevertheless, as the animating principle
of inductive investigation. The business of sound philosophy is not to
extinguish this spirit, but to restrain and direct its efforts " (524, 525).

DAVIS: RIVER TERRACES IN NEW ENGLAND. 327

IV. Observations of River Terraces in New England.

Valley of the Westfield River, Mass. Eastern Section. This
branch of the Connecticut rises among the hard-rock Berkshire hills
of western Massachusetts, the round remnants of the uplifted and
dissected Cretaceous peneplain of the Appalachian province, and thence
flows eastward part way across the broad valley lowland that has
been excavated in the weaker Triassic formation during later Tertiary
time. Between the eastern base of the crystalline uplands and the
ridge formed on the main sheet of extrusive trap within the Triassic
area, the stream has excavated a fine series of terraces invthe uncon-
solidated drift deposits that have been so abundantly spread over the
Triassic lowland by the Connecticut and its tributaries.

The village of Westfield lies near the middle of this terrace system
and serves to mark the separation of its unlike eastern and western di-
visions. In the eastern division, Westfield river, re-enforced by Little
river, a branch which leaves the hills two miles south of the main
stream, has opened a broad basin at an elevation of about 140 feet.
The basin floor is nearly everywhere enclosed by the strong scarp of a
single high terrace whose plain stands at altitudes of 240 to 280 feet.
The plain is not of simple origin. On the southeast, its surface is roll-
ing, as if consisting of morainic and kame-like deposits. On the north,
it is smooth and its sands are tine enough to have been raised in occa-
sional dunes; here the plain falls off southwestward to the valley of
Powdermill brook in a series of lobes, whose intermediate depressions
are too large to have been excavated by local drainage : hence it is prob-
able that this part of the plain is a delta front in one of the areas of dep-
osition described by Emerson (650-653). South of the main basin, the
smoother part of the high plain (Poverty plains) is regarded by Diller
(265) as an extension of the plain on the north; the originally con-
tinuous surface having been formed by the flooded Connecticut. West-
ward up the Westfield valley, the high plain ascends towards the I11II3
and is of much coarser materiajs than elsewhere ; this part seems to
have been capped by the local outwash from the high ground during the
period of aggradation. As the coarse upper gravels lie on fine sands
and silts, this high plain is probably, like the one on the north, a delta
surface, built up in standing water.

The strong scarps, B, Figure 39, by which the high drift plains
vol. xxxvm. — no. 7. 4

328 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

descend to the main low-level basin, everywhere present concave re-
entrants, whose curves unite in cusps — usually two-sweep cusps — of
greater or less acuteness ; and this shows that the streams have re-
peatedly swung against the terrace scarp, under-cutting it and pushing
it back, after the present grade had been essentially reached. The
curved re-entrants are of somewhat larger radius on the north than on
the south, as if they had been scoured by the Westfield and Little
rivers, respectively. With the significant exception of certain points on
the cast and west, to be described below, all these cusps, at least twenty-
four in number, are free, undefended by ledges. We have, therefore,
here an example of a vigorous stream with a good-sized branch working
in a broad deposit of loose drift, and free to sweep, swing, and wander
over a large area. A late stage of terracing has been reached, for the
wide plain is nearly or quite i-educed to grade with respect to a rela-
tively permanent local baselevel in the trap-ridge notch on the east.
The detention of further degradation by the trap barrier is a factor of
importance ; for many recent swings of the streams must have on this
account tended to destroy earlier terraces by reducing them all to one
level, instead of tending to make new ones at lower levels. Whatever
flood plains may have been produced during the excavation of the
present basin floor, the streams have now so well taken advantage of
their opportunity for lateral corrasion or " sapping " that terraces at
high and intermediate levels are nearly everywhere obliterated, and even
the low terraces are as a rule destroyed by the broad swinging of the
streams at their present grade.

Westfield river is at present nowhere working against the base of the
high terrace on the north ; its actual course lies about half a mile to
the south of the scarp, but several of its former courses along the ter-
race base are clearly revealed in a series of shallow swampy troughs,
the remains of channels from which escape seems to have been effected
by repeated short-cuts or cut-offs. The river is now engaged at several
points in grading down to modern flood-plain level a broad and low
terrace, parts of which are not yet destroyed.

Little river was in 1901 sweeping against its high terrace on the
south at two points a little east of the New-Haven and Northampton
Railroad. Here the usual cover of vegetation has been removed from
the scarp, the sands are under-cut, and the face of the scarp is sliding
intermittently into the stream. Small sand dunes are formed at the
top of the sliding bank by the northwest winds which sweep the sand
up from below. It is evidently enough by a repetition of sweeping and

DAVIS : RIVER TERRACES IN NEW ENGLAND. 329

swinging of this kind that the high terrace has been worn back to its
present outline.

Where the rivers have withdrawn from the high-scarped terrace, flat
fans have been formed at the outlet of the minor lateral valleys of small
brooks, or beneath little gulleys of wet-weather wash. The fan of
Powdermil] brook, for example, forms a low barrier, X, Figure 39, across
a deserted channel of Westfield river, and thus determines a swampy
depression just northeast of Westfield station. The further course of
the brook follows the marshy deserted channels of Westfield river at
the base of the scarp for over a mile.

It would be difficult to find better illustrations of the deductions
presented on page 310 than are offered by this beautiful basin. The
two chief streams, far from exhibiting any incapacity to open their
valley floors, have now widened them to a greater breadth than ever
before. Whatever decrease of capacity may be due to decrease of
stream volume and of stream slope, and whatever increase of work may
be due to the more active wash of side streams on account of gain in
height of valley sides, the main streams are certainly more competent
to corrade laterally now than they have ever heen, and there is every
probability that they will in the future continue to widen their basin
still further by intermittent attacks on its border until restrained by
defending ledges or by the hand of man. Indeed, so nearly complete
is the obliteration of all terraces above the level of the present basin
floor, one might be tempted to conclude that the Westfield and Little
rivers never produced any extended series of flood plains in this division
of their course at higher levels than those of modern times, until an ex-
amination of the western division of the Westfield terraces proves that
flood plains must have been produced at various levels in the eastern
division as well as elsewhere.

Evidently then, as far as this example goes, it affords no evidence
that the production and preservation of terraces is due to any incom-
petence arising from decrease in the volume or from other changes in
the habits of our New England streams. Terrace preservation must be
due to some control external to the streams ; and of this we find imme-
diate proof on looking at the eastern and western enclosure of the broad
basin just described.

The basin is enclosed on the east by the approach of a defended spur,
A, Figure 38, on the north towards a free spur, B, on the south, beyond
which a subordinate basin, C, is again opened. The defended spur
carries a terrace plain at a height of 200 feet, and the highest plain

330

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

rises further uorth by a faded scarp of gentle slope. Sandstone ledges
are abundant along the western base of the spur ; they are unusually
steep, in part because of the eastward dip of the strata, and in part be-
cause of a certain amount of under-cutting by the Westfield when it ran
beneath them. The eastern side of the spur is not trimmed close to the
defending ledges, but illustrates the uusymmetrical relations shown in
Figure 31. Widely as the river has swung from side to side in the
basin further west, it was here strongly constrained. Not only so :
Westfield river has been somewhat impelled northward by the

Fig. 38.

entrance of Little river from the south (west of the area shown in
Figure 38) ; and it is probably in part at least on this account that the
basin has been so well broadened northward ; }ret on every sweep or
swing against the sandstone reef, the river was not only restrained
from further northward conquest at that point, but was deflected
obliquely southward across the valley. It is very probable that the
excavation of the subordinate basin, C, is due to this cause, for it is
opened further to the south than to the north. Three strong south-
ward loops of the river, D, E, F, (including the present one) are here
recorded ; and it can hardly be by chance that the river has thus

DAVIS: RIVER TERRACES IN NEW ENGLAND. 331

repeatedly turned southward on its way to the fixed node, G, in the
trap-ridge notch.

Nearly opposite to this well-defended spur, hut a little further
westward, the free spur, B, rising to the full height of the drift
plain, separates the subordinate eastern basin, C, and that part of the
main basin, H, which has been scoured out chiefly by Little river.
Unlike the defended spur on the north, the free spur is not a rela-
tively permanent feature of the valley ; it will be removed without
difficulty if Little river takes a fancy to trim away its western base.
Nevertheless, its occurrence to-day does not appear to be altogether a
matter of chance, for it seems to illustrate the systematic features
described on page 322.

The main basin is enclosed on the northwest by a well-defended
spur, known as Prospect hill, A, Figure 39, just west of Westfield sta-
tion • this will be further described with the terraces of the western
division of the valley. On the southwest, Little river is held from
swinging at present levels by superposition on a transverse sandstone
ledge, to which brief reference will be made further on. The contrast
between the openness of the main basin, excavated where the streams
have not been restrained by ledges, and the narrowness of the entering
valleys where ledges have been encountered, is most striking.

Western Section. The western division of the "Westfield terraces,
occupying the valley for about four miles from Westfield village to the
base of the hills, is of greater interest than the eastern, inasmuch as it
preserves the records of river work at many levels between the highest
and the lowest plains. I have prepared a somewhat detailed account of
it for publication in the " American Journal of Science," and hence shall
here refer only to such features as confirm the deductions of earlier para-
graphs.

The chief features of this interesting locality are shown in bird's eye
view in Figure 39, as if looking northeast from a height of several
thousand feet above the left front corner of the diagram. The Boston
and Albany railroad runs through the view for a distance of about a
mile and a half ; the foreground scale is larger than that for the back-
ground. Heights are exaggerated. Outcropping ledges are black.

