Bulletin of the Museum of Comparative Zoology at Harvard College

t

K

m

^^^^^^:

^

''rr ' iiriiiiitiiiiia

ifc- ^

M^5

HARVARD UNIVERSITY

LIBRARY

OF THE

Museum of Comparative Zoology

BULLETIN

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT

HARVARD COLLEGE, IN CAMBRIDGE.

VOL. XVI.

(Geological Series, II.)

I If: f' r \,

CAMBRIDGE, MASS., U. S. A.

1888-1895.

Reprinted with the permission of the original publisher

KRAUS REPRINT CORPORATION

New York

1967

iv CONTENTS.

Page

No. 13. — Reports on the Dredn-ing Operations off the West Coast of Central
America to the Galapagos, . the West Coast of Mexico, and in the Gulf
of California, in charge of Alexander Agassiz, carried on by the U. S.
Fish Commission Steamer " Albatross," Lieut. Commander Z. L. Tanner,
U. S. N., commanding. V. Report upon Rocks collected from the Galapa-
gos Islands. By George P. Merrill. July, 1893 235

No. 14. — Contributions from the Petrographical Laboratory of the Harvard
University Museum. VI. A Basic^Dike in the Connecticut Triassic. By
L. S. Griswold. (1 Plate.) August, 1893 239

No. 15. — Notes on^the Geology of, the Island of Cuba, based upon a Recon-
noissance made for Alexander Agassiz. By Robert T. Hill. (9 Plates.)
April, 1895 . , 243

No. 1. — Contributions from the PetrograpMcal Laboratory of the
Harvard University Museum^ in Charge of J. E. Wolff.

On the PetrograpMcal Characters of a Dike of Diabase in the Boston
Basin. By William H. Hobbs.

In the region north of Boston occurs a most interesting series of mas-
sive rocks, which break through the slates and sandstones, and include
granite, quartz-porphyry, quartzless-porphyry, elaeolite-syenite, diorite,
porphyrite, diabase, augite-porphyrite, and gabbro. They have been
studied in greater or less detail by many observers, prominent among
whom are W. 0. Crosby, M. E. Wadsworth, and J. S. Diller.

The rocks which have afforded material for the present study belong
to a single dike, and may be seen in a series of exposures in Medford
and Somerville. They are coarsely crystalline rocks, and have borne
the names "syenite "and "diorite." They have in general been care-
fully distinguished from similar rocks of finer texture known as " green-
stones," which were shown by Wadsworth in 1877 to have about the
same composition as the coarser rock, and were considered by him as
identical with it. The finely crystalline rock seems to be more widely
distributed than the coai-se variety. In the present paper it has been
studied only at a few localities, where there was some promise of de-
ciphering its relations with the " diorite."

The age of these rocks has not been accurately determined, though they
have generally been considered post-triassic on account of their lithologi-
cal resemblance to the diabase of the Connecticut Valley. The slates
through which they have broken are probably identical with the Lower
Cambrian argillite of Braintree. Diller has furnished evidence to show
that the finely crystalline diabase ("greenstone") is the youngest of the
eruptives of this region, its dikes cutting those of the other rocks.^

Many mistakes have been made in determining the composition of
both the so-called " diorite " and " greenstone." The " diorite " was long

1 Bull. Mus. Comp. Zool. at Harvard College, VII. 179.

VOL. XVI. — NO. 1 . 1

2 BULLETIN OF THE

ago described from the Granite Street locality in Somerville "by J. F.
and S. L. Dana/ and by J. W, Webster,^ as made up essentially of the
minerals feldspar and hornblende. This was supported by Professor
Hitchcock, in his Survey of Massachusetts.' The mistake, which con-
sisted in taking augite to be hornblende, was farther repeated by W. W.
Dodge ^ and W. 0. Crosby.^ Professor "Wadsworth^ was the first to
apply the microscope to the study of this region, and was enabled to
determine correctly the general composition of tlie rocks. His study
included all the eruptives exposed in Somerville, and part of those out-
cropping in Medford, Maiden, Brighton, and Brookline. He pointed out
the connection of the outcrops, and indicated their general trend. He
showed that the feldspar is plagioclase, and that the prevailing non-
feldspathic constituent is augite. On the basis of an observed identity
in mineral composition he concluded that the fine-grained " greenstone "
is identical with the coarser " diorite." Professor Crosby, in his " Con-
tributions to the Geology of Eastern Massachusetts,"'' has quite accu-
rately outlined on his map the area in which these rocks are exposed.

The present article, being restricted mainly to the coarsely crystalline
diabase, repeats to some extent the work of Professor Wadsworth.
"While the results are in the main confirmatory, there are still important
points of difference, and some new facts have been determined.

Leaving out of consideration the vicinity of tlie Old Powder House
in Somerville, the field study has yielded but little. The attempt has
been mainly to add something to our knowledge of the petrographical
characters of the so-called " Mesozoic diabase of the Atlantic border."
Those who desire a full literature of the subject treated in this paper
should refer to the above-cited work of Professor Wadsworth.

Field Notes.

The dike of diabase which is here considered extends from Granite
Street in Somerville to Spot Pond in Stoneham, and probably beyond,
although no examination in the field was carried beyond that point.

1 Memoirs Amer. Acad., 1st series, IV. 16-3 (1818).
'^ Boston Jour. Pliil. and Arts, XL 282 (1824).
8 final Kep. Geol. of Massachusetts, 640-663 (1841).
* Proc. Boston Soc. Nat. Hist., XVIL 415 (1875).
^ Occasional Papers of Boston Soc. Nat. Hist., Vol. III.

^ Notes on the Petrography of Boston and Vicinity, by M. E. "Wadsworth. Proc.
Boston Soc. Nat. Hist., XIX. 217 (1877).

^ Occasional Papers of Boston Soc. Nat. Hist., Vol. III.

MUSEUM OF COMPAKATIYE ZOOLOGY. 3

No outcrops of the coarse" rock have been found south of the Granite
Street quarry. Professor Crosby has included the fine-grained diabase
"which crops out at the Pumping Station in Brighton, and similar rocks
in Brookline and Xewton, as a part of this dike ; but the great change
of strike required, and the long intervening distance without exposures,
are opposed to the supposition. From the Granite Street quarries to
the Old Powder House in Somerville, (a distance of about one and a half
miles,) the strike of the exposures is N. 25° "NV. From that point to
Spot Pond, the trend is N. 10° E. In Medford and Somerville the
country rock is argillite, Avhich has been thrown into gentle folds, the
dips of which seldom exceed 35°. A notable exception to this state-
ment is seen at the old slate quarry on Professor's Row, College Hill,
where beds strike ±N. 95° E. and dip ±72° to the south. The area
of coarse diabase, which has an average width of about two thousand feet,
is never found in contact with the slate. The exposures of diabase al-
most invariably show the well-known weathering to boulders in situ,
though this is best observed at Pine Hill. North of High Street in
Medford the areal geology is complicated by the occurrence of granite
and felsite, for the mapping of which very detailed field-work v^'ill be
necessary.

The arrangement of exposures of coarse and fine grained diabase in
the vicinity of the Old Powder House seems to show a gradual passing
of one rock into the other. In the immediate vicinity of the Powder
House is an extensive outcrop of coarse rock, like that at the Granite
Street quarries and Pine Hill. About four hundred feet northeast of the
Powder House on Harvard Street the texture is much finer, though not
sufficiently fine to be ranked with the normal " gi-eenstone." About six
hundred feet S. 20° W. of the Powder House the rock is somewhat finer
than at the last-mentioned locality. Again, at the corner of Elm and
Morrison Streets, which is about one thousand feet west-southwest of the
Powder House, the normal " greenstone " occurs in slate. Moreover, on
Willow Avenue, about fifteen hundred feet along the strike to the south
from the Harvard Street locality, the rock is practically identical with that
at the latter place. From this it seems probable that the coarsely crystal-
line rock at the Powder House is near the middle of the dike, where the
cooling was slow, and that the gradual diminution in the size of the
grains in going from that point is owing to more rapid cooling near
the contact.

The wide distribution of the " greenstone " has made it impracticable
for me to make a complete examination of it, but the few localities which

4 BULLETIN OF THE

have here been studied agree so well with each other, and with Professor
Wadsworth's notes on the other localities where the same rock is ex-
posed, that a complete study of it seems unnecessary to prove its identity
with the " diorite."

Petrographical Notes.

Having shown by field observations that the so-called " diorite " is
equivalent to the so-called " greenstone," the rocks will be distinguished
according to their mineral composition, without regard to coarseness of
texture. Microscopic examination shows the rock to be either a diabase
or diorite. The diabase is the principal rock, being found at the quar-
ries on Granite Street (Nos. 207, 209, 209 a, 214), and at the very ex-
tensive quarries on Pine Hill (Xo. 222). The exposure on Harvard
Street in Somerville is also of this rock (Xo. 210). The diorite forms
a facies of the diabase, and includes all the outcrops in the vicinity of
the Old Powder House (Xos. 208, 216, 218) except Xo. 210, already
mentioned as belonging to the diabase, and the hill north of High
Street and east of Highland Avenue (Xo. 203), Aside from the
amphiboloid mineral, the two rocks appear to be almost identical.

Diabase. — This rock is quite uniform in character, and occupies a large
part of the area of the dike. A macroscopic examination shows that it
is composed mainly of a plagioclase feldspar, and an amphiboloid min-
eral with more or less biotite and pyrite. The proportions of these
minerals vary considerably, causing the specific gravity to range from
2.98 to 2.65. In weathering, the amphiboloid mineral is first de-
composed, causing a complete disintegration of the rock to a coarse
feldspathic sand. The rock does not show the slightest evidence of lam-
ination, and the feldspars, which have been little interfered with in
crystallizing, illustrate well the divei-gent strahlig-kornig arrangement of
Lossen, the remaining space being largely occupied by hypidiomorphous
crystals of augite. The feldspar grains are more or less lath-shaped ; of
a white, pink, or green color ; generally striated ; and have an average
length of about 8 mm. The specific gravity of cleavage pieces from
Xo. 222 was found to be 2.638 and 2.643 by determinations with the
Thoulet solution, though these results are probably much affected by an
incipient alteration. In a single specimen (Xo. 214), the cleavage of
the amphiboloid mineral is so well developed that the mineral can be
identified as augite in the hand specimen.

Under the microscope, feldspar and augite are found to be always

MUSEUM OF COMPAEATIVE ZOOLOGY. 5

present. The feldspar is shown by twinning striations to be plagioclase,
which exhibits when fresh beautiful zonal phenomena. In a number of
sections the method of Professor Pumpelly ^ or M. Michel Levy ^ was
applied for the determination of this feldspar. As is well known, this
method consists simply in a determination of the maximum extinction
angle in the zone of the macro-pinacoid and base, sections whicli are in
this zone being characterized by symmetrical positions of extinction in
the two sets of twins, with reference to the twinning plane. Results
were obtained as high as 27°, requiring the presence of a feldspar as
basic as labradorite. In the classical work of Pumpelly above cited,
crystals of feldspar from the Granite Street locality were determined by
this method, combined witli a modihcation of Des Cloiseaux's method
for determining the size of the basal extinction angle. The highest re-
sult obtained by the first method was 16°, and by the second 3° to 4°,
though, owing to the sections being inclined to the base, the latter re-
sults were more or less unreliable. He concluded that the feldspar was
probably albite or oligoclase.

Mechanical separations of the constituent minerals have been made in
a number of cases by the Thoulet solution. In every case feldspar was
removed with each separation between the specific gravity limits 2.76
and 2.6, and often a considerable portion came below the inferior limit.
The grains were found to be seldom pure, and the wide range in specific
gravity is doubtless, in part, to be referred to decomposition products.
The portion separated below the limit 2.6 was in several cases subjected
to microchemical tests by both Boricky's ^ and Behrens's^ methods, after
careful washing to remove all iodide of potassium. Potassium as well
as calcium being always detected in this powder, that derived from No.
222 was subjected to quantitative chemical determination, which yielded
4.16% of oxide of potassium. The products of alteration of the feldspar
are calcite, and a mineral which is probably kaolin. Considerable
green chloritedike material is often contained in the feldspar grains ; but
it has apparently been derived from the biotite or augite by alteration,
and has found the way to its present position in the feldspar throiigli
the cleavage cracks. This substance is the viridite of Professor Wads-
worth, which he considered an incipient alteration of the feldspar.

1 Metasomatic Development of the Copper-bearing Rocks of Lake Superior.
Proc. Am. Acad., XIII. 253.

2 Mineralogie Micrographique, p. 227.

3 Archiv der Naturw. Landesdurchforschung voa Bohmen, III. Band, 5 Abth.,
Prag, 1877.

* Mikrocliemische Methoden zur Mineral Analyse, Amsterdam, 1881.

6 BULLETIN OF THE

From what has now been said, it seems certain that a feldspar as basic
as labradurite exists in the rock. This is attested both by the high ex-
tinction angles in the zone of oP and oo Poo , as well as by the presence
of calcite as a significant product of alteration. That a feldspar less
basic than oligoclase, and probably as acid, is present also, is shown by
the zonal structure and wide range in specific gravity. The potassium
obtained from No. 222 may be derived either from alteration products
(uiuscovite 1) or from the feldspar itself. The analysis of this rock
m ioto, which may be found in the sequel, shows by the small amount
of hydration that little alteration has taken place. It therefore seems to
be certain that the potassium is derived from the feldspar itself, either
from orthoclase or from a potiish plagioclase.

The principal non-feldspathic constituent is augite, which is much
broken up by feldspar crystals, though rarely the feldspar is penetrated
by augite, showing the nearly contemporaneous formation of the two
minerals. The augite, which has generally a rose color, displays a faint
dichroism, the ray parallel to b being pink, and that parallel to a pink-
ish yellow. Both prismatic cleavages are generally well developeil, and
intersect on the basal plane with the pyroxene angle 87°. Parting par-
allel to CO Poo and occasional twins according to the same plane, though
not constant features, are observed in the rock from some localities.
Zonal structure is not uncommon, aud rarely the hour-glass structure is
well developed. The principal alteration is to uralite, which is found in
rims completely surrounding many grains, while with others it has gone
farther, and found its way to the centre along the cleavage cracks. This
mineral occurs in its usual form in scales or sheaves, and is easily dis-
tinguished by its high double refraction, small extinction angle, and
strong pleochroism, the ray vibrating parallel to the long axis being
dark green, while that vibrating perpendicular to this direction is either
green or bright yellow. It is probable that the uralite further changes
to chlorite ; but since biotite was generally to be found in the vicinity
passing into chlorite, it could not be definitely determined. The distri-
bution of the calcite shows that it is derived from the pseudomorphism
of the augite, as well as from tlie decompositon of the feldspar.

Compact green and brown (basaltic) hornblende, though noticed once
or twice, are extremely rare.

The biotite when present is generally either in plates, from its pene-
tration of feldspar and augite, clearly original, or in fine scales or aggregate
masses in association with augite. The former variety contains inclu-
sions too small to be determined, with the characteristic pleochroic zones

MUSEUM OF COMPARATIVE ZOOLOGY. 7

about them. Both varieties have suffered alteration to the ordinary
product, chlorite.

Professor Wadsworth considered the biotite secondary to the augite,
chlorite being an intermediate stage in the process of alteration. It
seems, however, much more probable that this form of the biotite, if
indeed secondary, is derived directly from the augite, and that the fur-
ther alteration of biotite to chlorite sufficiently explains the occurrence
of the latter between biotite and augite. The occurrence of biotite as a
pseudomorph after augite has been described by Blum,-^ Eichthofen,^
Tschermak,^ Eohrbach,* and Brauns.^ On chemical grounds, without
assuming a high degree of metamorphism, the change from chlorite to
biotite is difficult to conceive. In some sections, particularly Xo. 207, a
large part of the chlorite can be referred to the diabantite of Hawes.®
In many slides chlorite occurs in clearly defined hexagonal sections sur-
rounded by one, or more frequently four or five, concentric rims of mag-
netite. In other cases biotite can be seen in these basal sections in the
act of changing to chlorite, the centre of the crystal being biotite, about
"which is a wide or narrow rim of chlorite. Figure 1 is taken from sec-
tions No. 202 and No. 203, and shows the different stages in this process
of pseudomorphism.

Apatite is found as a constant constituent, in unusually large clear
crystals, cutting all other minerals. A very small amount of quartz is
present, which, in sorae cases at least, is of secondary origin. Pyrite, mag-
netite, and ilmenite are present in varying amounts. Magnetite is either
in hexagonal sections or more or less irregular masses. These masses are
often elongated parallel to blades of chlorite, and are then evidence of
secondary origin. A case of this kind is shown in Figure 2. Ilme-
nite appears in sections, generally hexagonal, like the magnetite, but is
easily distinguished by its change to leucoxene or titanite. In a section
from the Granite Street quarries (No. 207) this change has been com-
plete and the only vestiges of ilmenite are the masses of white, more or
less opaque, highly refracting leucoxene. In other specimens (Nos. 202,
208, 209 a) the decomposition has been less complete, but has taken place
in bands, which have three directions parallel to the sides of the rhom-

^ Pseudomorpliosen, I Nachtrag, p. 30; III Nachtrag, p. 93.
2 Wien Akad., XXVII. 335. Blum, Pseudomorpliosen, III. 96.

* Porphyrgesteine Oesterreichs, Wien, 1869.

* Min. u. petr. Mitth., VII. 27.

* Neues Jahrbudi, V Beilage Bd., 275.

6 Mineralogy and Lithology of New Hampshire, p. 120.

8 BULLETIN OF THE

bohedral sectious. (See Figure 2.) This structure has been described
by many observers and figured by De la Vallee Poussin and Eenard,^
and by Teall.^ The structure has been explained by Teall as due to
intergrowths of magnetite and ilmenite, according to the fundamental
rhombohedron. Since the Gleitfldche of ilmenite is E, which is also the
normal-solution plane,^ these may be due to decomposition along the
normal-solution plane. From No. 222 the heavy portion separated in
the Thoulet solution was subjected to treatment with the electro-magnet.
Material was thus obtained so magnetic that, when removed from the
poles, the grains clung to each other like magnetized iron filings. Treated
with concentrated hydrochloric acid, this material was strongly attacked,
but did not entirely dissolve even by continued digestion.

Professor Wadsworth has described the occurrence of prehnite as a com-
mon product of the alteration of the feldspar and augite. This mineral
occurs in veins at the Granite Street quarries, and to determine its char-
acters a section was made from the mineral obtained from one of these
veins. The columnar crystals by macroscopic examination seem to have
their vertical axes, in general, perpendicular to the walls of the fissure.
In the slide, sections parallel to tlie long axis (c) always showed a sheaf-
like grouping of individuals having perfect cleavage, both parallel and per-
pendicular to the vertical axis. These sections afforded no interference
figure. Another series of sections (basal) had nearly equal dimensions,
with two equally perfect cleavages (» P) cutting each other at about
100°. These sections gave also, in converging polarized light, a very per-
fect biaxial interference figure, with high positive double refraction and
orthorhombic dispersion. The optic angle when measured in air was
found to be 83° 30', which is much smaller than the results obtained by
Des Cloiseaux with prehnite from other localities. The plane of the optic
axes bisects the obtuse angle between the cleavages. The prismatic cleav-
age is very perfect, hardly less so tlian the basal. No evidence of twinning
like that noticed by Des Cloiseaux* in some specimens, or that found by
Professor Emerson^ in the prehnite of the Deerfield dike, was observed.

The only section of rock from the region under consideration in which

1 Memoires sur les Caracteres mineralogiques et stratigraphiqnes des Roches
dites Plutoniennes de la Belgique et de I'Ardenne franfaise. Me'm. Couronpe's
de I'Acad. Roy. de Belgique, XL. 50, 74.

2 Quart. Journ. Geol. Society, XL. 640.

8 Cf. Judd, On the Relations between the Solution Planes of Crystals and
those of Secondary Twinning. Min. Mag., December, 1886.
* Manuel de Mineralogie, p. 430.
6 Am. Journ. Sci., (1882,) XXIV. 270.

MUSEUM OF COMPARATIVE ZOOLOGY. 9

prehnite was discerned is No. 215 (Bell Eock, Maiden) where it was
found filling a minute fissure vein. The chalcodite which Professor
Wadsworth has described was not nut.'ced in any of the slides, and it
seems certain that the important part which he assigned it, of completely
taking the place of both feldspar and augite, is wrong.

A typical and unusually fresh specimen of the diabase (No. 222, Pine
Hill, Medford) has been subjected to a quantitative chemical analysis by
R. C. Sweetser, B. S., Assistant in Chemistry at the Worcester Poly-
technic Institute, to whom I would here express my great obligation.
Though fuUy engaged with other duties, he kindly offered to do the work
and obtained the results given below in column I. Column II. contains
the results of an analysis of a diabase from the Lenneschiefer at Bochteu-
beck by Schenck, which shows considerable more decomposition, but is
otherwise nearly identical.^ Column III. is an analysis of diabase by
Teall from Cauldron Snout, Durham, in the Whin Sill.''

I.

n.

m.

SiOj

48.75

48.42

51.22

Al^O,

17.97

17.59

14.06

Fe^O,

0.41

1.05

4.32

FeO

13.62

8.36

8.73

CaO

8.82

7.73

8.33

MgO

3.39

4.30

4.42

MnO'

0.91

0.16

K3O

2.40

3.07

1.25

Na^O

1.63

5.15

2.55

H2O

0.60

2.24

1.28

TiO,

0.99

2.23

2.42

P2O5

0.68

0.28

0.25

CO3

tr.

0.08

0.19

FeSj

tr.

0.15

0.49

100.17 100.65 99.67

Sp. Gr. 2.985 2.919

Schenck considered orthoclase as probably present in II. The analysis
as well as the extinction angle shows the feldspar to be more acid than
that of I. Augite is changed to viridite, and ilmenite occurs and alters to
leucoxene along the Gleitjldchen. The rocks I. and II. are thus shown to
be very similar. The mineral composition as well as the chemical com-
position of III. is also nearly identical with that of I.

1 Adolf Schenck. Die Diabase des oberen Ruhrthals und ihre Contacterschein-
angen mit dem Lenneschiefer. Diss., Bonn, 1884, p. 20.

2 Teall. Quart. Journ. Geol. Soc, XL. 640.

10 BULLETIN OF THE

Angite Diorite. — This rock, though quite sitnilar to the diabase, differs
from it in a loss of the ophitic structure, and in the appearance of brown
(basaltic) hornblende as the principal non-feldspathic constituent. In
some sections augite does not appear, owing to complete uralitization.
At other localities it comes into prominence, and there the rock may be
known either as a diorite or a diabase.

The hornblende is for the most part the massive brown variety, which
is well characterized by its color, perfect cleavage, large optical angle, and
strong pleochroism. The absorption may be written c = b>>a. The
sections have in general distinct outlines parallel to the fundamental
prism and the clino-pinacoid. A common feature of the honiblende crys-
tals is the occurrence within them of cores of augite, which seem to show
either that the hornblende is derived from the augite by pseudomorphism,
or that the two minerals crystallized originally in their present relations.
Such pseudomorphism was first noticed by Streng ^ in 1877, and subse-
quently by Hawes,^ Irving," Van Hise,* Sjogren,^ and Von Lasaulx."
Remarkable instances of this change have been described by Professor
Williams,'' from the Cortlandt Series on the Hudson River, and by
Schenck,® from the diabase of the Upper Ruhrthal in Westphalia. The
former has shown the gradual passing of the augite into brown horn-
blende. The latter has described a further change of the brown to green
hornblende, while Von Lasaulx found in the diabase of Kiirenz that
the change of the augite was first to uralite, then to brown hornblende.
In the diorite which we are considering, the contact of augite and horn-
blende is a sharp line. No evidence of a gradation from one mineral to
the other was anywhere observable. The hornblende is in general very
fresh, while the augite alters readily to chlorite, so that in many cases only
a few scattered fragments of augite can be seen (Figure 2). It seems prob-
able, therefore, that these combinations are the result of parallel growth.
Teall^ has figured such growths in the Whin Sill, and Rohrbach"

1 A. Streng. Neues Jahrbuch fiir Mineralogie, etc., 1877, p. 133.

2 G. "W. Hawes. Mineralogy and Lithology of New Hampshire, pp. 57, 206,
Plate VII. Fig. 1.

3 R. D. Irving. Geology of "Wisconsin, III. 170.

* C. R. Van Hise. Am. Journ. Sci. [3], XXVI. 29.

5 H. Sjogren. Neues Jahrbuch fiir Mineralogie, etc., 1884, 1. 82 (Ref.).

•^ A. V. Lasaulx. Verb. d. Naturh. Vereins d. pr. Rheinl. u. Westf., 1878, p. 171.

T G. H. Williams. Am. Journ. Sci. [.3], XXVIII. 259.

8 A. Schenck. Die Diabase des oberen Ruhrthals, etc. Diss., Bonn, 1884.

» Quart. Journ. Geol. Soc, XL 653, Plate XXIX. Fig. 3.

10 Min. u. petr. Mitth., VII. 1, Plate I. Figs. 1-7, 1886.

MUSEUM OF COMPARATIVE ZOOLOGY. 11

in the Cretaceous formation of Silicia. The figures of the latter show a
sharp line of contact between the two minerals. He was also able to
show that, in the majority of cases at least, the minerals were in parallel
position. Chemical analysis showed an essential difference in the com-
position of the augite and hornblende. It was also observed that decom-
position had seldom progressed to the same point in both minerals when
together, though neither seemed to offer iu all cases more resistance to
decomposition than the other.

A remarkable instance of mechanical deformation is exhibited in sec-
tion E"o. 202. A large crystal of brown hornblende has been bent until
it has the sliape of a letter S. The optical properties are anomalous, as
would be expected, and a crystal of apatite has been bent about the horn-
blende crystal, which has been attended with crushing, and optical dis-
turbances, so that the apatite crystal is extinguished in a mosaic. This
must be referred, however, to motions wliich existed in the partially
consolidated magma, as we would expect to find anomalies in the optical
behavior of the plagioclase grains if it were due to the action of oro-
graphic forces.

Section No. 208 (corner Elm and Morrison Streets) is porphyritic, the
base being difficult to resolve. The porpliyritic crystals are feldspar and
parallel growths of augite and hornblende.

The clilorite of No. 202 is often tilled with belonites of a green color,
arranged in three parallel directions, cutting each other very precisely at
angles of 60°.

Summary and Conclusions.

"What has been noted in the preceding pages may be summed up in
the following statements.

The dike under consideration includes, not only the exposures of so-
called " diorite," but outcrops, in the vicinity of the Old Powder House
in Somerville, of rock intermediate in texture between the normal
" diorite " and normal " greenstone," as well as the " greenstone " itself.
The coarseness of texture is in general dependent only on the position of
the specimen in the dike, the fine-grained rock being naturally found
near the contact. The general composition of the rock is that of a
diabase, though facies of augite-diorite occur.

The diabase has in general a more or less ophitic structure, and is
characterized by the original constituents, plagioclase, augite, biotite, apa-
tite, ilmenite, and magnetite ; apatite and the ore minerals comprising

12 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

the first generation, while feldspar, augite, and biotite crystallized nearly
contemporaneously and form the second generation. The secondary
minerals are uralite, chlorite, biotite (in part]), leucoxene, kaolin (]),
magnetite (in part), calcite, pyrite, and quartz. The plagioclase is
somewhat variable in composition, owing to zonal structure, but has
probably an average composition corresponding to andesine. It is also
probable that orthoclase is present, though the potash obtained in the
analyses may be derived from a plagioclase containing a considerable
per cent, of potassium. The augite-diorite differs from the diabase in
that the ophitic structure is wanting, and that the brown hornblende,
which now comes into greater prominence than the augite, is generally
in idiomorphous crystals. The diorite is characterized by very perfect
instances of the parallel intergrowth of augite and hornblende.

In both the diabase and diorite the change of the augite has been
uralitization, though in the diorite it has in many cases changed directly
to chlorite.

In conclusion, I have to acknowledge obligation to my instructors,
Mr. J. E. Wolff, of Harvard University, and Dr. George Huntington
Williams, of the Johns Hopkins University. The greater part of the
microscopical examination in connection with this paper was made in the
laboratory of Mr. Wolff, and I am indebted to him for much advice and
suggestion. Dr. Williams has examined most of the slides, and assisted
me in countless ways in the preparation of this paper. I am also much
indebted to jSIr. R. C. Sweetser, of the Worcester Polytechnic Institute,
for a complete chemical analysis.

November, 1887.

EXPLANATIOX OF PLATE.

Fig. 1. Illustration of the change of hexagonal plates of biotite to chlorite, with
separation of magnetite in concentric rims. From sections No. 202
and 203.

Fig. 2. Illustration of the intergrowth of augite and brown hornblende, and
alteration of the former to chlorite. The alteration of ilmenite to leu-
coxene along the Gleifjldchen is also shown. From sections No. 202
and 208.

Ho BBS, Diabase

Figl

Figr2.

B Meisel. lith

No. 2. — On the Geology of the Cambrian District of Bristol County,
Massachusetts. By N. S. Shaler.

[Published by Permission of the Director of the U. S. Geological Survey.]

Preliminary Note.

For a number of years I have been engaged in an incidental manner
in studying the geological structure of the great synclinal district to
which I have given the name of "the Xarragansett Basin." This
geological field, extending from the southern part of Narragansett Bay
in Rhode Island to the region of the granitic hills which includes the
Blue and Sharon Hills of Massachusetts, and eastwardly to the region
occupied by the town of Hanover in Massachusetts, is mainly under-
laid, as is well known, by rocks of Carboniferous age. My principal aim
has been to determine the geological history of this Carboniferous sec-
tion. Incidentally, it has been necessary to make some study of the
deposits which lie below the level of the Millstone Grit. In these latter
inquiries I found it necessary to do a good deal of work on the exten-
sive series of more or less metamorphosed ancient rocks which lie be-
tween the western border of Rhode Island and the western edge of the
Coal Measures, from Greenwich, R. I., to Wrentham, Mass. This
inquiry, although incomplete, has developed certain facts of consider-
able interest, which it appears to me should be made public before the
preparation of my final report on the Narragansett Basin, which can-
not be finished for some time to come. The most interesting of the
many results which I have obtained in this Pre-Carboniferous series
of rocks consists in the discovery of an extensive series of Cambrian
deposits, containing a tolerably abundant and fairly determinable set of
fossils. The discovery of these beds not only enables us to fix the age
of an extended section of rocks, but to ascertain a number of facts
which have a great importance with reference to the general history of
this portion of the continent.

Several geologists have observed the fact that between Providence,
R. I., and Wrentham, in Massachusetts, we have an extensive develop-

VOL XVI. — NO. 2.

14 BULLETIN OF THE

ment of interbedded conglomerates, shaly slates, and sandstones, the
whole separated from the other detrital deposits of this region by
peculiarities of color. While the Coal Measure conglomerates have
generally a grayish or blackish hue, these more western deposits of
puddingstone are generally of a red color, wliile the intercalated shales
vary in hue from a brilliant red to an olive-green. Various conjec-
tures have been made as to the age of these deposits. They have
been thought by one observer to resemble the Trias, while others,
owing partly to their position, have assigned them to the Devonian age.
Until I began my studies upon this district, the strata had aflbrded no
fossils, and the determinations above noted were purely conjectural.

As it appears to me that there is a lesson of some value affoi'ded by
the conditions of my inquiry, I venture to set forth the methods under
which it was pursued. After carefully traversing all the roads in this
district, with the hope of obtaining geological data on those lines, I
became convinced that results of value could not be yielded by track-
ing these paths, for the reason that here as elsewhere in an accidented
country the public ways avoid the outcrops. Inspecting a portion of
the field, I found that almost all the available exposures were covered
by brushwood, and thus hidden from a hasty glance. I therefore re-
solved to trace the country on foot in such a manner that I should
obtain sight of every exposure. A preliminary iiKjuiry showed that,
even where the rocks were covered by drift, a careful consideration of
the fragmentary matter in the glacial deposits would give very clear
evidence as to the nature of the deposits below the covering. Pursu-
ing this latter inquiry in a methodical way, I found that at any point
whatsoever the boulder clay as distinguished from tlie kame dejjosits
was to the extent of at least three fourths its mass composed of mate-
rial which had not been carried for a greater distance than lialf a mile.
Pursuing my further inquiries in this close manner, I found in a very
short time that these apparently barren deposits of shales and con-
glomerates afforded at certain points a good number of fossils. Within
an area of a single square mile, three important localities have already
been discovered from which, as will be seen in the sequel, we have ob-
tained over a score of recognizable fossils. It is likely that further
research on these beds will increase the list of organic remains to thirty
species or more.

My studies on the Narragansett field were begun in 18G5. The first
of the fossil localities was not found until 1883. The pressure of other
work and the lack of good topographic maps made it impossible to work

MUSEUM OF COMPAKATIVE ZOOLOGY. 15

up this field at an earlier date. lu 1884, my service as Geologist in the
United States Geological Survey enabled me to command more time for
these explorations, and the rapid advance of the topographic work in
this area done by that Survey in co-operation with the State of Massa-
chusetts has afforded a sufficient topographic basis for the inquiry.

I am indebted to my assistant, Mr. August F. Foerste, for a certain
amount of aid in the preparation of this report. He has worked out a
part of the boundaries which arc delineated in the accompanying map,
and has collaborated with me in the preparation of the descriptions
contained in the second part of this I'eport.

General Structure of the Field.

The area indicated in the first of the accompanying sketch maps
evidently contains two distinct series of stratified rocks, besides the
numerous and peculiar injected materials, which are not to be discussed
here. On the east we have the Coal Measures of the Narragansett
basin. The western boundary of this series cannot be exactly traced,
fur the reason that it is altogether hidden by drift deposits mostly
belonging to the class of kame and terrace accumulations, and therefore
unfit to afford the basis of any determinations as to the subjacent beds.
West of this boi'dor, the position of which cannot be at any point fixed
within the limits of some hundreds of feet, we have the area of Cam-
brian deposits. This strip has an average width of not far from two
miles. Although its contact with the Carboniferous deposits is not
seen, it is likely that it belongs to the class of erosion overlies, that
is to say, the Carboniferous rests upon the worn surface of the steeply
inclined Cambrian beds. Evidence in favor of this supposition is also
afforded by the conglomerates of the Coal Measures, which contain
more or less detrital material brought from the Cambrian series, which
was evidently exposed to erosion at the time when the lower portion
of the Coal Measure deposits were formed. Moreover, as my exten-
sive studies on this district have adequately shown, few faults occur
in the field. The habit of accident is that of folding rather than
faulting.

On the west of the Cambrian lies another field of rocks, which I am
compelled at present to consider as of Pre-Cambrian age. The deposits
in this section consist in the main of gneissoid rocks of varying compo-
sition and a great area of dark greenish chloritic deposits which appear
in part at least to be metamorphosed conglomerates and shales. In the

16 BULLETIN OF THE

region west of Pawtucket and Valley Falls on the south side of Black-
stone Eiver these rocks contain a number of areas of crystalline lime-
stone. As yet it has proved impossible to determine the sections in
this district with any satisfactory degree of definition. At Mannville
on the eastern or right bank of the river, about two hundred feet in
thickness of these supposably Pre-Cambrian rocks appear as rather
distinct beds of what seems originally to have been clay slate, now
changed to a gneissoid material. In these gneissoid rocks near Cum-
berland Hill there are extremely interesting mineralogical localities.
The remarkable deposit of ilmenite. the iron ore composing Iron Hill,
has long received much attention from mineralogists. The hill near
Sneech Pond is said to have yielded a cei-tain amount of metallic
copper in the various explorations which have been made upon it.
Near by, and in the same district, is an interesting and extensive vein of
pyrolusite.

The imperfect sections of this district which have thus far been ob-
tained indicate that the total depth of the deposits probably amounts
to more than five thousand feet, and may attain to twice that amount.
However, as the region has been much affected by mountain building
forces, and as the metamorphism has gone so far that little trace of the
original bedding is now discernible, it is very diflBcult to obtain a satis-
factory account of the series. The nature of the contact between these
evidently ancient rocks and those of Cambrian age has not yet been
well determined. There are some reasons to suppose that it may be
by faulting, but the fact that pebbles of the supposed Pre-Cambrian
series are found in the rocks of Cambrian age is rather against this
supposition. Actual contact is at no point traceable, the surface being
too deeply covered with detrital materials. It may be noted, however,
that the line between the two formations is much more direct than
that which separates the Cambrian from the Carboniferous, and we are
therefore more justified in supposing that faulting may have taken place
at this point.

Whatever be the age of these strata we have termed Pre-Cambrian,
it is evident that they were formed long before the Cambrian deposits
themselves, and this for the reason that the measure of metamorphism
which has affected the two regions is extremely diverse. The Cam-
brian rocks exhibit very little sign of metamorphic action. The shales
indicate a slight amount of infiltration, and in the conglomerates the
pebbles all retain essentially th^ir original character, save that they are
sometimes slightly indented one into the other. The cement of the

MUSEUM OF COMPARATIVE ZOOLOGY. 17

mass is not more altered than is usual with our unchanged conglomer-
ates. Its general character, save for its reddish color, is undistinguish-
able from the deposits of Millstone Grit age in the neighboring coal-tield.
This wide difference in the measure of metamorphism of the rocks which
are traced almost in contact with each other is of itself sufficient evi-
dence of great disparity in age. Too much importance, however, should
not be given to this consideration, for the reason, as I shall have here-
after occasion to show in the final report on the Narragansett field, that
the Carboniferous rocks in the southern portion of the basin appear to
have undergone a very extensive regional metamorphism in which the
sandstones have taken on a gneissic shape, the black shales been con-
verted into otrolitic schists, and the conglomerates also metamorphosed,
the cement taking on a gneissoid form, and the composition in many of
the pebbles being similarly changed. This alteration, however, seems
to come about gradually as we go from the north southward, while the
corresponding change in passing from the Cambrian to the subjacent
rocks is of a sudden nature.