From Westfield to a small rural settlement known as Pochassic Street,
two miles to the west, many small ledges are exposed, and man}7 step-
ping terraces occur along the northern side of the valley. Few ledges
are seen on the southern side, and there the valley is generally bordered
by a strong upper terrace, with a few low terraces beneath it. On the

332 bulletin: museum of comparative zoology.

northern side there are four groups of defended terrace cusps, forming
what may be culled the Pochassic (just to the left of Figure 39), Perry's,
K, Brown's, F, and Prospect spurs, A, while curved re-entrants have been
excavated between the spurs where ledges are rare or wanting. The
re-entrants show that the river has everywhere attempted to widen its
valley, while the terraces on the defended spurs show that the widening
has been locally prevented by the outcropping ledges. Wherever free
cusps occur they exhibit the patterns deduced as of common occurrence
on page 308. Xone of the combinations there deduced as rare are
found. The cusps are usually more closely trimmed on the up-valley
than on the down-valley side. It would be difficult to imagine a more
complete confirmation of Miller's theory than is here presented.

Special mention may be made of a few features. Just east of Po-
chassic Street a series of at least nine terraces, H to M, may be counted.
They range in height from eight to fifteen feet, and thus suggest a
rough measure for the amount of valley deepening during a swing of the
river southward across the valley and back again. This maximum num-
ber is evidently dependent on the numerous ledges here discovered at all
levels from highest to lowest. Although no other part of the valley
shows so many terraces, it must be concluded that flood plains, con-
tinuous with the remnants here preserved, were made far up and down
the valley ; and hence that the river was essentially at grade during the
whole process of valley degradation. Two terraces at the top of this
flight, in the re-entrant east of Pochassic Street, exhibit minor re-en-
trants of small radius and large arc near H and H', comparable to the
curves of the present river, thus indicating that no significant change of
volume has occurred since the work of terracing began. A broad
terrace plain stretches back of Perry's spur, K, and four low terraces
rise above it to higher levels, showing that four broad northward swings
were here executed. The fifth terrace (counting from the top of the
series) runs forward to Perry's spur, because the highest ledge of that
spur was discovered when the river was making its fifth northward
swing. It is worth noting that several defending ledges in this spur
would be unseen but for road and railroad cuts. The fourth terrace
swings forward in a long sweeping curve to the apex of Brown's spur,
because the summit ledge was there found by the fourth northward
swing. Only two distinct terraces occur on the high plain back of
Prospect spur, because the ledges in that spur rise still higher than in
Brown's spur. In a word the river has always shown a capacity for
broad swinging until it became hampered in its movements by coming

DAVIS : KIVER TERRACES IN NEW ENGLAND. 333

upon previously buried spurs. Brown's spur is peculiar hi being closely
trimmed on the down-valley side as well as on the up-valley side.
Prospect spur has a terraced re-entrant, C, scoured out at mid-height
with small radius and large arc, far back on its up-valley side ; that is,
a meander of the river has there been twice compressed against defend-
ing ledges, after the style of Figure 30. Elsewhere the meanders
seem to have slipped past the defending ledges, after the style of
Figures 25 to 28.

The terraces on the south side of the valley are in several cases de-
termined indirectly by the ledges on the north side. This is most
distinctly the case where "the river formerly swept forward from the
lowest and furthest forward of the Pochassic ledges, M, and consequently
cut out one of the deepest re-entrants on the south side of the valley, P.
A single scarp now descends from the high-level plain into this strong
recess. Similar but less manifest relations are suspected elsewhere ;
thus K', K", K " on the north may correspond with S', S", S'" on the
south. Conversely, a number of low-level terraces remain on the south
side of the valley south of Brown's spur, perhaps because the repeated
northward swings of the river into the largest northward re-entrant,
that between Brown's and Prospect spurs, have not required their
removal. The numerous free cusps here found exhibit the features al-
ready deduced as of common occurrence. It is intended to make
a close measurement of the slopes of these terrace plains in the hope of
correlating the now separate remnants of single flood plains, and thus
tracing the history of the terracing process in some detail.

Little River. A few words may be said about Little river, although
the southern side of its valley has not been closely studied. The valley
of this stream is divided into three sections by two barriers of sandstone,
next up-stream from which are considerable bodies of till. The till has
been cut down to grade with the sandstone barriers, but the valley in
the till is held to a small width, practically without terraces. Relatively
few terraces are found even where the valley is bordered by stratified
drift. In explanation of this it should be noted that Little river is
smaller than the Westfield, and that a small stream must be hurried in
attempting to keep pace with the degrading action of its master. Hence
the smaller stream will have little opportunity for lateral swinging and
terracing so long as it runs through loose drift to a more actively degrad-
ing master stream. There are two conditions under which opportunity
for lateral swinging will be presented to the smaller stream. First, when
the master stream has effectively ceased degrading its valley. This is

334 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

now the case with the Westfield, because it has cut clown upon a hai'd-
rock barrier in the trap-ridge notch ; and it is probably for this reason
that the lower section of Little river has swuug so broadly and opened
the extensive valley floor already described as forming the southern
part of the open basin east of Westfield. Naturally enough, then, the
enclosure of the broad valley floor on the south — where Little river is
alone responsible for the form of its border — is nearly everywhere a
single high-scarp terrace with numerous one-sweep or two-sweep cusps.
In other words, Little river is swinging on its present flood plain
more broadly than it has at any earlier time during the process of deg-
radation. Second, whenever the smaller stream becomes superposed
upon a rock barrier, its work in the next up-stream stretch proceeds
at its own rate, entirely independent of that of the master stream.
Hence the valley floor in such a stretch tends to widen and thus to
under-cut all the narrower flood plains formed in earlier stages of deg-
radation. This is the case with both the second and third sections of
the Little river valley. It has a well-opened valley floor usually enclosed
by single terrace scarps that rise to the full height of the upper
plain, so far as I have followed them. The simple scarps have well-
developed re-entrants and cusps, showing an active lateral swinging of
the stream at present grade, but not giving indication of an equivalent
swinging at any higher flood-plain level, and hence not giving support
to the opinion that the river is to-day of an enfeebled constitution.

Valley of Saxtons River, Vermont. Saxtons river enters the Con-
necticut from the west at Bellows Falls, Vermont, and shows a beautiful
variety of terrace foi*ms for some three miles above its mouth. Figures
40 and 41, separated by an unrepresented interval of about half a
mile, give rough illustration of these features. Careful survey would
undoubtedly show that the sketch-maps need many changes in de-
tails, but it is believed that the relative positions of terraces, with
their free and defended cusps, are shown with sufficient accuracy for
the purposes of the present discussion. The chief points here illus-
trated are as follows : —

Western Section. In the up-stream or western section, Figure 40,
there are numerous ledges, but none of them have acted as local base-
levels. The present valley floor is graded with respect to a heavy rock
barrier a little east of the limit of Figure 40, and at the western border
of Figure 41. In the three strong ledges, M, Q, R, Figure 40, on the
north side of the valley, the rocks are schists, with strong dip to the
northeast, and hence with bold outcrops to the southwest. The stream

Fig. 89. Diagram of the Terraces of Westfield river, in bird's eye perspective,
looking northeast.

<3$sn2®gflnr

-FICLD

wt*5

DAVIS : RIVER TERRACES IN NEW ENGLAND. 335

has swung against the steep face of these ledges, sweeping them prac-
tically free from drift on the up-valley side down to modern flood-
plain level, but fine flights of stepping terraces are preserved on the
down-valley side of each ledge, where the trailing remnants of suc-
cessive flood plains have been defended from stream attack. At least
ten different terrace levels can be counted adjoining ledge M. The
third and fourth levels from the top are pleasantly shaded by a pine
grove, and are used as a picnic ground, access to which is conveniently
given by an electric railroad on the valley floor. Some of these
terraces may have beeu carved by a small stream that here enters
from the north, but in any case they have all been developed with
respect to graded flood plains of the main stream. Their vertical
interval ranges from five to ten feet, which may be taken here, as in
other cases, to represent the amount of deepening that the valley floor
suffered between two northward swings of the stream. The value of
the ledges is most manifest ; they defended the upper terraces from being
consumed when the lower terraces were cut by the returning stream.

Ledges Q and R present similar features in flights of eight and six
steps respectively. The river is to-day swinging vigorously against the
base of ledge M. The modern flood plain reaches the base of Q and R,
and is opened northward between M and Q in a space that seems to be
comparatively free from ledges. The ledges here outcropping on a low
terrace at N and 0 seem to have served the double purpose of stopping
the northward swinging of the main stream, and of limiting the east
and west swinging of the side stream at that level. I have not closely
examined the terraces up-valley from M, but at least one of the blunt
cusps there seems, when seen from the terrace on the opposite side of
the valley, to be defended by a ledge at present flood-plain level. Down-
valley from R, the valley side is heavily wooded for quarter of a mile.
Then it closes in as numerous ledges and boulders make their appear-
ance about S, near the main road bridge.

Low-scarp terraces are wanting at high levels on the south side of the
valley. The upper plain descends by a single strong scarp, twenty feet
or more in height. It presents a number of sweeping re-entrants
between the defended cusps, A, B, C, D, and E. The A-B re-entrant is
floored by a rather uneven plain in which several indistinct terraces have
been cut on what seems to be at least in part a mass of till, for large
boulders are seen thereabout ; and this plain is cut off in front by two
terraces, whose blunt cusps from F to G appear to be in part determined
by ledges, in part by boulders. The small tributary stream that crosses

336 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

this re-entrant from the south has formed a fan on the high terrace plain
and again on the floor of the re-entrant, but it is now dissecting the
fans. No B-C re-entrant has been carved out, perhaps because till was
there discovered. Several ledges were encountered at lower levels
between G and H, against all of which the stream has swung most
faithfully. The valley floor would surely be wider to-day, had these
ledges not existed. A fine re-entrant wa3 swept out between the
defended cusps C and D, when the river ran at a height about ten feet
over the modern flood plain, and another effort was here made to widen
the valley floor at its present level, but as ledges are now discovered at
H and J, further forward than C and D, the lower re-entrant has not
quite consumed all of the earlier flood plain. A low terrace, caught on
ledges J and K, stands in front of the re-entrant between D and E. The
projection of the strong but low cusp at J as compared to that of the
blunt but high cusp at D is one of the best illustrations of the effect of
ledges that is found in this little valley. The river must have slipped
past the ledge at D, as well as past most other defending ledges here-
abouts ; but a compressed meander must have been caught for a time
on the ledge at J. Down-valley from E, a modern swing of the stream
has under-cut all the earlier terraces, and a full-height scarp is the
result.