The rocks of apparently Pre-Cambrian age, possibly to be assigned to
the Huronian period, which lie to the west of the Cambrian field, differ
in their attitudes from those of the Carboniferous as well as the Cam-
brian series. The prevailing strikes in the valley of the Blackstone are
from northwest to southeast. There are some local variations which
give other directions, but there can be no question that, considered as a
field, the highly tilted rocks pretty regularly extend in a northwest and
southeast direction. Thus the limestone belt which extends from near
Valley Falls to Harris's Quarry, about four miles west of that point,
has a tolerably uniform trend in the above-mentioned direction. At
the Dexter Quarry they are locally thrown from the prevailing strike, so
that the axis is nearly north and south ; but the general direction of the
limestone 'belt is nearly that above described. On the other hand, the
rocks of the Cambrian, as well as those of the Carboniferous, have a toler-
ably uniform northeast and southwest trend, the strikes varying from
north to north 45° east, thus following the general course of the disloca-
tions along the Atlantic coast. It therefore appears that there must
have been a change in the character of the tension and consequent dis-
ruption which have affected this country in Pre- and Post-Cambrian
times, the more ancient rocks having undergone -extensive displacements
in a peculiar axis before the later deposits were accumulated.

In this connection it may be interesting to note tha't the beds of
Tertiary age on Martha's Vineyard, about fifty miles to the southeast of

VOL. XVI. — NO. 2. 2

18 BULLETIN OF THE

Attleborough, also exhibit northwest and southeast strikes. This matter
I have considered in some detail in my Report on the Geology of Mar-
tha's Vineyard, now in press in the United States Geological Survey.

The age of the rocks which I have terraed Pre-Cambrian must for the
present remain doubtful. The absence in the section of any beds like
those containing the Paradoxides of Braintree raises the presumption
that they do not belong in the Braintree Cambrian series. The massive
limestones which occupy a portion of the section are also to a certain
extent evidence to the same effect. As a whole this section reminds me
more of that which occurs at Eockport, Maine, than any other deposits
known to me on the coast ; still I think there is nothing which can be
evidence to prove the likeness in age of these beds.

General Character of the Cambrian Rocks.

As the object of this paper is to set forth the phenomena of the
Cambrian series with no other attention to other deposits than is neces-
sary to make them comprehensible, I shall now proceed to give in some
detail an account of the deposits which appear to belong in this portion
of the section. S^: far fossils have been found in rocks of this section
which probably do not in the aggi-egate include more than one hundred
feet or so of the total section of the Cambrian series. However, as these
deposits are of the same aspect as all the red slates and conglomerates
of the area, it appears at present reasonable to include all rocks of this
description with the above-mentioned series. The total thickness of the
section which I have termed Cambrian is not accurately determinable.
It probably amounts to not far from two thousand feet. In the main it
consists of thin-bedded shaly layers which occasionally pass into moder-
ately thick fine-grained greenish and i-eddish slates. Intermingled with
these in several levels we have a number of layers of conglomerate, per-
haps as many as half a dozen distinct beds, varying in thickness from
two hundred to three hundred feet. In all cases these conglomerates
are frequently interrupted by thin layers of shale or sandstone. The
pebbles are mostly of small size ; none have been observed exceeding a
foot in diameter, and few above six inches in thickness. The pebbles
are in most cases rather angular. At certain points they have a very
high degree of angularity, so that they assume the form of a breccia.
The rocks from which the pebbles were taken are mainly identifi-
able in the western portion of the field before described. Xo distinct
traces of cross-bedding have been observed in the deposit. Indeed, in

MUSEUM OF COMPARATIVE ZOOLOGY. 19

the layers of conglomerate a thickness of from twenty to thirty feet
may often be found in which only slight evidences of bedding are ascer-
tainable.

This section has been subjected to two classes of accidents. One has
resulted in the general tilting of the deposits in an easterly direction,
the angle of inclination not usually exceeding about 20° of slope. At
certain points in the field there have been disruptions of the rocks, with
possible faulting on the lines of breakage, attended by the extrusion of
massive dikes of hornblendic granite, which appear in the form of elon-
gated somewhat lenticular-shaped ridges, which taper abruptly at either
end. These ejections vary in diameter from a few hundred feet to as
much as a mile, and their greatest length in one or two cases may
amount to three miles or more. On either side of the injected syenite
(or hornblendic granite) the bed rocks are folded abruptly upward into
vertical attitudes, which sometimes continue for a considerable distance
from the face of the dike material.

It is an interesting feature connected with these intrusions of gran-
itic matter, that in no case do they appear to have brought about any
very conspicuous metamorphism in the sedimentary deposits with which
they have come in contact. The change is rarely apparent at more than
two or three feet from the dike. Actual contact has been seen but at
one point, southeast of locality No. 3 on the map, where the slates are
found in almost immediate juxtaposition with the hornblendic granites.
At locality No. 1 we have an extensive area exposed within one hun-
dred feet of the contact with the hornblendic granite mass, more than
half a mile in diameter. At this point we find no perceptible meta-
morphic influence on the sedimentary strata.

The topographic features within the limits of the Cambrian field are
in a large measure determined by the resistance to erosion aff'orded by
these elongate domes of ejected matter. In a less determined way the
ridges of conglomerate influence the shape of the country. The horn-
blendic granites are but rarely exposed to the eye, for the reason that,
wearing evenly, they form a uniform surface on which the drift material
rests as a blanket. The conglomerate ridges, wearing irregularly, often
appear as sharp peaks too steep to retain any considerable coating of
glacial detritus.

It is probable that these granitic ejections took place before the depo-
sition of the Carboniferous rocks, for, although a very marked feature in
the Cambrian district, they have nowhere been observed penetrating
through or into the coal measures which overlie them. Indeed, as I

20 BULLETIN 05 THE

shall hereafter note in the discussion of this district, one of the most
remarkable features in the Coal Measures is the entire absence of dike
materials in this wide area, a feature in which it is in sharp contrast with
all the neighboring portions of New England. The fact appears to be
that the dikes which intersect the rocks of Southern New England were
formed before the Carboniferous age, or if formed after for some reason
never penetrated the deposits of the Coal Measure series.

Although the intrusive rocks of this area are a baffling element in
the effort to unravel its structure, the principal difficulty arises from the
drift coating which covers at least ninety-five per cent of the surface.
In passing over the country, the student is, on account of the distribu-
tion of this drift coating, led to give too great structural importance to
the conglomerates and to the hornblendic granites. The fact is, that
the softer shales almost always occupy the lowest parts of the area.
Almost all the stream beds course upon them, and it is only by a care-
ful study of the drift materials that the preponderance of these slates
becomes evident.

Origin of Sediments, and Conditions of Deposition.

The sediments composing this Cambrian section appear to have been
derived f-om rocks substantially the same as those which now lie in the
field west of the area. Although fossils have been found in a small part
of the section, close study makes it plain that by far the greater portion
of the strata are clearly azoic. The frequent return of conglomerate
layers and the coarseness of the pebbles show that during most of the
time when the beds were accumulating the region was near shore ; so,
too, the large amount of sandy matter even in the slates affords a pre-
sumption that the region was not remote from the coast line. About
one hundred feet of shale beds have been subjected to a very careful
search for organic remains. The total thickness of the deposits in which
any trace of life has been found probably does not exceed one hundred
feet, and even in this section only a small part of the rocks actually con-
tain fossils. As before remarked the rocks of this Cambrian series are
very little metamorphosed. "VVe therefore cannot attribute the absence of
life to secondary changes, but must regard it as an original characteris-
tic of these sediments. The great abundance of conglomerates, the
considerable size of their pebbles, the fact that none of these have a
beach-worn character, but are in general form like the pebbles con-
tained in the neighboring glacial deposits of a stratified character, afford

MUSEUM OF COMPARATIVE ZOOLOGY. 21

a basis for the presumption that these rocks were accumulated during
an ice epoch.

The glacial origin of these sediments is made more probable by the
fact that they contain a large amount of ferruginous material. My ob-
servations on the recent drift of New England show that at a hundred
localities, representing all the States except Vermont, the drift contains a
large amount of such material. The conditions of glacial erosion, the
rapidity with which the process goes on, and the absence of acids pro-
duced by decaying vegetation in the rocks, cause glacial deposits formed
of detrital materials originating in crystalline rocks to contain large
amounts of iron, which under ordinary conditions of decay would be
oxidized and borne away in the dissolved state.

The distribution of Cambrian fossils in these beds, where they occur in
thin layers, appears to indicate that life was present in the sea at some
distance from this shore line, and that it occasionally, in the interruption
of the conditions which made the rest of the beds non-fossiliferous, won
its way to this field. Precisely similar invasions of life took place dur-
ing the last glacial period along the shores of this part of the continent.

Characteriacics of Life.

The organic fossils obtained from the Cambrian beds of Attleborough
show very clearly that the section in which they lie belongs in the earlier
divisions of that age. This is indicated by the general correspondence
of the organic forms with typical sections elsewhere, particularly those
in the region about the Hudson valley. It will be noted that no trilo-
bites of the Olenellus group have been found in this section, though the
total number of specimens of this order observed is considerable. The
fact that one species of Paradoxides occurs in these beds appears to in-
dicate that the fauna has rather close affinities with the Braintree Cam-
brian horizon. It is interesting to note that this surviving member of
the Paradoxides series is very small. I believe it to be one of the most
minute forms which has yet been described. Although this fossil is so
far represented by a single specimen, it affords ground for the presump-
tion that the group was at this time imperfectly developed.

The most interesting feature connected with these fossils is the ample
representation of the group to which Salterella and Hyolithes belong.
By far the greater number of the individual fossils which were found at
the three localities belong to one or another of five species described in
the following account of the fossil remains. Indeed, at locality No. 1,

22 BULLETIN OF THE

at least ninety-five per cent of the recognizable fossils are members of
this group. Some of the layers at that locality which have an aggre-
gate thickness of half a foot are in good part composed of these remains.
It is interesting to note the fact, that certain of these species appear to
have found lodgment in the empty shells of their predecessors. In no
other way can we so well explain the fact, that from one to four of the
cones are often found packed into the larger shells in the manner indi-
cated in the diagrams of the descriptions of fossils from this section.
If this view of the relations of these included cones be correct, we have
in this horizon perhaps the first evidence of a habit of a somewhat intel-
lectual nature which is known through the history of the rocks.

It is perhaps worth while to note that one of the Ptychoparias found
in this section is clearly rolled, as is the fashion with many of the forms,
such as the Calymenes, in higher horizons. This peculiar habit has been
supposed to be of a protective nature, the trilobite thereby securing im-
munity from danger when assaulted by enemies. This indeed seems at
first sight a very probable interpretation of this habit, and of the peculi-
arities of form which make the means of rolling the body into a ball
possible. The difficulty, however, is to see what was the nature of the
enemies from which the creature had to defend itself. The rocks of this
horizon are not known to contain any creatures capable of threateniug
the safety of the trilobites. So far as our knowledge goes, they were
themselves the only highly organized forms in this horizon. The other
creatures appear to have been relatively weak ; none of them, so far as
we know, were able to menace the trilobites, nor does it seem likely that
the trilobites could have assailed each other in a serious manner.

Relation of this Deposit to Cambrian Problem.

The position of these Attleborough beds with reference to the fauna of
the Paradoxides section is one of extreme interest. As yet these two
horizons have never been found in definite relations with each other, so
that it may be affirmed which of the two is the earlier. The Scandi-
navian geologists claim that in their country the Paradoxides zone oc-
cupits a higher position than that of the Olenellus group. On this
account I have taken much pains in seeking for any indication of
beds which could be referred to the Paradoxides zone. So far, I have
not succeeded in finding any trace of rocks whicn would serve to es-
tablish the relation between the two horizons. As is well known, a
considerable mass of strata, having an aggragate thickness of some

MUSEUM OF COMPARATIVE ZOOLOGY. 23

hundreds of feet, belonging to the Paradoxides section of the Cambrian,
exists on the southern shore of Massachusetts Bay in the township of
Braintree. This deposit probably extends, as a continuous mass or as
an isolated section, as far as the Neponset River in Quincy, a distance
of about four miles. Although no distinct fossils have been found, save
at Braintree, a number of distinct remains occur near the Neponset
River, in beds having much the same aspect, and apparently at about
the same distance from the syenites, as those at Braintree. It there-
fore, on account of the large extent of the Paradoxides section about
Massachusetts Bay, seems possible that the Braintree section may be
represented somewhere iu the Attleborough Cambrian district.

Although I spent a good deal of time searching for rocks which should
have a physical likeness with those at Braintree, I have not yet been able
to discover any such in the Narragansett field. The conditions under
which the search was made render it difficult to make sure that such
deposits may not yet be found in that vicinity. A search for the Attle-
borough series in the Boston synclinal and in the neighborhood of the
Braintree beds has likewise been unavailing. No deposits of conglom-
erates or saudstones having the peculiar hue of that series have been
found in any part of the Boston basin. I therefore regretfully conclude
that the probability of determining the relative position of these two sec-
tions in this field is small. The absence of one of these members of the
Cambrian series from the Boston basin and from that of Attleborough
may be fairly attributed to the large amount of erosion to which both
regions have been subjected. The Paradoxides beds of Massachusetts
Bay are evidently a mere remnant of a sheet which once overspread a
large part of that area. The extensive conglomerates belonging to the
Roxbury series, with their associated slates and the argillaceous deposits
of Cambridge and Somerville, are probably of Cambrian age, and may
possibly belong to the lower portion of that section, along with the Para-
doxides bearing strata. But it is barely possible that they may repre-
sent the same age as the conglomerates and shales which lie above the
level of the Attleborough fossiliferous horizon. The wide difference in
the mineralogical character especially of the slates makes this view, how-
ever, improbable.

Although the relation of these two horizons is not determinable by a
comparison of the Massachusetts Bay and Narragansett deposits, it is
possible that it may be elsewhere determined. Fragments of sections
containing these horizons may well be foimd along other portions of our
Atlantic coast.

24 BULLETIN OF THE

Discussion of the Evidence afforded by the Attleborough

Series.

The facts as given above concerning the rocks of the Attleborough sec-
tion and the neighboring parts of Rhode Island carry our information con-
cerning the condition of the Atlantic coast line much further than might
at first sight be supposed. In the first place, they prove that the Atlantic
coast line was during the Cambrian period not far removed from its pres-
ent position. The great thickness and general character of the conglom-
erates appear to me to be abundant evidence on this point. Whether the
formation of these conglomerates was due to glacial action or not, it is
clear that they were deposited near the coast line. Only by the action
of water moved by strong currents could we have had the stratification
induced which appears in many of these pebbly sections. Such rapid
movements of water are only possible in shoal regions. The fact that
the pebbles have apparently all been derived from rocks in the immedi-
ate neighborhood, those which lie to the westward of the Cambrian
deposits, indicates that, while the Cambrian region was sea, the neigh-
boring district was in a condition to yield detritus to erosive forces, and
was therefore presumably laud. We thus fix the marine shore line of
the continent in this area close to the present coast.

It may be here remarked, in passing, that we have now determined
four stages in the history of this part of the continent, in which the
coast line was near its present position. These are as follows : the
Cambrian, which we are now considering, the Carboniferous, which
immediately succeeds it in the same field, the Triassic conglomerate of
the Connecticut valley, and the probable Miocene conglomerates which
appear at Gay Head on Martha's Vineyard. There are two other hori-
zons pretty well determined in which fragmental materials formed along
the coast line exist, viz. that of the Roxbury puddingstone, which prob-
ably belongs in the Cambrian age, possibly in the horizon of the Para-
doxides beds, and the coarse sandstones of Cretaceous age which appear
on Martha's Vineyard. If we add to these the glacial conglomerate of
the last ice period, we have a total of seven stages in the earth's his-
tory from the Lower Cambrian to the present day, in which the shore
of the continent has appeared near its present position. When we re-
member the amount of evidence going to show great erosion in this
field since the earliest geological ages, an erosion which may have re-
moved the evidence of coast line deposits of many difi'erent ages, we
are struck with the fact that we have here proof as to the permanence

MUSEUM OF COMPARATIVE ZOOLOGY. 25

in the relation of the continent to the sea in this portion of the earth's
surface.

The same evidence which enables us to affirm the frequent presence of
the coast line at this point, serves also to indicate that this portion
of New England has from a very early date possessed and retained its
present mineralogical character. The conglomerates of this Cambrian
horizon contain substantially the same kinds of rocks as make up similar
detrital deposits of the drift period. So far, I have been unable to dis-
cover any varieties of rocks in the one which are not contained in the
other, with the single exception of the hornblendic granites, such as
are intruded in the form of dikes amid the Cambrian deposits. It
appears likely that these materials did not appear in this district until
after the Cambrian had been deposited.

The hornblendic granites which are intruded into the rocks of this
Cambrian field have a general likeness to those which appear in the re-
gion of the Sharon and Blue Hills. Although in the form of detached
masses, they are scattered in a somewhat linear fashion, as in those
fields of granite. It is not improbable that their ejection may be of
the same date as that of the similar rocks to the northward, but as yet
there is no sufficient evidence to make any affirmation in the matter.

The evidence afforded by the Attleborough series as to the history of
the Narragansett Basin, taken in connection with the other facts which
I have ascertained in my study of this district, is of a very interesting
nature. On the eastern side of the Narragansett synclinal, north of Fall
River, the Carboniferous deposits lie immediately upon syenites. On
this side of the field the Lower Carboniferous strata are composed, to the
thickness of a hundred feet or more, of consolidated waste derived from
these crystalline rocks. This waste is so little changed, that at first
sight the section appears to be composed of decayed granitic matter. It
was only on finding fossils in the deposit at Steep Brook, Mass., where it
is quarried for fire-clay, that I became convinced of its Carboniferous
age. It thus appears that while on the western part of the basin the
Carboniferous series rested upon the great section of Cambrian and
Pre-Cambrian rocks, it was bounded on the east by ejections of crystal-
line materials.

This fact enables us in a general way to determiLe something con-
cerning the time when these granitic deposits appeared on this part of
the continent. They evidently were injected after the formation of the
Cambrian, and before the formation of the Carboniferous. At the time
when the Coal Measures were deposited these hornblendic granites had

26 BULLETIN OF THE

beeu worn down to something like the form in which they now appear.
If we are ever able to determine the age of the Roxbury conglomerate,
"we can place the period of the extrusion of some of these hornblendic
granites in a yet more accurate manner, for it is evident that those
of the Blue Hill region, as before remarked, were subjected to erosion
before the deposition of those conglomerates.

The foregoing account of the Cambrian localities of the Attleborough
district will, it is hoped, be suflBcient to attract the attention of geolo-
gists to this important district. Although this report is in its nature
preliminary, enough has been set forth to show the importance of the
field with reference to many problems in American geology.

-N\-.x;/-./ ^■■''>-?^"-^"1aj.<<v;->,\^-;v' %•''

^S' 4 .C' •^'":^^;^l|??^x^iM^S>i^^^^

;;;] li

MUSEUM OF COMPARATIVE ZOOLOGY. 27

Preliminary Description of North Attlehorough Fossils. By N. S. Shaler

and August F. Foerste.

1. Obolella crassa, Hall, var.

Plate I. Fig. 1.

Shell oval or sub-circular, the beak projecting a little beyond the general
outline of the shell. The surface is marked by numerous concentric, lamellar
striee, and also by rather strong radiating striae. The latter, although usually
continuous throughout their Avhole length, frequently become more or less dis-
jointed and laterally displaced in passing across certain of the more marked
concentric striae. The radiating striae also vary at such points in their relative
prominence and distinctness.

The interior of the shells differs considerably from that of typical specimens
of this species. The cast of the interior of the dorsal or anterior valve ex-
hibits two short triangular elevations at the beak, which represent the cardinal
area, and a depression between which corresponds to the cardinal tooth. On
either side are additional larger elevations, this pair representing the scars of
the cardinal muscles. Immediately above the second pair, the general surface
of the casts is strongly elevated, the elevation decreasing in distinctness to-
wards the margin. That part of this elevation which lies nearest to the hinge
margin is quite abrupt, and marks the position of the lateral muscular scars.
Along the median line of this elevation is a depression extending to above the
middle of the shell, the more or less distinct sides of which are known by
some writers as central muscular scars.

The cast of the ventral valve shows a median elevation, narrow and promi-
nent, at the beak, which represents very likely a notch in this part of the
cardinal area of the original shell. On either side of this elevation are two
laterally directed notches, in front of which is an elevation representing cardi-
nal muscular scars, and the elevated portions immediately behind represent
lateral scars. No satisfactory central markings could be distinguished. Where
it seemed that these could be detected, closer examination has shown them to
be too faint for determination.

Locality and position. — Stations Nos. 2 and 3, North Attleborough, Mass.,
Cambrian, 160 specimens; also at Troy and Schodack Landing, N. Y. ; St.
Simon and Bic Harbor, Canada.

2. OboleUa?

Plate I. Fig. 8.

Shell almost circular in outline, moderately convex, with no prominent
beak. The exterior surface is marked by concentric (exfoliated) striae of

28 BULLETIN OF THE

growth, a moderate distance apart, and distinct. The interior cast of the
dorsal valve is in general moderately convex, at the edges heing more finely
and less distinctly striate than the exterior surface. The margin along the
beak is flat. The cardinal scars in the cast follow the outline of the shell, and
are well defined along their exterior outline, but not along their interior. The
reverse is true of the casts of the lateral scars. The lateral scars unite with
the central scars, forming a figure comparable with that of a reversed W,
which is distinctly outlined along the outline facing away from the beak, but
is indistinct along the outline facing the cardinal scars. The diameter of the
shell is 5 mm.

Locality and position. — Station No. 2, North Attleborough, Mass., Cam-
brian, one specimen.

3. Fordilla (Troyensis, Barrande?).

Plate I. Fig. 4.

Cast of left valve 7 mm. long and 4 mm. broad, moderately convex along the
border, quite strongly convex near the hinge line. The broadest part of the
shell is slightly anterior to the middle of the length of the shell. Posteriorly
the shell decreases rapidly in breadth. The posterior extremity is rounded,
but more attenuate than in specimens figured by Walcott. Anteriorly the
border of the cast is narrowly indented; in consequence of the indentation,
the border is produced as a small lobe, and forms the anterior extremity of the
shell. A rather broad, shallow groove runs along the shell near the margin.
The cast shows no striae.

Compared with typical specimens of this species, the North Attleborough form
is largei', more attenuate posteriorly, and more strongly arched near the hinge
line. Walcott, in his Second Contribution to Cambrian Faunas,* figures, on
Plate XI. fig. 3 b, a cast which forms a connecting link between the form
here described and the typical forms, which have a broader posterior outline.

Locality and position. — Station No. 1, North Attleborough, Mass., Cambrian,
one specimen; also at Troy and Schodack Landing, N. Y.

4. Lamellibranch ?

Plate I. Fig. 5.

A single specimen of entirely unknown relations has at least the genei'al
outline of a Lamellibranch. It is 9.5 mm. long and 3.7 mm. wide. The
border is gently curved ; the hinge line almost straight. Along the hinge line
is a flat, strongly inclined narrow field, with fine striaj almost perpendicular
to the hinge line. The rest of the shell forms a surface almost perpendicular
to the hinge area. It is finely striated, the curved strise following the outline
of the border. The broadest part of the shell is about one third the dis-

* Bulletin U. S. Geol. Survey, No. 30.

MUSEUM OF COMPARATIVE ZOOLOGY. 29

tance from the anterior extremity of the shell. If indeed a Lamellibranch,
the specimen is the left valve of the shell.

Locality and position. — Station No. 1, North Attleborough, Mass., Cambrian,
one imperfect specimen.

6. Scenella reticulata, Billings.

Plate I. Fig. 6.

Shell small, conical. The aperture is almost circular; but there is a great
convexity of curvature on one side, giving rise to a low, indistinct carina, and
to a slightly oval outline at the aperture. The apex of the shell is slightly in-
curved towards the carinated side. The length of the aperture is 6.5 mm.,
the breadth 6 mm.; the height of the shell is 3.9 mm. The surface is faintly
wrinkled transversely. True concentric and radiating striae are not shown in
the specimen. The absence of radiating and concentric striae would at once
separate this specimen from the types of the species, but the markings are
so delicate that their preservation in this decayed rock would be extremely
improbable. In other respects, however, it is very much like the type
specimens.

Locality and position. — Station No. 2, North Attleborough, Mass., Cam-
brian, one specimen; also at Topsail Head, Conception Bay, Newfoundland.

6. Stenotheca rugosa, var. pauper.

Plate I. Fig. 7.

Shell small, decreasing rapidly in size towards the apex. Apex strongly
incurved. Shell corrugated into from five to eight rounded ridges, passing
transversely around the shell. These are crossed by very fine, closely set
striae, passing longitudinally along the shell. In the casts, the transverse
ridges are less distinct, and the longitudinal striae are not seen at all. When
not crushed, the apex is broad oval in outline. Diameter of the aperture
2.5 mm.; height, the same. Specimens are often smaller.

Locality and position. — Station No. 1, North Attleborough, Mass., Cam-
brian, 20 specimens. It is of some interest to note that this form does not
occur at Station No. 2, where two other varieties are found.

7. Stenotheca rugosa, var. abrupta.

Plate I. Fig. 9.

Shell small, decreasing rapidly in size towards the apex. The apex never
strongly incurved; usually within a moderate distance of a line vertical to
the base at its centre. Shell never more than slightlv curved. Shell corru-
gated into four or five rounded ridges, passing transversely around the shell,
verj' strong below, decreasing rapidly in size towards the apex. Greatest di-
ameter of the aperture, 4 mm. ; height of the shell, the same. Longitudinal
striae very fine and closely set.

30 BULLETIN OF THE

Mr. Walcott's note on the type of Stenotheca pauper recalls this variety ; bnt
our specimens are larger than is indicated by the original description of Bil-
lings, and the apex can scarcely be said to be incurved. The ridges are coarse,
and not small.

Locality and position. — Stations No. 2 and 3, North Attleborough, Mass.,
Cambrian, 30 specimens; also at Troy, N. Y., and Bic Harbor, Canada? Va-
rieties pauper and abrupta are not found connected by intermediate forms at
North Attleborough, and may be distinct species.

8. Stenotheca ciirvirostra, sp. n.

Plate I. Fig. 8.

Shell small, rather elongate; the lower part gently curved, the curvature
more marked, especially at the beak; the beak always considerably elevated
above the aperture of the shell. The transverse ribs are narrow and sharp;
from ten to eighteen are found on a single shell; the interspaces are broad
and flat. The longitudinal striae are fine and closely set. Diameter of the
aperture of the shell in the largest specimen found, 4 mm. ; height of the
shell 5 mm.

Locality and position. — Station No. 2, North Attleborough, Mass., Cam-
brian, 5 specimens.

9. Platyceras primaevum, Billings.

Plates I. and II. Fig. 10.

Shell very small, whorls two in number. Seen from above, the whorls Ue
very nearly in the same plane; they increase rapidly in size, the second be-
coming comparatively very large, and all being throughout evenly rounded.
Seen from below, only the last whorl is visible, and the whorls have a some-
what spiral form ; the ascent of the spire increases rapidly towards the aper-
ture, at that point partly overlapping the first part of the whorl. There are
faint traces of transverse striae ; this characteristic ornamentation of the
species would not be well preserved in the decomposed material in which
the North Attleborough specimens occur. Width of the shell 2.8 mm. ; height,
1.5 mm.

Locality and position. — Station No. 2, North Attleborough, Mass., Cam-
brian, a dozen specimens; also at Troy, N. Y., and Bic Harbor, Canada.

10. Pleurotomaria (Raphistoma) Attleborensis, sp. n.

Plate II. Fig. 11.

Shell small, flattened, composed of three whorls. The first whorl Ls very
small; the succeeding ones increase rapidly in size. The surface in general
slopes at a low angle from the apex of the shell to the sides. In the last whorl

MUSEUM OF COMPAEATIVE ZOOLOGY. 31

of the cast, the outside margin of the coil thickens a little, forming an indis-
tinct border along the margin of the shell, which becomes more evident as it
approaches the orifice. The edge of the whorl is compressed and rather nar-
rowly rounded. The surface of the shell is marked by fine, transverse, closely
set striae, which apparently are directed backward towards the earlier formed
parts of the shell, but in reality indicate various stages of growth of the shelL
The internal cast does not show these fine striae ; but broader and more widely
separated elevations, having the same direction as the striae. The diameter of
the shell is 3.2 mm. ; the height is a little less than 1 mm. Owing to the shape
of the shell it is difficult to measure its height accurately.

Locality and position. — Station No. 1, North Attleborough, Mass., Cam-
brian, one specimen. The discovery of another coiled gasteropod in this divis-
ion of the Cambrian of America is of interest, and makes the sudden influx of
coiled genera in the Upper Cambrian less inexplicable.

11- Hyolithes quadricostatus, sp. n.

Plate II. Fig. 15.

Shell straight, elongate, tapering gradually to an acute point ; apical angle
17°. The external cast of the type specimens is 20 mm. long; including an
additional length represented by an internal cast of the same specimen, but
extending farther from the apical extremity, it is 25 mm. long. As the end of
cast is broken, a length of 30 mm. may be presumed for the entire specimen.
The supposed dorsal side is broad and flat or slightly concave along the centre ;
when depressed, a low elevation may occur along the median line; in all cases,
the surface retains a rather flat appearance. The lateral angles are rounded.
The supposed ventral side is as usual flattened along the median line for about
half the vridth of the shell. The sides of this flattened surface are more or
less elevated, giving it a slightly concave appearance along the median line.
Immediately beyond the flattened surface on either side is a more or less dis-
tinct groove. The result is, that, in addition to the two lateral angles, there are
two angles or ridges on the ventral side, gi^Tng the entire shell the quadri-
costate appearance indicated by the specific name. The two ventral ridges in-
crease in distinctness as they recede from the apical extremity, and are usually
more distinct, or at least less rounded, than the lateral angles. At a distance
of six or seven millimetres from the apical extremity, the shell is crossed by
an apparently imperforate septum. The cast of this septum from the upper
side had the appearance of a flattened surface with a slightly elevated border
around the margin. The surface of the shell is marked by fine transverse
striae. The longitudinal ridges are less prominent on the interior cast of the
shell than on the exterior.

Taken by itself this species would appear to uc very distinct from the usual
forms of Hyolithes, but in reality it forms only the extreme of a series of
intermediate species, which begins with specimens characterized by numerous

32 BULLETIN OF THE

fine longitudinal striae, these striae increasing in size and diminishing in num-
ber until we have such forms as Hyolithes hexagonus, Barrande, with only four
ridges in addition to the normal two lateral angles. In the specimens here
described, this number is reduced to two additional costee. The character
of the variation is quite distinct from that aiforded by a more acute or salient
ventral median line alone.

Locality and position. — Station No. 2, North Attleborough, Mass., Cam-
brian, rare.

12. Hyolithes communis, var. Emmonsi, Ford.

Plate II. Fig. 17.

Shell slender, straight, gradually tapering. In the specimen here described,
the part from the apical extremity to the imperforate septum is lost. It is fair
to presume that the length of this unseen part was about 15 mm.; the re-
mainder of the specimen is 38 mm. long, so that an occasional total length of
55 mm. or 60 mm. would not be too high an estimate. The apical angle is 11°.
The dorsal side is flattish, more or less depressed along the median line, the
depression becoming more distinct at a distance from the apical extremity.
The ventral side is quite evenly rounded, and is strongly convex; the lateral
angles are also rounded, their position chiefly defined by the depression along
the median line of the dorsal side. The cast of the septum warpedly curved;
the convexity turned towards the apex of the shell; otherwise smooth, with a
faint, raised margin. The surface of the shell is ornamented by fine transverse
striae.

This species varies greatly in size, and the specimen here described is one
of the largest forms. The depressed median area of the dorsal side is most
characteristic.

Locality and position. — Station No. 2, North Attleborough, Mass., Cam-
brian, one specimen. The type specimens were found near Troy, N. Y.

13. Hyolithes Americanus, Billings.

Plate II. Fig. 16.

Shell straight, triangular, tapering gradually to an acute point. There is no
trace of a septum in the specimen at hand. The apical angle is 20°. The
dorsal side flattened or gently convex, with a slightly increased curvature at
the lateral angles. The lateral angles are but slightly rounded. The ventral
side is composed of two flat surfaces, which meet each other at the median line,
giving a triangular outline to the cross sections of the shell. The angle formed
at the median line is in type specimens never rounded; but there is a slight
tendency in some specimens to form a more or less distinct elevation or incip-
ient wing alone; this line. The flattened surfaces of the ventral side, meeting
sharply at the median lire, are characteristic of this species. The character of

MUSEUM OF COMPAKATIVE ZOOLOGY. 33

tlie ornamentation of the surface cannot be determined from the specimens at
hand. Mr. Billings describes them as being finely striated, "the stride curving
forwards on the dorsal side, then passing upwards on the sides at nearly a right
angle, curve slightly backwards on the ventrum." The specimen described
here is 11 mm. long, and is one of the smaller specimens of the species.

Locality and position. — Station No. 2, North Attleborough, Mass., Cam-
brian, one specimen. It occurs also at Troy, N. Y., and at Bic and St.
Simon in Canada.

14. Hyolithes princeps, Billings.

Plate II. Fig 25.

Shell large, straight, gradually tapering, very thin. The shells occur in
large numbers throughout the section at Locality No. 1. Their outline is
readily seen on almost any fracture of the rock in the bed in which the species
is found ; but owing to the irregular fracture of the rock and the large size of
the species, as well as to the very frail nature of the shell, entire specimens
are not found. The following description is dxawTi up from abundant frag-
mentary material.

The shells vary greatly in size, reaching at maturity a diameter of 9 or even
11 mm. and a length of 100 mm. or more. The dorsal side is flattened or mod-
erately convex. The ventral side is decidedly convex; usually the convex-
ity is more marked on one side of the ventral surface than on the other, the
latter side being often almost flat. The median line is rounded, yet usually
distinct enough to be recognized. The apical angle is very moderate. In some
specimens it is as low as 6°. In the original description of the species, it is
said to be as high as 15°. The lateral angles are also more rounded than in
typical specimens. The surface is marked by fine transverse strife.

It will be noticed that the specimens here described do not closely accord
with the type. The lateral angles are not prominent enough, and the apical
angle is lower; but they agree with those forms in size and in general appear-
ance. At any rate, the amount of variation seems insufficient to be ranked as
specific. One interesting feature of these specimens is the frequency with
which the shells of different individuals are found loosely inserted in each
other so that three or four shells are successively sheathed one within the
other, or they may be inserted side by side in a large individual. Small
slender shells apparently belonging to Hijolithellus micans also occur in this
position. Whatever may be the conditions of this sheathing, it does not
suggest any structural connection between the difl'erent shells at the time of
fossil ization.

Locality and position. — Station No. 1, North Attleborough, Mass., Cam-
brian, 200 specimens.

VOL. XVI. — NO. 2, 3

34 BULLETIN OF THE

15. Hyolithes Billingsi, Walcott?

Plate II. Fig. 20.

Shell small. The dorsal side flattened; the lateral edges distinct, but nar-
rowly rounded. The ventral side moderately convex or composed of two
flattened surfaces meeting along a rounded median line. Shells are often tri-
angular in cross section and are then distinguished from Hyolithes Araericanus
by the rounded character of the median line on the ventral side, and their
smaller size. In the specimens figured by Walcott, the dorsal side is slightly
curved, this concavity being almost filled up again by a low, broad, median
elevation. This feature has not been detected in the North Attleborough
specimens. The identification of this form is entirely unsatisfactory, owing
chiefly to the imperfect material at command.

Locality and position. — Station No. 1, North Attleborough, Mass., Cam-
brian, 6 or 7 specimens. Also at St. Simon and Bic Harbor, Canada, and
Silver Peak, Nevada.

16. Hyolithellus micans, Billings.

Plate II. Fig. 23.

Shell very slender, often 22 mm. long, tapering gradually to a pointed ex-
tremity. Cross-sections are circidar, unless disturbed by pressure. The surface
is marked by fine transverse striae. These strise may be either of approximately
equal size or at more or less regular intervals may have intercalated single
striae of larger size. The apical angle is 4° or 5°.

The identification here made is based chiefly upon the slender form and cir-
cular outline of the shell. The operculum has not been found.

Locality and position. Station No. 1, North Attleborough, Mass., Cam-
brian, 400 or 500 specimens. Also at Bic and St. Simon, Canada, and
Troy, N. Y.

17. Salterella curvatus, sp. n.

Plate II. Fig. 22.

Shell short, curved, rather rapidly tapering. Cross-sections circular. The
curvature of the shell can usually be referred to one plane, but sometimes it is
slightly irregular, ha\dng what might be called an incipient spiral structure ;
the apical angle varies from 8° to 12°. The surface is smooth, or ornamented
hy faint, scarcely vi.-;ible transverse striae, in no manner comparable with the
much stronger striae of S. pulchella, Billings.