These terraces are even better than those of the "Westfield for purposes
of field illustration, inasmuch as defended cusps here occur in abundance
on both sides of the valley. The narrowing of the interscarp space, as
the valley floor was degraded to lower and lower levels, is manifestly
due to the presence of the ledges. That the river was continuously
acting as a graded but degrading stream is sufficiently proved by the
fine flight of stepping terraces at M. That the preservation of the
successive terraces is not due to any shrinking of the stream from its
first intention as to valley widening, is proved by the vigor with which
it has opened the modern flood plain to as great a width as the numerous
ledges permit.

It was on seeing — in October, 1900 — the relation of the defended
cusp of the little terrace at F to the corresponding defended cusp of the
next higher terrace a little farther back at A, that the value of ledges
in determining terrace pattern and in preserving the upper terraces from
later attacks of the stream first came to my mind. The manner in
which this explanatory idea first took shape was as good an example of
the sudden invention or birth of theory as I have ever experienced, for
the theory was essentially complete at the moment of its first conscious

DAVIS : RIVER TERRACES IN NEW ENGLAND. 337

appearance ; since then it has only been confirmed by finding that it had
already been born to Miller, and by deducing its more minute conse-
quences as presented in Part 111. of this essay in order to confront them
with numerous examples of actual terrace forms, some of which are
described on these pages.

Eastern Section. The lower stretch of Saxtons river, Figure 41, gives
beautiful illustration of terraces produced by a stream that has oscillated
between two fixed nodes. At the upper node, the stream is narrowly
held by ledges at A and G. A little further up-stream is a rocky gorge
with cascades, from which the stream is diverted for water-power. The
lower valley becomes somewhat more open as the space widens between
the ledges B, C, on the south and J-H, L-K, on the north. The small
re-entrants between these ledges nearly everywhere bear the marks of
having been energetically swept back as far as possible by the stream
at various levels during the erosion of the valley. The stream has
swung northward at least nine times on the J-H group of ledges, and
southward at least seven times on the B group, where till seems to
supplement the restraint of rock.

On leaving the cascade and the rapids below it, the stream has graded
its course with respect to the eastern rock node between M and F-E ;
none of the ledges encountered on the way have had other effect than
in limiting the breadth to which the successive flood plains have been
opened during the degradation of the valley. That the degradation wa£
gradual, giving the stream abundant time for broad swinging and
wandering, right and left, is abundantly proved by the terrace remnants
of flood plains at various levels.

Passing the narrows at C, L-K, there is a broad stretch comparatively
free from ledges until the heavy ridge of rock, M, F-E, is encountered
close to the junction of Saxtons river with the Connecticut. The ridge
is now cut through by a narrow gorge, with falls on the down-stream
side whei'e the road and railroad bridges cross the stream : whether this
gorge is entirely the work of postglacial time, I cannot say.

An oval plain, known as the Basin farm, has been opened between
the upper and lower narrows, its smooth fields uniting with the curving
terrace scarps iu a most graceful and pleasing landscape. The Basin
plain probably had twice as great an area at the level of its mid-height
terraces as it now has at the level of the modern flood plain ; but this
reduction of area is not to be wondered at, iu view of the increasing
constriction imposed upon the swinging stream by the mutual approach
of the ledges C and K as lower and lower levels were reached.

338 . bulletin: museum of comparative zoology.

A few special features deserve mention. A southward deflection of
the stream from ledge K, and an increasing southward meander of the
stream in consequence of this deflection, with a southward swinging of
such a meander, has probably been responsible for some of the large
re-entrants on the south side of the basin ; but it is not yet apparent
why the re-entrants on the north should have been repeatedly worn
farther from the line between the nodes than on the south. The vari-
ous combinations of two-sweep cusps in the group of nine stepping
terraces next down-valley from ledge C is in every respect confirmatory
of the deductions (page 309). It is notable that remnants of terraces
at intermediate heights form little recesses in the cusps of later origin,
as might have been expected. The terraces in the wooded slope south
of C have not been traced out ; but a little farther east four low-scarped
terraces rise at D above the plain that forms the top of the nine-step
flight, thus making a series of thirteen steps. Curiously enough, these
four upper terraces, forming the upper members of the longest flight
that I have yet counted, swing northeastward from what seems to be
a free cusp, thus apparently imitating conditions similar to those of
Figures 15 and 17. No ledge is visible in the front of this group, but
as the scarp of the ninth terrace, here descending to the seventh, also
bows a little forward directly in front of the apparently free cusp, I am
much inclined to think that there is really some defence here, now
masked by a slipping drift cover. A soil augur test is proposed to settle
this doubt. A detailed map of this basin and its terraces would be well
worth preparing.

The Connecticut below Bellows Falls, Vermont. The Connecti-
cut river at Bellows Falls is superposed on a large body of rock, on
whose down-valley side the river is narrowed in rushing cascades and
rapids. A mile farther down-stream, three fine terraces are developed
at A, Figure 42, on the west side of the valley just below the mouth of
Saxtons river, S, and all seem to be defended by ledges ; the upper two
by ledges of the large rock ridge at the mouth of Saxtons river (these
are marked N, in Figure 41), the lowest by a ledge that stands several
hundred feet farther forward and is seen in the railroad cut a quarter of
a mile south of Saxtons river. This group of terraces is shown as seen
from the southeast in Figure B, Plate VIIL, of Russell's " Rivers of
North America." A little further south, the two lower terraces are
cut away westward in the formation of a broad valley floor, B. A full-
height scarp rises at the back border of the floor, and near its middle is
a blunt cusp determined by a strong ledge. The southern end of this

Fios. 40 and 41. Portions of Sax-
ton's river valley, Vermont.

Scale, about four inches to a mile.

DAVIS : RIVER TERRACES IN NEW ENGLAND.

339

Fig. 42. The Connecticut Terraces from Bellows Falls to Walpole. Scale, about
an inch and a half to a mile.

340 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

open section of the valley is enclosed by a high free cusp, C, at whose
apex the river is now working. Another open valley floor, I), follows the
free cusp, and is limited by a second free cusp, E, of less height but of
greater forward reach than the first. The meaning of these two free
cusps will be considered below.

Returning to Bellows Falls and following down the east side of the
valley past the entrance of Cold river, R, from the northeast, ledges are
found to be more numerous. A broad mid-height terrace was opened
until a ledge, F, was discovered in the base of the uppermost terrace
just south of the Alstead road ; the further southward extension of the
high terrace has not been followed. The mid-height terrace, followed
by the upper north-and-south road, was cut back by a much later swing
of the river near present flood-plain level, until a high ledge, G, was dis-
covered an eighth of a mile south of Cold river : the lower north-and-south
road skirts its base. Nearly a mile further south is a group of admir-
ably defended terrace cusps, H, up-valley from which the river has swept
out some vigorous curves, and on one of which — where the mid-height
terrace first advances near the Fitchburg railroad — the river was nearly
superposed ; the rapids that occurred here for a time were abandoned
as the river slipped off the northwest slope of the ledges. Near this
point the lower terrace advances to the river bank, on account of
the farther forward reach of other ledges, J ; one of them now out-
crops in the river bank and thus insures the enduring protection of
at least part of the low plain on which the railroad is here laid.1 It
is thus evident that the meander belt of the river has here been
constrained to take a more and more westward course as it cut
deeper and deeper ; and it is probably on this account that the first
large western re-entrant below Saxtons river has been so thoroughly
scoured out at a low level.

The lower eastern terrace is gradually cut back down-valley from
the foremost defending ledge ; and a broad low plain, K, is thus
opened to a half-mile width, after which it narrows towards the
bridge between Walpole (W) and Westminster (X). The mid-height
terrace continues down-valley from the abandoned rapids, first show-
ing an apparently free two-sweep cusp ; then a defended cusp, L, the
defending ledge of the latter being disclosed in a shallow railroad
cut at the base of the terrace ; after this the terrace is cut some

1 Tie shading to indicate a low terrace along the east bank of the river is acci-
dentally omitted in Figure 41 for three-fourths of an inch up-stream and half an
inch downstream from J.

DAVIS : RIVER TERRACES IX NEW ENGLAND. 341

distance back in a low-level re-entrant enclosed by a scarp concave
southward or down-valley. A little farther on, an isolated hill, M, trav-
ersed by the main valley road and crowned by a mansion, is sepa-
rated from the main slopes of the eastern side of the valley by a
deep -trench, N, of large sweeping curvature to the northeast and for
the most part apparently cut in till. The trench has a rather strong
under-cut slope on the outer side of its curved course, and a gently
terraced slope on its inner side. There can be no doubt that it
marks a former path of the river around a lobate spur, and that the
river was diverted from the trench at a comparatively modern,
though unrecorded, date by wearing through the narrow neck of the
spur at P, a little up-stream from Walpole bridge. The second one,
E, of the two free cusps above mentioned on the west of the river
is the unconsumed remnant of the neck of this spur, west of . the
cut-off; the isolated hill, M, is the terminal part of the spur north-
east of the cut-off. The following explanation of the relation between
the two free cusps, C and E, may be suggested.