Locality and position. — Station No. 1, North Attleborough, Mass., Cam-
brian, 20 specimens. Also at L'Anse au Loup, Labrador, and Point Levis,
Canada.

MUSEUM OF COMPARATIVE ZOOLOGY. 35

18. Aristozoe ?

Plate II. Fig. 18.

What appears to be a hinge line is quite straight, 4.3 mm. lung ; valve
oblique, the longest diameter from the anterior end of the hinge line to the
opposite border is 7.6 mm, and is inclined to the hinge line at an angle of
about 50°. A large tubercle with broad base and almost pointed extremity is
situated at a slight distance from the middle of the valve, being closer to the
posterior extremity of the hinge line than to the opposite parts of the valve.
Between the tubercle and the hinge line is a depression. A slight depression
is found between the tubercle and the posterior border of the shell. A strong
groove borders the tubercle along its anterior border and extends along its side
in the direction of the longest diameter of the valve. Half-way between the
tubercle and that part of the border directly opposite to the posterior extremity
of the hinge line is a ridge which is well defined towards the border by a
depression. Two low grooves cross this ridge transversely, connecting with the
groove at the base of the tubercle. The valve is in a fair state of preservation.
The aflBnities of this form are very doubtful.

Locality and position. — Station No. 2. North Attleborough, Mass., Cam-
brian, one specimen.

19. Microdiscus belli-marginatus, sp. n.

Plate II. Fig. 19.

Head semicircular, 4.3 mm. long and 5 mm. broad. The glabella is oblong,
strongly convex, slightly narrowed in front. It is well defined by a deep, dis-
tinct groove, which continues around the sides and anterior part of the glabella ;
it is not connected at the front with the groove which lies along the border of
the head and within the rim. The marginal groove is deep, broad in front,
gradually growing narrower towards either side ; it gives a high relief to the
rim. The marginal rim has very nearly the same breadth throughout its
length ; it is beset with small tubercles, usually sixteen or eighteen in number,
which lie near the interior margin of the ridge. The tubercles directly in front
of the glabella are often indistinct or obsolete. The occipital furrow behind
the glabella is low, and not very distinct. It serves chiefly to bring into greater
prominence a tubercle on the middle of the occipital ring. This tubercle is
directed backwards, varies in size, and is often low, and again may become a
large sharp-pointed tubercle in the form of an incipient nuchal spine. The
extension of the occipital furrow along the posterior part of the cheeks is very
deep and marked, giving high relief to the cheeks. The posterior rim is very
narrow, but sharp and distinct, and the postero-lateral extremities of the head
have very s;niall acute terminations, without which they would appear some-
what rounded. The cheeks are connected in front by a narrow, sharply rounded
ridge, which lies a short distance from the glabella, along its anterior border.

3C BULLETIN OF THE

The pygidium is of an oval form, and is about 5 mm. broad and 4.3 mm.
long. The middle lobe is strongly divided from the side lobes by grooves. It
is very convex, and is also curved antero-posteriorly, giving the pygidium a
strongly convex outline from front to rear as well as from side to side. It is
divided into nine or ten segments ; along the median line is a series of tuber-
cles, very distinct on the anterior segments, diminishing in size near the poste-
rior extremity. The sides show no traces of segmentation. They are connected
posteriorly by a narrow ridge .similar to that connecting the cheeks. The rim
is sharp and distinct, being well defined by a furrow which lies between it and
the side lobes. The specimens are u.sually of the size above noted, but one al-
most entire pygidium found at locality No. 2 must, when perfect, have been at
least 8 mm. long.

Locality and position. — Stations No. 2 and 3, North Attleborough, Mass.,
Cambrian, thirty specimens.

20. Microdiscus lobatus, HaU.

Plate II. Fig 13.

Head minute, 2.2 mm. long. The glabella is cut transversely by two fur-
rows, giving rise to three lobes, of which the anterior one is considerably larger
than the rest. From the occipital ring to the first lobe the glabella grows
narrower. The first lobe itself is again larger. The occipital groove is also
well marked and the occipital ring has the efl'ect of another lobe to the gla-
bella. The grooves separating the glabella from the cheeks are deep and dis-
tinct. The cheeks are prominent and strongly convex, bordered distinctly by
the deep continuation of the occipital furrow. Anteriorly the border is rather
broad, becoming narrower along the sides. A moderate groove defines the in-
terior of this border anteriorly ; it decreases much in breadth along the sides.

Locality and position. — Station No. 2, North Attleborough, Mass., Cam-
brian, one specimen.

2L Paradoxides "Walcotti, sp. n.

Plate II. Fig. 12.

A single specimen was found showing the under side of the integument which
covered the head, 2.8 mm. long, and 3.8 mm. broad. The cheeks are in posi-
tion, and the facial suture is barely indicated by a faint line running from the
anterior extremity of the palpebral lobe forward, bending at first a little out-
ward, then more rapidly inward near the margin, which it cuts; posterior to
the palpebral lobe it almost immediately cuts the margin in a slight outward
curve. The outline of the head forms a curve, which would be semicircular
were it not for its disproportionate breadth. There are faint indications of a
spine at the postero-lateral extremities. The glabella is broad in front, the
posterior half with incurved sides, narrowing to half its anterior width. The
occipital furrow is distinct, and the occipital ring has a distinct tubercle at

MUSEUM OF COMPARATIVE ZOOLOGY. 37

the middle. The glabella is marked by three pairs of shallow furrows, with
perhaps a fourth scarcely discernible pair. The second and third pairs are not
seen to meet across the median line; but owing to the position of two very low
and rather indistinct tubercles, one anterior and one posterior to the first or
posterior pair of furrows along the median line, these furrows seem to meet in a
curve bending slightly backwards along the middle of the glabella. The palpe-
bral lobes are large and prominent, beginning a little anterior to the third pair
of furrows, and curving around to within a very short distance of the exten-
sion of the occipital furrow across the cheeks. The curve along the anterior
border of the head is regular. A shallow groove runs within a short distance
of the anterior border, gradually becoming deeper and broader and receding
more from the border on the sides of the head, so that the rim here becomes
broader. A faintly discernible shallow pit near the anterior extremity of the
glabella may in this case be only accidental.

Paradoxides tenellus, Billings, is in size like this species, but otherwise
very distinct. It is interesting to find a Paradoxides in the Olenellus Cam-
brian, since its occurrence there diminishes the importance of the Paradoxides
Cambrian as a Paradoxides division.

Locality and position. — Station No. 2, North Attleborough, Mass., Cam-
brian, one specimen.

22. Ptychoparia mucronatus, sp. n.

Plate II. Fig. 31.

Glabella in small specimens very convex, the degree of convexity decreasing
with the increase of size; general form oval, the anterior part becoming some-
what narrowed. The occipital furrow is always well defined. There are three
pairs of glabellar furrows. These vary greatly in distinctness, being as a rule
less prominent in young individuals and more marked in large specimens.
The anterior pair is also usually much less distinct than the other two. The
occipital ring is well rounded, except where interrupted by the nuchal spine,
which is generally present, but usually small, being a mere terminal tubercle
directed backwards and often more prominent in young specimens than in
older individuals. From this are all variations to that of a moderate-sized
spine. In one specimen, a fragment, the head of which may possibly have
been 14mm. long, the spine extended for a distance of 3.2 mm. beyond the
general outline of the occipital ring, the base being broad, narrowing suddenly
to a short slender spine. Many large specimens are found in which this nuchal
spine is not seen. Examination, however, always indicates that in these cases
the posterior extremity of the occipital ring has been injured, so that the ab-
sence of a nuchal spine cannot be definitely asserted. They agree perfectly in
every other respect with the spined forms. Considering that we have positive
evidence of large forms with spines, the absence of the same in specimens all
apparently more or less injured seems to mean little. One specimen 15 mm.

38 BULLETIN OF THE

long has a process of the usual type, namely, a small mucronate tubercle or
spine at the posterior extremity. The species occasionally attains a length of
head of 20 mm. The average size is within 10 mm. Anterior to the glabella
is a groove which separates the anterior border of the head from the glabella
and the ocular ridges. Within the border is a broad, shallow groove. It is
of medium size, rounded and curved. The distance between the extremities
of the border, at the facial suture, is slightly less than the distance between
the grooves defining the palpebral lobes. In line with the anterior margin
of the glabella, or slightly behind the same and parallel with the anterior
border, are the ocular ridges, increasing in prominence with the size of the
individual, joining laterally the anterior end of the palpebral lobe. The
groove which more or less distinctly defines the posterior margin of the ocular
ridge joins the more distinct groove which separates the palpebral lobe from
the fixed cheeks. The palpebral lobes are obliquely curved, having a postero-
lateral direction. The facial sutures anterior to the palpebral lobes bend
slightly outwards to meet the anterior margin of the head. Posteriorly they
curve towards the side and backwards, cutting the posterior edge within the
postero-lateral angles. The cheeks are more convex in young specimens; in
larger individuals they are only moderately curved. The extension of the
occipital groove over the sides of the head is quite deep and distinct. Numer-
ous specimens of free cheeks show that the postero-lateral extremities of the
head were quite strongly spined.

Three specimens have been found preserving most of the segments of the
thorax, the posterior ones being more or less injured. One of these specimens
shows thirteen segments, but there may have been fourteen or fifteen in the
complete individual. The pygidium, judging from the specimens at hand,
must have been relatively very small, perhaps about the size of that of
Ptychoparia Piochensis. The pygidium has not been found.

The side lobes of the thorax are moderately broader than the axial lobe.
The middle lobe is strongly convex, and marked with a median row of mu-
cronate tubercles, or small spines. These in the individual best preserving
them were more prominent along the middle segments, being of moderate size
anteriorly and practically obsolete in the last three or four segments. The
species, as already noted, is quite variable; but the variations are none of them
of any marked character, and all are abundantly connected by intermediate
specimens. It takes the place of the series of species from the Vermont sec-
tions known as Ptychoparia Adamsi, P. Teur.er, P. Vulcanus, and the type of
fossils in which the border is separated only by a short interval from the gla-
bella, as figured by Walcott under P. Adamsi (Bulletin U. S. Geol. Survey,
No. 30, PI. XXVI. fig. 1 c). These specimens would have been placed under
P. trilineata, Emmons, had not such a good observer as Walcott decided,
from a personal observation of the types, that the species was properly a
Conocoryphe, which our specimens decidedly are not.

Locality and position. — Stations Nos. 2 and 3, North Attleborough, Mass.,
Cambrian, 300 specimens.

MUSEUM OF COMPARATIVE ZOOLOGY. 39

23. Ptychoparia Attleborensis, sp. n.

Plate II. Fig. 14.

Head small, often minute; in the largest specimen, 4mm. long. The usual
size is about 2.6 mm. The glabella is oblong or slightly attenuate anteriorly.
The occipital groove is low or indistinct. The occipital ring extends beyond
the general posterior outline of the head, and apparently forms part of the
glabella before it. The glabella is sometimes intersected by faint lateral
grooves, of which there are three pairs, the anterior pair scarcely visible.
Oftener these grooves are obsolete, and the glabella may, in case the occipital
groove is very slight, appear as a continuous undivided body as far as the poste-
rior margin of the head. The glabella is always convex, and considerably ele-
vated above the general level of cheeks. There is in some specimens a very
slight trace of an ocular ridge, which runs from the anterior end of the gla-
bella laterally, and slightly posteriorly, joining a similar slight trace of the
palpebral lobes. The most marked feature of the fixed cheeks is the existence
of a depression along their postero-lateral outline. The anterior border is
proportionately very broad. About the character of the rim little can be said.
Near the lateral margin of the border, or rather near the facial suture, there are
sometimes two or three low tubercles visible. There is also in some specimens
a faint trace of a sufficient elevation of the border to indicate an incipient
marginal rim. A careful comparison of these specimens with published figures
of P. subcoronata, Hall and Whitfield, a specimen of similar size, shows numer-
ous differences, which are too marked to permit the Attleborough specimens
to be placed under the same species.

Locality and position. — Station Xo. 2, North Attleborough, Mass.,
Cambrian, 20 specimens.

40 BULLETIN OF THE

EXPLANATION OF PLATES.

PLATE I.

Fig. 1. Obolella crassa, hc.il, var. ; a, dorsal valve ; b, interior cast of the same,
the features of the central area exaggerated to explain theoretical views ; c, the
same, in its normal state ; d, the interior surface of the dorsal valve diagrammati-
cally represented ; e, the interior cast of the ventral valve ; /, the interior surface
of the ventral valve diagrammatically represented.

In the diagrammatic figures: x, cardinal area; a, cardinal muscles; <I, lateral
muscular scars ; c, central muscular scars ; p, a notch in the cardinal area.

Fig. 2. Obolella ; a, ventral valve ; b, interior surface of the same ; c, a diagram-
matic representation of the same.

Fig. 3. Prol)ahly a cast of an operculum of some species of HyoUthes.

Fig. 4. FordiUa Troyensis, Barrande ?

Fig. 5. Lamellibranch ?

Fig. 6. Scenella reticulata, Billings ; a, lateral view ; b, outline of base.

Fig. 7. Stenotheca rugosa, var. pauper.

Fig. 8. Stenotheca curvirostra, sp. n.

Fig. 9. Stenotheca rugosa, var. abrupta ; a, normal type ; b, a single larger speci-
men.

Fig. 10. Platyceras primcevum, Billings ; a, seen from above ; 6, an umbilical view
of the same, specimen laterally compressed.

PLATE II.

Fig. 10. Platyceras primcevum, Billings ; c, a specimen of normal type.

Fig. 11. Pleurotomaria (Raphistoma) Attleborensis, sp. n.

Fig. 12. Paradoxides Walcotti, sp. n.

Fig. 13. Miscrodiscus lobatus, Hall.

Fig. 14. Ptychoparia Attleborensis, sp. n.

Fig. 15. HyoUthes quadricosfatus, sp. n. Ventral view ; a section near its larger
extremity, and a basal view of the septum.

Fig. 16. HyoUthes Americanus, Billings ; ventral view, and cross-section near its
larger extremity.

Fig. 17. HyoUthes communis, var. Emmonsi, Ford ; dorsal view, with cross-section
of its larger extremity, and basal view of the septum.

Fig. 18. Aristozoe?

Fig. 19. Microdiscus belli-marginatus, sp. n. ; a, head ; b, pygidium.

MUSEUM OF COMPARATIVE ZOOLOGY. 41

Fig. 20. Hyolithes Billingsi, Walcott ? three specimens with cross-sections.

Fig. 21. Ptychoparia mua-onatus, sp. n. ; a, glabella, with very marked terminal
spine, smaller type ; b, glabella, larger type ; c, movable cheek associated with the
same ; d, specimen preserving thoracic segments.

Fig. 22. Salterella curvatus, sp. n. ; two specimens, the lower slightly coiled.

Fig. 23. Hyolithellus micans, Billings.

Fig. 24. Not numbered in the plate, a movable cheek, relationship unknown.

Fig. 25. Hyolithes princeps, Billings ; a, ventral view ; b, sections showing in-
vagination of different individual shells of same species ; some of the smaller with
more circular outlines may also be those of Hyolithellus micans, Billings.

Fig. 26. Microdiscus speciosus, Ford ; head ; figure introduced as a means of com-
parison with 21. belli-marginatus. From Olenellu? Cambrian of Troy, N. Y.

Shaleri- Foersle.
Al'leboi'o h'ossils.

la

I'lnlf 1 .

Id

Jf

I
I

r

62.

/-

\

"fa

IV a

c

101

AFF

Shaler&- Foerste.
Alt!ehoi-o Fossils.

Plate .2.

11 I

a

l^a.

JO

'<t^^.. J

/f^

^^J^^i ^^

(7

i3

/^FF

No. 3. — Specimens of Fossil Plants collected at Golden, Colorado,
1883, for the Museum of Comparative Zoology at Cambridge,
Mass., examined and determined by Leo Lesquereux.

[Returned to the Museum, July 17, 1884.*j

CRYPTOGAMOUS PLANTS.
Lycopodiaceae.

1. Selaginella Berthoudi, Lx. 2 specimens.

Filicaceae.

2. Sphenopteris Lakesii, Lx. 62 specimens.

3. Sphenopteris membranacea, Lx. 2 specimens.

4. Hymenophijllites confmus, Lx. 5 specimens.

5. Ptcris pseudopennceformis, hx. 2 specimens.

6. Pteris suhsimplex, Lx. 20 specimens.

7. Pteris erosa, Lx. 1 specimen.

8. Pteris undulata, sp. nov. Leaves large, linear-lanceolate, regularly deeply
nndulate-crenate especially in the upper part; secondaries thin, distant, de-
clined in joining the rachis, open in passing toward the borders; forking once
at base, sometimes once again near the borders, very distinct.

Much like P. suhsimplex, Lx., U. S. Geol. Surv. of the Terr., VIL p. 52, PI.
TV. tig. 5; but with the secondaries thinner and the borders undulate. The
leaves, about 12 cm. long, 3^ cm. broad, are coriaceous with polished surface.
The angle of divergence of the veins is more acute, 45? to 50°, and their dis-
tance 1^ mm. 2 specimens.

9. Woodxoardia latiloba, Lx. 53 specimens.

10. Gymnogramma Haydenii, Lx. 8 specimens.

Equisetaceae.

11. Physagenia, species. Tubercles attached to filaments diverging in rows
from a central point, composing the rhizoma of some Equisetacece. Central
point exactly round, 2 mm. in diameter; tubercles oval, 12 mm. long, 6 mm,
broad in the mitldle, strangled to 2 mm. at the point of union, and forming a
chain of which two of the tubercles are seen in chjse connection. They are

* The manuscript of this memoir, as it was delivered in 1884, is copied without
any correction. — Editob.

VOL. XVI. — NO. 3.

44 BULLETIN OF THE

deeply irregularly ■wrinkled lengthwise, of the same size and aspect as those of
P. Parlatoni, Heer, Fl. Tert. Helv., p. 100, Plate XLII. figs. 6 a and 13. A
large number of fragments of stems are mixed with the tubercles distributed in
most of the specimens. 21 specimens.

PHANEROGAMOUS PLANTS-
Palmae.

12. Geonomites Goldianus, Lx. 10 specimens.

13. Geonomites graminifolias, sp. nov. Broken or separated rays of Palm,
varying in width from 5 to 10 mm., marked in the middle by a broad nerve,
the borders thinly regularly striate by 10-15 thin veins scarcely distinct even
with a lens. These fragments may belong to leaves of Calamopsis Danaiy Lx.,
or to a species much like it. One leaf in a better state of preservation has thi
rays narrowed to a point of connection to a broad thinly lineate rachis, united
3 to 5 together, either connate part of their length, or disconnected to the base,
like the fragments described above. Though the likeness to Calamopsis Danai,
figured and described in Trans. Am. Phil. Soc, XIII. p. 411, Plate XIV.
figs. 1-3, is marked, I have never seen the rays divided in narrow laciniae as in
this species, generally found in small narrow linear segments. 7 specimens.

14. Geonomites, species undet. 1 specimen.

15. Falmocarpon commune, Lx. 13 specimens.

16. PcUmocarpon lineatum, sp. nov. Seeds small, oval or oblong, obtuse at
apex, subtruncate at base, regularly thinly but distinctly striate.

The seeds 4 mm. long, 2^ mm. in diameter, are very numerous, apparently
derived from racemes. 8 specimens.

17. Palmocaiyon truncatum minor, Lx. 8 specimens.

PiperacesB.

18. Piper Heerii, s^). nov. Leaves subcoriaceous, round or oval, very entire,
palraately nerved from the base ; lateral nerves very curved, the outer follow-
ing the borders up to the middle of the leaf, the inner aerodrome.

The fragment, the half of one leaf cut lengthwise or along the medial nerve,
is, in all its characters, identical with the leaf described by Heer as Piper
antiquum, Beitr. zum Foss. Fl. v. Sumatra, p. 11, Plate I. fig. 7. As the leaf is
fragmentary, the petiole being absent, I cannot well identify it with that of
the Miocene of Sumatra, though I am unable to see any marked difference.
The leaf described by Heer is unequdateral and long petioled, and none of
these characters can be seen from the American fragment. The nerves are
distinct ; the outer primary follows the borders at a small distance, its branches,
nearly at right angles, forming, by anastomosing curves, a series of areoles along
the borders from the middle downward, and the areas are traversed by ner^-illes
at right angles. 1 specimen.

MUSEUM OF COMPARATIVE ZOOLOGY. 45

AMENTACE^. (APETALE^.)

Betiilacese.

19. Betula fallax, sp. uov. Leaves rhomboiJal-ovate, cuneate to the base and
narrowed in the same degree from the middle to the apex, penninervate ; lower
pair of secontlaries attached to the midrib above the base of the leaves and op-
posite, the others, 5-6 pairs, parallel, at variable distance, at an acute angle of
divergence, somewhat curved in passing toward the borders ; border distantly
dentate in the upper part of the leaves, the teeth short and turned up, being
marked only at the apices of the lateral nerves which are mostly simple, the
lower ones only with few branches. The species resembles in the form of the
leaves Betula nigra, Linn. The leaves vary from 3| to 5^ cm. in length;
from 2^ to 3| cm. in width in the middle. The angle of divergence of the
nerves is only 30°, and therefore the leaves are only dentate in the upper part,
at the points entered by the secondaries. 32 specimens.

20. Betula Schimperi, sp. nov. Leaves small, as broad as long, round or
subcordate at base; ovate, acute, simply or doubly dentate; lateral nerves thick,
5 pairs, the three lower opposite, the lowest more or less branching outside,
all craspedodrome, as well as the divisions, entering the larger teeth directly
when simple, and the intermediate short ones by branches; curved in travers-
ing the areas at a broad angle of divergence. The leaves measure 2 to 3 cm.
across, both ways, and are deltoid-acute at the apex. The teeth, like the veins,
are somewhat distant, turned outside, sharply pointed, triangular. The base
of the medial nerve is pressed into the stone, and thus the leaves appear sub-
cordate ; but the base is really rounded and abruptly turned down to the ped-
icel. The surface is rugose by the deep impression of the nerves and ner-
villes, these being simple or branching in the middle. The relation of the
species is with Betula angulata, Goepp., Shoss. Fl., p. 10, Plate IIL fig. 3.
2 specimens.

21. Betula. Species not determinable. 1 specimen.

22. Alnus rugosa, sp. nov. Leaves membranous, elliptical-ovate, narrowed
to the base, undulate on the borders, pinnately nerved; lateral nerves equi-
distant, parallel, straight, craspedodrome, with few branches; surface rugose.
The leaves are about of the same size as those of A. nostratum, Ung., as figured,
Chlor. Protog., Plate XXXIV. fig. 1, and the nervation is of the same type;
but the leaves are narrowed, not rounded to the base. 2 specimens.

23. Alnus carpinifolia, sp. nov. Leaves comparatively small, ovate-acute,
narrowed and abruptly short-decurring to the petiole, entire to the middle,
denticulate above ; lateral nerves equidistant, the lower pair much branching,
all parallel, at an acute angle of divergence, craspedodrome.

The relation of the leaf is with that of Alnus cycladum, Ung., Fl. v. Kumi,
p. 23, Plate III. fig. 19, differing essentially by the base which is rounded,
and abruptly turned down to the petiole. The lower lateral nerves are joined
to the midrib a little above the base of the leaf. 1 specimen.

46 BULLETIN OF THE

Cupuliferse.

24. Quercus Haydinc/eri, Ett. 2 specimens.

25. Quercus straminea ? Lx. The leaf which I refer to this species is oval,
apparently denticulate near the apex, and narrower than those figured in the
U. S. Geol. Surv. of the Terr., VII. , Plate XIX. figs. 6, 7. It may therefore
belong to a different species. The facies and nervation are the same, and in
these last two leaves the upper part is destroyed. 2 specimens.

26. Quercus viburnifolia, h\. 11 specimens.

27. Quercus j^^eudo-alnus, Ett. 1 specimen.

28. Quercus celastrifolia, sp. nov. Leaves subcoriaceous, oval, equally nar-
rowed at both ends, obtusely pointed; secondaries very oblique, distant, par-
allel from the base, curving in passing toward the borders, ascending along
them, nearly simple, passing upward u.uder the teeth and joined to them by
short branches.

The teeth are acute, turned upward, one at the end of each of the secon-
daries which are subopposite in 5 or 6 pairs ; nervilles thin, at right angles to
the medial nerve, except near the Ijorders, where they turn upward like thin
tertiary nerves. The leaves are 5-6 cm. long, 3-3| cm. broad, the angle of
divergence 30-35°. 3 specimens.

29. Quercus color adensis, sp. nov. Leaves subcoriaceous, entire, oblong,
obtuse, rounded at the base and abruptly decurring in joining the petiole
(broken) ; secondaries 5-6 pairs, at an acute angle of divergence, campto-
drome, the lower ones branching, all connected by distinct though thin ner-
villes and running high up along the borders which are parallel in the middle.
By the size and form of the leaves, 5 to 6 cm. long, 3 to 4 cm. broad in the
middle where they are somewhat contracted, the species is related to Quercus
(Jreadmn, Sap., Fl. de Cumi, Ann. Soc. Nat., Plate II. fig. 11. 2 specimens.

30. Quercus JFhitei, sp. nov. Leaves of medium size, membranous, ovate-
lanceolate, cuneiform to the base, short-petioled, regularly more or less deeply
dentate ; secondaries straight, oblique, equidistant, sini])le, parallel, each en-
tering one of the teeth which are gibbous on the back.

This fine species is closely allied to Q. clymodrys, Massal., Foss. Fl., PL
XXII. , XXIII. , fig. 10, 12, esj^ecially with the variety described as CastclUnensis
by Capellini, Form Gessosa, p. 52, Plate V. fig. 1. It is also allied to Q. fur-
cinervis, Rossm., differing from both by the subdentate or umbonate teeth,
resembling those of Q. plaionia, Heer. The secondaries at an angle of 40° are
straight, the upper ones only slightly curved. The teeth, short upon some
leaTes, as long as i cm. upon others, are always distant and bossed on the back.
The leaves average 7 cm. long, and 4 cm. broad in the middle. 6 specimens.

Salicineae.

31. Poj)ulus Ungeri, Lx. The specimen referred to this s])ecies is a frag-
ment, the lower half of a leaf, apparently round, entire on the borders and

MUSEUM OF COMPARATIVE ZOOLOGY. 47

with the nervation of the species as figured in U. S. Geol. Surv. of the Terr.,
VII., PUite XXIV. fig. 5. The size is also the same. In this fragment the
basilar border is abruptly turned downward, and slightly decurrent to the
petiole, as in P. subrohmdata, fig. 8 of the same plate. The petiole is inflated
below the border, as in some species of Ficus. But that is apparently a result
of comjDression. 3 specimens.

32. Pupulus monodon, Lx. 1 specimen.

33. Populus mutabilis, Hr. 1 specimen.

34. Populus Gaudini, Heer. A small leaf, ovate in outline, narrowly long-
acuminate, truncate at base ; pinnately nerved ; nerves thin.

The leaf is of the same size as that of the species in Heer, Fl. Tert. Helv.,
Plate LXIV. fig. 3. It has also the same form ; except that it is narrowly
acuminate, like fig. 6 of the same plate and same species. The leaf is 5 cm.
long, 3-^ cm. broad near its base, and abruptly curved to the slender petiole.
1 specimen.

35. Populus Zaddachi, H. 1 specimen.

36. Populus arctica, Heer. The species is represented by a number of speci-
mens, all small leaves, 2-4 cm. long, crenulate on the borders, more or less
enlarged transversely, five-nerved from the base, coriaceous.

Except that the leaves are generally more distinctly crenulate, nothing in
the characters indicates a difference from those which I have figured in U S.
Geol. Surv. of the Terr., VII., Plate XXIII. and Plate XLVI., or of those in
Heer, Fl. Arct., I., Plate IV. fig 6 a. The nervilles are generally strongly
marked. 15 specimens.

37. Populus Nebrascensis, Newby. The species is represented by a very large
number of specimens, some of them with the characters indicated by the author
in " Illustrations and Notes on the Extinct Flora of North America," p. 62, Plate
XII. figs. 4, 5, while others are definite varieties, which could be considered as
species closely allied to P. arctica, P. Zaddachi, and P. Richardsoni of Heer.
I have separated the varieties by short diagnoses. The true P. Nebrascensis of
Newberry has the teeth of the borders unequal, always obtuse. It differs from
P. arctica by the absence of transverse nervilles, and the leaves longer, generally
ovate-lanceolate, nerved to the base. 145 specimens.

38. Populus Nebrascensis, vav. grandidentata,, Lx. Leaves broader, rounded
and undulate toward the base ; borders cut from the middle upward in large
deltoid obtuse gradually longer teeth. Some of the leaves are subtruncate at
apex with long irregular teeth ; others are rapidly narrowed to an obtuse apex.
The nervation is the same in the varieties as in the normal form, 3-5 palmate
from the base, with the inner pair of primary nerves curved inward and ascend-
ing to near the apexes and the secondary ones at a great distance from the base.
85 specimens.

39. P. Nebrascensis, var. rotundata, Lx. Much like the preceding, differing
from it by the broader leaves, broadly round and enlarged at the base. The
teeth are obtuse and large. 48 specimens.

40. P. Nebrascensis, var. acute-dentata, Lx. Leaves oval, narrowed at base,

48 BULLETIN OF THE

lanceolate above, generally palmately trinerved, teeth of the borders large,
equal, sharply pointed. 13 specimens.

41. P. Nebrascensis, var. longifolia, Lx. Leaves large, oblong-ovate, rounded
at base, five-nerved ; lateral primary nerves ascending to above the middle,
scarcely curved inside, much branching ; secondary nerves three or four pairs,
at a great distance from the base, camptodrome or craspedodrome, with their
divisions effaced at the borders which are cut in obtuse large teeth. The leaves
are longer, lanceolate from below the middle, 7^ to 11 cm. long, 4 to 7 cm.
broad below the middle ; the teeth are large, roimd, equal, marked from near
the base ; the petiole is long and slender. 15 specimens.

42. Pofulus tenuinervata, sp. nov. Leaves comparatively small, rouud or
ovate, broadly cuneate or rounded at base, palmately five-nerved ; inner pri-
mary nerves cur\ing inward and ascending near the apex ; those of the out-
side also curving and ascending to the middle, all camptodrome, thin but
distinct ; borders irregularly dentate, the teeth unequal and pointed ; nervillea
distinct, obliquely joined in the middle.

The leaves resemble those of a Ficus, being, in their facies, like those'of Ficus
crenata, Ung., which, however, has not distinct nervilles. They have a still
more marked likeness to those of Populus latior-tratisversa, Heer, as figured
by Ludwig in Paleont., V., Plate XXVI. fig. 3. The petiole is thick. 34
specimens.

43. PopHliis crenata, Ung., considered by Schimper a var. of P. mutabilis,
Heer. The leaves are referable to the species as it is figured by the author,
Fl. of Sotzka, Plate XV. fig. 6, being small, oval, rounded, and trinerved at
base, the primary lateral nerves at an acute angle of divergence, not much
curved inward, the borders with large unequal teeth, the substance coriaceous
and the surface polished. 17 specimens.

44. Populus attenvAita, Al. Br. 2 specimens.

Plataneae.

45 and 46. Platanus Guillelmce, Goepp., and P. aceroides, Heer. It is ex-
tremely difficult to separate the species. The description, of Heer, Fl. Alask.,
p. 473, merely defines the leaves of P. GuillelnuB as undivided or merely sub-
lobate, acutely dentate, narrowed into a short petiole ; secondary nerves at an
acute angle of divergence ; and he remarks in the explanation, that the leaves
of Disco which he refers to the species are tapering to the petiole, slightly lobed
or not at all, and have short teeth. According to this, most of the very numer-
ous specimens obtained at Golden are referable to P. Guillelm.<B. I refer to
p. aceroide those with more open primary lateral nerves and of a more solid
consistence. The determination is rendered more uncertain by the generally
fragmentary state of the specimens. 76 specimens.

47. Platanus Haydenii, Newby. A beautiful specimen, No. 508 of the list,
appears referable as a variety to this species. The leaf is oval in outline, tri-
nerved from above the decurring base ; the lobes, five, short, obtuse at the apex

MUSEUM OF COMPAKATIVE ZOOLOGY. 49

of the three primary nerves and of the two lower secondary ones which come
out of the midrib very obliquely, far above the base of the primary nerves or
nearly in the middle of the leaf. The borders of the leaf are obtusely dentate
between the lobes, the teeth being short, turned outside, sej^arated by flat
sinuses. The leaf is 9 cm. long, 'ndthout the petiole, of which only 1 cm. is
preserved and 5 cm. broad in the middle, cuneiform to the base, decurring
under the primary nerves, and tapering upward in narrowing about in the
same degree. This fine leaf is apparently of the same kind as that in Newberry,
*' Illustrations," Plate XIII. fig. 1, named Populus nervosa, var. elongata. It
is however distinctly trilobate, and referable to Platanus, on account of the
decurring base of the leaves under the lateral primary nerves. 16 specimens.

48. Populus Eaynoldsii, Newby. 23 specimens.

49. Populus rhomboidea, Lx. 6 specimens.

50. Populus, species undeterminable. 7 specimens.

Urticaceae- — Ulmaceae.

51. Ulmus quercifolia, Ung., Iconogr., p. 43, Plate XX. fig. 23. The speci-
men merely differs from the European species as figured by Unger in the
narrower more elongated base of the leaves. The borders are sharply dentate,
the lateral nerves distant, oblique, parallel from the base, passing in a curve
toward the borders, where they become effaced. 1 specimen.

52. Ulmus antecedens, sp. nov. Leaves small, thickish, oblong-lanceolate,
acute, subcordate and subequilateral at base, doubly or triply dentate ; teeth
short, curved upward ; secondaries thick, parallel, strong and straight, gener-
ally simple, sometimes forking in the middle, with thick oblique nervdlles.

The leaf has the same character as those of Ulmus crassifolia, of Texas. The
substance is thick, the size is the same, 4 cm. long, 2 cm. broad in the middle,
the widest part ; the lateral nerves 12 or 13 pairs. 1 specimen.

Moreae.

53. Ficus (Dombeyopsis) grandifolia, Ung. Considered by Schimper a syno-
nym of Ficus tilmfolia, Al. Br., differs by the coarser texture of the leaves and
the larger size. Of the leaves which represent this species, one, preserved en-
tire, is 15 cm. long from the base of the petiole, 18 cm. broad in the middle,
with the base prolonged downward into two auricles, descending 4 cm. lower
than the base of the medial nerve. 8 specimens.

54. Ficus tilicefolia, Al. Br. 18 specimens.

55. Ficus Bcrthoudi, sp. nov. Leaves thick and coarse, broadly cordate at
base, ovate-lanceolate, acuminate above, entire, enlai'ging toward the base and
rounding to the petiole, descending lower than its top, sometimes auriculate,
the basilar border in one leaf overlapping the top of the petiole ; primary
nerves deep and broad ; lower lateral nerves opposite, the upper alternate, all
very deeply curving toward the borders and following them in a series of

VOL. XVI. — KO. 3. 4

50 BULLETIN OF THE

areoles ; nervilles deep, close, parallel, cut by branches at right angles, forming
a square distinct areolation. 4 specimens.

56. Ficus asarifolia, Ett. 1 specimen.

57. Ficus Andrcei, sp. nov. Leaves subcoriaceous, long petioled, elliptical-
oblong, slightly emarginate at the rounded base, lanceolate above, crenulate on
th eborders ; palmately five-nerved ; lower lateral nerves with fewer branches
following the borders, the inner ascending the borders, near the apex some-
what incurved, much branched outside, camptodrome ; secondaries three pairs,
at a great distance from the base ; nervilles strong, at right angles to the
nerves.

The leaf resembles in shape and size some of those of Populus Richardsoni,
Heer. It differs essentially by a strongly fibrillose nervation, the narrowly
oval oblong shape, not enlarged, but subcordate at base. The leaf broken at
apex is 8 cm. long, 5 broad in the middle. The preserved part of the petiole
is 2 cm. long. 4 specimens.

58. Ficus auriculata, Lx. 8 specimens.

59. Ficus suhtruncata, Lx. 4 specimens.

60. Ficus spectabilis, Lx. 7 specimens.

61. Ficus occidentaiis, Lx. 5 specimens.

62. FiciLS irregularis, Lx. 2 specimens.

63. Ficus protogea ? H. A fruit. 2 specimens.

64. Ficus, species undeterminable. 4 specimens.

65. Protoficus Zeilleri, sp. nov. Leaves of medium size, coriaceous, rugose on
the surface, enlarged and round-cordate at base, deltoid at the acute apex, pal-
mately three- or five-nerved from the top of the petiole; lateral nerves much
branched; borders crenulate. The leaves, deeply rugose by the impression of
strong nervilles, are 6 to 7 cm. long, 5 to 5^ cm. broad below the middle; pri-
mary nerves 3 or more, generally 5, the lower at a broad angle of divergence,
following the borders, the inner ascending in a curve somewhat inclined to
the midrib; secondaries two or three pairs at a great distance from the base.
The borders, mostly destroyed, are seen crenulate, at the few places where they
are preserved. 5 specimens.

Latiraceae.

66. Laurus socialis, Lx. 6 specimens.

67. Laurus primigenia, Ung. 4 specimens.

68. Laurus Smidtiana, Heer, Fl. Shakal, p. 51, Plate XV. fig. 8. Leaves
petiolate, coriaceous, obovate, very entire; secondaries camptodrome, dissolved
in the reticulation. The fragments which I refer to this species are not quite
satisfactory for positive determination, though the outlines of the leaves, their
size and nervation, fully agree with the description and figures of the author.
The areolation is peculiar, composed of thin close nervilles, crossing the areas
at right angles to the midrib, or obliquely to the lateral nerves, as in some
species of Pyrus. But one of the leaves preserved nearly entire has the base

MUSEUM OF COMPARATIVE ZOOLOGY. 51

decuning to a short petiole, and is evidently a LauTus. That leaf is oval, nearly
obtuse. 3 specimens.

69. Laurus, species undeterminable. 1 specimen.

GAMOPETALE^.

Lonicereae.

70. Viburnum marginatum, Lx. 15 specimens.

71. Viburnum anceps, Lx. 5 specimens.

Oleaceae.

72. Fraxinus denticulata ? Heer. A fragment only, the upper part of a small
oblong oval leaf, with boixlers slightly denticulate, the lateral nerves mostly
craspedodrome, entering the teeth. By the form, the size of the leaf, and
the nervation, the fragment is similar to that described by Heer, Arct. Fl.,
II. p. 118, Plate XLVII. fig. 2. The identity is not positively ascertained.
1 specimen.

Sapotaceae.

73. Styrax anthra, Ung. Leaves membranous, broadly ovate, round-trun-
cate at base and abruptly attenuated to the petiole, entire, pinnately nerved;
secondary nerves curved, following the borders, camptodrome, branching at the
apexes, the lower pair following the borders, joined to the medial nerve at a
short distance from the base; nervilles transverse, oblique, continuous.

One of the leaves referred to this species is preserved whole, except the apex,
and is conformed in all its characters like that figured by linger, SiUog., III.
p. 34, Plate XXIV. figs. 19, 20. 2 specimens.

74. Styrax Laramiense, sp. nov. Leaf subcoriaceous, smooth on the surface,
ov.il, very entire, equally narrowed in rounding downward from the middle to
a short petiole and upward to the point (broken), pinnately nerved; medial
nerve straight, narrow; secondaries at an acute angle of divergence, a little
curved in passing to the borders, which they follow in areoles.

One leaf only 5 J cm. long, 4 cm. broad; the lateral nerves at an angle of
divergence of 30°. It is related to »S. officinale, Linn. 1 specimen.

POLYPETALE^.

Araliaceae.

75. Aralia notata, Lx. 2 specimens.

76. Aralia Hercules ? Ung. 1 specimen.

Ampelideae.

77. Cissus tricusjpidata, H, 3 specimens.

78. Cissus primcEva ? Sap. A fragment, the upper part of one leaf, deltoid
acuminate, with four pairs of opposite lateral nerves diverging from a thick

52 BULLETIN- OF THE"

midrib, representing, in shape at least, the figure of that species in Saporta,
Sezanne Fl., Plate XI. fig. 2. The sandstone is coarse and hard, and no trace
of areolation is distinguishable. 1 specimen.

79. Cissus parrotecefolia, Lx. 6 specimens.

80. Cissus I'jbato-crenatUf Lx. 1 specimen.

81. Cissus corylifolia, sp. nov. Leaves thickish, ovate, blunt at apex, simply
or doubly short dentate, strongly pinnately nerved ; lateral nerves at an acute
angle of divergence, close, parallel, scarcely curved in passing to the borders,
the lowest much branched on the under side, the upper ones branching near
their ends, craspedodrome ; nervilles at right angles to the nerves, simple or
branched in the middle, deeply impressed.

The leaves, finely preserved, vajy iu length from 6 to 9j cm. and from 4 to
74 cm. in width, being broadest a little below the middle. They have a degree
of likeness to Parrotia pristina, Ett., as figured in Fl. v. Bilin., Plate XXXIX.
lig. 23. 3 specimens.

82. Cissus duplicafo-serrata, .'sp. nov. Leaves of various size, subcoriaceous,
ovate, taper-pointed, roundetl at base, palmately three, more generaDy five
nerved; primary lateral nerves diverging from the midrib at acute angles,
scarcely curving or not at all, entering the teeth which are prolonged into short
lobes at a distance below the apexes, mnch branched outside ; borders doubly
irregularly dentate, the teeth pointed; all the nerves and their divisions dis-
tinctly craspedodrome.

The leaves are referable to Cissus, though they have a degree of affinity with
some varieties of Populus Xebrascensis. They differ essentially by the primary
lateral nerves not incurved, much branched, all the divisions, craspedodrome,
the teeth acute, the substance of the leaves thick. 7 specimens.

83. Cissus specfabilis, Heer. Fl. Schakal., p. 45, Plate III. fig. 3 b. Leaves
oblong-ovate, subcordate-emarginate at base, unequally dentate on the borders,
very entire at and toward the base ; lateral nerves branching.

The above description is that of Heer, which fully agrees with the characters
of the leaf which I refer to the species. The leaf is merely smaller; the teeth
though unequal are not distinct, but mere crenulations, the same as seen in the
figure of Heer. 1 specimen.

84. Viiis, species undeterminable. 1 specimen.

Hamam-elideEe.

85. Parrotia fagifolia. Ett. Leaf broadly oval, irregularly undulate on the
borders; lateral nerves simple, alternate, distant, oblique, running straight to
the borders as in a leaf of Fagus, which it resembles. The leaf has the charac-
ters of the species, as figured in Fl. v Bilin., Plate XL. fig. 24, and i^ positively
identified. 1 specimen.

Corneae.

86. Comus Studeri, Heer. 2 specimens.

MUSEUM CF COMPARATIVE ZOOLOGY. o'6

Nysseae.

87. Nyssa Europcea, Unf^. A fragment of a leaf which by its form, the
lower part of it especially, its nervation, and the thick curved petiole, is remark-
ably similar to the figure of that species in Ung. Sillog. PL Foss., III. p. 73,
Plate XXIII. figs. 6, 7, 10. 1 specimen.

Magnoliaceae.

88. Magnolia tenuinervis, Lx. 3 specimens.

Nelumboneae.

89. Nelumbium Lakesii, Lx. 2 specimens.

Malvaceae.

90. Pterospermites grandidentatus, sp. nov. Leaves large, sometimes very
large, somewhat like leaves of Platanus, palmately sub-five-nerved; the outer
lateral nerves being generally thin and shorter, much divided outside; lower
secondary nerves opposite, at a distance from the base; borders sharply dentate,
the teeth acute, turned upward, entered by the primary nerves and their
branches, while toward the apex the secondaries curve in festoons along the
borders, joined to the teeth bj' small anastomosing branches; nerviUes strong,
at right angles to the nerves.

This definition is about the same as that given by Saporta of P. incequifolius,
Sez. FL, p. 402, Plate XII. figs. 3-5. One of the leaves of Golden is well pre-
served, and merely diff'ers from those described by the French author in the lateral
primary nerves somewhat incurved not quite straight. Two other specimens
represent merel}" the base of two leaves with five primary nerves around the
point of attachment of the petiole and two smaller ones declining downward
to the cordate base, as in the leaves of Ficus {Dombeyopsis) grandifolia, Ung.
3 specimens.

9L Pterospermites, species. A mere fragment, the lower half of an oblong 1
comparatively small leaf, membranous, rounded and slightly emarginate at the
base, palmately nerved, with two paii-s of more slender nerves under the base
of the primary ones. The nervation is the same as in P. spectabilis, Heer,
Arct. FL, II. p. 480, Plate LIII. figs. 2, 3 ; but the species apparently diff'ers by
the nerves being more straight and the leaf apparently smaller. 1 specimen.

Tiliaceae.

92. Tilia antiqua ? Newby. Probably the species, though the areolation and
nervation of one of the leaves are much like that of Greviopsis sidcefolia, Sap.,
FL Foss. de Sezanne, p. 407, Plate II. fig. 10. There are only two leaves, upon
the same specimen, and both are fnigmentary, the borders being mostly

54 BULLETIN OF THE

destroyed. The teeth appear rather short, turned outside as in Greviopsis ;
but the transverse nervilles, much more distinct, have the character of those
of Tilia. 1 specimen.

93. Grewia crenata, Ung. Leaf nearly round, subcordate at base, obtusely
and obscurely crenate ; palniately five-nerved ; nervation camptodrome.

The leaf is like that in Ett. Fl. v. Bilin., Plate XLII. fig. 7. The marginal
or lower pair of primary nerves is only less marked, though distinct enough.
The leaf is 4 cm. in diameter. 1 specimen.

94. Grewiojjsis tenuifolia, Lx. 1 specimen.

Aceraceae.

95. Negundo decuirens, sp. nov. Leaves compound, trifoliate. Terminal
leafiet apparently large, trinerved from above the decurring base ; lateral
nerves thick, branching on the lower side, lateral leaflets at a distance from
the base of the terminal one, undulate and sparingly dentate on the borders,
trinervate above the decurring base ; secondaries thin with few branches ; ner-
villes indistinct, oblique to the medial nerve.

Species relatetl to Negundo triloba, Newby., Later E.\t. Fl. p. 51, Illust., Plate
XXIII. fig. 5, differing In' the prolongation of the base of the lateral leaf-
lets along the petiole, not cordate. Their nervation is however the same,
and the borders also are marked by few sharp teeth. It may be a marked
variety. 1 specimen.

SapindacesB.

96. Sapindus caudatus, Lx. 1 specimen.

Celastreae.

97. Celastrus Gaudini, sp. nov. Leaves membranous, rugulose on the sur-
face, large, oval-oblong or broadly oval, crenate or obtusely serrate, abruptly
narrowed at base ; secondaries alternate, about 6 pairs, much curved upward
in passing to the borders and effaced ; tertiary nerves and nervilles obsolete.

The leaves 5^ to 6 cm. long, 4|to 5 broad, are apparently obtuse, the upper
part is broken in all the specimens. The species is closely allied to C. Heerii,
Sism. Contrib., p. 449, Plate XXIX. fig. 5. 2 specimens.

Iliceae.

98. Ilex sphenophylla, Ung. Chlor. Protog., p. 149, Plate L. fig. 9. Leaves
very small, short-petioled, obovate, cuneate or rounded at base, coriaceous,
spinose-dentate, penninerved.

The specimens are very good and the species positively identified. 2
specimens.

MUSEUM OF COMPAKA.TIVE ZOOLOGY. 55

Rhamneae.

99. Paliurus Zizyphoides, Lx. 2 specimens.

100. Paliurus tenuifolius, Heer. Leaf small, membranous, ovate, acute,
entire to near the apex, where it is marked with a few acute small teeth ; lat-
eral nerves parallel to the borders up to near the apex, where they join the
secondaries by anastomosing branches ; secondaries opposite, at a great dis-
tance from the primaries.

The leaf has the characters of the species as described and figured by Heer,
Fl. Tert. Helv., III., Plate CXXII. fig. 31, and by Saporta, Ett., I. p. 122,
Plate XII. fig. 5. Heer, however, says of the leaves that they are very en-
tire ; but the upper part of the leaf figured by him, loc. cit., is destroyed, while
Saporta describes the leaves as subundulate-denticulate. The one which I
refer to this species, considering it a mere variety, is preserved nearly entii-e.
1 specimen.

101. Paliurus Coloradensi.% sp. nov. Leaf small, obovate, obtuse, denticu-
late above, triple nerved from above the base, with a single pair of opposite
secondary veins above the middle of the leaf, parallel to the basilar ones,
aerodrome and branching outside.

The top of the leaf is somewhat obliterated ; its nervation is much like that
of P. ovoideus, Heer, Fl. Tert. Helv., Plate CXXI. figs. 58, 59, Plate CXXII.
fig. 3. Its size is intermediate between that of figs. 58 and 59. But it greatly
differs by its obovate form and the position of the intermediate pair of sec-
ondary nerves, as thick as the primaries and parallel to them. 1 specimen.

102. Zizyphus fibrillosus, Lx., U. S. Geol. Surv. of the Terr., VII. p. 276,
Plate LII. figs. 1-6. The specimens of this species are fine, and the petiole
of one of the leaves is preserved 2 cm. long. The base of the leaves is either
round, subtruncate, or subcordiforni. The largest leaf is nearly 10 cm. long,
and 7 cm. broad in the middle. 7 specimens.

103. Zizyphus hyperboreus, Heer. The leaf is like that described in U. S.
Geol. Surv. of the Terr., VII. p. 276, Plate LI. fig. 15, doubtfully referred to
Heer's species described from Greenland specimens. The texture of the leaf
is coarse, the nerves very prominent ; the nervilles uniting the lateral nerves
to the borders are at right angles, thick, and the medial nerve has two pairs
of branches in the upper part. 3 specimens.

104. Zizyphus, sp. undeterminable. 1 specimen.

105. Ehamnus Goldianus, Lx. 13 specimens.

106. Rhamnus Cleburni, Lx. 7 specimens.

107. Ehamnus crenatus, sp. nov. Leaf large, ovate-lanceolate, acuminate,
rounded or subcordate at the base (broken), minutely crenate ; lower lateral
nerves more open, the upper gradually more oblique, all much curved in pass-
ing toward the borders, and inclining to the midrib ; the two highest pairs
aerodrome ; nervilles numerous, close, parallel.

This fine leaf is 12 cm. long (base and apex broken), 6 cm. broad a little
above the base. It has the same form, size and nervation as R. grosse-serratus,

56 BULLETIN OF THE

Heer, Fl. Bornst., p. 20, Plate IV. fig. 10, diifering by the borders being
minutely crenate and the nervilles less distant. 1 specimen.

108. Ehamnus rectinervis, Heer. 1 specimen.

Juglandeae.

109. Juglans rhamnoides, Lx. 3 specimens.

110. Juglans, species undetermined. 1 specimen.

111. Carya antiquorum, Newby. 1 specimen.

112. Pterocarya retusa, sp. nov. Terminal leaflet large, ovate-lanceolate;
the lateral small, linear-oblong, blunt at apex, rounded-subtruncate at base;
lateral nerves close, parallel at a broad angle of divergence, branching near
their ends ; surface rugose.

The leaflets are thick, denticulate or crenulate on the recurved borders, the
lateral nerves, at an angle of 50°, are not more than 5 mm. distant in the
small leaves ; deeply marked. It is closely allied to P. Americana, Lx., U. S.
Geol. Surv. of the Terr., p. 290, Plate LVIII. fig. 3, diff'ering by the open prox-
imate nerves. By its rugose surface and its nervation also, the species is
related to Juglans corrugata, Lndw. Palseont., VIII., Plate LXX. figs. 11, 12,
differing especially by the leaves being blunt or obtuse, not acummate as in
the European species. 6 specimens.

Pomaceae.

113. Cratwgus betulcefolia, sp. nov. Leaves subcoriaceous, variable in size,
oblong lanceolate, or enlarged toward the base, pyramidal-lanceolate above,
deeply acutely or obtusely dentate, trinervate from the base, pinnately nerved
from the middle ; primary and secondary nerves craspedodrome entering the
teeth ; lateral primaries with few branches.

The leaves vary in size from 3^ to 7 cm. long, and from 2^ to nearly 4 cm.
broad below the middle. All the nerves are deeply marked, the teeth some-
what long, rather obtuse, either turned outside or upward. 21 specimens.

114. Gratcegus myricoides, sp. nov. Leaves membranous, small, ovate-lan-
ceolate, more or less deeply dentate, penninervate ; lateral nerves simple or
forking, oblique and straight to the borders, or curving inside before entering
the teeth ; teeth alternately larger, or large and bi-tridendate on the back.

The two leaves representing the species are only 2 to 3 cm. long, 1^ cm.
broad, fragmentary. They resemble leaves of Betula ; but the irregular ner-
vation is that of Myrica, the two lower pairs of nerves being longer and enter-
ing the teeth, the upper shorter, curved, and effaced near the borders, all at
unequal distance, not parallel. They have some likeness to the leaves of Gra-
tcegus oxyacanthoides, Goepp., Schoss. Fl., Plate XXXVI. fig. 1. 2 specimens.

115. Gratcegus Engdliardti, sp. nov. Leaves of medium and small size,
subtrilobate or simply ovate, enlarged above the base, rounded and abruptly
deflexed to the petiole ; trinerved at base or pinnately nerved, the lower lateral

MUSEUM OF COMPARATIVE ZOOLOGY, 57

nerves being opposite, from above the decurring base ; borders lobate, denticu-
late, the lobes short, denticulate around.

Species resembling much C. tomentosa, Linn., in the form of the leaves, the
divisions of the borders, the decurring base, and the nervation, diileriug by the
leaves being generally less deeply lobed, and the teeth shorter. The leaves
vary from 7| to 11 cm. long, 4 to 7^ cm. broad below the middle, the widest
part. One of the leaves has the petiole or part of it preserved, 4 cm. long.
4 specimens.

116. Cratcegus antiqua, Heer. As far as can be seen, the leaves are referable
to this species, agreeing with it by the form, size, and nervation. The borders,
however, are mostly destroyed, and the teeth, apparently obtuse, rather than
acute as in Beer's species, Fl. Arct., I. p. 125, Plate L. figs. 1, 2. 5 specimens.

117. Ainalanchier tijpica, Lx., var. A simple leaf a little smaller than the
one figured in U. S. Geol. Surv. of the Terr., VIII., Plate XL. fig. 11, from
the Green River Group, and with the teeth smaller and pointed. I refer it
to the same species, but it may differ. The nervation is somewhat obscure.
1 specimen.

SUMMARY.

The collection of fossil plants from which the above data have been
derived was made at Golden, Colorado, by Mr. A. Lakes, and contains
873 specimens. The number of fragments of vegetable examined for
determination is 1,044. They represent 118 species, or vegetable forms
considered as species, 28 of which are admitted as new species and de-
scribed above, and 32 as new for the Flora of the Laramie Group, but
known from other localities, making therefore for that Flora an addition
of 60 species. Of those already known from localities outside of the Lar-
amie Group, 11 are American (3 from Carbon, 3 from the Green River
Group, 5 from Evanston), and 21 are European. All are of Tertiary
age, mostly observed in the Lower Miocene ; two of them are identified
with Sezanne species or Eocene.

Of the species described above from Golden, as of those formerly
known from the Laramie Group, either by the publications of Dr. New-
berry or of my own, none is identified with any of those of the Middle
Cretaceous (Cenomanian) or of the Dakota Group. In two only, a
marked affinity has been recognized and mentioned ; Populus elliptica,
Newby., Later Extinct Flora, Illust., Plate III. figs. 1, 2, which the
author says has a striking resemblance in general form to that of P.
cuneata, is closely related to P. arctica, Heer, of the Arctic Miocene ;
and Platanus primcBva, Lesqx., U. S. Geol. Surv. of the Terr., VI. p. 69,
Plate VIL fig. 2, is not less closely allied to Platanus Guillelmce, or

58 BULLETIN OF THE

P. aceroides, Goepp., two species also abundantly distributed in the Ter-
tiary, especially the Miocene of both continents. As from the Upper
Cretaceous or Senonian measures no species of fossil plants have as
yet been recognized as identical with or even related to any of those of
the Laramie Group, the assertion that the flora of this last formation is
Tertiary in its character remains positive and as yet unrefuted.

Indeed, as it can be seen in looking over the table of distribution,
not only some of the more predominant species of the Flora of the
Laramie Group are Miocene in characters ; but some of them are iden-
tified with species of the present epoch, or at least closely allied to them.
Woodwardia latiloba, for example, represented in the collection by 53
specimens, is a near relation of Woodwardia Virginica, Smith, not rare
in the woody swamps of the Northern United States. Bettda fallax,
with 32 specimens, has the same degree of relation to Betula nigra,
Linn., the Red Birch, Populus Nebrascerisis, typically and closely allied
to P. arctica and P. Pichardsoni, two very common species of the Arctic
Miocene, is represented with its varieties by more than 300 specimens.
Platanus Guillelmoe, abundant in most of the localities where Miocene
plants have been found, is represented by 70, and the allied species,
Platanus aceroides, P. Haydenii, P. Raynoldsii, by 45. The authority
of Professor Newberry is still more conclusive on the subject ; for in
describing the plants of the Fort Union Group in his Notes on the
Later Extinct Flora of North America, he not only finds most of them
related to Miocene species, but he identifies four of them with common
plants of the present epoch : Onoclea sensibilis, Corylus rostrata, Corylus
Americana, and Amelanchier similis, this last considered as a form of the
very variable and common Amelanchier Canadensis, Tor. & Gr.

Formerly, or before the examination of the new specimens sent from
Golden was made, I did not consider the Flora of the Union Group as
of the same age as that of the Laramie, known as it was to me by the
plant remains obtained at Golden, Black Buttes, and Point of Rocks.
There were between the plants of these localities and those of Fort
Union and the Yellowstone River, some points of affinity, marked in
the profusion of Palm remains, especially Sabal, of w^iich the most
common species, ^S*. Campbellii, Newby., first described from large speci-
mens of the Yellowstone, was found equally abundant at Golden and
the Raton IMountains. There were also a few identical species found at
Golden and Fort Union ; Platanus Haydenii, P. Baynoldsii, two species
of Juglans, two of Cissus, and a Carpolithes. But the facies between
the groups appeared different, that of the Union Group being strikingly

MUSEUM OF COMPAEATIVE ZOOLOGY. 59

conspicuous by the presence of some of the most common species of the
Miocene, even of some species still living at the present epoch, as seen
above, which had not been observed at anv of the localities of the
Eastern States where the specimens I had for my first examination
had been found. These species, Populus arctica, P. Nebrascensis, P.
crenata, Platanus Guillelmce, P. acei'oides, Aralia notatn or Platamis
nobilis, Tilia antiqua, Negundo decurrens, closely allied if not a mere
variety of Negundo triloba, Gary a antiquorum, Amelanchier typica, all
Miocene species, now found in the collection made at Golden, show an
intimate relation between the flora of Golden and that of Fort Union,
which by new discoveries will probably become more apparent between
the different localities of the Laramie Group, and complete the evidence
of tlie unity of the characters of the flora.

The group of plants described here aff'ord a remarkable evidence of
the distribution of vegetable remains under peculiar circumstances.
The specimens w'ere obtained from a locality which has been visited
many times by members of the U. S. Geological Surveys of the Terri-
tories, often by Mr. Lakes himself, and twice by myself. Nevertheless,
not a single specimen of Populus Nebrascensis, nor of Platanus Guillelmce,
had been found there before ; yet in a locality at the same horizon and
at a very short distance, a few rods only, as far as I know, from the ex-
cavations formerly made, specimens of these two species and their varie-
ties have been obtained in such abundance that they constitute nearly
one half of the collection. This proves that the mode of distribution of
the vegetable remains results in some cases from the deposition of the
fragments at the place where the trees have grown, not from transporta-
tion by water. Though the specimens of Golden are very fragmentary,
they are generally flat upon the surface of shale apparently composed
of muddy deposits, seemingly pi-ogressing in their formation while the
leaves were falling from trees grown either around woody swamps or on
the borders of shallow lakes.

Recent explorations have brought on the discovery of a large number
of localities rich in remains of fossil plants, over the whole extent of the
Great Lignitic. The flora of the Laramie Group, wliich now counts
only 250 species, will therefore probably soon become better known, and
by the greatly increased number of its species will take an important
place in the history of the ancient vegetation of the earth.

No. 4. — The Faults in the Triassic Formation near Meriden,
Connecticut : A week's work in the Harvard Summer School of
Geology. By William Morris Davis.

In previous articles on the Triassic formation of the Connecticut val-
ley, I have expressed the opinion that faults occun-ed between the ad-
jacent trap ridges. The opinion was based on the repetition of similar
sequences of strata, the evidence of which will now be stated in greater
fulness, so that it may be easily followed in the field by those who wnsh
to examine this interesting region. During much of the work, which
was undertaken for the United States Geological Survey, I have had
the assistance of Mr. C. L. Whittle, who is now engaged in preparing
an account of the results of his microscopic observations on the con-
tact phenomena of trap and sandstone. After many visits to the region,
a plan of exploration of the Meriden-New Britain district was laid out in
1887 for the students of the Harvard Summer School of Geology whom
I accompanied as teacher, and as it has borne very well the practical
test of two seasons of field study, it is adopted for presentation here.
The localities were not discovered by any means in the order here de-
scribed ; indeed the district was traversed many times before the sys-
tematic repetition of its oblique block structure was clearly perceived ;
but when this is once made out, it may be best explained by presenting
descriptions of its features as daily excursions in a selected order, in
which they may be most readily appreciated by a new-comer.

Excursion 1. — Cross-Section of Lamentation Mountain.

Meriden serves as a convenient centre for excursions. Its position in
central Connecticut and in the southern portion of the Triassic area of
New England is indicated by the black square in Fig. 1. About a mile
north of the city, (1) on Fig. 2,^ the Meriden, Waterbury, and Connecticut
River Railroad (Waterbury and Cromwell Railroad on the sketch maps)
crosses over the Consolidated Road (New York, New Haven, and Hart-
ford Railroad), and here one may begin the construction of a section

1 The sketch maps that illustrate this paper are copied from town maps, pub-
lished in county atlases. Tiie roads are approximately correct, but the ledges have
been located without instrumental measurement. In some cases, the structural
lines as now drawn may have to be changed when accurate maps are prepared.

VOL. XVI. — NO. 4.

62 BULLETIN OF THE

crossing the strike to the east, and thus mounting the monoclinal
sequence of beds which descend eastward with much uuitbrmity. Coarse
sandstones and conglomei*ates outcrop in small ridges with abrupt slopes
to the west, and control the local topography east of the railroad, where
very little drift has been deposited. The attitude of the beds is toler-
ably constant; the strike averages N. 30° E., and dip, 12° or 15° to the
east. Low ledges are plentiful over much of the distance, until the
Meriden-Berlin road is reached, half a mile east of the railroad ; then a
short ascent up a wooded slope leads to the higher ridge of amygdaloidal
trap (2), lying anterior to, i. e., in front of the face of Lamentation
Mountain. This ridge should be followed north about one tliird of a
mile, to the fine exposure of its bed of ashes and volcanic bombs (3),^
now locally well known through the efforts of the Meriden Scientific
Association, at whose expense the base of the cliff has been opened by
blasting, to secure fresh specimens of its remarkable rock. A fine shaly
sandstone may be seen at a few points at the bottom of the cliflF, under-
Iving the ash bed. One of the bombs seems to have embedded itself
in the sandy mud by the force of its fall, like the examples described
by ^crope in the recent volcanic region of central France.

Climbing the cliff and crossing through the woods, an old quarry (4)
is found at the edge of a pasture that slopes to the east with the dip of
the beds. The sandstone is seen here, with strike N. 25° E., and dip
13° to the east. Two dark layers, consisting chiefly of small rounded
and irregular fragments of trap, were found here by Mr. Whittle, and a
few feet lower, the top of the lava sheet is seen in a little ledge in the
woods. The lava is vesicular, and the sandstone immediately overlying
it contains many of its fragments. A blast lately fired has disclosed
this fairly well. Going eastward down the pasture, occasional outcrops
of shaly sandstone (5) are found with normal strike and dip, forming
little ridges in the valley between the anterior trap ridge and the main
ridge of Lamentation Mountain. They also appear occasionally in
benches (6), ou the steep western slope of the latter. Percival speaks
(Geology of Connecticut, p. 365) of a bed of impure limestone some-
where in this valley, but I have not been able to discover it ; in former
years the thin beds of Triassic limestone were burnt at a number of
points, but with the improvement in transportation and the bringing of
better lime from elsewhere, this has been given up, and the old lime-
stone quarries are often covered over and lost to sight. After making

^ This interesting locality was discovered on an excursion during the spring
recess from college work, on April 9, 1887.

MUSEUM OF COMPARATIVE ZOOLOGY. 63

out the structure of the anterior valley, the observer should return to
the anterior or amygdaloidal ridge and follow it northward about a third
of a mile from the ash cliff; here a trap-conglomerate (7) replaces the
ash and lava bed, and gives additional evidence of the contemporaneous
extrusive origin of the volcanic rock. Thin beds of sandstone occur in
the conglomerate, with strike and dip conforming to the general mono-
cline. Half a mile farther north, the lava again appears and, in its
northernmost outcrop (8), forms a bluff twenty to thirty feet high in a
spindle-shaped space between two roads. From this point, one may
renew the cross-section work by going eastward and climbing over the
thick bed of lava that forms Lamentation Mountain (9) ; but the result
is accomplished more easily by following the road, which lies chiefly on
a drift plain, and passing over the low north end of the mountain on a
level (10), until Spruce Brook is reached coming from the eastern slope
of Lamentation and running north. Crossing a field a hundred yards to
the south, the stream is found emerging from a little rocky channel
(11), cut in the uppermost part of the Lamentation trap sheet, and run-
ning down over the sloping beds of sandstone that rest upon it. The
strata dip and strike in accord with the general monocline. Close to
the trap they contain numerous vesicular fragments derived from it ; in
some of these, the vesicles may be seen to contain little deposits of
sandstone. It is an ideal " locality." A pool in the vesicular portion
of the trap, shaded by overhanging trees, tempts one to a bath after a
dusty walk. A little farther east, a ledge of brown shaly sandstone
(12) crosses the road, with normal strike and dip ; its outcrop persists
for some distance north and south at frequent intervals. Half a mile
yet more to the east, and at the further base of an immense drumlin,
known as Stow Hill, another small trap ridge (13) is discovered; this
may be called the posterior ridge, following Percival's nomenclature.
The points thus far described furnish about as much work as one may
care to do on foot in a summer day.

Returning along the eastern side of Lamentation, the posterior ridge
may be followed southward with little interruption to its end, just be-
yond the Cromwell railroad. Where the railroad crosses it, a small but
valuable contact of the lava is exposed with the overlying sandstone,
which here contains an abundance of trap fragments for three or four
feet over the back of the trap sheet, clearly indicating the extrusive
origin of the sheet. The vesicles in some of these fragments are more
or less completely filled with sand, and under the glass the bedding of
these little deposits is seen to conform closely with the stratification of

64 BULLETIN OF THE

the surrounding sandstone, whose strike and dip still conform to the
monocline of the region. Further extension of the cross-section to the
east is not necessary for the present, and the return to Meriden may be
made by the road that passes around the southern end of Lamentation
Mountain, there known as Chauncy Peak.

A general section across the strike of the monocline is placed in its
proper position oil Fig. 2. A vertical section, constructed on the mar-
gin of Fig. 2 from the material gathered on this excursion, contains
about fifteen hundred feet of sandstones and conglomerates at the base,
then two hundred feet of amygdaloidal trap in the anterior ridge, fol-
lowed by six hundred feet of thin bedded sandstones and shales in the
anterior valley, four hundred feet of heavy trap in Lamentation Moun-
tain, and fifteen hundred feet of sandy shales before coming to the pos-
terior ridge, which may be one hundred and twenty feet thick. The
total thickness of the aqueous and igneous beds traversed is some five
thousand feet. It is at present still uncertain from all that I can learn
how many feet of deposits are below the base of this section before
reaching the crystalline foundation on which the Triassic formation
rests ; and the thickness of the beds that originally overlay the poste-
rior trap is equally indefinite.

Excursion 2. — Cross-Section of Shuttle Meadow.

In order to measure the thickness of the underlying beds or to deter-
mine their sequence, a second excursion may be made to some point far-
ther west, — as by rail from Meriden to New Britain, and then on foot
or wheels to Shuttle Meadow reservoir, three miles to the southwest.
From the valley south of the reservoir, the ground rising to the east
leads over several outcrops of brown and red sandy shales (1, Fig. 3),
with strike N. 30° E., and dip 10° eastward, before reaching the bold
western face of the High Rock trap ridge (2). If one should climb
over this rugged ridge and descend across the meadow at its eastern
foot, a small trap ridge would be found (11), trending parallel to the
main ridge, but much concealed by drift. No shale or sandstone is seen
accompanying it, and so small an observation is hardly worth the time
it will cost. Returning now westward to the valley south of the reser-
voir and ascending its western slope, the enclosing ridge is found to con-
sist of amygdaloidal trap (3), with a bold cliff, trending N. 10° E., and
facing westward over the Southington valley-plain. A mile to the south,
a bed of impure limestone is exposed in a quarry on the back of the trap.

MUSEUM OF COMPAKATIVE ZOOLOGY. 65

Below the cliflf, a few outcrops of sandstone or conglomerate may be
found. These data for a cross-section are less complete than those col-
lected on the previous day, but they suffice for a rough diagram added
in the margin of Fig. 3.

On the supposition that the whole valley consists of an unbroken
monocline, the two sections now constructed can be placed in their
proper relative positions by means of a map, which would result in show-
ing that the second belongs about four and a half miles westward across
the strike from the first, and therefore its beds stand about five thou-
sand feet below those of Lamentation. But, if not already noticed,
attention will soon be called to the similarity between the two sections;
and the question then arises, how can this be best explained. It may
be the result of similar processes repeated in a given order, whereby
similar sequences of beds were deposited at different times ; or it may
be the result of one or more strike-faults, by which portions of a single
sequence of beds are brought to the surface in different places. It re-
mains to decide between these two alternatives.

A general consideration of the problem will make it evident that the
explanation by repetition of similar processes becomes less likely, and
that the explanation by faulting becomes more likely, with the increase
in the number of beds in the repeated series ; with the lack of genetic
relation among the members of the repeated series ; and with the in-
crease in the number of times that such repetitions occur. It should
also be noted that the two explanations are not mutually inconsistent ;
both might apply in a single field.

The drift covers so much of the surface that detailed sections cannot
be constructed. The stratified beds are, on the whole, so much alike,
that the precise identification of equivalent beds by agreement in com-
position is impossible. While it may yet be found that fossils will serve
as a guide to the recurrence of repeated outcrops, this means of identifi-
cation cannot at the present time be applied in the region we are exam-
ining. All that can be done is to make the best of imperfect evidence.

We may first examine the argument based on the number of members
in the repeated series. The completed sequence of visible beds is : con-
glomerate and sandstone ; trap of moderate thickness ; thin limestone ;
shaly sandstone ; heavy sheet of trap ; more shaly sandstone ; a thin-
ner sheet of trap ; and, finally, more sandstone. It may be conceivable
that this sequence of beds was independently made at two different
times ; but it is not at all likely that so considerable an agreement
should be the result of the accidental repetition of a sequence of deposits.

VOL. XVI. — NO. 4. 5

66 BULLETIN OF THE

The second argument is based on the arbitrariness of succession or
lack of relation between the different members of the series. Here the
interpolation of contemporaneous lava flows at certain points in the
series is significant ; for as far as is known they bear no necessary rela-
tion in the time of extrusion to the deposit of conglomerate, sandstone,
shale, or limestone. Professor Emerson has, it is true, regarded some
of the limestones that occur associated with the trap sheets in Massa-
chusetts as the product of thermal springs that were excited by the
eruption of the trap. I cannot say anything as to the cause of the lime-
stone deposit in Connecticut, but whether connected causally with the
eruption or not, the association of limestone and trap is by no means
invariable ; the limestone occurs only, as far as known, on the back of
the first trap sheet, and this in itself is enough to make one lean toward
the explanation by faults. If all the beds were sedimentary, and their
succession were of the normal kind described by Professor Xewberry in
his essay on " Circles of Deposition," the repetitions might perhaps be
explained without faulting; but it has been seen that such is by no means
the case. Before proceeding to the third argument, it may be noted that
the explanation thus far given does not depend altogether on the contem-
poraneous extrusive origin of the trap ; it is important to note this, for
while extrusion has been well proved for the first or Lamentation section by
observations on the previous day's excursion, it is not yet directly proved
for the High Rock section, although as will appear in the sequel there
can be little doubt that all the sheets of the district are extrusive. But
if intrusive, it is difficult to conceive that intrusive sheets should have
taken their places among the bedded rocks as systematically as these
Meriden sheets have, unless they had been driven in between the beds
before they were tilted, and then faulted afterwards. Therefore, whether
the trap is intrusive or extrusive, the evidence thus far collected favors
the hypothesis of faults, and of the eruption of the trap before faulting.

The third argument by which decision is to be made between repeti-
tion by faults and by recurrent processes, depends on the number of times
the repetition occurs. If the two sections now described are the only
ones in the valley thus repeated, they might possibly be regarded as the
results of recurrent processes ; but if similar sections occur frequently
or habitually, with no more change in the corresponding members than
should be expected in different parts of beds of lava, sandstone, con-
glomerate, and shale, then there can be no question that the repetition
is due to faulting. It may be noted also that if faults are found in
some number and in systematic relation and accord, they may be ac-

MUSEUM OF COMPARATIVE ZOOLOGY. 67

cepted as proved on less complete evidence than would be reasonably
required to establish the existence of a single isolated fracture. In like
manner, if the hypothesis of the occurrence of faults enables one to
correlate a great number of otherwise apparently disconnected and
arbitrary facts, it is therefore to be regarded favorably. Finally, if it
lead to the detection of facts not before noticed, and thus gives the
power of prophecy, it can be considered as fully established.