At an early stage of the time during which the river was making
its great northeastward detour around the spur, EM, its course mav
be represented by the curve, a, a, a, a. The normal order of change
in these curves would develop a later course, b, b, b, b ; thus open-
ing out the two large westward re-entrants, B and D, and leaving the
free cusp, C, as yet unconsumed between them. Had this process been
normally continued, the free cusp, C, would have been in time worn
away by the down-valley sweeping of the first b meander ; but before
this was accomplished, the cut-off occurred at P. Now it may be shown
by a study of the detailed maps of the Mississippi River Commission that
when a cut-off occurs, a systematic series of changes is initiated, and that
these changes are extended up-stream as well as down-stream from the
cut-off. The essence of the changes is such that the straightening of the
course at the cut-off tends to straighten it elsewhere also, and that this
tendency is most active near the cut-off, and weakens with distance from it.
Following this principle, the course, b, b, b, b, will be changed toe, c, c, c,
and to d, d, d (the present channel). The river will thereby be with-
drawn from both of the westward re-entrants, around whose up-valley
curves it was probably flowing when the cut-off took plare. Thus the
free cusp, C, will be left unconsumed for a time at least. It may per-
haps be possible, by accumulating other examples of changes similar to
this one, to give some degree of verity to the rather hazardous explana-
tion here offered.

342 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

The Connecticut below Turners Falls, Mass. My field notes
here are hardly of sufficient detail to serve as the basis of a sketch map.
Suffice it to say, that for several miles down the river from Turners
Falls to the Fitchburg railroad bridge in Montague, there are numerous
examples of defended terraces, fully confirming the principles already
illustrated and suggesting some new ones, to which I hope to return in
a later essay. But concerning a stretch of the river southward from this
section, Emerson has written as follows : " The subsidence of the waters
of the Connecticut lakes to the present Connecticut river was very rapid.
... As a result, one goes down ■ — through the whole length of the
Montague Lake, which was well filled up in the flood time, except in its
southern portion — by a great scarp to the series of erosion terraces of
the modern river, the highest of which rise but a few feet above the level
of the flood plain" (725). 1

This conclusion as to the " very rapid " change of level by which the
erosion of the present valley floor in the former drift filling was initiated
seems to be based entirely on the feature here noted, namely, the great
scarp by which descent is made from the high drift plain to the low ter-
races of the modern river. The conclusion is so directly opposed to the
one that I have reached in the course of the present study that it has
been considered with some care. In the end I am led to doubt its
validity for the following reasons.

First, the occurrence of the single great scarp does not necessarily
prove that the river suddenly cut its channel down from the level of the
"lakes" to about that of the modern flood plain. The single great
scarp is here, as in other examples of the same kind, perfectly consistent
with a leisurely degradation by the river, and with the production of
numerous flood plains during the degradation, provided only that the
modern swinging of the river has been greater than the swinging at
higher levels, whereby all remnants of the earlier flood plains shall have
been destroyed. It has been shown that this is habitual with a number

1 The stratified drift deposits of the Montague and other basins are interpreted
as lacustrine by Emerson. Their fine texture and even stratification certainly
indicate deposition in quiet water, but in the absence of any well-proved barrier
between the basins of the middle Connecticut valley and the sea, there seems to be
a possibility that the water bodies were of the nature of narrow-mouthed bays,
with surface at sea-level, rather than normal lakes, with outflowing streams de-
scending to sea-level.

If the reader should consult the original text of this reference, he should note
that the first line of page 726 seems to have been misplaced from the bottom of
that page.

DAVIS : RIVER TERRACES IN NEW ENGLAND. 343

of streams wherever they are free to swing in loose drift of fine texture ;
and it appears from Emerson's description that such was the case with
the Connecticut while it was sweeping away the fine silts of the Mon-
tague " lakes."

Second, the leisurely lateral swinging of Saxtons and Westfield
rivers at high levels, as recorded by the upper members of the terrace
flights in the valleys of those tributaries of the Connecticut, show that
they were not hurried in the early stages of their work of degradation ;
yet hurried they must have been had the master river suddenly in-
trenched itself in the weak drift filling of its aggraded valley. There
are, to be sure, certain rock aud till barriers between these terrace
flights and the junction of their streams with the Connecticut, and such
barriers might separate a quickly degrading trunk river from slowly de-
grading tributaries ; but it is believed that the barriers are too low to
have been encountered until after the highdevel terraces had been
carved.

Third, there are certain points where the Connecticut itself exhibits
stepping terraces at altitudes of at least eighty or more feet above its
present level. The best of these are at East Deerfield, two miles south
of Turner's Falls, where the Fitchburg railroad crosses a group of ter-
races between the mouth of the Deerfield and the east-south bend of
the Connecticut. Here an eastern profile from the high terrace south
of the railroad station to the river crosses five terraces, the height of
whose scarps I have estimated at 30, 25, 15, 18, and 35 feet (total, 123
feet). If the profile be taken northward from the high terrace, seven
scarps are passed, including the descent from the lowest plain to the
river, with heights estimated at 30, 25, 10, 5, 15, 10, and 15 feet
(total, 110 feet). The locality of these terraces is openly connected
with that of the high scarps in the Montague silts ; and hence a lei-
surely process of degradation with repeated lateral swinging is probable
there as well as here. The only essential difference between the two
localities is that the terraces are for good reason preserved in East Deer-
field, where the river has become increasingly constrained by successive
discoveries of sandstone ledges, while they have been destroyed further
south where the river has been free to swing at modern levels.

The Connecticut near Springfield, Mass. There are few ledges
exposed hereabouts, and few stepping terraces. The high terrace that
is nearly continuous on the east side of the valley from Springfield up
to Chicopee, is now swept by the river for a part of its length, and
within this stretch a sandstone ledge is seen in the river bank. In
vol. xxxviii. — no. 7 5

344 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

the southern part of Springfield thei-e are several ledges and some ex-
posures of bouldery till, by which the opening of the valley to the east
has been restrained, and a little further south, by the mouth of Pecowsic
brook, a strong ledge deflects the river to the southwest.

No ledges are found in the terrace scarps on the western side of the
valley hereabouts, although several rather well-formed cusps project
forward towards the river : one at the grounds of the Country club ;
another at the old Meeting-house north of West Springfield and a third
between West Springfield and the Agawam. The Westfield river
enters the main valley in the re-entrant between the second and third
cusps, while south of the third cusp the Connecticut has repeatedly
scoured out re-entrants from which it has been withdrawn by short-cuts
or cut-offs. I am inclined to think that the Connecticut has been
pushed eastward by the action of the Westfield ; that it has therefore
repeatedly swung westward from the Pecowsic ledges, and that the
Agawam re-entrants are thus to be explained. If so, they fall into the
same class with those of the Westfield in the re-entrant next west of
the trap-ridge notch (page 330). The southernmost of the three free
cusps on the west side of the Connecticut valley, below the entrance of
the Westfield, would thus correspond with the free cusp on the south
side of the Westfield, below the entrance of Little river. The other free
cusps of the Connecticut may perhaps come to find an explanation in a
process similar to that suggested for the free cusps between Bellows
Falls and Westminster (page 341).

There is nothing in this stretch of the river to suggest a significant
diminution of volume since terracing began. The frequent occurrence of
high single scarps would on the other hand suggest that the river is
to-day demanding a breadth of swinging as great as or greater than it
ever did before.

The Merrimac between Concord and Manchester, N. H. The
Merrimac, near Concord, has opened a broad flood plain, on which a
number of former meanders are now represented by ox-bow lakes. On
the east the plain is commonly bounded by a single high scarp, which
the river is actively under-cutting at one point. On the west there is a
single high scarp bordering part of the plain north of the city ; but a
few ledges appear, and the scarp is divided into several terraces as the
city is entered. Passing down the valley (southward) ledges appear more
frequently, the breadth of the flood plain gradually decreases, and ter-
races appear in increasing numbers. The valley about Concord is one
of the best examples for illustration of the capacity of an unconstrained

DAVIS: RIVER TERRACES IN NEW ENGLAND. 345

river to open a broad flood plain enclosed by strong scarps, as typified
in Figures 2 and 22 ; while further south the valley exhibits the com-
plete control exercised upon a swinging river by the chance discovery
of ledges during the progress of degradation. This process is shown
to have been gradual by the preservation of flood-plain remnants at
various heights, wherever ledges are present to defend their bases ; yet
so complete is the destruction of the plains at intermediate levels just
above Concord that one might there infer that the river had had no op-
portunity to swing laterally until the opening of the present flood plain
was be<nin.

346 bulletin: museum of comparative zoology.

V. Bibliography.

Adams, C. B.

Second Annual Report on the Geology of the State of Vermont. Burlington,
1846.

Davis, W. M.

The Drainage of Cuestas. Proc. Geol. Assoc. (London), XVI., 1899, 75-93.

Diller, J. S.

Westfield during the Champlain Period. Amer. Journ. Sci.., XIII., 1877,
2(52-265.

Emerson, B. K.

Geology of Old Hampshire County, Massachusetts. U. S. Geol. Surv.,
Monogr. XXIX., Washington, 1S98.

Gilbert, G. K.

Report on the Geology of the Henry Mountains. Washington, 1877-

Hitchcock, E.

Illustrations of Surface Geology. Smithsonian Contr. to Knowledge. 1857.

Miller, H.

River Terracing; Its Methods and their Results. Proc. Roy. Phys. Soc.
Edinb., VII., 1883, 263-305.

Shaler, N. S.

Eluviatile Swamps of New England. Amer. Journ. Sci., XXXIII., 1887,
210-221.