One who goes on the ground as far as this nari*ative has now led him,
will, I think, find it already difficult not to lean too strongly towards
closing the question in favor of one of the suggested explanations instead
of maintaining an open mind in the question. As a working hypothesis,
let the faults be admitted, and a simple terminology adopted in accord-
ance with the explanation that they offer. Let the mass between two
faults be called a block ; or, if small, a chip. Let the thick sheet of
trap be called the main sheet, and the others, the anterior and posterior
respectively, in accordance with their position relative to the main sheet.
Certain deductive considerations also need attention before going farther
in the field ; in our school excursions, these were briefly discussed on
the ridge south of the reservoir, overlooking the Southington plain,
during the noon rest of the second day.

First, if faults are suspected, what must be learned about them before
they are completely known 1 The elements of a fault are : its outcrop
line, its direction, position, and le'ngth ; the hade of the fiiult plane ; the
throw of the fault ; and the width and other characteristics of the fissure.
Its depth, its date, and its cause are also subjects for more advanced
inquiry. With all these elements in mind as subjects for search, one is
more alert to discover them.

Second, if faults do occur, how can they be best recognized 1 Some
easily identified bed is the first requirement, such as the main sheet of
trap, which forms prominent ridges in the broad valley between slopes
of crystalline rocks east and west, and can be easily seen from a distance.
But, besides this, it is important to have also a number of thin hard
beds whose lines of outcrops are more sharply defined than those of the
broad main sheet, in order to determine the position of the fault with ac-
curacy. The anterior and posterior trap sheets serve this purpose nicely.

Third, if the faults occur, what will be their effect on the topography
of a well-denuded monoclinal mass, containing hard and soft beds 1 For
the sake of simplicity, the case of a single hard bed reduced nearly to
baselevel may be taken alone, and the dip of the monoclinal regarded
as constant ; the fault plane will for the time be considei'ed vertical.

68 BULLETIN OF TllE

Variations from these simple conditions can be easily introduced after-
wards. Several cases may be distinguished, depending on the relative
directions of the fault and the strike of the beds.

A. If the fault (/, Fig. 4) run parallel to the strike of the mono-
cline, and its outcrop lie behind a ridge, a, made by the hard bed, and
the heave, A, is on the side of the dip, d, then the hard bed will be
indefinitely repeated in a second ridge, b, parallel to the first. The
distance between the two ridge lines may be called the offset, t ; then
t =^h cotan. d. The fault may lie anywhere between the two ridges.
Such a fault economizes a formation in allowinpf a triven thickness to
cover a great width of country.

B. If the heave is on the side of the ascent of the monocline, some of
the beds will fail to appear at the surface. This is, in contrast to the
preceding, a wasteful arrangement.

C. If the fault, with heave of the same value and on the same side of
the fracture as in A, run oblique to the strike with an angle e between
the two lines. Fig. 5, the offset remains as before, but the two ridges
are not indefinitely repeated ; the north end of h overlaps the south end
of a, and the overlap p equals t cotan. e, or h cotan. d cotan. e.

D. If the fault run square with the ridge, Fig. 6, there will be an
ofiset as before, but no overlap.

E. If the heave be on the other side of the fault, e being less than 90°,
Fig. 7, there will still be an offset as before but in the opposite direction,
and instead of an overlap there will be a space in which there is no ridge.

All these are special cases, easily generalized. If the values of dip,
heave, and angle between strike of beds and fault-line are represented as
in C, the formula there given will apply to all cases ; a negative heave
meaning a change in the side of the uplift, and a negative offset mean-
ing a loss of visible outcrop, as in case B.

To the student, at least, there is here seen good reason for the ety-
mological connection of geometry and geology. In the geometrical
conceptions, the angles are to be seen all sharp and precise ; in the
geological occurrence they are rounded off and obscured.

Thus prepared by a deductive review of principles that have been
learned elsew^here and that may find application here, further explora-
tion can be most profitably undertaken.

A little north of the point where the anterior ridge of High Rock was
ci'ossed in the morning, it falls off and ends in low ground at the south-
ern point of the reservoir (4) ; but a little way to the west another and
very similar trap ridge (5) begins, with offset of a little more than a

MUSEUM OF COMPARATIVE ZOOLOGY. 69

quarter of a mile, and overlap of almost as much allowance, being made
for the missing acute southern point of the western ridge, which is here
rounded off. This suggests a fault of the pattern given in case C.
Knowing the dip of the anterior trap sheet to be about 10°, the strike
of the fault must be east-northeast ; and its heave, 300 feet or more on
the southeast. The surrounding topography is very suggestive of such
a dislocation ; a view of the oblique gap formed in the ridge as seen
when looking southwest from a hill (G) on the west side of the reservoir
is given in Fig. 8.

If the fault extend and maintain its displacement a mile or more to
the northeast, it must intersect and dislocate the main trap sheet in the
same manner as it has broken, the anterior. Looking along the calcu-
lated trend of the fault from the north end of the anterior, we see the
gap (7) by which we have come from New Britain, between the north
end of High Rock and the high trap ridge (8) next northwest of it, to
which no special name is attached. The outlet of the reservoir valley
runs through this gap. Assuming that the fault is straight, we have
now a better means of determining its direction by sighting the long
line from one gap to the qther ; the bearing thus found is N. 60°
E., which agrees satisfactorily with that determined by the offset and
overlap of the anterior ridge. In further confirmation of the fault, an
afternoon return trip may be made from the thrown portion of the an-
terior trap ridge, after following it northward half a mile (9), across the
strike of the beds toward the thrown portion of the main sheet (8) ;
and in so doing, the few shale outcrops that appear are in such close
accord with those seen in the morning in corresponding position in the
High Rock block that they may be regarded as equivalent beds. Stand-
ing on the high ground (10), west of the reservoir, and looking back
across the fault valley, the correspondence between the two monoclinal
blocks is very apparent, in topography as well as in structure. Rapid
advance may now be made on the foundation thus laid.

Excursion 3. — The Great Fault West of Lamentation

Mountain.

A third day may be spent looking for the fault that is supposed to
separate Lamentation Mountain from High Rock. It must lie some-
where in the country between them. The search may best begin at the
point where the cross-section was taken up on the first morning ; but
now exploration should be turned westward, descending in the mono-

70 BULLETIN OF THE

clinal series. A little preparatory consideration of the area! topography
of several blocks, Fig. 9, bounded by faults similar to the one in the
reservoir valley, will be of service. Each block, consisting of a mono-
clinal sequence of harder and softer beds, such as has already been
described as the result of the first day's walk, will possess its anterior,
main, and posterior trap ridges, with several intermediate and associated
ridges and valleys of sandstone and shale, in definite order. The ridges
formed on the successive beds will end without apparent cause on the
oblique fault lines that bou»id the blocks. If a block be narrow, a
quarter to a half a mile wide, for example, the posterior trap ridge will
have its south end to the north of the north end of the anterior ridge.
At first sight, these two members of the block-sequence would not seem
to belong together. Continuity in the monocline can therefoi-e be
found only by crossing the country within the limits of a single block,
and advancing parallel to its enclosing faults. A cross-section made at
right angles to the strike of the beds would soon traverse a fault line,
and would only confuse the observei*. This principle is all-important in
deciphering the topography and structure of the region. Eeference to
Fig. 2 will show that the section there made was traced out obliquely
for this reason, although no mention of it was made at the time.

The fact that there has been a long period (probably Jurassic and
Cretaceous) of deep-reaching erosion after the faulting took place,
facilitates our exploration by reducing the constructional form nearly
to a baselevel plain, still perceptible in the hard crystalline plateau east
and west of the Triassic area ; but the task would be still easier if this
great erosion had not been followed by a period of uplift and consequent
denudation (post-crctaceous) , in which the softer beds have wasted down
■well towards a lower baselevel surface where they are now generally
drift covered, leaving only the crest lines of the thickest and hardest
trap sheets to bear witness to the existence of the previous baselevel of
the region. When a ridge formed by the main sheet is cut by a fault,
the crest of the ridge gradually descends as it approaches the fault,
and the two lines intersect on low ground. The crest line curves gently
at the northern side of a block, and sharply at the southern, as in Fig. 9;
hence the strong bluffs at the south end of the ridges in the neighbor-
hood of Meriden. On first recognizing the existence of faults here, the
observer may be disposed to postulate a horizontal motion along the
fault line, in order to account for the offset of the corresponding ridges ;
but there is no necessity of this ; a vertical uplift followed by a base
levelling will serve as well, as has been indicated in the diagrams above.

MUSEUM OF COMPARATIVE ZOOLOGY. 71

Now returning to our excursion in search of the fault west of Lamen-
tation and going west from the crossing of the Cromwell railroad over,
the Consolidated Eoad, Fig. 2, we pass two conglomerate ridges, and then
find a strong bluff of trap (14), deeply quarried at its southern end.
It is the easternmost and lowest of the bluffs that constitute the Hanging
Hills. If it be a portion of the main sheet, then the fault may be in
the little hollow occupied by the quarry branch track (15) of the Con-
solidated Road, and the trend of the fault would in this case be in the
line from the south end of the quarry ridge to the north end of Lamen-
tation Mountain, or N. 55° E., and all the conglomerate ridges and the
anterior trap ridge of the first day's section must end as they run north
to this line. I have seldom found a more pleasing confirmation of a
theory in a predicted result than was afforded in tracing out this fault.
Every little ridge, trending about N. 30° E., runs with well maintained
continuity until it reaches the invisible fracture, and then, without ap-
parent reason, it promptly ends (16). The farther east the ridge, the
farther north it extends. The anterior trap ridge obediently follows the
same rule (8), and so does the little outcrop of shale (5), in the valley
between the anterior and the main sheets. Lamentation Mountain itself
falls away for no apparent reason ; its trap sheet seems to be as thick
here as anywhere, but it cannot cross the invisible line of dislocation.
Spruce brook, flowing north from locality (11) of the first day's walk,
runs on shales for a quarter of a mile after crossing under the road, and
then cuts down to the back of the trap for a little distance (17); the
shales soon reappear, but with abnormal dip to the northwest, and
finally end in a violently dislocated and crushed ledge (18). This is
undoubtedly close to the fault line. A short quarter of a mile farther
on, the course of the fault leads to a curious anticlinal, mentioned by
Percival, which like the last finds explanation by the drag of the fault.
Departures from the general eastward dip of the monocline are rare, and
it is interesting to see that they are associated with one another and
with other phenomena as the common results of a single controlling
cause. Some isolated knolls of trap near by may possibly be blocks
caught in the fault, but this is questionable.

The fault line from the quarry bluff northeastward is thus found to
maintain a tolerably direct course as far as it has been traced, and it
follows much the same course as the one discovered in Shuttle Meadow.
We may therefore expect it to be prolonged to the southwest also. In
walking in this direction from (15), Fig. 2, there is nothing decisive for
a mile or so, unless a valley followed by the Cromwell railroad oblique

72 BULLETIN OF THE

to the strike of the sandstones be so considered ; the valley broadens
farther southward beyond Meriden, and is bounded on the western side
by a strong ascent. On the face of this hill, not far north of the proba-
bly post-glacial trench that the Quinnipiac has cut through it from the
■west, there are numerous outcrops of reddish shaly sandstone with strike
N, 70° E., and dip 15° northwestward into the hill, and thus accounting
for its steep eastern slope ; but the westward dip itself is very unusual
in the Triassic area, and suggests some local disturbance. Standing on
the hill and facing about X. 50° E., one may look in the direction of
the quarry bluff by Meriden, and in the distance beyond it see the de-
scending northern end of Lamentation, showing that the hillside is in
line with the fault already traced. Just south of the Quinnipiac, there
are plentiful outcrops ; the beds are about level near the bridge, but
farther east steepen to a dip of 25° northwestward, beyond which the
outcrops suddenly end. This confirms the occurrence of the fault. As
its heave is on the east and of a value of several thousand feet, it is
natural enough to find the beds immediately west of it somewhat flexed
upward from the attitude that prevails generally in the monocline.
Their abnormal dip is satisfactorily explained by associating it with this
fault, as in the stream bed northeast of Lamentation, already men-
tioned.^ Still farther to the southwest, the hill is so broadly covered

1 The upward drag on this great fault, by which the dip of the beds on its west-
em side is here reversed from an eastward to a westward direction, is liomologous
with the drag that has flexed the posterior sheet (Percival's P 2, El) of Pond
Mountain, the southernmost member of the main trap sheet, east of New Haven.
This block is cut off by the heavy fault that limits the present area of the Triassic
rocks on tiie southeast, and brings up the crystalline base of the formation against
them ; the drag has in one small locality even overturned some of the beds, so that
the conglomerates that here belong beneath the posterior sheet seem to lie over it
at a steep eastward angle. That they are really overturned and normally belong
below the trap is shown, first, by the occurrence in the quarry near by to the north
of much larger outcrops of similar conglomerate, dipping westward under the trap ;
second, by the dense texture of the trap that apparently underlies the overturned
beds, this dense texture being elsewhere characteristic of the base of the trap sheet,
and in strong contrast with the loose vesicular texture of the upper surface of the
sheet ; third, by the absence of all trap fragments in the overturned conglomerate,
■while they are plentiful in the shaly conglomerate with westward dip that overlies
the trap in an exposure by a pond an eighth of a mile northeast ; fourth, by the
general abrupt eastern face and gentler western slope of the trap ridge, which
shows that its prevalent dip must be westward, and that the eastern dip of the
conglomerate is local and exceptional. Search has been made for cross-bedding in
the conglomerate, by which its upper and lower surface could be so nicely detected',
but it could not be found.

MUSEUM OF COMPARATIVE ZOOLOGY. 73

with drift that no further sign of the fault has been found there ; there
are, however, indications that it is extended a number of miles beyond,
even to the intrusive sheet of West Rock and Gaylord's Mountain. Be
this as it may, there can remain no more doubt as to the existence of a
fracture between the Lamentation and the Hanging Hill blocks. This
fracture we shall call the " Great Fault."

Excursion 4- — Faiilts south of Lamentation Mountain.

It was noted above that the occurrence of a number of faults systema-
tically arranged might be taken as further evidence in favor of the fault
theory as against the theory of repeated deposits. A fourth day may there-
fore be spent in searching for them. The first guide in the search will
be, as has already been suggested, the gaps in the line of the main trap
ridges. One of these gaps appears at the southern end of Lamentation,
dividing it from Chauncy Peak. If a fault run between them, it must
dislocate the anterior ridge also. Walk, therefore, to a point (1) Fig. 9»
on the anterior ridge of Lamentation, a little south of the ash-bed that
was visited on the first excursion. To the north, we know the ridge is
continuous as far as the Great Fault. To the south, a short walk brings
us to a little notch by which a small stream escapes from the anterior
valley, and beyond which the ridge is continued in the same line as be-
fore ; the notch must therefore be regarded as a simple transverse water-
cut. But on following the ridge a little farther, it ends (2) on the
northern side of an open meadow, and crossing here to the south noth-
ing but conglomerate (3) is found, and with such strike as would if pro-
longed northward carry it directly to the trap. The fault between Lam-
entation and Chauncy may therefore probably pass by the southern
end of the anterior thus determined. A bearing taken from the south-
ernmost point of the anterior trap to the gap (8) at the south end of
Lamentation reads N. 60° E., and this is within a few degrees the same
as the bearing of the faults already described. As a further test of
the occurrence of a fault here, the northern end of the anterior to the
Chauncy Peak block should be found somewhat to the north of the
southern point of the anterior to Lamentation. It is found (4), pre-
cisely as predicted, lying at appropriate distance in front of Chauncy
Peak, containing an ash-bed with bombs of lava (5) like those already
described in the anterior of the Lamentation block ; and it ends directly
under the southern blufi" of the main sheet (6) in the northern block.
A few shale outcrops are found on its back, and others appear on

•74 BULLETIN OF THE

the bold western slope of Chauncy Peak. The fault has an offset of
about 1500 feet and an overlap of similar value; its heave must be
about 300 feet. Although of smaller displacement than the Great Fault,
it manifestly belongs to the same system and contributes effectively to
the rational explanation of the Triassic structure and topography. It
probably determines the location of a ravine (7) southwest of the Berlin
road towards Meriden, but I have not been able to follow it beyond the
city, as outcrops are few and monotonous in that direction. On the
other side of Lamentation Mountain, if there were time to go there, it
probably causes a slight dislocation (9) in the posterior trap ridge ; the
displacement is so small that in speaking of this part of the posterior
ridge when describing the last stretch of the first day's walk, the ridge
was referred to as traceable with little interruption. If this identification
is correct, it is probable that the throw of the fault decreases to the
northeast, and that its line is somewhat curved, as indicated in the figure.
The wide valley between Chauncy Peak and the north end of Higby,
or Middletown Mountain, through which the return to Meriden was
made on the first day, suggests a fault with throw of value interme-
diate between that of the Great Fault and the one just described. Its
examination may be conveniently begun by following the anterior of
Chauncy Peak southward to its vague ending (10) near the Cromwell
railroad. This termination must be near the southeast side of Chauncy
Peak block. A wide swamp, in which the engineers of the Cromwell
railroad found much difficulty in making a steady roadbed; conceals all
outcrops for some distance, but by followiug the track for half a mile
east from the end of the anterior, and then, at a point opposite the
southern end of Chauncy Peak (11), crossing a little field to the south,
several ridges of conglomerate (12) will be found in the woods; they
strike N. 35° E. and their beds dip about 10°, and after crossing four
or five such outcrops, a low trap ridge (13) is found; it is soon identi-
fied as yet another appearance of the anterior sheet, for it stands in
proper position with regard to the main sheet, which rises in the high
mountain to the east ; it is very scoriaceous on the back, where it is
followed by a road ; and at a few places it shows the ash and bomb
structure that has already in two other blocks been fotind to characterize
the anterior trap. It must be followed north to its end (14), which is
indistinctly located shortly before reaching the Meriden- Westfield road.
Sighting back from here, we find a bearing of about S. 65° W. will carry
the fault line back to the southern point of the Chauncy Peak anterior ;
the accordance with previous (neasures is satisfactory enough. But, in

MUSEUM OF COMPARATIVE ZOOLOGY. 75

determining the course of a fault, care must be taken to select as guides
at least three points, which lie alternately on opposite sides of the frac-
ture. It is evident that an error may result from trusting too implicitly
to the apparent termination of a ridge, for the real termination may
be covered ; but if three ridges, of which the first and third are on one
side of the fault and the second is on the other, all terminate on the
same straight line, the presumption is very strong that they indicate
a straiglit fault and that the indication may be trusted. The case
in hand therefore needs additional ridge-endings before the fault line
can be established. The south end of Chauncy Peak (11) and of
its posterior (15) and the high north end of Higby Mountain (16)
serve abundantly for this purpose. The southern end (15) of the ridge
posterior to Lamentation and Chauncy is found a little south of the rail-
road cut to the east of Highland station, mentioned on the first day's
•walk as affording a good exposure of the sandstone overlying the lava.
The ridge ends in a little knoll back of a farm house and barn, north of
the Meriden-Westfield road. In sighting backward from this knoll, the
course of the fault is seen to be curved, and if the middle of the meadow
between this point and the north end of Higby be taken as the location
of the fault, its curvature is greater still \ but this is hardly more than
might be expected : a straight line fault is too rigid to be natural. On
continuing the walk to the northeast, a reverse curvature of the fault line
is required, in order to leave the long descending ridge (17) of Higby
Mountain on its eastern side. The northernmost low end of Higby is
found at High Falls (18), where the trap suddenly ends. The little
gorge opened by Falls Brook discloses much breccia in fractures running
northeasterly, and evidently associated with the strong fault close by.
No other outcrops appear for some distance, but a cut on the Cromwell
railroad, just west of Westfield station and about a mile and a half north-
east of High Falls, reveals strong disturbance in the dip of the shales
there exposed, as well as two faults of indeterminate throw. It is likely
that these dislocations are associated with the fault that we have been
tracing. It is interesting to notice that the course of the fault thus
traced curves somewhat in the neighborhood of the north end of Higby
Mountain, and that the curvature is closely conformable to that found
between the Lamentation and Chauncy Peak blocks.

On returning to Meriden, a superb view of the valley may be gained
by an easy walk up a path leading to the terminal bluff' of Chauncy
Peak from the road below it. The strong range known as Beseck Moun-
tain, formed on the main sheet, may be traced many miles southward

76 BULLETIN OF THE

from Higby with little interruption. The broad back of the same sheet
faulted down in the district of the Hanging Hills, all heavily wooded,
rises on the farther side of the wide valley to the west. Descent from
the cliff may be made hy t. little crevice, in which a narrow band of brec-
cia is seen ; its trend is northeasterly. A southward view of Lamenta-
tion with Chauncy at its farther end, and of the north end of Higby,
from a point several miles to the north, is given in Fig. 10. Some dis-
tance up Falls Brook from High Falls, the sandstone lying on the back
of Higby (19), Fig. 9, contains numerous fragments of vesicular trap;
stones showing such inclusions are common in the stream bed.

Attention should here be called to the apparent double ridge at the
north end of Higby Mountain ; as if the main sheet were slightly dis-
located by a strike fault, or as if it consisted of two lava beds, separated
by a weaker stratum of some kind. The main sheet, which forms the
bold west bluff of the mountain, ends south of Highland station (16);
but a second ridge, a little lower than the first, comes into sight from
behind the mountain, and extends a mile or more farther to the north-
east (17, 18). This great extension of the second ridge, although so
close to the first, suggests a change in the course of the fault, as already
indicated. Another example of the double form of the ridge will be
found on the excursion for the morrow.

Excursion 5.— Faults in the Hanging Hills.

The Hanging Hills, one of the most picturesque districts in the state,
may be visited on the fifth day, beginning at the Quarry Ridge, along
the margin of which ran the Great Fault discovered on the third day.
Approaching the ridge by the lane east of the Fair Grounds, Fig. 11, it
is significant that the ledge of conglomerate there followed strikes di-
rectly toward the trap bluff and ends on reaching the fault valley.
When the quarry (1) is reached, attention will soon be taken by the
variety in the structure of the trap, here so well exposed by quarrying.
The greatest part is indeed rather uniformly dense and of medium tex-
ture, but in the upper platform along the west side of the opening and
on the lower bench at the southwest end, much vesicular trap is found,
and its relation to the dense trap brings up an interesting problem.
Close examination will discover that the dense and vesicular masses are
separated by a rather sharply marked surface of gently undulating form,
inclined to the eastward at an angle of about twenty degrees, and hence
about parallel to the prevailing dip of the monocline. Below this sur-

MUSEUM OF COMPARATIVE ZOOLOGY. 77

face of Beparation, the vesicular trap may be seen to extend downward
for a thickness of ten feet or more, becoming gradually denser below.
The upper mass, on the contrary, maintains its dense texture to the top
of the quarry, a height of sixty feet above the vesicular mass below it.
The only way in which these facts can be explained seems to be to re-
gard the two masses as separate lava flows ; the lower one showing its
upper vesicular surface, which, at the few points where it happens to be
stripped bare, resembles the form of ropy lava, such as is called " pa-
hoe-hoe " in the Sandwich Islands ; the upper mass revealing only the
dense under part of a later flow, from which the original vesicular upper
surface has here been worn away. This is entirely in accord with what
has been learned elsewhere in the Triassic formation ; for if the lavas of
this portion of the district are extrusive, it is the most natural thing in
the world that the sheets should be composed of successive lava floods.
The time between the outpouring of the two sheets here must have been
short, for they are not separated by any deposit of sandstone or shale,
nor does the upper surface of the lower flow manifest signs of wearing
away, as it might if it had been long exposed to the weather above water
level. It may be here mentioned that a sheet of trap exposed in the
railroad cut between Springfield and Westfield in Massachusetts bears
evidence of at least three successive flows, the thinnest being only a few
feet thick.

Another feature of the trap quarry is found in the bands of fragmental
material that traverse it from one end to the other, trending N. 57° to
68° E. One of the clearest of these runs along the margin of the upper
platform and extends across the high face of the quarry. It is from
two to four feet thick, stands nearly vertical, with a slight hade to the
west, and consists of angular fragments of trap of all sizes contained in
a matrix of what looks like sandstone, although it bears no distinct
marks of stratification. The trap walls of the band are of the same me-
dium texture as the rest of the upper flow, and are sharply defined ; the
trap fragments are of similar texture, without change from their margin
into the centre. They vary in size from minute grains to great blocks,
three or four feet across. Slickensided surfaces abound, generally
parallel to the walls and sometimes extending into blocks of trap, which
are slightly dislocated thereon. In short, these bands are fault breccias,
the trap fragments coming from the adjoining rock, while the sandstone
has been washed down the fracture from the beds that once overlay the
trap, but which at this point are now worn off". As their bearing is the
same as that of the faults already found from topographic evidence, it is

78 BULLETIN OF THE

manifest that they furnish us with dissections of small fractures similar
to the greater cues that control the structure of the region. The quarry
could not have been placed more advantageously for geological results.

The faults of the region have been found to have their heave on the
east ; if the dislocations revealed by the breccias belong to the same
family as those that dominate the topography, we should expect them
to present the same relative movements. The only opportunity to test
this is found in the western part of the quarry, where the surface of
contact between the upper and lower sheets affords a recognizable layer
for identification on the two sides of the fracture. The displacement
thus determined is of small measure, about eight or ten feet, but it
is of the same order as the larger ones already determined, having its
iiplift on the eastern side.

Four fault breccias may be found in the quarry ; their average bear-
ing is N. 63° E., their hade averages 71° with much constancy. Near
the eastern foot of the quarry a broad breccia is seen much weathered ;
it has probably been but little stripped of t*he cover that it had before
the quarry was opened. The little hollow, along which the branch track
is laid from the Consolidated Road, undoubtedly marks the site of the
Great Fault, and if opened would be of much geological interest. Walk-
ing across the hollow to the sandstones with barytes veins on the
eastern side, we have stepped down almost two thousand feet, for the
Great Fault which sets Lamentation ^Mountain over a mile back from this
portion of the main sheet can hardly have a less throw than that
amount. A transverse section of the quarry would, if fully worked
out, probably appear as in Fig. 12, and the occurrence of step faults as
there indicated goes far to explain the reason for the easy opening of
other fault lines into gaps such as characterize the region.

It may be noted that fault breccias have been found in several other
localities, and that they accord in strike and dip with the system here
described. Percival called them " clay dikes," and examples will be
mentioned below.

An instructive view is opened by climbing to the top of the quarry
bluff by its western slope. A northeast valley (2), Fig. 11, separates
the quarry ridge from other similar but higher ridges (3), and indicates
a fault. Isolated ledges in the valley suggest chips of trap broken from
the adjoining blocks. The quarry ridge therefore belongs to a very
narrow block, and its anterior trap can only be found by keeping care-
fully within the limits of its enclosing faults. Walking southwestward,
a small trap outcrop (4) is found in the roadway west of the Fair

MUSEUM OF COMPARATIVE ZOOLOGY. 79

Grounds, and although not visibly traversing even the small measure
of the narrow block, it may be fairly identified as the anterior sheet, by
reason of its position. Its failure to make a continuous ridge may be
attributed to weakness resulting from the numerous small faults that
"were seen in the quarry. Farther along in the block, ledges of con-
glomerate and sandstone (5, 5') are found in appropriate position ■ but
their strike is seen to turn somewhat west of north, departing thirty
or forty degrees from the strike prevalent in the Lamentation fault
block, and thus helping to account for the abnormal trend of the
southern face of the Hanging Hills. The fault bounding the quarry
block on the northwest may be found by walking from the conglomerate
ledges (5) towards a wooded ridge (6), which is soon discovered to con-
sist of trap, and which must be regarded as the anterior sheet of another
block by reason of its attitude betwen the strong bluffs of Cat Hole
Peaks on the north and the conglomerate ledges (8) on the south. It
is overlain with fine red shales (7), whose strike is N. 45° W. and dip
20° N.E. ; no contact with the trap has been found here, though it
might be discovered by a little digging. The several breaks in the front
of the anterior ridge probably indicate small faults, and may in part be
associated with corresponding notches in the main sheet. This may he
called Cat Hole block, taking the name from -the deep pass in the main
sheet on its western side. The view of the block from the round hill of
conglomerate and sandstone (8) is especially valuable ; no point in the
district illustrates more clearly the necessity of working but the struc-
ture of every block by walking parallel to its length, instead of as usual
at right angles to the strike of the beds. The several members of the
quarry block can be located : the main sheet in the quarry ridge (1), the
anterior (4) alongside the Fair Grounds ; the conglomerate and sand-
stone ledges (5, 5') below. Cat Hole block is equally distinct ; the
main sheet in the castellated knobs at the end of the long ridges
(3 — 3) ; the wooded ridge of the anterior trap with the red shales
behind it ; and the conglomerate ledges below it, where we stand.
Farther west, Xotch Mountain block can be as well interpreted ; the
main sheet in its superb cliffs surmounting a long talus ; the anterior
(9), wooded again, west of us; and the lower sandstones in the rolling
ground farther south.

Oblique valleys, undoubtedly located on fault lines, enclose Cat Hole
block on either side. Taking the anterior sheet as a guide to the dis-
location on these faults, and regarding its position in Cat Hole block
(6 — 6) as normal, we find it thrown to the northeast in the Quarry

80 BULLETIN OF THE

Ridge block, and to the southwest in Notch Mountain block ; and from
this it is apparent that the faults are of the usual pattern, with heave
on the southeastern side. The movement on each one may be one
hundred or more feet.

A pleasant spot for lunch is found at the Cold Spring, in the ravine
near the Poorhouse. This spring is fed in part by melting snow or
ice hidden under the great trap blocks that have fallen from the cliff of
Notch Mountain, and long ago attracted notice.^ A water-cure sanita-
rium was built near it on the bench of the anterior sheet ; but proving
unsuccessful, it' was bought by the city of Meriden for a Poorhouse;
it is not likely that any other similar institution possesses so delightful
a view as is here spread out to the eastward, even as far as Higby
Mountain. Before following the wood road to the reservoir, the open-
ings in the western end of Cat Hole anterior should be examined, as
they show a peculiar structure, perhaps indicative of ropy lava flow.
Sandstone is seen lying close over the anterior just south of the Poor
House in Notch Mountain block, and a small piece of vesicular trap was
found enclosed therein.

Broad views may be had by climbing the face of Notch Mountain in-
stead of following the road through the wood below. A noteworthy
feature is the furrowing of the top of the mountain by several ravines,
parallel to the general direction of the neighboring faults. These and
the crevice mentioned in the terminal bluff of Chauncy Peak are the
topographic expression of small fractures similar to those dissected in
the quarry bluff. The view from the southwestern bluff of Notch
Mountain opens the West Peak block to easy inspection. The fine
cliff of its main sheet rises from the farther side of the reservoir, and its
anterior sheet forms a broad bench (11) to the south. On descending
to the reservoir an exposure of red shales (10) is seen in the roadside,
a little above the vesicular back of the anterior trap ; its beds strike N.
30° W, and dip 6° N.E.

The reservoir valley has a little more northerly course — about N.
15° E. — than that followed by the faults hitherto met. For some
time it appeared to be an exception to the general rule ; but closer ex-
amination has led me to conclude that the valley is not coincident with
the fault. This is certainly a hazardous conclusion, and not to be
lightly accepted. The following facts lead to it. Toward the northern
end of the reservoir, a hill of heavy trap (14) stands below the western
face of Notch Mountain, and more in accord with the attitude of the
^ Silliman, Amer. Journ. Science, 1st series, iv. 1822, 174.

MUSEUM OF COMPARATIVE ZOOLOGY. 81

West Peak sheet. The trend of the western face of Notch Mountain is
N. 50° E., and it may be traced in a more or less distinct bluflF or ledge
(12) for a mile in this direction. On following up the intercepting
canal cut in the trap to bring the streams from the back of West Peak
into the north end of the reservoir, a band of breccia about a foot wide
may be found traversing it, trending N. 60° E. ; and the back of West
Peak is furrowed with ravines (13) trending N. 55° or 60° E., showing
that faults of the normal direction occur here as well as to the east of
the reservoir. The general topography leading to the above conclusion
may be perceived from a knoll (15) a good half-mile northward. The
long northwestern face of Notch Mountain is from here clearly seen to
be independent of the detached portion (14) of the West Peak block.
A low trap ridge (16), a little to the west of the knoll, is probably to
be identified as the posterior sheet of the same block.

The abnormal position of the reservoir valley finds no sufficient ex-
planation. It may be located on a branch of the chief fault ; but in
such case the chief fault ought to be the site of the chief valley, and
not merely of a little ravine. Perhaps a more likely explanation will
some day be found' by regarding the reservoir valley as the abandoned
course of an old river, whose direction was taken during the pre-
cretaceous base-levelling of the region, and maintained for a time after
the post-cretaceous elevation, until some other stream, which en-
countered no heavy trap sheet and therefore deepened its channel
quickly, captured and led away the head waters of the reservoir river ;
the reservoir notch would thus fall into the class of wind gaps derived
from water gaps, not uncommon in the Appalachians. But diflFerent ob-
servers may well have different opinions here.

While on the knoll (15), the double form of Notch Mountain will be
observed. A second trap sheet (17) seems to lie on the back of the
first. (It is rather too distinctly drawn in Fig. 11.) The same thing
might have been noticed from the back of the anterior sheet of Cat Hole
block, where the roads form a little triangle. A rough walk through the
woods around the base of the upper sheet (17) to Cat Hole shows the
back of the lower sheet to be highly vesicular ; a rocky talus hides
the contact between the two. I have interpreted this as the topo-
graphic expression of the two lava sheets disclosed in the quarry blufi" :
the vesicular upper part of the lower sheet acts as a soft bed between
the denser parts of the two flows, and the mountain crest is therefore
doubled. The same double form may be seen in the mountains of
Medina sandstone in central Pennsylvania, and for a similar reason.

VOL. XVI.— NO. 4. 6

82 BULLETIN OF THE

The other trap mountaius about Meriden do not show the double form
so distinctly ; the second ridge on Higby has already been mentioned ;
on the back of West Peak the upper sheet may perhaps be identified in
certain ridges near the base of the wooded slope ; and if the observer is
ambitious of hard scrambling over rough trap ledges and waste branches
of felled trees, he may attempt to work out the faults and tlie double
sheet of Cat Hole ridges (3) to northeast of the Peaks ; but this is not
to be recommended as an easy return from the short walk of this busy
day.

The eastward turn of the main sheet bluff from West Peak to the
Quarry bluff deserves a word. It seems to depend on three causes.
The strike of the beds changes from the general trend of N. 20° E. to N.
20° or 30° W. or more, and this alone accounts for much of the turn ;
the displacement on the faults accomplishes something in the same direc-
t.ion ; and finally the accelerated recession of the cliff faces where the
faults are numerous accomplishes the rest. The moderate altitude
maintained by the main sheet in the Cat Hole and Quarry ridge blocks,
where the fractures are numerous, bears witness to the effectiveness of
the last cause.

Excursion 6. — North of West Peak.

A final excursion may be made along the range north of West Peak.
I'or reasons that will appear later, the walk may be best begun at Cook's
Gap, where the New York and New England Railroad crosses the trap
range about three miles west of New Britain ; thence southward we
shall pass the Shuttle Meadow fault of the second day's excursion, and
afterwai-ds approach the northern side of the West Peak block. Taking
early train from Meriden to New Britain, connection may be made with
a local train on the New England Road, which on proper presentation of
the case may probably be induced to stop at the western side of Cook's
Gap, at a road crossing (1), Fig. 13.

Cook's Gap is unlike most of the others of the region in crossing the
trap range almost at right angles, and thus indicating its independence
of the fault system. It may be an abandoned river course, like the
Reservc'r Notch. Farmington River comes out from the crystalline up-
lands on the west about opposite to this gap.

The road at which we left the train follows the anterior ridge south-
ward : sometimes it approaches the bluff, from which an extended view

MUSEUM OF COMPAKATIVE ZOOLOGY. 83

of the Southington plain is obtained ; it is broadly drift-covered. No
interruption in the ridge is discovered for a mile and a half ; then near
the outlet of. Southington reservoir (2) two small notches are found, in-
dicating dislocations vpith heave of seventy and eighty feet; these are
probably connected with the indentations in the main ridge to the
northeast, which bear about N. 50° E. from the notches. Another mile
•without interruption brings us to the Shuttle Meadow fault, Fig. 3 be-
ing repeated in Fig. 13; the bluff of the northern member of the an-
terior should be followed around its edge into the fault valley in order
to appreciate the regularity of its curve. Three small dislocations ap-
pear in the North High Rock block, next beyond : and near the third one,
fragments of vesicular trap are found in the shaly beds in the road (3).
Advancing a little farther, an oblique valley (4) between North and
South High Rocks is disclosed ; if it is located on a fault, and if the
fault belong to the prevailing system, a dislocation in the anterior ridge
should be found when we have gone far enough southward to give the
oblique valley a bearing of about N. 60° E., and the dislocation should
be of the Shuttle Meadow pattern. At the expected point, the anterior
bluff curves around and ends in a ravine (5), across which another bluff
of the same form begins at a little higher level, indicating an uplift of
say a hundred feet. The oblique valley between the High Rocks cannot
be distinguished until the ravine is followed up towards the road ; then
its bearing is found to be closely parallel to that of the faults near by.
The impure limestone that is fQund at a number of points in the region
on the back of the anterior is exposed in an old quarry close by on the
roadside.