The following Publications of the Museum of Comparative Zoology
are in preparation : —

Reports on the Results of Dredging Operations in 1877, 1878, 1879, and 1880, in charge of Alex-
ander Agassiz, by the U. S. Coast Survey Steamer " Blake," as follows: —

E. EHLERS. The Annelids of the " Blake."

C. HARTEAUB. The Comatuhe of the " Blake," with 15 Plates.

H. LUDWIG. The Genus Pentacrinus.

A. MILNE EDWARDS and E. L. BOUVIER. The Crustacea of the "Blake."

A. E. VER1UEL. The Aleyonaria of the " Blake."

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in charge of
Alexaxder Agassiz, on the U. S. Fish Commission Steamer "Albatross," from August,
1899, to March, 1900, Commander Jefferson F. Moser, U. S. X., Commanding.

Illustrations of North American MARINE INVERTEBRATES, from Drawings by Burk-
hakdt, Sonrel, and A. Agassiz, prepared under the direction of L. Agassiz.

LOUIS CABOT. Immature State of the Odonata, Part IV.
E. L. MARiC. Studies on Lepidosteus, continued.

" On Arachnactis.

R. T. HILL. On the Geology of the Windward Islands.
W. McM. WOODWOUTH. On the Bololo or Palolo of Fiji and Samoa.
A. AGASSIZ and A. G. MAYER. The Acalephs of the East Coast of the United States.
AGASSIZ and WHITMAN. Pelagic, Fishes. Part II., with 14 Plates.
J. C. BKANNER. The Coral Reefs of Brazil.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission Steamer
"Albatross," Lieutenant Commander Z. E. Tanner, U. S. N., Commanding, in charge of
Alexander Agassiz, as follows: -

A. AGASSIZ. The Pelagic Fauna. H. LUDWIG. The Starfishes.

" The Echini. j P. McMURRlCH. The Actinarians.

The Panamic Deep-Sea Fauna. E. L# MARK. Branchiocerianthus.

K. BRANDT. The SagittsB. JOHN MURRAY. The Bottom Specimens.
" The ThalassicolaB. _ ■ • • . _ , , TT .

.„;'', P. SCHIEMENZ. The Pteropods and Hete-
C. CHUN. The Siphonophores.

" The Eyes of Deep-Sea Crustacea. ropods.

W H DALE. The Mollusks. THEO. STUDER. The Alcyonarians.

H. J. HANSEN. The Cinipeds. M. P. A. TRAUTSTEDT. The Salpid* and
W. A. HERDMAN. The Ascidians. Doliolidse.

S. J. HICKSON. The Antipathids. E. P. VAN DUZEE. The Halobatidae.

W. E. HOYLE. The Cephalopods. H B WARD. The Sipunculids.

G. VON KOCH. The Deep-Sea Corals. H y WIESON. The Sponges.

C. A. KOFOID. Solenogaster. w McM ^OODWORTH. The Nemerteans.
R. VON EENDENFEED. The Phospho- „ The Annelk,s

rescent Organs of Fishes.

PUBLICATIONS

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletin Vols. I. to XXXVII. ;
of the Memoirs, Vols. I. to XXIV.

Vols. XXXVIII., XXXIX., XL., and XLI. of the Bulletin,
and Vols. XXV., XXVI., XXVII. , and XXVIII. of the Memoius,

are now in course of publication.

The Bulletin and Memoius are devoted to the publication of
original work by the Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural Histoiy, and of work by specialists based
upon the Museum Collections and Explorations.

The following publications are in preparation : —

Reports on the Results of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander C. D. Sigsbee, U. S. N., and Commander J. R. Bartlett, U. S. X.,
Commanding.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission
Steamer " Albatross," Lieut. Commander Z. L. Tanner, U. S. N„ Com-
manding, in charge of Alexander Agassiz.

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in
charge of Alexander Agassiz, oil the U. S. Fish Commission Steamer
" Albatross," from August, 1899, to March, 1900, Commander Jefferson E.
Moser, U. S. N., Commanding.

Contributions from the Zoological Laboratory, Professor E. L. Mark, Director.

Contributions from the Geological Laboratory, in charge of Professor N. S.
Shaler.

These publications are issued in numbers at irregular inter-
vals ; one volume of the Bulletin (8vo) mid half a volume of the
Memoirs (4to) usually appear annually. Each number of the
Bulletin and of the Memoirs is sold separately. A price list
of the publications of the Museum will be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

31^ Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 8.

THE FORAMIXIFERA AXD OTHEfl OffrMMISMS 1-^ THE
RAISED REEFS O* FIJI.

By R. L. Sherlock.

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

PRINTED FOR THE MUSEUM.

March, 1903.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXVIII.

GEOLOGICAL SERIES, Vol. V. No. 8.

THE FORAMINIFERA AND OTHER ORGANISMS IN THE
RAISED REEFS OF FIJI.

By R. L. Sherlock.

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

PRINTED FOR THE MUSEUM.

March, 1903.

Iff :.

No. 8. — The Foraminifera and other Organisms in the Raised
Beefs of Fiji.1 By R. L. Sherlock.

The rocks described were obtained from various islands in the Pacific
Ocean, but chiefly from the Fiji and Tonga Groups.

The specimens from Eua, Tonga Tabu, and Vavau in the Tonga Group,
Makatea and Niau in the Paumotus, and Guam in the Ladrones were
collected by Mr. Alexander Agassiz, and those from Niue by Prof. T.
Edgeworth David, of Sydney, N. S. W. The remainder from Mango,
Ngillingillah, Vatu Vara, Nainuka, Yathata, Kambara, and Singatoka (in
Viti Levu) all in the Fiji Group, were collected by Mr. E. C. Andrews
for Mr. A. Agassiz.

The structure of the islands has been described by Mr. Agassiz in his
" The Islands and Coral Reefs of Fiji." Bull. Mus. Coinp. Zool., XXXIII.,
1899, and in a series of letters published in the American Journal of
Science for 1898 and 1900. Mr. E. C. Andrews has given an account
of the Fiji Group in his " Notes on the Limestones and General Geology
of the Fiji Islands, with special reference to the Lau Group." Based
upon Surveys made for Alexander Agassiz. Bull. Mus. Comp. Zool.,
XXXVIII. , 1900. Mr. J. J. Lister gives some account of the Tonga
Group in his " Notes on the Geology of the Tonga Islands." Quart.
Journ. Geol. Soc, London, 1891.

Raised terraces of limestone occur in each of the islands, and the ques-
tion has arisen whether the limestones are composed of recent reef-build-
ing corals, or are of Tertiary age and of a different origin from the reefs
now growing round the islands.

The chief interest of the terraces lies in the light they may be expected
to throw on the question of the origin of atolls. On the theory of slow
subsidence, the raised limestones would be composed throughout essen-
tially of shallow-water reef-building corals. If, however, the limestones
have not this composition, some other theory must be found to explain
their formation in these particular cases.

The Tertiary age of certain of the rocks has been proved in two cases
by the discovery of Orbitoides, and in one of these cases the rock is

1 Contribution from the Geological Laboratory of the Royal College of Science,
London.

vol. xxxviii. — no 8 1

350 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

shown by the specific characters of the organism to be Miocene. Although
a few of the rock-sections are composed of coral, and corals are also pres-
ent in some others, yet in the majority of cases they are absent ; Algae
and Foraminifera make up the bulk of the rocks.

As regards the origin of the islands, this appears to have been some-
what as follows, according to Andrews (Bull. Mus. Comp. Zool.,
XXXVIII.) : First of all in Tertiary times bedded limestones were
deposited on the ocean floor on which, in many cases (e. g. Mango),
volcanic materials, bursting through, were heaped. Wherever the
accumulations reached sufficiently near the surface, corals began to grow
and to form a reef. Intermittent elevation followed, each pause allow-
ing the corals to form a new reef growing seawards on its own talus, and
which, on further elevation, was transformed into a raised terrace. A
succession of terraces is formed in this manner, amounting in the case
of Yathata to as many as six, whilst the modern reefs, should elevation
be repeated, will form another. The foundations of the terraces may
be the bedded limestones and volcanic accumulations, or very often
masses of rubble representing the talus derived from the terrace itself.
The rubble at Xiue is bedded and dips towards the sea at an angle of
about 40 °. It will be seen from the mode of origin that whilst the
oldest terrace is the highest one, and the rubble foundations are of the
same age as the terrace immediately above, the bedded limestones form-
ing the core of the island follow the ordinary rule that the oldest bed
is at the base.

The organisms found comprise fifteen genera of Foraminifera, with seven
other genera doubtfully present, besides algae, corals, echinoderms, mol-
lusca, including Tunicata and Polyzoa, with an occasional annelid. The
genera of Foraminifera are as follows :. —

Miliolina, Orbitolites, Textularia, Gaudryina, Globigerina, Planorbulina,
Discorbina, Truncatulina, Carpenteria, Polytrema, Tinoporus, Gypsina,
Amphistegina, Heterostegina, and Orbitoides.

The genera doubtfully present are : Ammodiscus, Hastigerina,
Sphoeroidina, Spirillina, Pulvinulina, Calcarina and Anomalina.

For aid in the determination of doubtful forms I have to thank Mr.
F. Chapman, who has given me valuable assistance.

FIJI. — Mango. An island of composite structure composed partly
of limestone and partly of volcanic rocks. It has been upheaved to the
extent of about 500 feet, and there are remains of a terrace at about 200
feet, showing a pause in the elevation.

FORAMINIFERA IN THE RAISED REEFS OF FIJI.

351

303, No. T, ht. 6 ft. The only recognizable organism is Litho-
thamnion. This is quite fresh although the rock has undergone
dolomitization.

305, No. B, ht. 250 ft. Indistinct corals are the sole visible organ-
isms. The rock is greatly altered.

302, No. C, ht. 280 ft. Lithothamnion is well preserved. Other organ-
isms have been almost obliterated, but Halimeda can be distinguished.