If the vague conception of the Triassic structure with which Shuttle
Meadow was entered on the second day's walk be now recalled and com-
pared with the definite conception that has slowly grown up as the to-
pography has been deciphered and the structure interpreted from it, the
student will find that the alternative hypothesis of repeated sequences
of deposition has no longer any claim on his attention. The hypothesis
of repetition by faulting has found continued confirmation since it was
first tested at Shuttle Meadow. Every method devised for testing the
occurrence of faults has been applied, and no doubt whatever can re-
main of their occurrence. The members of the repeated sequence are
sufficient in number to make a good case, and succeed one another too
arbitrarily to be regarded as products of a single process ; repetition is
frequent, and when once perceived it becomes a prominent characteristic
of the region ; the faults by which the repetition is produced are strik-

84 BULLETIN OF THE

ingly systematic in direction, and all agree in having their uplift on the
southeast ; the complicated topography of the region is reduced to sim-
plicity ; and a limited power of prophecy is gained, as in the case of the
notch in the anterior ridge, just described. The key to the structure of
the region is discovered. When the walk southward along the back of
the anterior ridge is resumed, and the terminal bluff of South High
Rock is seen in its characteristic form, with Short Mountain rising be-
yond it, he must be indeed a sceptic who is not ready to predict that
yet another notch in the anterior with normal offset and overlap will be
found corresponding to this break in the main sheet. The notch (6) is
soon reached, but on looking eastward as usual for the heaved continua-
tion of the anterior, it is not to be found. The ground is low and open
to the foot of the main sheet of Short Mountain, and it is only to the
westward that there is any ridge (7) that may correspond to the ante-
rior. Just as our generalization was to be established we meet a de-
parture from it. The case is certainly of great educational value as well
as of geologic interest; and one must approach it with an excellent
geometric understanding of the several patterns of faults described on
the second excursion, if he would not be puzzled by its departure from
the topography of the faults thus far encountered.

Turning westward instead of eastward, a ridge (7) is found that cor-
responds in every way with the anterior, and a brief consideration of its
position will show that it is an example of the case of Fig. 7, in which
the heave is on the northwest of the fault and the downthrow on the
southeast, the reverse of our usual style of dislocation. The offset is to
the west, or negative, instead of to the east ; and there is a lapse of
bluff front instead of an overlap. After perceiving this there is no
further difficulty. The main sheet of Short Mountain is seen to stand
farther west than the same sheet in the next block to the north, thus
confirming the conclusion derived from the anterior ; and when looked
at from the east, the backs of the two portions of the main sheet readily
disclose their relative altitudes : the southern is depressed compared to
the northern. The bearing of this reversed fault, determined by sight-
ing from the notch in the anterior through the pass in the main sheet,
is N. 65° E. It may be noted that if this course be turned a little to
the left, as if the fault curved to the north, it would lead in about three
miles to a peculiar fault breccia in the sandstones ; one of Percival's
" clay dikes," disclosed in a post-glacial stream channel, a mile south of
New Britain, at the eastern base of a great drumlin. The peculiarity
of this fault is that the deformation of the bedding on either side, shown

MUSEUM OF COMPARATIVE ZOOLOGY. 85

in Fig. 14, indicates an uplift on the northwest ; it is probably there-
fore an extension or a branch of the fault just described north of Short
Mountain.

The southern boundary of the Short Mountain block presents nothing
unusual, for both the anterior (8) and the main sheets (9) of the next
block — West Peak — are offset normally to the eastward about a third
of a mile. The course of the fault determined by sighting from the
south end of the anterior of Short Mountain block to the north end of
the West Peak main sheet is about N. 60° E. An old " paint mine "
(10) lies on this line ; the heap of refuse about it consists of a b..eccia of
vesicular and dense trap cemented by barytes and other minerals. The
sartie line, carried several miles northeastward, runs to a normal dislo-
cation in a trap ridge (probably a second posterior) a little distance
southwest of Berlin Junction station ; shortly before reaching this dis-
location, a "clay dike" (Geol. Conn., 378) is seen in the banka of the
Mattabesick ; its position places it on the fault line ; its direction, about
N. 40° E., accords fairly with that of the fault ; its structure shows it
to be a breccia ; and the deformation in the bedding on either side,
Fig. 15, shows that its heave and throw agree with the rule of the re-
gion. Midway on the same line, where the road from Cat Hole to New
Britain crosses a stream by an old burnt mill between two ponds, the
posterior trap is exposed in the stream channel, and close west of the
road there is a four-foot breccia of trap and sandstone, bearing N. 50°
E., with a hade of 15° northwest of the vertical, and slight uplift on the
east as indicated by apparent repetition of the scoriaceous upper por-
tion of the trap. This is probably a small fault, associated with the one
that bounds Short Mountain on the southeast.

There are no other significant faults till the. Reservoir Notch ig
reached ; and the day's walk may be ended either by following the road
(11), Fig. 11, that runs around the curve of West Peak, or by a shorter
cut (12) leading through Cat Hole to Meriden.

Review.

All the chief faults from Cook's Gap to Higby Mountain — ten in
number — have now been worked out. They accord fairly well" in di-
rection, as appears in the general map, Fig. 16, corresponding to the
black square of Fig. 1. Here the several sketch maps, Figs. 2, 3, 10,
11, and 13, are outlined in their proper relative positions, and indicated

86 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

by numbers 1 to 5 ; the main sheet is shaded, and the anterior and pos-
terior are located on either side of it ; the faults are drawn as broken
lines. Finally, a section drawn at right angles to the prevalent trend
of the faults is given in Fig. 17; it is not constructed closely to scale,
but indicates the general structure of the region. The abnormal Short
Mountain fault, with downthrow on the southeast, which when first
found seemed to endanger our generalization, is seen to be only a single
exception to a well marked rule. No other explanation of the structure
of the region than that by faulting seems admissible.

It has been difficult and in most cases as yet impossible to trace the
faults far to either side of the three trap ridges ; elsewhere, the surface
is so heavily drift-covered, and the sandstone or shale ridges are so
monotonous, that dislocations cannot be demonstrated. But it is in the
highest degree probable that the faults are not limited to the belt
of trap ridges ; the uniformity of direction and of throw over the con-
siderable district where we have traced them indicate their extension
over a much larger area, where more patient search may yet detect
them.

The Triassic monocline must therefore be regarded as composed, in
the Meriden-New Britain district, of a number of long and relatively
narrow blocks, whose direction is oblique to the strike of their beds, but
is in a most striking way accordant with the trend of the fundamental
schists, where they are exposed to the southwest and northeast, I find
it impossible to resist the conviction that this accordance is not due to
chance, but that it points to physical dependence of the superficial on
the deeper structure, as has been suggested in my earlier papers.

Cambridge, Mass., March 27, 1889.

MUSEUM OF COMPARATIVE ZOOLOGY. 87

EXPLANATION OF PLATES.

PLATE L

Fig. 1. Outline map of southern New England, showing the Triassic area of the
lower Connecticut Valley (dotted), and the area described in this paper (black
square).

Fig. 2. Sketch map for first and third day's excursions around Lamentation
Mountain.

PLATE IL

Fig. 3. Sketch map for second day's excursion around Shuttle Meadow
Reservoir.

Figs. 4 to 7. Diagrams illustrating the topographic displacement produced by
faults running at divers angles with the strike of the faulted beds.

Fig. 8. View of fault gap in the anterior trap ridge, southwest of Shuttle
Meadow Reservoir.

Fig. 9. Diagram illustrating the general scheme of interpretation of the faulted
Triassic monocline about Meriden.

PLATE III

Fig. 10 a. Sketch map for fourth day's excursion from Lamentation to Higby
Mountain.

Fig. 10 b. Distant view of Lamentation and Higby Mountains from the north.

PLATE IV.
Fig. 11. Sketch map for the fifth day's excursion in the Hanging Hills.

PLATE V.

Fig. 12. Generalized section of the Quarry ridge, near Meriden.

Fig. 13. Sketch map for the sixth day's excursion, from Cook's Gap to Short
Mountain.

Fig. 14. Section of fault with heave on the northwest.

Fig. 15. Section of fault with heave, as usual, on the southeast.

Fig. 16. General map of trap ridges in the Meriden-New Britain district ; the
several sketch maps already referred to being located by rectangles numbered from
1 to 5.

Fig. 17. Generalized cross-section of the Meriden-New Britain district.

DAVIS- MERIDEN

PLATE 1.

m

-123.

'^fe.

DAVIS- MER I DEN

ric 6

I^IOHTH EAID OF EASTERN K'DOE

A TAUl-T I
GAP

SOUTH END Of WESTERN RIDGE

r/G.e.

DAVIS- MERIDEIN

PLATC 3.

HiGBY AND LA/~7ErjrAT;ON Mou/s/ta/ns.

r/& /0,b

DAVIS -MERIDEN

PLATE ^

DAVIS -MERiDEN

PLATE 5.

NW

SCCT/OA; orrni MdRIOCN dunnny f-i^ I

^^

r/o i4

FIG It

fiC. 16.

Ge-NCfi^L. CKOSJ Sccr/oiM.

FIG. I

riG. 13.

No. 5. — On the Occurrence of Fossils of the Cretaceous Age on the
Island of Martha's Vmci/ard, Mass. By N. S. Shaler.

It has long been known that fossils apparently pf Tertiary age occur
in the peculiar rocks about Gay Head, on the western extremity of Mar-
tha's Vineyard. These strata of alternating clays, sands, and occasional
lignites occupy the western half of Martha's Vineyard, in which region
they rise above the sea level. They probably underlie the glacial de-
posits throughout much of the area of the island, and may have a yet
wider extension. Even where the Tertiary beds lie above the sea level,
they are generally covered by a thick coating of glacial debris. Where
this debris has the character of true shoved moraine, the accumulations
are often a hundred feet or more in depth. Where the detritus exists
in the form of a sheet, it is less continuous, but nevertheless covers the
greater part of the underlying rocks, which are only exhibited in a clear
manner along the gulf-like shores.

In 1870, while engaged in some studies on the erosion of the coast
lines of this island, I found several fragments of a coarse sandstone at va-
rious points in the drift material, which contained extremely obscure mol-
luscan fossils. Among these were specimens of what appeared to me to
be Exogyra, a genus which, as is well known, does not extend to the Ter-
tiary period, and is practically limited to the lower portion of the Creta-
ceous. Although the evidence was extremely imperfect, it was enough
to warrant a careful search of the island, with the hope of finding in
place the beds whence the fragments were derived. I spent more than
a month in this systematic inquiry before attaining to any results what-
ever. At length I discovered two localities where these sandstone frag-
ments with imperfect molluscan remains were tolerably plenty. These
positions are indicated in the descriptive sketch, Plate T. One of them
lies on the western shore of what is called Lagoon Pond, immediately
west of Cottage City. At this point a skilled collector may in the
course of half a day discover half a dozen fragments scattered in the
drift, which are clearly referable to Cretaceous rocks. A second and
more important locality lies near the centre of the northern shore, at
the distance of about three fourths of a mile from the coast line on the

VOL. XVI. — NO. 5.

90 BULLETIN OF THE

track of the last locality known as the "Woods Schoolhouse." The
schoolhouse of the name has disappeared, for its foundations only remain ;
but the explorer can readily find his way to the spot by passing from the
new schoolhouse on the Cedar Tree Neck road westwardly along the ser-
pent kame, the only deposit of this nature on the island, until he passes
a stone wall, a little to the west of which, in the roadway and on the
bare ground thereabout, he may find an abundance of fragments of this
peculiar sandstone. Circumstances prevented my undertaking any care-
ful study of this place until seventeen years after its discovery. In 1887
I returned to the locality, and with the help of my assistant, Mr. Foerste,
undertook a careful collection of the abundant fragments which I found
at this point, as well as a systematic study of all the area of the island
which gave promise of affording similar material. The search for other
localities was fruitless, and as this is the only one on the island which
has afforded fossils in condition for identification, I shall hereafter limit
my account of the bed to what is exhibited at this point.

As is shown in the accompanying section, the Cretaceous fragments
found at this locality occur only within a small area. They have been
found over a surface having an east and west extension of about 300
feet, and a north and south length of about 200 feet. The position is
immediately to the south of a shoved moraine, which extends up to and
probably includes this part of the drift accumulations. In this area,
from the surface to the depth of four or five feet, or as far as the exca-
vation penetrated, by far the larger part of the fragments are composed
of the deposit in question. The rock consists of a very coarse sandstone
abounding in quartz pebbles, containing indeed little other material save
quartz fragments from an inch in size downward. The largest of the
fragments containing fossils are about three feet across and a foot thick ;
the greater part of them are extremely angular, showing by their form
that they have been transported for a very short distance. Moreover
the extreme softness of the material would make it impossible for it to
endure any distant ice carriage. The sand in which the fragments con-
taining fossils were embedded appears to be to a great extent derived
from the destruction of the same rock. This fact is indicated not only
in the physical aspect of the sands, but in the character of the vegeta-
tion which grows upon them. Generally, in this morainal district, the
decomposition of the pebbles containing large amounts of feldspar and
mica affords a moderately fertile soil, which maintains grass. In the area
where these fragments abound, the sand is evidently far more siliceous
than elsewhere in the area of the moraine, and is too lean to support

MUSEUM OF COMPARATIVE ZOOLOGY. 91

plants. In part it is covered by a growth of lichens, and in part alto-
gether bare of vegetation.

This assemblage of facts makes it seem clear that the locality whence
this Cretaceous material is derived is not more than a few hundred feet
to the north of the site where these fossils are found. Were it farther
away, there could be no such concentration of the Cretaceous waste.
The hypogene material would be more extensive than it is. I have
therefore no hesitation in saying that we have in this immediate vicinity
a deposit of Cretaceous age. It is probable that this deposit of small
area is girdled about by strata of the same age as those about Gay
Head. This is indicated by the fact, that, at various points in every
compass direction from this locality, the drift contains large amounts of
bright-colored clays such as give the name to Gay Head. These clays
are not seen in their natural position, but are commingled with the
glacial waste, the fact being that when the glacier overrode this area it
ground up and commingled a good deal of bed-rock clays over which it
moved with the morainal material brought from a distance. A careful
study of all the exposures on Martha's Vineyard containing Tertiary
clays has failed to show any distinct fragments of Cretaceous rock. This
assemblage of facts has led me to the conjecture that some small remnant
of the Cretaceous beds projecting through the enveloping clays of later
age is the source whence these fragments containing fossils have been
derived.

The foregoing conjecture is more probable, for the reason that it will
explain in a satisfactory way the origin of much of the sedimentary
matter contained in the beds of the Gay Head section. That section is
remarkable for the very large amounts of siliceous matter contained in
its sandy and pebbly beds. This material is substantially what would
be obtained from the erosion of the Cretaceous strata such as are found
at this point, and the reassortment of the materials.

The physical conditions of the fragments of Cretaceous rock appear to
indicate that the beds were deposited near a shore line. The rock is of
a very coarse texture, showing faint indications of cross bedding ; the
clay element is scanty, and the quantity of lime is very small. It is
evident that the bed containing the fossils was accumulated with con-
siderable rapidity, and that only in certain levels was the. organic life
developed in sufficient quantity to make the bed fossiliferous. All these
conditions indicate that the deposit was formed near the coast line.

This opinion is borne out by the character of the fossils. The great
prevalence of oysters, and the fact that the two valves are generally

92 BULLETIN OF THE

found separate from each other, in some cases appearing to have been
worn by wave action before they were fixed in the strata, is almost con-
clusive proof that the deposits were made in shallow water. Although
the Exogyras differ in a certain measure from our ordinary oysters, their
distribution in this and other countries is always consistent with the
hypothesis that, like their living kindred, they did not inhabit the
deeper parts of the sea, but were dwellers in the shoal water.

The existing condition of the Cretaceous fragments affords us some
light as to the condition of the rock before it was disrupted by glacial
action. All the fragments containing fossils are extremely ferruginous,
the lime of the shells having been replaced by limonite. This is the
ordinary [result of atmospheric action on superficial deposits of this
nature. It appears to me quite evident that this replacement of the
lime was effected while the material was in its original position, and this
for the following reasons. The fragments of Cretaceous rock were in
many instances found lying upon the surface of the soil, or only partly
bedded within it. In these cases the limonation cotild not have oc-
curred since the fragments came to their present position, for the
reason that there would have been no source whence the iron could have
been derived. There has evidently been no considerable degradation of
the drift on this region since it was abandoned by the glacier. Owing to
the position of the deposit it was not subjected to any water erosion. It
is evident that the corrosive work since the disappearance of the glaciers
has not taken away more than a few inches, if as much, from the surface.
If the fragments had come to their present position without having ex-
perienced the processes of decay, the replacement of the lime could not
have been effected. I therefore am forced to the conclusion, that this
material had decayed in its original bed, before it was disrupted by the
glacier, and that the iron was derived from superjacent beds in the origi-
nal stratification. Although this conclusion is hypothical, it is of certain
interest, for the reason that it combines with the other known facts to
indicate that the glacial erosion which has taken place in this region
has been of slight amount. If it had been great in quantity, if several
hundred feet of the section had gone away, we should not only have had
this detrital material of Cretaceous age distributed over a larger field,
but the fragments would probably have come to this point in an unoxi-
dized condition.

A very severe rain-storm which occurred in the month of September,
1888, disclosed a portion of the sections in the neighborhood of the point
where Cretaceous fossils are found. Although the sections are obscurely

MUSEUM OF COMPARATIVE ZOOLOGY. 93

exhibited, a tolerable notion can be formed as to the character of the
materials along a line having a length of about 2,000 feet. The position
of the beds is in general indicated in the accompanying diagram, which
gives a somewhat generalized section from the north shore of the island,
a little west of Cedar Tree Neck, across to the small brook on the Cedar
Tree Neck road. On the north shore the beds occasionally exposed
after severe storms consist of grayish green sands, with occasional iron
concretions resembling those found in the " Wood Schoolhouse " locality.
Some of the fragments closelv resemble the material containing fossils
at the last mentioned locality, and in one fragment an unrecognizable
species of oyster was observed. At the highest point delineated in the
section, the shoved frontal moraine is partly interrupted, so that the un-
derlying rocks are exposed. Here we find a section having a length of
about 300 feet, showing a deposit of grayish gi-een sands alternating
with red and white clayey sand, the dip of the beds being to the north-
west, the angle varying from 45° to 60° of declivity. Proceeding south,
we find 300 feet of section in which the beds are concealed from view ;
then a small exposure of red clayey sand with an obscure dip, not more
than 60 feet in thickness of beds being exposed to view ; then 140 feet
of measures hidden by the covering of drift ; following that, 70 feet of
red and white clayey sand, very micaceous, dip obscure, but apparently
in the same northwest direction. Again, southward, a covered section of
about 100 feet in length, in which the drift is more or less churned up
with grayish sands presumably derived from the underlying beds. This
is the point where the fragments containing species of fossils described
in this report were obtained. Farther on, 600 feet of the section is un-
exposed ; then, for 180 feet, we have mainly greenish gray sands, having
a total thickness of about 80 feet, with traces of yellow and white sands
above and below them. At this point the dip is clearly shown. It is
to the northwest, at an angle of from 35° to 50°. Following to the
southward, 150 feet of the section is concealed; then for the distance of
about 125 feet, to near the margin of a small brook, the reddish clays
appear at the surface, but the dip is not clear. It appears to be in the
same northwest direction. At several points in the greenish measures,
some compact ferruginous layers resembling those containing fossils are
found ; but in none of these beds have I as yet been able to obtain
organic remains. They serve, however, to indicate that the material
containing fossils is really derived from this section.

Until fossils are actually found in a bedded condition in the deposit,
it will not be possible to assert in a positive manner that this section is

94 BULLETIN OF THE

of Cretaceous age. It still seems possible that the fragmeiits containing
fossils may be in their nature exotic, as are the fossiliferous materials in
the beds at Gay Head. It is to be noted, however, that the greenish
gray sand in the section containing the fossils differs considerably in its
general aspect from the beds at Gay Head. Moreover, there appears to
be an absence of lignites in this portion of the Vineyard series.

The prevailing northwest dips of this section are in contrast to the
attitude of the Gay Head series. I have carefully examined the bedding
with reference to the theory that the dislocation is due to glacial thrust.
I find it impossible to accept this view, for the following reasons. In the
first place, the dips are everywhere tolerably uniform, except within a
foot or so of the glaciated surface. In this uppermost part of the section
the thrusting and dragging action of the ice is distinctly exhibited in the
somewhat sharp flexure of the beds, as well as the considerable contor-
tion which they present. It seems to me impossible to believe that a
steadfast dip such as is shown by these beds could have been produced
by the thrust of a glacial sheet. If the dislocation were due to the direct
forward movement of the ice, we should have to explain these dips by the
supposition either that the beds originally horizontal were thrown into
an arched form, and that we have here the northwest side of the anti-
clinal, or that the beds were completely overturned in order to produce
the existing dips. There is no trace of such an arch exhibited in the
section. Indeed, the presence of such a fold is contra-indicated by the
fact that the dips increase to the southward, and the hypothesis of a
complete overturn finds no support whatever in the facts. Last of all,
we observe that the surface of this district apparently retains its pre-
glacial topography. A system of stream valleys is traceable over all the
section where the Vineyard series of deposits rise above the sea level.
The persistence of a pre-glacial topography, manifest even in the details
of the surface, — a topography on which the drift materials are simply
imposed, — is overwhelmingly against the supposition that the disloca-
tions are in any measure due to the action of the ice-sheet.

In a memoir on the Geology of Martha's Vineyard, prepared for
the Annual Report of the Director of the U. S. Geological Survey for
1885-86, I have endeavored to show that the deposits at Gay Head,
probably of Miocene or early Pliocene age, were formed in a delta at
a time when the level of the shore was perhaps not more than 200
feet below its present position. If the evidence from the fossils and
the physical condition of these Cretaceous deposits is to be trusted, it
indicates that in a much earlier time the shore on this part of the

MUSEUM OF COMPARATIVE ZOOLOGY. 95

North American coast was also not far from its present altitude. These
fragments lie at the height of about 150 feet above tide-water ; the most
elevated point at which I have traced the Tertiary deposits of the Gay
Head series is about the same height. Taken together, these two sec-
tions afford interesting evidence of the existence of the shore line close
to the present coast in two remote stages of the earth's history.

The occurrence of Cretaceous deposits in this part of New England is
particularly interesting, for the reason that it indicates the former exten-
sion of the deposits of this age to points much farther north than they
have hitherto recognized on the eastern versant of the continent. Hith-
erto, no beds of Cretaceous age have been known on the Atlantic coast
north or east of New Jersey. This new locality establishes the existence
of such beds about 100 miles farther north, and about 200 farther east
than those which occur in the New Jersey area.

The fact that the Martha's Vineyard Cretaceous material was appar-
ently deposited near a shore line, appears to indicate that the Cretaceous
section, at least that part which belongs in this particular horizon, never
covered the general surface of New England. Therefore we cannot
fairly attribute to erosion the absence of this portion of the Mesozoic
deposits in the New England area and the region to the northward. It
seems to me more likely that these beds were never deposited on that
portion of the continental surface.

Palaeontology of the Martha's Vineyard Cretaceous.

Although there can be hardly any question as to the general geologi-
cal position of the beds whence these fossils were derived, their precise
place in the section is not readily determinable with the material at
hand. As will be seen from the appended figures of fossils, the remains
are very imperfectly preserved, revealing the existence of four species
which are characteristically Cretaceous in their aspect. The species,
however, differs so far from that which has been found in more southern
portions of the continent that no sufficient identification of the particular
horizon is possible.

The most abundant fossil is the ExogjTa, shown in Figs. 19 and 20 of
Plate II. The specific differences between the several species of Exo-
gyra are rarely if ever sharply defined. The Martha's Vineyard form
appears to be sufficiently distinct from any others which have been de-
scribed to warrant the application of a new specific name. So far, no
distinct Exogyras have been found above the horizon of the Cretaceous

96 BULLETIN OF THE

period. Indeed, it seems likely that this genus does not range up to
the \evy summit of that series, but passes out of existence shortly after
the Middle Cretaceous section. Therefore the occurrence of this fossil of
itself affords fair ground for concluding that the deposit does not belong
to a higher level than the Middle Cretaceous. The species of Camp-
tonectes has never been found above the middle of the Cretaceous series.
Indeed, it appears to be characteristic of the lower portion of that sec-
tion. It may be taken as evidence, that the beds in question do not ex-
tend below the horizon of the Cretaceous. The other fossils which are
described and figured are less determinative in their value. They are,
however, so fiir as it has been possible to identify them, not inconsistent
with the hypothesis that these beds are of Cretaceous age, and that
they probably belong in the lower portion of that period.

In the present stdte of our knowledge concerning the field from which
these Cretaceous fossils are derived, it does not appear worth while to
undertake any description of the species. Although the material is in
fair condition for such work, it seems to me likely that further study of
the field will develop much better specimens. I have therefore sought
to do no more than refer these species to their genera, with suggestions
as to the apparent affinities of certain forms. In preparing tliis list, I
have been so fortunate as to be able to confirm my general determi-
nations by the advice of Dr. C. A. White, Palaeontologist of the U. S.
Geological Survey. My thanks are due to him, and also to my assist-
ant, Mr. Aug. F. Foerste, for a careful search of the island of Martha's
Vineyard, in order to determine wliether localities other than those I
had found existed on the island. Although the result of this was purely
negative, it has been of value to the investigation. I am also indebted
to Mr. Foerste for the preparation of the drawings figured on Plate II.
of this report.

The foregoing report is intended as a preliminary statement concern-
ing the Cretaceous rocks of this interesting locality. I hope to explore
the field by systematic excavations, and thus secure more complete and
accurate information than has here been presented.

MUSEUM OF COMPARATIVE ZOOLOGY. 97

EXPLANATION OF PLATES.

PLATE I.

This plate gives a sketch map of the island of Martha's Vineyard, intended to
afford in mere outline sufficient indications as to the position of the Cretaceous
localities which have so far been determined. The principal locality where these
fossils are found is shown by the line indicating the position of the section given
at the bottom of the plate. From this point the fossils were obtained which are
mentioned in the text and figured in Plate II. The locality on Lagoon Pond lies
on the eastern face of that sheet of water from one third to one half a mile south
of the entrance to the pond. The fossils from this locality are extremely imper-
fect, and are found in occasional fragments of Cretaceous rock involved in a thick
section of drift. A third locality, where a single fragment of a fossil oyster was
observed, is on the southern part of the island of Chappaquiddick, which lies to
the east of Edgartown.

At various points to the eastward of a line drawn from Great Tisbury Pond to
Lumbard's Cove the drift is frequently stained with ferruginous sandstone waste,
which is probably derived from Cretaceous deposits. It is possible that a portion
of the stratified rock deposits lying to the westward of the above mentioned line
may also be of Cretaceous age.

For a further account of the geology and topography of this district, see my
Memoir on the Geology of the Island of Martha's Vineyard, in the Seventh An-
nual Report of the Director of the U. S. Geological Survey.

PLATE n.

Fig. 1.

1 a, 1 b. New genus ? Compare Myoconcha.

Fig. 2.

Plicatula or Ostrea. Compare PL instabite, Stol., and 0. lugubris, Conrad.

Fig. 3.

Tellina (linear ia)!

Fig. 4.

Cardium ?

Fig. 5.

Pteria.

Fig. 6.

Lucina ?

Fig. 7.

Turritella (nerina?).

Fig. 8.

Camptonectes Burlingtonensis, Gabb.

Fig. 9.

Camptonectes parvus (?), Whitfield.

Fig. 10.

Chemnitzia.

Fig. 11.

Lucina.

Fig. 12.

Ceiithium.

Fig. 13.

Anomya ?

Fig. 14.

Turritella.

Fig. 15.

Nuculana.

Fig. 16.

Ostrea or Exogyra 9

Fig. 17.

Modiola.

Fig. 18.

Modiola ?

Fig. 19, 20. Exogyra. Compare E. ostracina, Lam.

No. 6. — The Intrusive and Extrusive Triassic Trap Sheets of the
Connecticut Valley. By William Mokris Davis and Charles
LivY Whittle.

[Published by permission of the Director of the United States Geological Survey.]

Contents.

1. Introductory.

2. Means of distinguishing Intrusions and Extrusions.

3. General Features of Intrusive and Extrusive Sheets in Connecticut.

4. Special Accounts of the more important Localities.

5. Conclusions.

1. — Introductory.

The outcrops of conglomerate, sandstone, and shale in the Triassic
formation of tlie lower Connecticut Valley are generally inconspicuous,
and alone would hardly afford means of deciphering the structure of the
region ; but they are accompanied by ridges of strong relief, marking
the resistant edges of trap sheets whose close conformity to the adja-
cent sedimentary beds has long been recognized. It was noticed by the
elder Hitchcock that some of these sheets were extrusive. Manifestly
these are of great stratigraphic value, for after taking their places in
the stratified series, they constitute truly conformable members of the
mass, and may be used as guides to the deformatioris that the whole has
suffered. 1 Attention was called to their value in this respect by the
senior author of this essay in 1883,^ and since then something of the
structure of the region has been worked out ^ for the United States
Geological Survey by means of the dislocations of the sheets that are
regarded as extrusive. The field about Meriden has also been found an

1 Chamberlin and Irving. Bull. 23, U. S. G. S., 1885, pp. 100, 101.

'^ Amer. Journ. Science, XXIV., 1882, p. 347. Bull. Museum Comp. Zool.,
Geol. Ser., L, 1883, p 249.

3 Amer. Journ. Science, XXXII., 1886, p. 342. Amer. Assoc. Proc, XXXV.,
1886, pp. 224-227. Seventh Ann. Rep. U. S. G. S., 1888. Bull. Museum Comp.
Zool., Geol. Ser., IL, 1889, pp. 61-87. Amer. Journ. Science, XXXVII., 1889, pp.
423-434. Meriden Scient. Assoc. Proc, 1889.

VOL. XVI. — NO. 6.

100 BULLETIX OF THE

excellent training ground for the Harvard Summer School of Geology.
There is, however, still difference of opinion as to which of the trap
sheets are of extrusive origin, and it has therefore seemed advisable
to examine all the evidence thus far collected which bears on this
question.

2. — Means of Distinguishing Intrusions and Extrusions.

Our belief is that the eastern traps are extrusive sheets, which were
poured over the floor of the Triassic estuary from various and undis-
covered vents at several times during the deposition of the bedded
members of the formation ; that three of the sheets attained areas of
many square miles, — perhaps of several hundred square miles, — the
second of the three being the sheet now seen in the main line of ridcres
from Branford northward to Meriden and beyond to the Massachusetts
line, while the first and third constitute the anterior and posterior ridges
respectively. It is probable, also, that certain other eruptions occurred
later, although the outcrops of their flows are not yet well correlated.
If such be the facts, we should expect from our knowledge of existing
lavas to find many indications of the contemporaneous origin of these
sheets. Deposits of ashes and bombs may reveal the locus of eruption.
More or less disturbance may have been created in the unconsolidated
sediments as the lava flood advanced over them at the bottom of the
estuary. Successive flows or intermittent advances of a single flow may
have quickly followed one another, forming a composite sheet of lava.
While the middle part of a flow would be relatively dense, the upper
part would be vesicular, after the fashion of modem flows, and the sur-
face might exhibit the ropy or clinkery character of lava streams. After
the eruption, the igneous sheet would be gradually buried by the con-
tinued deposit of sediments that settled slowly down in all the cavities
and inequalities of the surface, thereby acquiring a stratification in mi-
nute accord with all its irregularities. Where the waves and currents of
the ancient estuary were strong enough, clinkery fragments may have
been moved about on the surface of the sheet from the more exposed
situations, and carried to the deeper, quieter water, there settling down
with finer detritus from a more distant source.

On the other hand, if the lava sheets that we have nictured as extru-
sive were in reality intrusive, nearly every feature would be changed.
The contrasted features of the two kinds of sheets must surely be
distinct enough for preservation and detection. We have therefore

MUSEUM OF COMPARAXn^E ZOOLOGY. 101

searched the region carefully for all the outcrops and openings that
might give opportunity of testing these deductive possibilities, and we
now present the result of this search.

During the progress of our field and laboratory studies, the latter
having been carried on by the junior author, we have looked for the
results of similar studies in other regions. It appears from this that
the question as to the intrusive or extrusive origin of lava sheets is
seldom discussed in detail ; as a rule, it has been settled by the citation
of a few facts, without going through the greater labor of making com-
plete diagnoses. We cannot therefore always determine whether all the
criteria of intrusion or extrusion are present in the examples referred
to. Opportunity for observation is often limited ; search for outcrops
is frequently hasty ; but the criteria that are cited are as a rule dis-
tinctive. Putting all these together, we find that the facts indicative
of an intrusion are as follows : —

An intrusive sheet is not confined to a single horizon, but may break,
across the adjacent strata.

The lower and upper portions of an intrusion are nearly identical.
Offshoots may traverse the superincumbent beds for some distance from
the main sheet.

The texture of the mass is, with small exception, dense throughout,
being uniformly and coarsely holocrystalline in the middle, but becom-
ing very close-grained and glassy close to the upper and lower surfaces,
with the development of marked porphyritic structure and of minerals
not observable in the middle, and non-polarizing action immediately at
the contact.

A cellular or amygdaloidal texture is rarely developed, and when
occurring seems to be confined to the upper portion of the sheet. The
microscope generally does not discover a definite boundary or a tangen-
tial arrangement of feldspar crystals around tlie walls of these pseud-
amygdules, and their cavities are therefore ascribed to replacement.

The porphyritic crystals of the upper surface are arranged tangen-
tially to the inequalities of the enclosing rock, showing the former to be
secondary to the latter.

Enclosed fragments of the country rock may be found near the upper,
as well as near the lower, surface of the sheet.

The overlying rocks, as well as the underlying, are fractured and dis-
turbed, and friction breccias are sometimes formed along the contact
surfaces, the fragments from the intruded and the enclosing rocks being
mutually and mechanically commingled.

102 BULLETIN OF THE

The beds above the sheet, as well as below, may be altered by heat.
The alteration is commonly seen in change of color, induration, pro-
duction of new minerals, or the development of a local prismatic habit.

Strongly contrasted with all these are the features characteristic of
extrusions : —

An extrusive sheet lies conformably on the surface over which it was
poured.

The lower and upper portions are strongly unlike.

The upper surface sometimes manifests a ropy flow structure, and
sometimes consists of a mass of clinkers.

Vesicular or amygdaloidal texture is very common, especially near
the upper surface, and sometimes within the mass.

A composite structure, as of two or more flows, is not uncommon.

Vesicles are often drawn out in a common direction, parallel to the
adjacent surface, and indicative of motion ; but greatly elongated
" spike " amygdules stand at right angles to the neighboring surfaces.
These amygdules are commonly characterized by a definite boundary,
and by a tendency to an arrangement of the adjacent feldspar crystals
parallel to their walls, and are therefore regarded as the product of
expanding gases. Pseud-amygdaloidal cavities are also common.

There is a marked tendency to the development of a porphyritic
structure throughout the whole mass.

The overlying beds show no evidence of alteration by heat.

The overlying sediments are arranged conformably with the upper
surface of the sheet ; open vesicles and the spaces between clinkers are
more or less completely filled with sediments, deposited conformably
with the surface on which they rest.

A stratified mixt\xre of clastic materials and trap fragments, the latter
more or less water-worn, overlies the sheet.

Extrusive sheets may be associated with ash beds and volcanic bombs,
and with beds of volcanic conglomerate, more or less water-worn, in a
horizon nearlv continuous with the lava sheet.

It may be added, that the effects of heat and of mechanical disturb-
ance in the imderlying beds are features common to sheets of either
intrusive or extrusive origin ; and that absence of induration and ap-
parently complete conformability with adjacent beds cannot be taken as
proving extrusion.

Induration is one of the most commonly quoted effects of the action

MUSEUM OF COMPARATIVE ZOOLOGY. 103

of igneous masses on adjacent sedimentaries. Percival makes frequent
reference to it in his Eeport ou the Geology of Connecticut. Yet, of all
the above mentioned signs of intrusive sheets, it is perhaps the most
difficult one to recognize. Simple induration is easily enough deter-
mined with a hammer ; but it is another thing to decide whether it
results from well advanced cementation by minute deposits of calcite
or quartz brought by infiltrating waters, or from baking by heat. The
sandstone overlying Saltonstall or Pond Mountain at its northern end
is excessively hard ; but its hardness is due to secondary deposits of
calcite, and not in the least to fmsion or baking. Moreover, it fre-
quently happens that the beds overlying undoubted intrusions or ad-
joining dikes are not hardened : this is commonly the case with
sandstones, as, for example, on the back of Gaylord's Mountain. Shales
are more affected by a dehydration of their clayey constituents, new min-
erals being formed when the temperature is higher and water abundant.
Sections cut from ordinary biscuit-ware show under the microscope no
essential difference from the hydrous kaolinite from which the ware was
made, excepting a greater compactness. The argillites of Somerville,
Mass., manifest little local alteration near their abundant dikes; as if
the general metamorphic process which changed the original clay-beds
into argillite had been so complete that the comparatively slight local
influence of the dikes was not sufficient to carry tlie change any further.
The argillites of Quincy, Mass., contain small garnets close to the large
intrusions of the Blue Hills. The shales overlying the Palisade Range
have been changed in color and texture so as to resemble hoj-nstone ;
biotite, hornblende, and epidote have been locally developed.

The induration of the sedimentary rocks immediately overlying the
trap sheets has not been neglected in the study of the ridges ; but while
simple induration is associated in some cases with unquestionable signs
of intrusion, it is found in other cases with equally decisive indications
of extrusion, and we have therefore been driven. to the belief that mere
induration is by no means of constant occurrence or definite association,
and that it must be regarded as of little determinative value, at least
for the Connecticut eruptives.

Our search for evidence of the origin of the trap sheets has been car-
ried from the coast of the Sound, by New Haven and Branford, along
the greater part of the various trap ridges, to Cook's Gap, west of Xew
Britain. Attention iias been given chiefly to the back of the sheets,
for the upper contacts are much more significant than the lower ; but,
although the upper contact lines must altogether amount to one or two

104 BULLETIN OF THE

hundred miles in length, the number of exposures upon them is very
small. Upper contacts are generally found in streams that descend the
back of the ridges, and these have therefore been examined most carefully.
The list below embraces all that we have yet discovered. The localities
are numbered to correspond with the figures on the map of Plate I.,
and are arranged according to the sheet to which we suppose them to
belong, beginning with the trap range near the western border of the
formation, and proceeding with the anterior, main, and posterior sheets
farther east ; these being interpreted as has been explained in earlier
articles.^ Some specimens from the Palisade Range of New Jersey, col-
lected in 1883, are described with those from the western range of the
Connecticut Triassic. The several smaller ridges, not correlated with
any of the sheets above named, have not been closely examined, and are
not here referred to, except iu locality 26. The pages in Percival's
Report on the Geology of the State, where he describes the localities here
mentioned, are added to our list, for the sake of convenient reference.
Our descriptions are made as concise as possible, in order to shorten the
necessary repetitions ; several of the more interesting and instructive
localities are given more space in special accounts further on. Mention
is made in certain cases of peculiarities of structure that do not bear
directly on the question under investigation, partly in order that ob-
servations might not be lost, and also in the hope that the details thus
collected might iu time lead to new generalizations. Certain micro-
scopical variations in the trap naturally resulting from differences in the
conditions of solidification are added to those which have a direct bear-
ing on the question of origin ; not that they are criteria in themselves,
but that they have become recognized as commonly accompanying the
two kinds of eruption. For example, the occurrence of porphyritic
crystals in an eruptive rock does not establish its extrusive origin, but
extrusive sheets are notably more porphyritic than those solidifying
beneath the surface. So, too, a holocrystalline structure does not war-
rant us in saying that a rock is undoubtedly intrusive ; but intrusives
are more frequently holocrystalline and extrusives more frequently
glassy.2 But we have not attempted to give a complete petrographic
account of the specimens that have been examined. It seems advisable
to postpone this until samples from all the Triassic basins of the Atlan-
tic slope can be studied together.

1 Seventh Ann. Report U. S. G. S., 1888.

2 See, on the otlier hand, the account of recent lavas from Kilauea, in which glass
is rare or wholly absent. E. S. Dana, Amer. Journ. Science, XXXVII., 1889, p. 461.

MUSEUM OF COMPARATIVE ZOOLOGY. 105

3. — General Features of Intrusive and Extrusive Sheets in

Connecticut.

GROUP I. WESTERN RIDGES.

Locality 1. Section numbers, 85-89. Local name, East Rock. Percival's Report,
pp. 395-398. Percival's notation, W. S. 1. (1).

General Account. — The southwestern face of this fine mass is well
exposed in a strong palisaded cliff on the border of New Haven, below
which the underlying sandstones can be seen at several points. The over-
lying sandstone close to its contact with the trap was found on the
northeastern slope, in the woods, about a third way down from the sum-
mit. Tliis rock is regarded as a part of the West Rock sheet, from
which it is thought to have been separated by a fault ; similar faults
of smaller throw are supposed to account for the notches in the south-
eastern extension of East Rock itself.

Sections cut from specimens taken from the upper contact and from
four feet below it cannot be distinguished from sections similarly selected
from the base of the sheet.

The trap is wanting in vesicles of expansion throughout its mass, and
is hclocrystalline except at contact with other rocks. Extremely close-
grained and glassy at the tipper contact, where it shows microscopic
flowage parallel to surface of junction with the overlying sandstone.
Sandstone directly above does not contain fragments of trap ; hand
specimens appear much more dense than from beds distant from the
trap sheet.

Locality 2. West Rock. Percival's Report, pp. 394-396. Percival's notation,

W. S. I. (4).

The general features of this ridge are like those of East Rock ; but
no exposure of the upper contact has been found on its back. As far as
seen, it is of dense texture, even in the uppermost parts exposed. The
southern end of the ridge, where the underlying sandstone is quarried
and exposed in contact with the trap, may be reached by the West
Haven horse cars from ]^ew Haven ; the remainder of the ridge is
wooded and less easily examined.

Locality 3. Section numbers, 194-199. Gaylord's Mountain, Roaring Brook. Perci-
val's Report, pp. 402-404. Percival's notation, W. S. II.

Gaylord's Mountain is a slightly dislocated continuation of the West
Rock range ; on its back, Roaring Brook has cut a picturesque ravine,

106 BULLETIN OF THE

well known in the neighborhood and easily reached by a walk of two
miles and a half from Cheshire station of the New Haven and North-
ampton Railroad, or by a less distance from the station of the same
name on the Meriden, Waterbury, and Connecticut River Railroad. It
gives the only good exposure of the overlying strata known to us on the
back of the western trap sheet, and deserves careful examination.

The trap here is without vesicles throughout its mass; holocrystal-
line except at contact with other rocks ; at itL' upper contact it is
extremely fine-grained and glassy ; flowage action is seen in the micro-
scopic arrangement of the feldspar prisms parallel to upper line of
junction. U])per surface of sheet obliquely traverses the beds of the
overlying sandstones and shales ; several small offshoots of fine texture
extend into the overlying rock (Fig. 12). Pebbly sandstone directly
above the sheet does not contain fragments of trap, and is not per-
ceptibly affected by the igneous mass even close to the junction ; the
shales that elsewhere approach the sheet are apparently indurated.
See special account.

Section numbers, 45-55. Palisade Range, New Jersey.

The easternmost or lowest trap sheet of the New Jersey Triassic area
seems to correspond with the lowest or westernmost sheet of the Con-
necticut area, and is therefore referred to here in order to extend the
number of examples quoted. Its base is finely exposed in contact with
the underlying sandstones at the Hamilton-Burr duel ground in "Wee-
hawken, on the bank of the Hudson, opposite New York City ; this out-
crop is well figured in Plate IV. of the Annual Report of the New Jer-
sey Geological Survey for 1882. Other exposures of the xmderlying
sandstone are common up the west bank of the Hudson, but contacts
are relatively rare. The only upper contact known is one pointed out
some years ago. by Professor Cook (Geology of New Jersey, 1868, p. 201),
in Englewood, about a mile south of the station of that name on the
Northern New Jersey Railroad, in a brook channel a few hundred feet
west of a road.

The trap of this sheet is dense throughout, as far as examined at
numerous outcrops. Its texture is rather coarse in the middle of the
sheet, but becomes very fine at lower and upper contacts. The adja-
cent bedded rocks are distinctly altered from their original condition,
with the development of new minerals. No fragm'ents of trap are
found in the overlying beds.

Under the microscope the trap is seen to be almost identical with that

MUSEUM OF COMPARATIVE ZOOLOGY. 107

from Gaylord's Mountain, but more olivine is present in the holocrys-
talline portions. Approaching the upper and lower contacts, there is a
gradual disappearance of the augite and a decrease in the coarseness of
texture; the augite disappears at the contacts, porphyritic crystals of
olivine become abundant, and the rock is extremely fine-grained and
glassy. Occasional pseud-amygdaloidal areas occur in the trap ; but
no vesicles due to the expansion of occluded gases have been observed.

GROUP II. EASTERN TRAP RIDGES.

Division I. Anterior Ridges.

Locality 4. Section numbers, 26-28. Anterior at northern end of Totoket Mountain.
PercivaVs Report, pp. 344, 345. Percival's notation, A. 1. N. of E. II.

The ridge anterior to Totoket has few strong outcrops ; the one here
referred to is at the north end of the main sheet in a stream bank,
east of S. W. Loper's, South Durham (Fig. 3). Best reached by stage
from New Haven to North Guilford. Base of sheet for a thickness of
eight feet consists of a breccia of scoriaceous trap and clastic material,
cemented together by quartz and calcite ; upper part extremely vesicular ;
uo upper contact found. Lower portion glassy and porphyritic.

Locality 5. § mile S. E. of East Meriden. Peravnl'i; Report, pp. 302-305. Percival's

notation, A. 1. of E. III. (3).

The anterior to the long Durham range is traceable for many miles,
but is often heavily covered with drift. The bluffs of the ridge are of
the ordinary dense trap, and its back is as usual vesicular. About a
quarter of a mile south of Black Pond, near East Meriden, there is a
faint depression in its back, and here the ground is covered with
numerous fragments of sandstone containing pieces of vesicular and an-
gular trap (Fig. 13). A shallow opening would secure excellent speci-
mens. It seems as if there was here a depression in the surface of the
sheet, into which local fragments of trap were washed with sand from a
more distant source.

Locality 6. West of northern end of Higby Mountain. Percival's Report,
pp. 362-365. Percival's notation, A. 1. of E. III. (4).

The gap between Iligby Mountain and Chauncy Peak is followed by
the Meriden, Waterbury, and Connecticut River Eailroad, and by the
highway from Meriden tc Westfield (Fig. 4). A road branches from
the latter in the gap, and runs south on the amygdaloidal back of the

108 BULLETIN OF THE

anterior ridge. Following it about a third of a mile, and then turning
west into the woods, a few ledges are found consisting of ashes and
bombs, such as are more fully described under locality 8. Half a mile
farther south, the sandstone lying on the back of the highly vesicular
trap is exposed in the roadside. Numerous vesicular fragments of trap
are included in the sandstone. Clastic deposits are seen in many of the
vesicles in these fragments.

Locality 7. Southwest and west of Chauncy Peak. Percival's Report, p. 364.
PercicaVs notation, A. E. III. (5).

The road from Meriden to Westfield crosses this anterior ridge about
half a mile southwest of Chauncy Peak, and the above-mentioned ash
and bomb structure is visible in roadside cuts (Fig. 5, locality 7'). A
farm road follows the vesicular back of the ridge to the northwest, and
the ledges to the west of it show the same structure again, locality 7.

Locality 8. Section numbers, 83, 84 a, 209-212. Anterior of Lamentation Mountain.
Percival's Report, pp. 265, 266. Percival's notation, A. of E. III. (5).

The road from Meriden to Berlin follows the base of the ridge ante-
rior to Lamentation Mountain for some distance (Fig. 5). About two
miles north of Meriden, a curious bluff of volcanic ashes and bombs is
seen in the face of the ridge, locality 8. The tmderlying sandstone is
first seen at the foot of the bluff ; the overlying sandstone is found by
crossing the ridge to its eastern slope, locality 8', passing several trap
ledges in the woods on the way.

The trap is underlain by a bed of fine lapilli, about thirty feet
thick, containing numerous roimded blocks or bombs of dense trap,
from six inches to three feet in diameter ; one of these blocks is half
imbedded in the underlying sandstone. This basal ash bed is un-
doubtedly the. same as the one mentioned in the two preceding local-
ities, but it is not seen much farther north ; half a mile in that direction
there is a local trap conglomerate in the same horizon with the anterior
sheet ; vesicular and water-worn pebbles are here interbedded with sand,
as if this point were not far distant from a wave-beaten margin of the
anterior lava sheet. The trap of the ridge is frequently cavernous
and amygdaloidal, and remarkably so near the upper surface. No local
closeness of grain at upper contact ; overlying sandstone deposited
parallel to inequalities of trap surface ; fissures and vesicles near sur-
face filled with sand, connecting upwards with overlying sandstone.
Fragments of vesicular trap and abundant grains of water-worn glassy

MUSEUM OF COMPARATIVE ZOOLOGY. 109

trap in sandstone at contact ; two thin tufa beds a few feet above trap
sheet. See special account.

Locality 9. Anterior to Cat Hole Peaks. Percival's Report, pp. 375, 376.
Two small openings in the anterior ridge east of the Meriden poor-
house, a mile and a half northwest of the city (Fig. 6), expose the lower
part of the sheet. It is generally of dense structure, but presents ex-
tremely irregular forms, as if consisting of ropy masses of flowing lava ;
the spaces between these masses are filled with a much weathered
loose material that may perhaps be lapilli ; there are numerous " spike "
amygdules (see special account of locality 13) near and at right angles
to the convex surfaces of the lava masses. The upper portion of the
same sheet, where seen on roads on the back of the ridge, locality 9', is
highly vesicular.

Locality 10. Anterior of Notch Mountain. Percival's Report, pp. 375, 376. Perci-
val's notation, Ant. to E. IV. 1 (3).

A hundred feet southwest of the Meriden puorhouse, the sandstone
appears a little above the trap of the anterior sheet to Notch Mountain
(Fig. 6) ; a small piece of vesicular trap was found in it. The same
anterior sheet, where exposed in the Eeservoir Notch, a third of a mile
to the west, is extremely vesicular in its upper part.

Locality 11. Anterior to Shuttle Meadow Mountain.^ Percival's Report,
pp. 375, 376. Percival's notation. Ant. to E. IV. 1 (4).

A few poor exposures in the road on the back of this anterior, half a
mile south of Shuttle Meadow Reservoir, reveal weathered fragments
of vesicular trap in the sandstone overlying the sheet. Some of the
vesicles in these fragments contain clastic deposits.

Locality 12. Anterior to Farminprton Mountain. Percival's Report, pp. 375. Per-
cival's notation, Ant. to E. IV. 1 (9).

An excellent exposure of this anterior is found about a mile east
of Farmington, directly north of Stetson's house (Fig. 7). Middle of
sheet dense ; bottom sparingly cavernous ; upper portion generally
sub-amygdaloidal to cavernous ; very vesicular at upper surface, where
numerous vesicles are filled with indurated bitumen ; ^ surface of sheet
very uneven, with sandstone conformably filling hollows and open vesi-
cles; intimate mixture of trap fragmenfs and sand grains on upper
surface.

1 Called "North High Rock" in Bull. Mus. Comp. Zool., 1889, No. 4, Fig. 13.

2 Percival, Geol. Conn., 1842, p. 375.

110 BULLETIN OF THE

Locality 13. Section numbers, 175-182 b. Farmington River Gap, Tariffville. Per-
cival's Report, pp. 391, 393. Percivai's notation, A. to E. IV. 2 (2).

The main and anterior ridges are traversed by the Farmington River
at TariflFville (Fig. 8) ; the Connecticut Western Railroad passes through
the gap and exposes the complex structure and the upper surface of
the anterior ridge in a long cut a quarter of a mile east of the vil-
lage. The upper surface is seen again on the east bank of the river,
just above the road bridge.-^

A double sheet, as if of two flows. Lower sheet generally dense ;
sub-amygdaloidal, very porphyritic and glassy toward upper surface ;
upper portion very vesicular, and near surface contains " spike " amyg-
dules. No local close grain in trap at top of sheet; sand grains
conformably stratified in vesicles and small irregularities of surface ;
mixture of large and small fragments of trap with sand over surface,
this mixture passing laterally into a tufa bed ; trap fragments often
rounded as if water-worn.

Upper sheet compact at the base ; sub-amygdaloidal and vesicular
in upper portion ; generally very porphyritic and originally possessing
a glassy base ; overlying sandstone not seen in railroad cut, but well
shown on opposite river bank below, locality 13', where it carries
numerous trap fragments. See special account.

Division II. — Main Ridges.

Locality 14. Section numbers, 1-4, 73, 76. Saltonstall Mountain. Percivai's Report,
pp. 323, 324. Percivai's notation, E. I.

Saltonstall or Pond Mountain is the southernmost member of the
eastern main trap range ; it forms a well marked crescentic curve, with
Saltonstall Lake lying along the inner side. An under contact, lo-
cality 14', is found in the cut of the Shore Line Railroad, a quarter
of a mile east of Fair Haven station, and an upper contact is almost
revealed at the eastern end of the same cut. The back of the sheet
is very scoriaceous all along the shore of the lake, but no upper con-
tact is found until the northeastern end of the ridge is nearly reached,
when it is exposed in a little gully in the woods on the back of the
sheet over a pasture, locality 14 (Fig. 2).

The trap is porphyritic and was originally glassy ; at lower contact
with sandstone, the trap is brecciated, fine-grained, and glassy ; slightly
vesicular ; vesicles elongated, indicating flowage action. Very vesicular

1 W. North Rice, Amer. Journ. Science, XXXII , 1886, pp. 430-433.

MUSEUM OF COMPARATIVE ZOOLOGY. Ill

and irregular texture near its upper surface ; stratification of sandstone
conformable to irregularities in the upper surface. Intimate mixture
of sand grains and trap fragments along and above line of junction ;
surfiice fissures and vesicles filled from above with sand grains, distinctly
stratified parallel with the sandstone bed above. The hardness of the
overlying sandstone is due to induration by infiltrated calcite, etc., and
presents no evidence of being derived from baking by heat. See special
account.

Localities 15, 16. Section numbers, 5-17, 77. Totoket Mountain, inside south and
north hooks. Percival's Report, pp. 336-338. Percival's notation, E. II.

Totoket }kIountaiu is a well formed crescent, next north of Saltoustall
Mountain. Exposures of the upper contact with the sandstone were
found in a stream, locality 15 (Fig. 2), half a mile northwest of North
Branford, in the southern hook of the crescent ; and again in a stream-
bed inside of the northern hook, locality 16 (Fig. 3). Another stream, a
mile southwest of the last, locality 16', cuts a channel in what seems to
be a bed of clinkers.

The trap is porphyritic, and originally possessed a glassy base ; upper
surfoce very vesicular and irregular ; sandstone lamination conformable
to uneven contours of surface ; intimate mixture of rounded (water-
worn) trap grains and sands at contact ; occasional trap fragments in
sandstone for a few feet above ; clastic grains of trap, quartz, etc., fill
vesicles, with lines of deposition parallel to the stratification of the sand-
stone above ; sand in vesicles is connected with the sandstone above by
narrow necks. The overlying sandstone, locality 16, is indurated by
cementation, and shows no signs of baking.

Locality 17. Section numbers, 204-207. Higby Mountain. Percival's Report, p. 351.

Percival's notation, E. III. [i].

The eastern base of Higby Mountain, south of the road from IMeriden
to Middlefield, is followed by the upper course of Fall Brook, which at
a point about a quarter of a mile south of the road lays bare a valuable
exposure of sandstone lying on the trap, locality 17 (Fig. 4). A second
exposure is found a little farther south, locality 17'. Numerous frag-
ments of vesicular trap enclosed in sandstone are found in the stream
for some distance northward.

The trap is porphyritic, and originally glassy ; upper surface very ve-
sicular, much decomposed, and uneven ; not excessively fine-grained at

112 BULLETIN OF THE

upper contact. Sand grains fill vesicles and irregularities of surface,
conforming closely to their shape ; intermixture of sand and numerous
large and small trap fragments along line of junction ; occasional
rounded (water-worn) fragments of amygdaloidal trap even five feet
above trap sheet.

Locality 18. Section numbers, 132-134. Lamentation Mountain. Percival's Report,
pp. 351, 352. Percival's notation, E. IIL (5).

A road passes the north end of Lamentation Mountain and bridges
Spruce Creek, that flows northward from the back of the mountain.
Exposures of sandstone on the trap are found up and down stream from
the bridge ; the best locality is about an eighth of a mile up stream,
south (Fig. 10), where the exposure is of much interest.

Trap porphyritic and glassy, particularly at upper surface ; upper
contact not locally of close texture ; upper portion of irregular texture,
highly vesicular, with uneven, rolling surface ; sand grains fill fissures
and vesicles near surface of trap ; narrow necks filled with the same
clastic material connect these vesicles with the sandstone above ; inti-
mate and complicated mixture of sand and trap over the upper surface
(Fig. 15) ; stratification of sand in vesicles and above sheet conformable
to surface, and generally parallel.

Water-worn fragments of vesicular trap occur in sandstone for two
or three feet above surface of sheet. The vesicles in these fragments
often contain small particles of trap mixed with quartz and muscovite
grains.

Locality 19. Section numbers, 136-150. Meriden City Quarry. Percival's RepoH,
pp. 370, 371. Percival's notation, E. IV. 1 (1).

The small easternmost ridge of the Hanging Hills group (Fig. 5
or 6) has been deeply quarried for railroad ballast and road metal at
its southern end, and now presents an excellent dissection of a complex
trap sheet, — the most instructive quarry in the region. It is about
a mile north from the centre of Meriden. The trap of the quarry con-
sists of a lower and an upper portion, separated by a well defined surface,
inclined to the eastward with the general dip of the Triassic monocline.
The lower sheet is exposed for about ten feet below the surface of sepa-
ration ; the upper, for sixty or eighty feet above it. Lower sheet ex-
tremely porphyritic, vesicular, and glassy ; upper part scoriaceous, of
rolling, ropy surfiice, showing evidence of normal weathering previous
to quarrying. A small amount of foreign clastic material occurs mixed

MUSEUM OF COMPARATIVE ZOOLOGY. 113

with scorise at contact with upper sheet. No local close-grained texture
at upper contact.

Upper slieet dense as far as exposed in quarry ; becomes somewhat
fine-textured at contact with lower sheet ; its original upper surface not
seen in the quarry, but half a mile northeastward on the east side of the
ridge, locality 19' (Fig. C), the trap becomes vesicular. Several lines of
fault breccia traverse the quarry, consisting of large and small angular
fragments of trap contained in apparently unstratified sandstone ; often
slickensided ; the trend of these breccias agrees with that of the neigh-
boring faults, as determined by stratigraphic evidence. See special
account.

No other significant exposures of the main sheet have yet been found
in its further northward extension in Connecticut.

Division III. Posterior Ridges.

Locality 20. Section numbers, 34-37, 74, 75. First ridge posterior to Saltonstall
Mountain. PercivaVs Report, p. 324. Percivai's notation, P. 1, E. I.

The upper surface of this posterior ridge is exposed only near its
northeastei'n end, at a road crossing, about a mile northeast of Salston-
stall Pond (Fig. 11). Elsewhere the outcrops are generally dense, but
sometimes vesicular on the back of the ridge.

Upper portion of sheet very vesicular and glassy ; not locally close-
grained at junction with overlying sandstone ; sand' grains and trap
fragments occur together at upper contact ; sand fills vesicles in trap ;
occasional water-worn fragments of trap in the sandstone a foot or more
above the sheet ; base of sheet sub-amygdaloidal.

Eidges of very coarse trap conglomerate occur in the neighborhood,
but their relation to this sheet is not yet clearly made out.

Locality 21. Sectton yuimher, 18-23, 187-19.". Second ridcre posterior to Saltonstall
Mountain. Percical's Report, p. u25. Percivai's notation, V. 2, E. I.

According to our interpretation of the stratigraphy, this ridge is a
second outcrop of the sheet already seen in the first posterior, here
showing a western dip, as if on the eastern side of a synclinal ; its base
is open in several small abandoned quarries near a road crossing, lialf a
mile northwest of Brauford station. Shore Line Railroad, locality 21
(Fig. 11) ; and its upper surface, with something of the overlying sand-
stone, is seen an eighth of a mile north of these quarries, on the eastern

VOL. XVI. —NO. 6. 8

114 BULLETIN OF THE

side of a small pond, locality 21'. The great fault that uplifts the crys-
tallines on the eastern border of the Triassic formation passes close to the
southeast of this ridge, and is probably the cause of the reversed dip of its
sheet and of the local fracture and overturning that it exhibits.-' Trap
generally porphyritic and glassy ; dense at the lower contact ; several
exposures of ftiult breccia with the sandstone (Fig. 17); trap sends
minute tongues of pure glass into lower sandstone, and occasionally en-
closes grains of quartz and feldspar. Highly cellular at upper surface
on northwestern slope ; its junction with sandstone above is not marked
by local close texture ; sandstone immediately above contains numerous
fragments of vesicular trap ; intercalated beds of shale and trappy con-
glomerate occur near base of sheet.

Locality 22. Section numbers, 31-33, 40, 73, 78. Ridge near Middlefield Station,
Air Line Railroad. Fercival's Report, pp. 355, 356. Percival's notation, P. 2 (S),
E. m. (3j.

This posterior is traceable for several miles on the east of Durham
Mountain, but the only satisfactory exposui'e is in a railroad cut, a little
way west from Middlefield station, Air Line Railroad. Base of sheet
sub-araygdaloidal as a whole, and locally very vesicular and uneven ;
subordinate intercalated layers of trappy shale and irregular masses of
abundantly vesicular trap near base ; some vesicles filled with clastic
grains of quartz, feldspar, muscovite, and fragments of glassy trap.
Upper surface very vesicular. Trap generally glassy and porphyritic.

Local'itij 23. Srrtion numbers, 24, 25, 72. Falls of tlie Aramamit River. Percival's
Report, pp. 354, 355. Percival's notation, P. 2 (N), E. III. (4).

This is probably on the same posterior ridge as the preceding, although
its direct connection has not been traced. Eock Falls Station of the
Air Line Railroad is close by (Fig. 9). Trap generally glassy and por-
phyritic, and not locally close-grained at junction with overlying rock.
Upper surface extremely vesicular, with many vesicles filled with clastic
material connecting with the main mass of sandstone above by narrow
necks. Trap grains mixed with trap fragments at contact and for sev-
eral inches above. A beautifully water-worn pebble -of ti-ap was found
imbedded in the sandstone several feet .above the sheet. Drift boul-
ders in railroad cut near by show contacts and mixture of trap and
sandstone.

1 Amer. Journ. Science, XXXII., 18SG, p. 347 ; Bull. Museum Comp. Zool., Geol.
Series, II., ISbO, p. 72.

MUSEl'-M OF COMPARATIVE ZOOLOGY. 115

Locality 24. Section numbers, 120-123. Highl.i'ml Lake. Pcrcim!'.<i Report, pp. 336-
3C3. Fercicid's notation, P. 4, E. III. (5).

The rid;^'o posterior to Chauucy Peak is cut near its southern end
by the Mcriilcri, Watcrbury, and Connecticut Kiver Raih-oad, a quarter
of a mile east of Hi-hland station (Fig. 4). x\.n excellent exposure.
Under contact not shown. Trap generally dense ; originally glassy and
porphyritic ; not locally close-grained at upper contact ; upper portion
extremely vesicular; sand grains filling vesicles and fissures, their lines
of deposit conforming to the irregularities of the trap surface (Fig. 14) ;
these deposits connected with the sandstone above by necks ; inequali-
ties in upper surface of trap covered by conformably stratified sandstone.
Numerous angular, vesicular, large and small fragments of trap lying
above the sheet ; spaces between these filled wifh n-regularly bvit con-
formably stratified sandstone ; vesicles in fragments filled with sand ;
some of the vesicles only partly filled, and in such cases the upper sur-
face of the filling is parallel to the dip of the Triassic monocline.

Localiti/ 25. Section numbers, 12J-1-"1. IlartforJ Ave. and N. Stanley St., New Brit-
ain. Percival's Re/jort, pp. 3S1, 3S4. Percivuf's notation, P. (e), E. IV. I. (7).

The overlapping ends of the small trap ridges on the northeastern
border of New Britain are regarded as faulted portions of a single pos-
terior sheet ; a small stream flows between them. The eastern ridge is
quarried, and discloses the base of the sheet ; the upper contact is found
where the stream runs on the back of the western ridge.

Trap generally dense, but containing local amygdaloidal areas, sur-
rounded by dense trap, as if produced by intermittent flowing; very
vesicular at upper surface, and originally possessing a glassy base ; sand-
stone immediately above contains water-worn grains and fragments of
much decomposed trap.

Locality 2(3. Section 7iumbcrs, 152-1-50. Near Trinity College, Hartford. Percival's
Report, pp. 385, 3SC. Percival's notation, 1\ (i"), E. IV. I. (9).

This ridge is of doubtfttl relationship : it may be a second posterior
sheet, and therefore not directly comparable with the previous examples.
Its middle portion and base are well exposed in large quarries. The
trap is generally dense; triangular areas between the feldspars contain
a little glass ; the lower portion is brecciated and extremely scoriaceous ;
obsidian-like grains of trap in shale immediately under trap ; upper
portion vesicular, but overlying sandstones not seen. See special
account.

116 BULLETIN OF THE

4. — Special Accouuts of the more important Localities.

The following more extended descriptious of certain selected localities
are added, to give a better understanding of the fulness of evidence on
the question in discussion than could be obtained from the foregoing
summary. AVe thus present examples of what we interpret as an in-
trusive sheet at Koaring Brook, on Gaylord's Mountain ; a bed of vol-
canic ashes and bombs, presumably near the locus of eruption of one of
the cxtiusive sheets, in the anterior ridge of Lamentation Mountain ;
the base of an extrusive sheet, at Hartford; the top of an extrusive
sheet in Saltonstall Mountain ; and extrusions of complex structure at
Meriden and Tariffville.

Roaring Brook, GaylorcVs Monntain. Locality 3. — On entering the
ravine of Roaring Brook from the drift plain at the eastern foot of Gay-
lord's Mountain, outcrops of sandstone are soon encountered with dip of
40° to the eastwai'd. These ai-e followed for several hundred feet up
stream until the rock in the stream bed is found to consist of fine-grained
trap, the line of contact having been passed unnoticed. A little search
is needed to discover it, but when once made out it can be followed
Avitb some distinctness. In a gcnei'al way, the trap sheet thus disclosed
lies parallel with the beds above it, but on tracing its surface up the I'a-
vine, it is seen to depart significantly from perfect parallelism and comes
in contact successively with dilTerent beds. jNloreover, it gives forth
very distinct branches or leaders (Fig. 12), one of which extends for
twenty feet into the overlying strata. The margins of these oiFshoots,
as well as the edge of the sheet itself, ai'e tolerabl}' even, in marked con-
trast with the excessive irrcgulai-ity of the upper surface of the trap
sheets of the eastern langes. The overlying beds give not the least
sign of trap fragments which so generally characterize the beds lying on
the back of the eastern sheets. Taking all these features together, and
placing them in contrast with those of tiie sheets on the eastern side of
the valley, there can be no question that their consistent differences
are due to some fundamental ditlerence in the manner of eruption of
the lava. We are forced to the conclusion, that the Avesteru sheet has
been driven in between the previously deposited beds of sandstone and
shale, while the others have been poured out on the surface of certain
beds, and afterwards buried under others of latei- date. Study with the
microscope confirms this conclusion. The trap of West Rock, a con-
tinuation of Gaylord's Mountain to the south, has been described petro-

MUSEUM OF COMPARATIVE ZOOLOGY. 117

graphically by Hawes, and classed by him as a dolerite.^ Sections from
near the middle part of the trap sheet forming Gaylord's Mountain do
not appear to diifer materially in their uvcroscopic characters from those
of West Eock. The trap is holocrystalline far from its upper and lower
junction with the sandstone or shales, and, as has been pointed out by
Hawes, is much less altered, and coiatains fewer hydrated minerals, the
products of decomposition of the augite, feldspar, etc., than the erup-
tive masses forming Saltonstall Mountain, or the Durham range, to the
east. Hawes believed this difference to be connected with geographical
location, and thought it had nothing to do with geological age.^ Ac-
cording to J. D. Dana,^ the great alteration of the trap in the eastern
range took place at the time of ejection, and depended on the en-
countering of subterranean waters which the molten rock took up in
its passage through the sandstone strata. Hawes followed this view,
and thought the eruptive magma might in such a way assume the
diabase type, while under less humid conditions the same magma on
consolidating would form a dolerite.

It appears, however, that the difference in the hydration of the east-
ern and western traps can be better accounted for by original structural
and mineralogical differences incident to the very different conditions
under which the several trap sheets solidified. This will be referred to
again in the special account of Saltonstall Mountain.

In the trap from Gaylord's Mountain, on approaching the overlying
sandstone, there is a gradual fining of the texture and an increased ten-
dency towards a porphyritic structure, the porphyritic crystals there
being set in an undifferentiated, non-polarizing base. The augite occurs
more rarely in well-outlined individuals, and constantly tends towards a
granular structure. Olivine, which has been detected in minute grains
in the same rock to the south, has once been abundant at the Roaring
Brook contact, in well-outlined porphyritic crystals, but is now mostly
altered to a fibrous grass-green to yellowish-green serpentine, or entirely
replaced by pseudomorphous calcite or dolomite. The augite occurs
less and less plentifully upwards, and at two inches from the junction
with the sandstone it cannot be found even in grains. Accompanying
the loss of augite and the increase of olivine, there is, especially at the
contact, a development of a non-polarizing base in which are scattered
innumerable acicular ledges of feldspar, some porphyritic, showing an

1 Amer. Journ. Science, IX., 1875, p. 186. 2 ibid., p. 190.

3 Ibid., VI., 1873, p. 107

118 BULLETIN OF THE

arrangement parallel to the adjacent surface of the sandstone. The
glassy base with its accompanying dots of ferrite is best shown in sections
from the narrow leaders running into the overlying sandstone (Fig. 12).
These leaders penetrate the sandstone for a distance of several feet ;
the largest, which is three inches wide at its beginning and over twenty
feet long, is seen under the microscope to be nearly pure glass, in which
minute double refracting areas are abundant ; the smallest leaders are
mere threads, and in composition are essentially glass.

Although as a whole the western trap is little changed, marked al-
teration and hydration are shown in the upper surface of the trap of
Gaylord's Mountain, and in the leaders ; and it is to be noticed in con-
nection with the much greater hydration of tlie Saltonstall ranoje, that
this zone of glassy trap corresponds to the general glassy base of the
extrusive sheets. By the association of the intrusive trap at Roaring
Brook with the coarse sandstone immediately above, it has probably been
brought into contact with water to a greater or less extent, and part of
its alteration may be attributable to this cause. No amygdules occur
in tlie trap, except rarely one of a pseud-amygdaloidal character; there
is no tendency towards a mixtui'e of the two rocks along the line of
junction, either of the kind seen above the extrusions or like the brec-
cias known with certain intrusions.

The microscope affords no evidence that the conglomeratic sandstone
has been indurated by heat. The sandstone is much decomposed, owing
to alteration of its feldspathic constituents, and its grains are somewhat
incoherent. This failure to show induration does not, however, militate
against tlie intrusive origin of the trap. Similar sandstone at the base
of Saltonstall Mountain exhibits no greater evidence of heat induration,
although it was surely subjected to a high temperature.

As far as Vjoth microscopical and field evidence go, there can be no
doubt that in the case of Gaylord's Mountain we have a well marked
example of an intrusive sheet. No observers have given it a different
interpretation.

The Ash-bed in the Lamentation Anterior. Locality 8 (Fig. 5). — Two
miles north of Meriden, near the road leading to New Britain, the fol-
lowing section is exposed in the ridge anterior to Lamentation Mountain.
The base of the bluff on the upper slope of the ridge shows a small out-
crop of fine-grained, brownish red sandstone ; immediately above this
there are twenty or move feet of tufa-like material, containing oval and
discoidal areas of close-grained trap that we have interpreted as volcanic

MUSEUM OF COMrAEATIVE ZOOLOGY. 119

bombs. Above the tufaceous deposit is a sheet of very amygdaloidal
trap, overlain by a dark pinkish gray sandstone, carrying two thin sub-
ordinate layers of trappy material a few feet over the contact. In the
hand specimen and under the microscope, this sandstone appears identi-
cal with the fine matrix of a trap conglomerate noted by Percival as
occurring half a mile to the north, and presumably forming the strati-
gi'aphical equivalent of the tufaceous deposit at this point. The se-
quence of outcrops here disclosed is one of the most valuable that it has
been our fortune to discover, and has attracted much local attention
since it was found in the spring of 1887. It will well repay attentive
examination. The following account refers in greater part to its micro-
scopic structures.

Under the microscope the material of the bluflf enclosing the volcanic
bombs is found to be made up of small fragments of trap, generally very
fine-grained and much altered. Small greenish brown areas dotted
thickly with ferrite are non-polarizing as a whole ; these appear to be
volcanic glass. A few porphyritic ledges of plagioclase occur in them.
Most of the eruptive grains have been altered to chlorite and quartz,
and are intimately mixed with granular calcite. The microscope fails
to discover any grains of water-worn quartz or other clastic material,
although it is probable that moi-e or less normally deposited sediment
occurs thinly scattered through the mass. No stratified arrangement
of the trap grains is noticeable in the microscopic sections, except an
orientation of chloi'ite plates parallel to the stratification of the sand-
stone on the back of the ridge, and to a rude lamination brought to sight
in the face of the tufaceous bed by weathering. Following Geikie, this
bed would be called a tufa, consisting of a shower of lapilli. It appears
to have been deposited rather rapidly in a body of water, and probably
at no great distance from a point of eruption, as it soon disappears to
the north and west. It is traceable a mile and a half to the southeast,
in localities 6 and 7.