Foraminifera : 1 Gypsina.

290 ft. This rock also is much altered. Lithothamnion is still dis-
tinct but is altered marginally.

Foraminifera : Orbitolites.

313, No. 23, ht. 298 ft. Lithothamnion and Halimeda are abundant.
The rock has undergone considerable recrystallization, and even the Litho-
thamnion has been affected by the change.

Foraminifera : Polttrema. A small fragment of this form growing
on Lithothamnion is the only foraminifer visible.

306, No. a, ht. 300 ft. The organisms are very indistinct in conse-
quence of dolomitization. Lithothamnion, Halimeda, and corals occur.

Foraminifera : Polytrema planum Carter. A large mass of this
occurs fairly well preserved.
Amphistegina lessonii d'Orb.
Orbitolites.
Textularian form.
310 ft. This section is the most interesting of the series. It shows
numerous well-preserved specimens of Orbitoides sumatrensis, thereby

Fig. 1. Mango, 310 ft. Orbitoides sumatrensis Brady, X 38. The organism is
of a pale straw-color.

proving the Miocene age of the rock. The genus has only been found
in one other rock-specimen, that from Namuka, 30 ft. Polyzoa occur.
Foraminifera : Orbitoides sumatrensis Brady.

352 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Amphistegina lessonii d'Orb.

Polytrema MiNiAGEUM Pallas, sp. intergrown with
Lithothamnion.

MlLIOLINA.

Textularia.

1 Gaudrtina.

320 ft. A coral rock with Lithothamnion.

Foraminifera : 1 Gypsina inhosrens Schultze, sp. A long wander-
ing form which may possibly be a Polytrema.
350 ft. No organisms are recognizable with certainty.
370 ft. Fragments of echinoderms are abundant; also Lithothamnion ;
Gasteropoda are present. The rock is fragmentary and the organisms
are for the most part in a fresh condition.

Foraminifera : Gypsina inhcerens Schultze, sp.

Carpenteria ; numerous fragments.

? DlSCORBINA.

316, No. B, ht. 400 ft. The organisms are very indistinct in conse-
quence of doloraitization. As usual Lithothamnion has proved most
resistant to alteration. Echinoderm fragments, corals, and Lithotham-
nion are recognizable besides the following
Foraminifera : Amphistegina lessonii, d'Orb.
Polytrema.
1 Heterostegina. Small fragments probably belong

here.
1 Carpenteria. A fragment.

Ngillingillah. The island is apparently composed of coral rock
500 ft. thick, but Mr. Andrews finds that the island consists of bedded
limestone, covered by coral rock and reef-debris at most 200 ft. thick.
There are signs of three, or possibly four, periods of upheaval.

318, No. B, ht. 25 ft. The section consists of dolomitized coral, the
crevices of which are filled with a black mud composed of organic frag-
ments amongst which Lithothamnion, echinoderms, and a rotaline foram-
inifer are recognizable. Tunicate spicules are also probably present.

308, No. 43, ht. 215 ft. The rock has undergone dolomitization and
the only organisms recognizable with certainty are corals.

301, No. 77, ht. 475 ft. The slide shows Lithothamnion and Poly-
trema growing in alternating layers. Other organisms have been obliter-
ated by dolomitization. .

Foraminifera : Polytrema planum Carter.

FORAMINIFERA IN THE RAISED REEFS OF FIJI. 353

Vatu Vara. Of purely limestone composition. The island forms a
steep pyramid 1050 feet high, surrounded by a broad flat. At least
four periods of elevation are shown by terraces (three) and beach lines
at the 800 ft., 600 to 700 ft., 350 ft., 25 and 15 ft. levels.

500 ft. Organisms are few in number and very indistinct. Corals, a
gasteropod, Lithothamnion, and Carpeuteria fragments are recognizable.

Foraminifera : Carpenteria.

309, No. F v. v., ht. 650 ft. The organisms have been quite de-
stroyed by dolomitization. Lithothamnion is indicated by certain dark
patches.

1000 ft. Lithothamnion and echinoderm fragments. Corals.

Foraminifera : Gypsixa.

Nummuloid form. There are numerous shattered specimens of un-
certain genus, possibly Heterostegina.

Namuka. A small island composed entirely of limestone. It must
not be confused with Nomuka, which is in the Tonga Group.

30 ft. This rock is of considerable interest as, with the exception of
that from Mango, 310 ft., it is the only one containing Orbitoides.
The specimen is too indistinct to enable the species to be determined,
but the rock is shown by its presence to be not younger than the
Miocene. Lithothamnion occurs. All the organisms are altered.

Foraminifera : Orbitoides.

Orbitolites, one very indistinct specimen.
? Anomalina.

75 ft. This rock much resembles the last. It contains Lithotham-
nion and the following : —

Foraminifera: Orbitolites.
Rotaline form.

162 ft. A much altered rock. Lithothamnion and Polytrema are
the only organisms recognizable.

Foraminifera : Polytrema.

Yathata. Partly of volcanic origin. The flat top of the island is
840 ft. above sea level and is the highest of six terraces.

800 ft. The organisms are in a fragmentary condition. They
include corals, echinoderms, Lithothamnion, and Foraminifera.

Foraminifera : Polytrema, growing attached to Lithothamnion.
Carpenteria.

354 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Kambara. The main mass of the island is composed of volcanic
rocks. It is surrounded by a flat. There were at least two periods of
upheaval.

250 ft. A fragmentary and fairly fresh rock. It contains Litho-
thamnion, numerous spines and fragments of the tests of echinoderms
and a few fragments of Polyzoa.

Fig. 2. Kambara, 250 ft. Lithothamnion (basal view), a fragment, X 85.

Foraminifera : Polytrema, a large mass growing on Lithothamnion,
besides fragments.
Carpentebia, abundant and very well preserved.

C» *^*?>»v-

''J

Fig. 3. Kambara, 250 ft. Fragment of Carpenteria, X 85. The organism is
of a pale straw yellow color.

Amphistegina lessonii d'Orb.

Gypsina vesicularis var. monticulus Chapman.

Heterostegina depressa d'Orb, young specimens.

Orbitolites, small fragments.

Rotalian form.

FORAMINIFERA IN THE RAISED REEFS OF FIJI. 355

Singatoka. This place is situated on Viti Levu, which is the main
island of the Fiji Group and of a considerable size. It is largely com-
posed of volcanic rocks, and is much older than the smaller members of
the group. At Singatoka hard blue limestones, with a dip of 50°, under-
lie more friable limestones with a much lower dip (15°). The total
thickness of the beds approaches 1500 ft.

317, No. A, ht. reef. The organisms are not numerous, but of consid-
erable variety, including echinoderms, Lithothamnion, and Foraminifera.
Foraniinifera : Planorbulina larvata? P. & J.
Truncatulina.

Polytrema miniaceum Pallas, sp., one small fragment.
Carpenteria.

Amphistegina lessonii d'Orb.
304, No. C, ht. 251 ft. Contains small fragments of Lithothamnion
and Polyzoa.

Foraminifera : Discorbina.

Planorbulina, several specimens.
? Calcarina, the spurs are not seen.
Gypsina vesicularis P. & J., sp. var. monticulus
Chapman.

'_■ , : ■ • ' .. ■" i . •. "- "

Fig. 4. Singatoka, 251 ft. Gypsina vesicularis P. & J., sp. var. monliculus
Chapman, X about 50.

Carpenteria.

Amphistegina lessonii d'Orb.
314, No. D, ht. 268 ft. Similar to the last. Fragmentary Litho-
thamnion and Polyzoa.

Foraminifera : Truncatulina.
1 Planorbulina.
Carpenteria, small fragments.
Gypsina.

Amphistegina lessonii d'Orb.

1 Gaudryina, a textulariau form, probably belongs to
this genus.

356

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Niue (Savage Island). This island is intermediate in position be-
tween the Fiji and Tonga Groups. It contains three raised terraces at
about 80 ft., 100 ft., and 200 ft., respectively, which are composed of
coral rock dipping at 6° to 10°. Banks of coral debris, bedded, and
dipping seawards at about 40°, form the foundations of the terraces.

Fig. 5. Singatoka, 268 ft. Amphistegina lessonii d'Orh, X 85. The organism is
of a pale straw-yellow color.

315, No. 3, ht. 18 ft. This rock is very fresh and the organisms well
preserved, though fragmentary. Lithothamnion, echinoderms, Mollusca,
and the spicules of tunicates occur.

Foraminifera : Carpenteria.

PoLYTREMA MINIACEUM Pallas, Sp.

Gypsina.

m^'^^M

i'"'' ^ifr

.

'•'^'m'g

"**^5-,

y '•' 'y, \ '_

Fig. 6. Niue, 53 ft. Halimeda, X about 50.

310, No. 9, ht. 53 ft. Halimeda and alcyonariau spicules abundant.
Lithothamnion and tunicates are present.

FORAMINIFERA IN THE RAISED REEFS OF FIJI.

357

Foraminifera : Carpentaria, fragmentary.

Orbitolites, one small fragment.

Polytrema MINIAOEDM Pallas, sp., small fragments.

Fig 7. Niue, Vailoa, 15 ft. Reef rock. Orbitolites, X about 45. The organ-
ism is of a burnt sienna tint.

307, No. 11, lit. 65 ft. Composed of corals with mud-filled cavities.

311, No. 12 A, ht. 70 ft. Very rich in organisms. It contains
Halimeda, Lithothamnion, Gasteropoda, echinoderm fragments, and
tunicate spicules.

Foraminifera : Carpexteria.

Tinoporus baculatus Montfort, sp. abundant.

Fig. 8. Niue, 70 ft. Tinoporus baculatus Montfort, sp., X 90.

Orbitolites.

Polytrema, one specimen encrusts two tunicate
spicules.
1st terrace, 80 ft. Portion of a mass of coral.

Upper terrace, 120 ft. Abundant Halimeda. Also echinoderms,
Polyzoa, tunicate spicules, and Lithothamnion.

35S

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Foraminifera : Okbitolites, abundant.

Carpenteria, small fragments.
1 Polytrema or 1 Gypsina, fragments.
20 A, near top of 2d terrace, 190 ft. Vailoa Nine. Tunicate
spicules :ire very numerous. Halimeda is abundant and Lithothamnion,
Gasteropoda, Polyzoa, and echinoid spines
also occur.

Foraminifera : Miliolina.

Orbitolites, abundant.

Textularia.

Spirillina (or Ammodis-

cus).
Truncatulina.
Planorbulina.
Hastigerina 1.
Heterostegina depressa
d'Orb, numerous speci-
mens.
Polytrema, fragments.
312, Xo. 208, ht. 200 ft. Contains corals, Halimeda, echiuoderms,
and Lithothamnion. Tunicate spicules are very abundant.

Fig. 9. Niue, Vailoa, 190
ft., near top of 2d terrace.
Tunicate spicule (Leptocli-
num). Large specimen,
showing zonings, X 300.

Fig. 10. Niue, 200 ft. Globigerina
and spicules of Tunicata (Leptocli-
nuin), X 85.

Fig. 11. Niue, 200 ft.- Tunicate spi-
cule (Leptoclinum), showing secondary
growth, X 300.

Fig. 12. Niue, 200 ft. Tunicate spicule (Leptoclinum), a
very small specimen drawn to show the lower limit of size,
X300.

FORAMINIFERA IN THE RAISED REEFS OF FIJI. 359

Foraminifera : Orbitolites.
Globigerina.
Polytrema, encrusting.
Also rotaliue forms.

TONGA ISLANDS (FRIENDLY ISLANDS). Tonga Tabu.—
This is the largest of the Tonga group, and is twenty-two miles long.
It is crescentic in shape, the convex side looking south and terminating
in low cliffs. The concave side to the north is low-lying, and its shore-
line represents the border of the old lagoon. The island is entirely
calcareous. (Cf. J. J. Lister.)

Mt. Zion Hill, 50 ft. Contains echinoderm fragments, Gasteropoda
and Polyzoa.

Foraminifera : Globigerina.
Gaudryina.

1 Heterostegina (or 1 Amphistegina).
Carpenteria.

Polytrema mintaceum Pallas, sp.
Also rotaliue forms.

Eua. Eua is composed of two ridges running north and south, with
a valley eroded between them. On the eastern side there is a limestone
cliff rising in places to 1000 ft., with terraces on the projecting points.
They are three in number in the north, and five in the south. The
western ridge, which is also of limestone, is lower, and shows three ter-
races. The island is not an atoll, but an eroded platform (cf. A. Agassiz).
2d terrace, 120 ft. A few almost opaque fragments of Lithothamnion
and the

Foraminifera : Gypsina.

Carpenteria.

Amphisteginia lessonii d'Orb, very abundant.
Miliolina, a number of specimens.
Also a broad, short, textularian form.
3d terrace, 250 ft. Halimeda, Lithothamnion, Gasteropoda. None
of these are abundant.

Foraminifera : Polytrema.
Miliolina.
Carpenteria.
Inner Valley, 600 ft. Echinoderm fragments abundant. Lithotham-
nion and Polyzoa are present.

360 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Foramiuifera : ] Planorbulina or 1 Gypsina.

AMPHISTEGINA LESSONII d'Orb.

Yavau. — Vavau is elliptical in outline, highest in the northern por-
tion, where it reaches 500 ft. The southern shore is deeply indented
and denudation has separated off numerous small islands from the main-
laud. Four terraces are visible.

1st terrace. Organisms somewhat indistinct. They consist of Polyzoa,
echinoderms, and the following

Foramiuifera : Miliolina ferussacii d'Orb, sp.

Carpenteria, somewhat indistinct.

Globigerina.

1 Gaudryina, a textularian form, probably belongs to

this genus.
Gypsina.

Polytrema miniaceum Pallas, sp.
% Gypsina globulus Reuss, one indistinct specimen.
2d terrace, 1 310 ft. Lithothamnion, Polyzoa, Gasteropoda.
Foramiuifera : Textularia.

Gypsina, numerous specimens.
Carpenteria.

Polytrema miniaceum Pallas, sp.
Miliolina.
2d terrace, ? 350 ft. Polyzoa, echinoderm spines, Lithothamnion.
Foramiuifera : Miliolina, several specimens.
Carpenteria.
1 Globigerina conglobata Brady (or possibly Sphce-

roidina bulloides), one specimen only.
Amphistegina lessonii d'Orb, one specimen.
2d terrace, 360 ft. Numerous fragmentary organisms embedded in a
crystalline matrix. Lithothamnion is the most abundant organism.
Echinoderms, Gasteropoda, and annelids (?) are also present.
Foraminifera : Polytrema miniaceum Pallas, sp.
Carpenteria.

Amphistegina lessoni d'Orb.
Globigerina, one specimen enclosed in an annelid (?)

tube.
Gypsina.
3d terrace. Highest point, north shore. The only organisms are
alcyonarian spicules, and of these the rock is mainly composed.

FORAMINIFERA IN THE RAISED REEFS OF FIJI. 361

PAUMOTUS GROUP. Makatea (Mbtia or Aurora I.). An ele-
vated mass of coraliferous limestone with a comparatively narrow shore
platform cut from the base of the cliffs. It is the highest member of the
group, reaching 230 ft.

2d terrace. No organisms.

Basin, 150 ft.— 175 ft. Lithothamnion and Mollusca, the former
abundant.

Foraminifera : Polytrema miniaceum Pallas, sp.

Polytrema plaxum Carter. The two species of
Polytrema are encrusting forms, the P. planum is
here growing on P. miniaceum, and the latter on
Lithothamnion. The three organisms make up the
bulk of the rock.
Carpentaria.
Gypsina.

1 Amphistegina, one. specimen, which may possibly be
Heterostegina.
3d terrace, 200 ft. A coral rock which has undergone alteration.

Niau. — This island, like the others of the group, has been greatly de-
nuded. The Tertiary limestones do not occur at a higher level than
about 16 ft. It is the only member of the group which possesses a
lagoon shut off from the sea.

20 ft., half-way across the rim. A coral rock, a fragment, apparently
of a mollusc, occurs, but alteration has rendered it indistinct.

THE LADRONES. Guam. — The northern half consists of raised
limestones, which form vertical cliffs on the eastern side, 109 to 300 ft.
high. In the southern half, volcanic rocks, in places, burst through the
limestones. This part of the island reaches 1000 ft.

Mt. Makawia, near the summit. A fragmentary rock made up of or-
ganisms broken to a uniform small size. They comprise Lithothamnion,
Polyzoa, and

Foraminifera : Miliolina, fairly numerous.
Textularia.

Truncatulina, numerous specimens.
1 Spirillina, one doubtful specimen.
1 Carpenteria, numerous, rather indefinite

fragments.
Gypsina.
Polytrema.

362 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Of the forty-seven rock-sections examined, Lithothamnion occurs in
thirty-five, and is decidedly the most abundant and most widely dis-
tributed organism. Many of the slides in which it is not found have
been dolomitized, and any organisms present have been removed. Prob-
ably Lithothamnion was present in some of these rocks originally. It
should be mentioned that the term Lithothamnion is used throughout
in the older and wider sense, and not as a generic name.

Halimeda, the other calcareous alga present, is much less widely dis-
tributed. It occurs in nine sections, coming from Mango, Niue, and
Eua. It is much less stable than Lithothamnion, and this may account
to some extent for its more limited distribution.

Echinoderm fragments (largely echinoid spines) occur in seventeen of
the specimens, and take an appreciable part in the composition of the
rocks. Polyzoa were found in twelve slides from eight of the fourteen
islands. They are most abundant in the Vavau rocks.

Corals occur in fifteen sections, being represented in two cases by
alcyonarian spicules only. Since they undergo alteration rather easily,
they may have been present originally in some of the altered rocks,
but even taking this into account, they play a much less important
part in the formation of the limestones than might be expected. More-
over, a large proportion of the rocks were obtained from terraces
where the corals seem to be especially abundant, yet the actual bulk
of the terraces compared with the rest of the limestones is not
great.

Of the Foraminifera, Carpenteria occurring in nineteen sections, and
Polytrema in twenty-one, are the most widely distributed and also the
most numerous individually. Carpenteria has probably had an even
wider distribution than appears, as it seems to undergo alteration fairly
easily. The genus Polytrema is commonly represented by P. miniaceum
Pallas, sp., but P. planum Carter also occurs. Both forms are found at
times intergrown with Lithothamnion, forming a mass consisting of layers
of the different organisms alternating. Polytrema, however, encrusts
various objects besides Lithothamnion.

Gypsina was found in eleven sections, and is often very abundant.
Orbitolites occurs in ten sections, but is only abundant in the Niue rocks.
As a rule, not more than one specimen occurs in a slide. In sections it
possesses a burnt sienna color which is characteristic, although in some
cases where alteration has taken place it is block and opaque. Miliolina
occurs in eight sections, but does not make up a considerable part of any
of the rocks. It is commonest in Eua and Vavau sections.

FORAMINIFEEA IN THE RAISED REEFS OF FIJI. 363

Amphistegina occurs in eleven slides, and is often abundant, as is the
case in the rock from Eua, 120 ft.

Although rotaline Foraminifera occur in many of the slides, no form is
common. The most generally occurring genera are Planorbulina and
Truncatulina. Globigerina occurs in four slides. Orbitoides is of espe-
cial interest, as it is not found in rocks younger than the Miocene. In-
dividuals are very numerous in the section from Mango, 310 ft., and it
is noteworthy that the genus is entirely absent from the other Mango
specimens examined.

The Niue rocks are characterized by the great number of tunicate
spicules which they contain. The first described occurrence of these
organisms as fossils was in the Pliocene of St.
Erth, Cornwall, where Messrs. Kendall and Bell
found spicules referred by Dr. G. J. Hinde to
Leptoclinum.1 Professor Herdman had pre-
viously pointed out that calcareous spicules oc-
curred in many genera of Tunicata, and were as
capable of preservation as the aragonite skeletons
of other organisms.2 As, however, with the ex-
ception of the St. Erth beds above-mentioned.
, , , , , . Fig. 13. Nine, \ ailoa,

they do not seem to have been found in the ^ ft Tunicate spicule

fossil condition, their occurrence in the coral (Leptoclinum), surface
reefs of Niue becomes of interest. view, x o00.

Although Niue (Savage Island) is situated between the Tonga and
Fiji Groups, and the rocks of all three places resemble each other in
structure and composition, the tunicate spicules appear to be confined to
Niue (with the doubtful exception of Ngillingillah). They also occur,
however, in the rocks from Christmas Island collected by Dr. C. W.
Andrews, and in the Funafuti boring described by Dr. G. J. Hinde.
In all three cases, Niue, Christmas Island, and Funafuti, the spicules
seem identical and belong to the same genus, Leptoclinum Milne
Edwards, as occurs at St. Ei'th.

The number of spicules present in the rocks also varies widely. All
the slides which are not simply sections of a coral mass, show some
spicules, but they are particularly abundant in the black mud. In
many cases, circular areas of clear granular calcite occur having the
usual size of the spicules, but showing no structure. These clearly
represent the spicules which have been altered into calcite and their

J Quart. Journ. Geol. Soc. Lond., XLII. pp. 201-215.
2 Proc. Geol. Soc. Liverpool, 1884-85, p. 42.
vol. xxxvin. — no. 8 2

364 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

structure destroyed. They are found, for example, in sections of the
talus, where one would expect alteration to take place readily from the
fragmentary and therefore porous nature of the rock. It is probable
from this that tunicates took a more important part in the composition
•of the reefs than appears at first sight, owing to the ease with which
aragonite organisms are obliterated. The value of tunicates as rock-
formers may be estimated in a particular case by finding the fraction of
the surface area of a section of the rock covered by the spicules.
Choosing a section in which the tunicates were especially abundant,
they were found to form approximately 2 per cent of the whole mass.
In most of the slides they form a smaller proportion than this, and about
.5 per cent may be taken as the average for the Niue rocks.

In the Niue specimens the spicules are circular in sections and show a
clear outer zone of radiating fibres surrounding a dark granular area.
In many specimens the centre of the granular portion has a distinctly
radiating structure, aud is more temslucent than the remainder of the
central area, but in others the granular area extends to the centre. The
composition of the spicules is shown to be aragonite, by staining them
with cobalt nitrate.1 By this method, in which a polished slice of rock
is boiled with ordinary cobalt nitrate solution for some time, aragonite
is stained pink, but calcite is not affected, unless the boiling has been
continued for a very long time, when it may stain blue. The granular
portion of the spicule stains pink readily, but the outer zone stains only
on long boiling. This shows that the whole is aragouite, but that there
is some difference between the zones, probably due to a difference in the
.state of aggregation.

From the fact that the sections are invariably circular, the spicules
must be spherical. In a few cases the surface is seen, and appears tuber-
culated, the bases of the relatively lai-ge but low tubercles being in close
contact.

The spicules vary very greatly in size, as is shown by the figures, which
are drawn to the same scale. Many of the smallest sections show only
the outer fibrous zone, and are, therefore, probably tangential sections
of large spicules. But others, such as the one figured, show the central
granulated area occupying the same proportion of the whole as in the
larger forms, and these must be small spicules. The diameter varies in
length from fa to Tihs mm- (anout ^^ to ysVff in.). In a few cases, as
shown by staining, the outermost fibrous zone is absent.

Some recent spicules of Leptoclinum from Anticosti (Canada), very
i W. Meigen. Centralblatt Min., 1901, pp. 577-578.

FORAMINIFEKA IN THE RAISED REEFS OF FIJI. 365

kindly lent by Dr. G. J. Hinde for comparison, are very closely similar
to the Niue specimens. The only difference is that theAnticosti spicules
show in optical section slightly projecting tubercles which are not seen
in the great majority of the Nine spicules. But in a few of the latter
there are indications of the tubercles, especially in those seen in surface
view, as already mentioned. Dr. Hinde has isolated spicules from the
sandy material of the Funafuti boring. A few of these have long spines
and closely agree with the spicules of Leptoclinum figured by Professor
Herdman in his Challenger Report. The remainder are covered by
short tubercles (like the recent specimens from Anticosti), but probably
belong to the same species as the spiny ones, for Professor Herdman has
shown that two varieties do exist in the same animal in the case of some
recent species. The spicules in the Funafuti rock-sections are identical
with the Niue forms, so that these latter may be taken as belonging to
the genus Leptoclinum.

The following Publications of the Museum of Comparative Zoology
are in preparation : —

Reports on the Results of Dredging Operations in 1877, 1878, 1879, and 1880, in charge of Alex-
ander Agassiz, by the U. S. Coast Survey Steamer " Blake," as follows: —

E. EHLERS. The Annelids of the" Blake."

C. HARTLAUB. The Comatuhe of the " Blake," nith 15 Plates.

H. LUDWIG. The Genus Peutacrinus.

A. MILNE EDWARDS and E. L. BOUVIER. The Crustacea of the "Blake."

A. E. VERR1LL The Alcyonaria of the " Blake."

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in charge of
Alexander Agassiz, on the IT. S. Fish Commission Steamer "Albatross," from August,
1899, to March, 1900, Commander Jefferson F. Moser, U. S. X., Commanding.

Illustrations of North American .MARINE INVERTEBRATES, from Drawings by BuRK-
hakdt, Sonrel, and A. Agassiz, prepared under the direction of L. Agassiz.

LOUIS CA BUT Immature State of the Odonata, Part IV.
E. L. MARK. Studies on Lepidosteus, continued.

" On Arachnactis.

R. T. HIL>. On the Geology of the Windward Islands.
W. McM. WOODWORTH. On the Bololo or Palolo of Fiji and Samoa.
A. AGASSIZ and A. G. MAYER. The Acalephs of the East Coast of the United States.
AGASSIZ and WHITMAN. Pelagic Fishes. Part II., with 14 Plates.
J. C. BRANNER. The Coral Reefs of Brazil.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission Steamer
"Albatross," Lieutenant Commander Z. L TANKER, U. S. N., Commanding, in charge of
Alexander Agassiz, as follows: —

A. AGASSIZ. The Pelagic Fauna. H. LUDWIG. The Starfishes.

The Echini. j P. McMURRICH. The Actinarians.

The Panamic Deep-Sea Fauna. E. L. MARK. Branobiocerianthus.

K. BRANDT. The Sagittae. */^,,>.t -,,ttT.t> . .r mi t> t<. c

6 JOHN MURRAY. Ihe Bottom Specimens.

" The Thalassieolae. _ „„„..„

_ ,,1ITTXT „„ „. , . P. SCH1E.A1ENZ. The Pteropods and Hete-

C. CHUN. The Siphonophores. '

" The Eyes of Deei>-Sea Crustacea

W. H. DALL. The Mollusks. THEO. STTJDER. The Alcyonarians.

H. J. HANSEN. The Cirripeds. M. P. A. TRAUSTEDT. The Salpidas and
W. A. HERDMAN. The Ascidians. Doliolidas.

S. J. HICKSON. The Antipathids. E. P. VAN DUZEE. The Halobatidse.

W. E. HOYLE. The Cephalopods. H. B. WARD. The Sipunculi.ls.

G. VON KOCH. The Deep-Sea Corals. H V. WILSON. The Sponges.

C A. KOFOID. Solenogaster. W. McM. WOODWORTH. The Nemerteans.

R. VON LENDENFELD. The Phospho-
rescent Organs of Fishes.

The Annelids.

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletin Vols. I. to XXXVII. ;
of the Memoirs, Vols. I. to XXIV., and Vol. XXVIII.

Vols. XXXVIII., XXXIX., XL., and XLI. of the Bulletin,
and Vols. XXV., XXVI. , XXVIL, and XXIX. of the Memoiks,
are now in course of publication.

The Bulletin and Memoirs are devoted to the publication of
original work by the Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural History, and of work by specialists based
upon the Museum Collections and Explorations.

The following publications are in preparation : —

Reports on the Results of -Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander C. 1). Sigsbee, U. S. N., and Commander J. It. Bartlett, 0. S.N.,
Commanding.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission
Steamer " Albatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Reports on the Scientific Results of the Expedition to the Tropical Pacific, in
charge of Alexander Agassiz, on the U. S. Fish Commission Steamer
" Albatross," from August, 1899, to March, 1900, Commander Jefferson F.
Moser, U. S. N., Commanding.

Contributions from the Zoological Laboratory, Professor E. L. Mark, Director.

Contributions from the Geological Laboratory, in charge of Professor N. S.
S baler.

These publications are issued in numbers at irregular inter-
vals ; one volume of the Bulletin (8vo) and half a volume of the
Memoirs (4to) usually appear annually. Each number of the
Bulletin and of the Memoirs is sold separately. A price list
of the publications of the Museum will be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

•Yi of'

Harvard MCZ Libra

llll II

3 2044 066 302 613

Date Due

KfC:ii%tr?

^™

-v

~ 'jtr

: 4\*

v y I

Sk v

u ^

3 x.

'Si Jt*ii