The volcanic bombs occurring with tlie lapilli give the face of the
bluflf a curious mottled appearance. They show no definite arrangement,
but are more ninnerous near the bottom of the bed, where one of them
seems to have imbedded itself in the underlying sandstone ; they are
remarkable for their non-vesicular character and their compact uniform
texture from the centre to the surface. The microscope detects no
variation in texture in any part except that due to a partial alteration
of the surface. It shows them to be extremely close-grained, with por-
phyritic crystals of augite set in a ground mass of minute plagioclase

120 BULLETIN OF THE

needles and brownish glass. As regards their origin the microscopic
study yields no solution, but the field evidence leaves little doubt in the
observer's mind. The thin trap sheet overlying the lapilli is, wherever
observed at this locality, more or less vesicular, and in many places
cavernous. The greatest vesicularity is at its upper surface, and in the
hand specimens from the contact with the sandstone above the sand is
seen to have minutely penetrated the cavities and fissures of the scoria-
ceous amygdaloid. The sand grains not only occupy surface vesicles,
but they have percolated along cracks and irregularities in the trap to
a depth of two feet below the surface ; in some cases, they apparently
lie between or surround large areas of amygdaloid. Irving speaks of
similar phenomena in connection with the upper surfaces of extrusions
in the Lake Superior region, and refers to them as sandstone " veins." ^
The lamination of the overlying sandstone is parallel to the surface of
the trap, conforming closely to its minor irregularities. Flowage action
is seen in the trap in the elongation of its vesicles.

An interesting point is the occurrence of two thin layers of tufa in the
sandstone just above the trap, each about an inch in thickness and about
a foot apart. These layers appear in the hand specimen of a rusty brown
color, composed of water-worn fragments of trap mixed with clastic quartz,
and have a much weathered appearance. Under the microscope their
tufaceous character is well shown ; vesicular poi-phyritic trap grains
abound, and others of non-polai'izing character are derived from yellow-
ish glass, now partially or wholly devitrified. IMixed with the trap frag-
ments, there are abundant grains of quartz, muscovite, and orthoclase,
probably derived from the crystalline rocks which surround the Triassic
formation. The tufas as well as the sandstone effervesce readily with
dilute hydioohloric acid, owing to the presence of secondary calcite.
The sandstone owes its dark color in a large part to the presence of
comminuted dust-like particles of extremely weathered trap, scattered
through it and now altered to earthy chloi'ite and fine dots of ferrite.

The several well-marked features of this interesting locality leave
no doubt tliat the trap sheet here is of extrusive origin.

Hartford City Quarry. Locality 2G. — One of the posterior sheets, as
yet not safely correlated with other outcrops, forms a ridge of moderate
height, witli strong western blufi", in the southern part of Hartford,
where it is extensively quarried for road material. Trinity College
stands on its eastern slope.

1 Copper-Bearing Rocks of Lake Superior, Monogr. V., U. S. G. S., 188.3, p. 292.

MUSEUM OF COMPAKATIVE ZOOLOGY. 121

The upper portion of the sheet is vesicular, but its upper contact is
not seen. The under contact is well revealed in the quarry, and afibrds
the best exposure for the study of the ^xiae of a sheet that we have
yet found. It is of interest also as being the locality described many
years ago by the elder Silliman.-^ Yet this particular contact is not
altogether characteristic of the under contact of most of the extrusive
sheets, for as a rule the junction of the trap with the shale is without
complication of any kind : one lies smoothly on the undisturbed sur-
face of the other.

The underlying shale of the quarry will be first considered. Four
inches below the trap, the shale locally consists of tufixceous mate-
rial. Eound and linear fragments of yellowish brown glass are seen
under the microscope, thickly sprinkled with minute particles of some
decomposition product of iron. These partially devitrified glassy areas
are undoubtedly the remains of obsidian-like fragments deposited as the
normal result of erosion from some volcanic flow, or as ejected matter
from a volcanic vent. In either case, volcanic vents sent forth showers
of ashes or flows of lava, presumably at no great distance from this
point, and at the time of the deposition of the sandstone.

The contact line between the bottom of the trap and the under-
lying shale is as a rule irregular and indistinct. The lower portion
of the trap for a distance of four feet presents a very vesicular and
scoriaceous appearance, not unlike the upper surface of the lower flow
exposed in the Meriden Quarry. The microscope shows portions of this
scoriaceous material thickly sprinkled with well marked gas cavities,
many of them having a linear arrangement, roughly parallel to the
upper surface of the shale, due to the flowing action of the trap while
in a viscous condition. The same parallelism is also well shown at the
upper surface of the first flow in the Meriden Quarry, locality 19.

The trap for a thickness of several feet is not only abnormally scoria-
ceous, but is extremely broken. Irregular and rounded areas of vesicular
trap are apparently cemented together by brown calcareous sandstone
possessing a lamination generally parallel to the stratification of the shale
below. The microscope shows these brown areas to be mixtures of
secondary quartz, calcite, and a little chlorite, arranged in layers ; they
must have been deposited by infiltrating waters. The texture of the
trap gradually increases in coarseness as we approach the central part
of the sheet, and then grows porphyritic and finer-grained near the up-
per surface. Careful search has failed to discover its upper surface in
1 Amer. Journ. Science, XVII., 1829, pp. 121-132.

122 BULLETIN OF THE

contact with shale or sandstone on the eastern slope of the ridge, but it
is generally very vesicular, and resembles in all particulars the upper
surfaces of all the well determined extrusions in the valley.

If the abnormal scoriaceousness and broken character of the under
surface of the trap be rightly interpreted as a result of the flowing be-
neath water, then its anomalous character, as compared with the lower
contacts of numerous other extrusives in the valleys, remains to be ex-
plained. We have little direct evidence on this point, but conclude, as
sufficient heat and moisture to form a scoriaceous texture at the bottom
of the flows were present in all cases, that some other factor must deter-
mine the variation between the considerable disturbance manifested
here and the lack of disturbance at the contact of sand beds and the
base of flows in other localities. The most available additional factor is
a variation of pi'essure, and this would be a minimum at the base of a
thin flow in shallow water. The Hartford sheet is probably not over
forty feet in thickness. Emerson has described a similar disturbance
and brecciation at the base of a rather thin flow in Massachusetts. It
may therefore be the case that thin lava flows in shallow waters develop
an unusually scoriaceous structure at their base as they advance.

Saltonstall Mountain. Localities 14 (Fig. 2) andli'. — The curved
outline of this ridge seems to be the result of a gentle folding after
the sheet had taken its place in the bedded series, rather than a conse-
quence of conditions attending the time of eruption ; the same may be
said of the larger and somewhat more irregular curve of Totoket Moun-
tain, next to the north. There is an almost intuitive hesitation before
the suggestion that anything so massive as a lava sheet could be folded,
but this must disappear on recalling the strong folds of the heavy sand-
stones of Pennsylvania, or the stupendous contortions of the gneissic
rocks on which the Triassic formation rests. If the sheet were intrusive,
it might, to be sure, have wedged its way in between the sedimentary
bods after they had been tilted and gently folded, thus accepting their
guidance as to the form its outcrop should present ; and this has been
currently believed, both here and in the case of the similar but larger
curves of the trap ridges in New Jersey. It is therefore of more than
local importance to determine whether the Saltonstall sheet is an intru-
sion or an extrusion ; for if the latter, it surely cannot have originally
taken its present form, but must have passively suffered deformation
from an initial horizontal attitude.

The small opportunity for observation of the contacts of this sheet

MUSEUM OF COMPARATIVE ZOOLOGY. 123

with the adjacent beds has already been mentioned. The base is seen
in the Shore Line Raih-oad cut, locahty 14'. The back of the sheet
has been carefully searched from one ei^d to the other with no success
except in the little gully in its northern hook, locality 14 (Fig. 2), but
the general uneven and scoriaceous texture of its upper portion is con-
tinuously visible for two miles or more as it dips under Saltonstall Lake ;
this is seen to best advantage by rowing along the shore in a boat,
which may be obtained at the southern end of the lake.

The base of the trap sheet for a distance of several feet is decidedly
amygdaloidal and close-grained ; and, owing to its broken character and
the subseqiient infiltration of secondary quartz and calcite, it locally
resembles a breccia. Under the microscope, the trap is seen to be very
amygdaloidal, and the vesicles are elongated by the flowing of the trap
conformably to the line of junction with the sandstone below. Speci-
mens of this breccia-like mass appear identical to the eye and imder the
microscope with those from the base of the anterior at the north end of
Totoket Mountain, locality 4.

Round areas of a brownish material resembling water-worn fragments
of sandstone are apparently enclosed by the trap near its junction with
the sandstone, but the microscope shows these to be secondary deposits
in vesicles, and to consist of quartz and granular calcite, products of
alteration, stained brown by iron. Similar areas are found at the base
of a trap ridge on the northeastern limits of New Britain, locality 25,
where Percival erroneously refers to them as consisting of dark red jas-
per, the product of the indurating action of the trap ; ^ also at the base
of the tufaceous bed of the anterior to Lam. tation Mountain, local-
ity 8, and at the Hartford City quarries, locality 19. A section of sand-
stone three inches below the trap sheet of Saltonstall shows water-worn
fragments of trap, and denotes that at the time of the deposition of the
sandstone layers there were bodies of trap undergoing erosion in the
neighboring region : they may have been derived fi'om the front of this
very sheet before it had advanced .so far as the locality in question.

The upper surface of the trap forming Saltonstall Mountain is ex-
tremely vesicular and irregular; the vesicles are sometimes well defined,
sometimes vague, indicating both gas expansion and replacement as
their cause. The texture shows a distinct decrease in coarseness as we
approach the upper contact, although the upper portion, as a general
rule, is more coarsely crystalline than tha lower portion in contact with
the sandstone. Pumpelly speaks of this fining of the texture on ap-

1 Geol. Conn., 1842, p. 383.

124 BULLETIN OF THE

proaching the upper surface as common to all amygdaloids studied
by him in connection with the copper-bearing rocks of Lake Superior;
and these sheets are well known to be extrusive.^ It is to be noted
here that coarseness of texture, even at the base of lava flows, presum-
ably depends, other factors being the same, on whether extrusion takes
place on land surfaces or under water ; so that we should expect the
trap to be much finer in gi-ain when extrusion takes place under water,
since texture is a function of rate of cooling.

When a section across the upper contact is examined under the
microscope, the lamination of the sandstone, which occupies the in-
equalities in the upper surface of the trap, is seen to conform to the
general contour of the hollows. This conformity is usually visible in
the thin section, even when not noticeable in the hand specimen ; it is
of common occurrence iu other localities along the eastern ranges, and
is highly significant of the deposit of the sandstone subsequent to the
eruption of the trap. Sections of the trap at the uj)per surface of the
sheet also exhibit vesicles, more or less open upwards, which are partly
or wholly filled with stratified clastic deposits, connected with the over-
lying sediments by narrow necks. In some cases the sand-filled cavi-
ties are apparently isolated in the trap, but tliis appearance is owing to
the fact that the thin section is transverse to the opening along which the
sand grains filtered into the cavities. The clastic grains occupying
the vesicles are usually of the most enduring minerals derived from the
ancient crystalline rocks, on the side of the Triassic estuary : these are
quartz, various feldspars including microcline, hornblende, and musco-
vite, cemented together by granular calcite stained red by ferric oxide.
Small fragments of vesicular trap occur here also, not the least interest-
ing of the constituents. The grains first deposited are generally ar-
ranged with their longer axes roughly parallel to the contour of the
lower portion of the vesicle ; grains later deposited appear approxi-
mately parallel not only to one another, but to the general stratification
of the main mnss of overlying sandstone, and also to the stratification
in a number of similar vesicles in the upper portion of the trap sheet
at this point. So highly specialized an occiu'rence of chistic mate-
rial in vesicles at the surface of a trap sheet can have but one inter-
pretation : the trap sheet is extrusive. Like tlie conformity of the
sandstone or shale to tlie upper surface of the trap, the clastic filling of
the surface vesicles is very characteristic of the eastern ranges, and is

1 Metasomatic Development of the Copper-Bearing Rocks of Lake Superior,
Proc Amer. Acad., XIIL, 1877-78, pp. 282, 283.

MUSEUM OF COMPARATIVE ZOOLOGY. 125

particularly interesting in the way it recalls the details of the slow
process by which these trap sheets were buried. Irving mentions the
occurrence of filled vesicles at the upper surface of the diabases and
diabase-porphyrites of the Keweenawau series of Lake Superior, and
cites it as one of the strongest proofs of the extensive origin of these
rocks. -^

Fragments of vesicular trap are not uncommon in the sandstone im-
mediately overlying the surface of the sheet ; their edges appear some-
what water-worn. It is of course possible that such fragments as these
might have been derived with the sand from some distant source, and
that they therefore do not in any way bear on the contemporaneous
extrusion of the subjacent sheet. In such a case we should expect to
find fragments of trap at various horizons in the Triassic series, showing
no definite association with the intercalated trap sheets, but this is not
the fact. The sandstones and shales throughout the valle}^ here and
there contain abundant fragments of trap, but, except in a few cases
that will be specified, the fragments occur only in the bed immediately
overlying some one of the sheets of the eastern traj) ranges ; the frag-
ments are commonly vesicular, and as such cannot have survived long
transportation ; they are moreover but imperfectly water-worn, if at all,
and are sometimes angular, and can therefore be referred only to a
source close at hand. It seems reasonable to conclude, on these several
grounds, that the trap fragments in the sandstones that rest on the trap
sheet of Saltonstall Mountain may be accepted as giving indication that
the sheet had been formed before the deposition of the sandstone above
it. The action of waves and tidal currents on the scoriaceous, irregular,
and fragmental surface of a lava flow would be entirely competent to
detach and transport relatively coarse pieces of the lava from more to
less exposed situations, and mingle them with fine sands derived from
more distant sources; and this process might continue with decreasing
activity until the last remaining knobs of lava were buried under the
growing sandstone cover. This interpretation is the onh' one that ap-
pears consistent with the facts here noted. The sandstone lying on the
back of this trap sheet is distinctly harder than is common in the region,
and our first impression was that its hardness was due to baking, and
that the trap sheet was intrusive ; but this is not in the least borne out
by more careful study. The hardness of the sandstone is due to
cementation by infiltrating calcite in chief part, and not at all to
change from the ordinary structure of sandstone. The sandstone on

1 Tlie ropper-Bearing Rocks of Lake Superior, Mouogr. V., U. S. G. S., 1883,
lip 79, \Z0, 140.

126 BULLETIN OF THE

the back of the Totoket sheet, locality 16, is similarly indurated, and
shows as little indication of bakinsf.

In review, it may be said that tlie absence of any tendency towards
a finer crystalline textnre in the trap immediately at the contact with
the larger sandstone areas, the highly vesicular texture of the upper
portion of the mass, the parallelism of the axes of the sand grains and
of the lamination of the deposits that they form, the connection of the
sand filled vesigles by narrow necks with the sandstone above, and the
occurrence of trap fragments in the overlying sandstone beds, all point
to the extrusive origin of the trap sheet, and to the subsequent depo-
sition of the sandstone upon it. After reaching the conclusion that
the Saltonstall sheet is extrusive, it may be prohtabl}' compared with
the West Rock and Gaylord Mountain sheet. The first contrast to be
mentioned, and the one most conspicuous in the field, is the presence of
numerous vesicles in the upper portion of the eastern sheet, and their
absence in the western : this we would refer to the small pressure
upon the surface of the extrusion at the time of its cooling. The
few vesicles near the base of the sheet may have been produced at
the front of tlie advancing flow, when its thickness was not so great as
afterwards. Next may be mentioned the general holocrystalline, n(jn-
porphyritic, and relatively coarse texture and the small degree of altera-
tion of the western sheet, while the eastern is more or less glassy and
porphyritic throughout, and greatly altered. The eruption into water
and the highly seoriaceous texture of the upper portion must have
favored quick cooling and subsequent alteration in the eastern sheet,
without noi'mal subaerial weathering ; the effect of the pi'esence of
much glassy base must also be considered, for this is peculiarly prone
to alteration ; it is now as a rule wholly devitrified to chloritic sub-
stances, microliths, ferrite, etc. But none of these factors could affect
the intrusive sheets ; their imprisonment between the beds deep be-
low the surface would allow them to cool slowly, and thus acquire
a coarse texture, and would decrease the rate of hydration and altera-
tion ; for these reasons we find them preserving in a great degree their
original characteristics. It should he noted, however, that inasmuch as
a thin extrusive sheet is vesicular through a greater proportion of its
mass than a thick sheet, thick lava flows may bo much less altered
than thin ones. Thus the heavy sheet of Mount Tom in Massachusetts
is practically anhydrous, while the thinner sheet of Saltonstall Mountain
contains 3.9^ of water.-^ Finally, there is a most marked and per-

1 Hawes, loc. cit.

MUSEUM OF COMPARATIVE ZOOLOGY. 127

sistent contrast between the features of the upper contact in the western
and eastern sheets. These need not be again stated ; suffice it to say
that the features of the western sheet demonstrate the trap to be sec-
ondary to the sandstone, while those of the eastern sheet are equally
conclusive in showing the sandstone to be secondary to the trap. It
does not seem too much to say that all the many peculiar features of
these two sheets find reasonable explanation as consequences of the
strongly different conditions of their origin.

The localities referred to above as yielding trap frtigmeuts, but not
lying on the back of a trap sheet, are the trap conglomerate of the
anterior to Lamentation Mountain, which is certainly the stratigraphic
equivalent of the adjacent trap sheet ; the heavy trap conglomerates
northeast of the first posterior ridge to Saltonstall Mountain, which are
perhaps to be associated with the posterior, although probably dislo-
cated from it by faults ; and a single case south of Durham, where one
fragment of vesicular trap was found in a conglomerate, distant from
any trap sheet, but near the eastern crystalline boundary of the for-
mation.

Meriden Quarry. Locality 19 (Figs. 5, 18). — The Meriden City
quariy, in the easternmost ridge of the Hanging Hills group, has been
attentively studied, and with much profit. Suites of specimens were
carefully collected from above and below the surface of separation be-
tween the upper and lower masses of trap which appear here, with a
view to examining the evidence of double flow presented. Numerous
specimens were also taken from the linear breccias of sandstone and
trap fragments which traverse the quai'ry, in order to compare them
with- fragments of sandstone included in trap, such as occur in a dike
at Mount Carmel, locality 27, several miles to the southwest, and to
discover if they should in any way bear on the intrusive or extrusive
origin of the Meriden sheet.

The lower mass of trap, a, a. Fig. 18, is seen beneath the upper, h, 5. h,
on the west side of the quarry, where abundant evidence may be found
to show that the two were produced by separate eruptions. They are
divided by a somewhat irregular surface, like that of rolling ropy lava,
and usually marked by a seam, more or less open to the weather. The
lower trap is changed to a reddish brown color for a depth of three feet
or more below its upper surface, and contains numerous amygdular areas
of chlorite, giving it a mottled appearance, simulating an altered sand-
stone to the eye. The reddish brown color gradually disappears down-

128 BULLETIN OF THE

ward, and at four feet below the junction it is replaced by an earthy
blue-green trap having abundant amygdules of chlorite and calcite, and
to the eye appearing much fresher than the reddish trap.

Numerous sections were cut from the red superficial portion of the
lower sheet, and from its contact with the dense trap of the upper sheet.
in order to detect any clastic material that might occur there. Very
little was found, but immediately upon the upper surface of the lower
sheet a thin layer was discovered consisting of rudely stratified grains of
clastic quartz and orthoclase, mixed with angular fragments of trap,
like that of the red seam. Some of the trap grains are glassy, non-
polarizing, and of a light green color, thickly sprinkled with minute
dots of ferrite. They are probably fragments of the puraice-like surface
of the lower sheet ; other grains are amygdaloidal, and contain small
ledges of some tricliuic feldspar. The whole is cemented together by
quartz and calcite. There is no marked tendency towards a stratified
arrangement of the grains, such as characterizes deposition in water.
The trap grains appear to have been the result of the comminution of
scoriae on the surface of the lower sheet during the ordinary progress of
subaerial erosion, while the occasional grains of orthoclase or quartz
may have been deposited by wind or stream action ; and from this we
have supposed that the thickness of the lower sheet was somewhat
greater than the depth of the water into which it flowed. Hitchcock
long ago noted that the reptilian tracks in the sandstones in Massa-
chusetts occurred chiefly in the beds closely overlying the trap sheets,
as if the depth of the Triassic estuary had been decreased for a time by
the lava that had flowed into it.

The lower trap of the quarry at ten feet below the red seam, where it
is the least altered as far as the quarry exposes it, is tine-grained, of
a dark greenish blue color, and of a uniform texture, containing abun-
dant amygdaloidal cavities. Mineralogically it is composed of extremely
altered porphyritic crystals of plagioclase in a ground mass of minute
crystals of the same, which are in turn set in a matrix of the unindi-
vidualized base. The- base in places is a yellowish green glass, and in
others is wholly devitrified. The augite that it undoubtedly contained
originally has been entirely removed by alteration. Calcite and secon-
dary quartz are abundant, the former so plentiful that the rock effer-
vesces readily, even with very dilute hydrochloric acid. Under the
microscope, the rock appears profoundly decomposed ; its numerous
amygdules being due to replacement, with the occasional exception of a
well-outlined cavity, the result of gas expansion. Admitting the original

MUSEUM OF COMPARATIVE ZOOLOGY. 129

presence of augite, the lower sheet would be classed as a glassy form
of angite-porphyrite.

The texture of the rock steadily grows finer, and the cavities due to
gas expansion more numerous, as we approach the surface of junction
with the upper sheet ; and there is at the same time a marked increase
in the amount of glass forming the base. At five feet below the contact
the vesicles occupy nearly one fourth of the space, and in slide 140 of
our collection they are seen to be elongated parallel to the surface of
contact, as if indicating flowing action. The origin of the vesicles bj
gas expansion is beautifully shown in this slide by the well marked
tangential arrangement of the feldspar crystals about the elongated and
tortuous amygdaloidal cavities, conforming even to their minor irregu-
larities. Sections from the red seam, just under the junction, show
this portion of the trap to have been blown almost to shreds by the
escaping gases. The scoriaceous character here cannot be doubted ;
fully two thirds of the rock is made up of secondarily deposited calcite
and quartz, filling the irregular cavities and vesicles of the porous mass.
The inter-vesicular areas consist of a greenish glass, thickly sprinkled
with hair-like microliths of feldspar and an occasional porphyritic crystal
of the same. The red color of the seam is due to the formation of iron
sesquioxide. Hawes noted that the oxidation of iron-bearing minerals
exposed to surface weathering is from the protoxide to the sesquioxide
state, while the change is from one protoxide to another when not thus
exposed, as is true of the eastern ranges.^ It therefore seems likely that
in this instance the red color of the surface of the lower sheet indicates
surface weathering before the upper sheet was erupted, thus confirming
the suggestion already made, tha*- the thickness of the flow was great
enough to raise its surface above water. It is rare that this red color is
seen in the traps of the Triassic area.

The scoriaceous character of the sheet at its upper surface is much
more strongly marked than in others thus far examined in the Connecti-
cut valley ; this is also thought to be connected with the appearance of
the surface of the sheet above the surface of the body of water into
which it flowed. Cooling under the air must have taken place much
less rapidly, and under much less pressure, than when below the water
surface, thus permitting a more complete expansion of the occluded
gases and the production of a highly pumiceous surface layer.

The trap of the upper sheet just above the red seam appears in the
hand specimen much less altered than that from below. Even -at the
1 Amer. Journ. Science, IX., 1875, p. 190.

VOL. XVI. — NO. 6 9

130 BULLETIN OF THE

contact with the scoriaceous upper surface of the lower trap, the rock
is sufficiently coarse to detect porphyritic plagioclase crystals ; but
amygdules are entirely wanting. Under the microscope a few pseud-
amygdaloidal areas are seen. The rock, shows evidence of an original
glassy base, seen in the triangular areas between ledges of feldspar; it is
made up of triclinic feldspar, magnetite, and occasionally a minute grain
of olivine. There is a slight local tendency toward a porphyritic struc-
ture ; but this is lost ten feet above the lower trap. Calcite and chlorite,
the usual decomposition products, occur at the base, the latter being
sufficiently abundant to give the rock a greenish color nest above the
red seam ; this is lost ten feet above the contact, and the great mass
of the upper sheet is of a very dark bluish color and holocrystal-
line. The mineralogical composition given above is that of a normal
diabase, the amount of olivine being so small that it can hardly be
classed as an olivine-diabase. While the upper surface of the lower
trap is abnormally scoriaceous, the base of the upper sheet is abnor-
mally coarse and free from vesicles, as compared with other trap
sheets resting on sandstones or shales. This can be explained by the
well known poor conductivity of volcanic scoriaceous substances, whose
presence here permitted the upper trap to cool and solidify slowly,
and produce a more complete crystallization. A practical illustration
of the low conductivity of such material is found in the use of scum or
slag irom iron furnaces as a packing foi' steam pipes.

The lower sheet may be confidently called an extrusion, but as far
as this quarry goes, there is notlihig to determine the origin of the up-
per sheet. This, however, is fully settled by the general field evidence
of the region, which correlates this whole mass with the heavy sheet of
Lamentation Mountain, and that sheet has been clearly shown to be
extrusive.

The field evidence here referred to concerns the occurrence of faults,
which, as is so generally the case, are at nearly all points buried under
surface waste. It is therefore of particular interest to examine the
bands of breccia {<'., c, c, Fig. 18) by which the quarry is characterized, as
they are best interpreted as small examples of the great dislocations b}'
which the structure and topography of the formation are deciphered.
The hreccias therefore deserve attentive study. The apparently unbed-
ded sandstone, of which they in good part consist, is best interpreted as
a fine clastic filling of the fault fractures, derived from above, where the
walls were of sandstone or shale, and gradually filtered down among the
large and small angular blocks of trap that were broken from the quarry

MUSEUM OF COMPARATIVE ZOOLOGY. 131

sheets ; but, on the other hand, the sandstone might also, until its con-
tinuity in bands across the quarry was noticed, be regarded as fragments
of sandstone picked up and included in the trap at tiie time of its erup-
tion : not that such inclusions would necessarily indicate intrusion, for
extrusive sheets are well known to contain fragments oft he adjacent
country rock.

The general attitude of the several bands of breccia negatives the
second interpretation. The bands all maintain a straight course
through the quarry ; a single band may cut the lower, as well as the
upper sheet ; the bands stand at right angles to the general extension
of the sheets ; they are parallel to one another and to the course of the
large faults by which the region is broken. The dividing surface be-
tween the lower and upper sheet in the southern end of the quarry is
seen to be dislocated by one of the bands, with small heave on the east,
this being the relative displacement of the large faults in the region.
Neglecting this sufficient series of indications of their origin, we ex-
amine their structure more closely, and discover that they are fre-
quently slickensided, and that the trap fragments that they contain are
sometimes broken since taking their places in the bands. Moreover,
these trap fragments are themselves included in the sandstone matrix
of the bands ; the fragments are angular, and show no variation of tex-
ture from centre to surface ; the sandy matrix contains small broken
grains of sandstone, as well as of sand. Again, if the sandstone which,
accompanies the trap fragments had been picked up and included in the
main mass of trap at the time of eruption, it should present evidence of
the action of heat, as in induration, or more likely in some alteration,
for the relatively small areas of sandstone in so large a mass of trap
must have long been subjected to intense heat. With this idea in
mind, a comparison was made of sandstone from the breccia bands with
a block of sandstone in a large dike a little north of Mount Carmel
station. New Haven and Northampton Railroad, locality 27, to which
Professor Dana had called our attention. The blocks of sandstone in
this dike are five or six feet long and two or more wide. When struck
with a hammer they give a ringing sound, characteristic of induration.
Sections of the sandstone show it to be principally composed of quartz
grains mixed with fragments of feldspar, and closely cemented by a
clayey material. While it exhibits no significant alteration in composi-
tion from ordinary sandstone, it cannot be doubted that its exceptional
density was the result of the dehydrating action of heat from the molten
dike on the kaolinite that formed the clayey cement. The contact of

132 BULLETIN OF THE

the dike and sandstone is sometimes blurred, as if they had been lo-
cally melted together ; and the texture of the dike becomes finer on ap-
proaching close to the included sandstone fragments, just as it does on
approaching its sandstone walls.

Returning to the quarry, we find that the sandstone from the breccia
bands has no indication of induration, except tbat resulting from the
moderate cementation of its clastic material by secondary quartz and
calcite deposited around the grains. Sections of the sandstone in con-
tact with the included trap fragments and with the main mass of the
trap sheet show a well marked laminated arrangement of the sand grains
Jiearly parallel to the walls of trap and to the faces of the trap frag-
ments ; this points decisively to the deposition of the sandstone poste-
rior to the eruption and Assuring of the trap. There is also a laminated
arrangement of the sand grains on all sides of the trap fragments, as far
as examined, which we do not fully understand, but which may be per-
haps interpreted as indicating continued motion of the faulted masses
while the breccia was still moist and soft, every trap fragment moving
as a whole and thus calling for an adjustment of the sand grains around
it. There is no change in the texture of the enclosing mass of trap on
approaching the breccia bands, such as would certainly appear if the
sandstones were inclusions. A change of texture is so characteristic of
rapid marginal cooling that it is often shown immediately at the bor-
der of large amygdaloidal cavities, as has been mentioned by Pumpelly,^
and as is well marked in our slide 141, from near the upper surface of
the lower trap in the Meriden quarry, and again still better in slide
218 from the Middlefield Railroad cut, locality 22, in which a nearly
spheroidal vesicle is surrounded by a layer of trichitic glass having an
area as large as the vesicle itself.

In order to apply this test carefully to the case in hand, several sec-
tions were cut from the trap in the quarry, on either side of the best ex-
posed breccia, at the contact, and one and four feet away. These show
no tendency towards a finer grain, or towards a development of porphy-
ritic crystals or glassy character on nearing the breccia ; the character
of the trap remains constant to the contact. Moreover, the angular
fragments of trap in the breccia are of uniform texture, and are identical
with the trap on either side, except for a little greater weathering in the
former. These fragments may therefore have been derived by fracture
directly from the enclosing walls ; but certain minute grains of very
fine-grained decomposed trap, also occurring in the breccia, appear to

^ Metasomatic Development of the Copper-Bearing Rocks, loc. cit., p. 283.

MUSEUM OF COMrARATIVE ZOOLOGY. 133

have been derived from the upper surface of the surrounding trap, or
from another trap mass above.

We therefore conclude, in reviewing the examination of the breccias,
that sand and sandstone grains and a moderate share of rounded grains
of close-textured normally eroded trap were all filtered together down
the fissures that traversed the sandstones and trap sheets, and that on
reaching the points exposed in the quarry they found a confused mass
of large and small angular fragments of trap, broken from the walls at
the time the fissures were made, the whole forming a highly character-
istic breccia. Such breccias ai-e not uncommon in the valley, as at
Branford, locality 21 (Fig. 17), where they are associated with the great
fracture by which the formation is bounded on the east ; and in the
Tariffville Railroad cut, locality 13, of minor importance. Percival knew
a few of them, and called them " clay dikes." ^ While our conclusion may
therefore be considered well supported, it must be remembered that
the breccias do not afford any evidence as to the intrusive or extrusive
origin of the trap sheets, and are therefore to be regarded as of secondary
importance in this essay, however valuable they may be structurally.

Tariffville. Locality 13. — One fourth of a mile east of Tariffville
station on the Connecticut W^estern Railroad (Fig. 8), a cut exposes a
valuable section of the anterior ridge.* The greater part of the cut is
in massive trap ; a narrow band of breccia occurs near its middle. At
the eastern end of the cut, the upper portion of the sheet shows a thin
bed of tufaceous material, which locally passes into a bed of trappy
sandstone along the strike ; and above this there is a second sheet of
compact trap of moderate thickness, with its upper surface lost in drift.
The two sheets together constitute the anterior ridge at this place.
There appears to be little if any lithological distinction between them ;
they are both glassy varieties of augite-porphyrite. The upper surface
of the lower trap, although generally amygdaloidal, is not so much so as
is usually the case. Immediately beneath the sandstone layer, the
amygdaloidal cavities have an aberrant character, being several inches
in length and generally about one fourth of an in inch in diameter, with
their longer dimension normal to the surface of the sheet. Amygdules
in such cavities have been described from one of the extrusive copper-

1 The relation of these breccias to the faults of the region is more fully dis-
cussed in a previous Bulletin of this volume, No. 4, p. 77.

"^ See an account of tliis locality by W. North Rice, in the Amer. Journ. Science,
XXXII., 1886, pp. 430-433, where it was first brought to public notice.

134 BULLETIN OF THE

bearing traps in the Lake Superior region, where many of them were
composed of native copper; hence the name " spike amygdules," as given
by Pumpelly.^ Irving also mentions them from the same locality, and
Hawes refers to similar ones occuring in the trap of Connecticut as
" pipe-stem " amygdules.^ N ason notes their occurrence in the trap of
the Watchung Mountains in New Jersey,^ which Darton thinks is of ex-
trusive origin,* and Winchell reports them in greenstone from Thes-
salon Point, Ontario.^ Their occurrence in the lower trap of the
Tariif villa cut is restricted to a zone of little depth near the surface of
the sheet, where it may be supposed that escaping gases found the
easiest direction of expansion to be toward the surface ; hence their
peculiar position. A fortunate breaking of the trap may liberate one
of these rod-like amygdules ; they are composed of concentrically de-
posited calcite with a chlorite centre, or more rarely the chlorite centre
is wanting and the amygdule is now hollow. An occasional amygdule of
ordinary form associated with the spike amygdules is beautifully banded,
with its lamination parallel to the stratification of the sandstone above,
and hence dipping with it at the same angle, about twenty-five degrees
southeastward. Under the microscope, the bands are seen to be com-
posed of granular calcite and secondary quartz, the banding being due
to fluctuations in the supply of ferric iron during the process of filling
the vesicles. The lower part of the amygdules is extremely granular and
ferruginous ; the upper part usually consists of composite calcite indi-
viduals, and is free from iron. Some amygdules near the surface contain
grains of clastic quartz or orthoclase lying in the calcite filling, as is so
common in the eastern sheets, and arranged with the major axes of the
particles parallel to thebedding of the sandstone and lamination of the
amygdule. Cavernous amygdules with banded structure were also
found in the Farmington anterior ridge, locality 12. Their only other
occurreuce in this country as far as known, is in the amygdaloidal mela-
physe at Brighton, near Boston, Mass., where the great number and
essential parallelism of the bands to one another, and to the bedding of
the overlying slates, has been taken to indicate deposition of some kind
guided by gravity.^ In all these cases it may be fairly argued that

1 Proc. Amer. Acad., XIII., 1877-78, p. 296.

2 Amer. Journ. Science, IX., 1875, p. 191.

3 Geol. Survey of N. J., Report for 1888, p. 37.

4 Amer. Journ. Science, XXXVIII., 1889, p. 134.

6 Geol. and N. H. Survey of Minn., XVII., p. 15, Plate I.
« Proc. Boston Soc. Nat. Hist., XX., 1878-80, p. 426

MUSEUM OF COMPARATIVE ZOOLOGY. 135

the accordance of the bands in the various amygdules with the bed-
ding of the adjacent sedimentary layers demonstrates the eruption of
the igneous sheet before the deformation of the whole mass ; but
manifestly it does not bear on the manner of its eruption.

The microscope reveals a marked decrease in the coarseness of the
texture of the trap upwards as the overlying sandstone layer is ap-
proached at tlie eastei'n end of the railroad cut, and a corresponding de-
crease in the freshness of the rock ; but the texture nowhere becomes so
fine as that on the back of Gaylord's Mountain. The intermediate
sandstone at the south end of the cut contains fragments of amygdaloi-
dal trap in abundance, often water-worn ; but a little distance to one
side, this mixture is replaced by a strongly marked tufa bed in the same
horizon, resembling in color and appearance the lapilli from the ash and
bomb deposit in the Lamentation anterior, locality 8 ; under the micro-
scope it shows decomposed fragments of glassy trap in a cement of
calcite and chlorite with occasional fragmental grains of quartz and
muscovite.

The upper trap sheet does not present significant features in the
railroad cut, but descending to the river and crossing by the road bridge,
where its upper surface is apparently found, several exposures occur a
little way up stream, in which there is the usual mixture of trap frag-
ments with the sands of the sandstone that overlies the sheet. This is
thought to be the upper surface of the upper anterior sheet, because no
other trap outcrop is to be seen imtil the base of the heavy main sheet
is reached.

The breccia in the middle of the cut resembles the breccias of the
Meriden quarry, but is much narrower, being only four to six inches
wide. It is a fissure in the trap, on which some slight faulting has
taken place, as is shown by slickensides ; it is filled with a mixture of
sand and angular trap fragments, and was undoubtedly formed posterior
to the production of the trap.

5. — Conclusions.

It is difficult for those who have become convinced of the correctness
of a certain conclusion to state in an impartial manner the evidence on
which the conclusion rests. We shall therefore not attempt to review
all the evidence presented above, but will briefly call attention to the
uniform association in the eastern trap ranges of the numerous chnrac-
teristics of extrusive sheets, while the western trap range as consistently

136 BULLETIN OF THE

manifests the several characteristics of an intriisive sheet. It must be
remembered, too, that of the numerous locaUties instanced on the east-
ern ranges, all (with one exception, Hartford) belong to only three
extrusive sheets ; and hence the evidence that is found at one point
supplements or confirms that found at another in a most satisfactory-
manner. All this seems to us to be beyond explanation either by acci-
dental coincidence or mistaken identification. While judgment might
well be suspended if our argument rested on single examples, or on nu-
merous examples confusedly arranged, it is difficult, even if necessary, to
maintain an open mind -in the face of evidence at once so full, so varied,
and so accordant. If all the trap sheets of the region were of one
kind, the ai'gument would be weakened ; for in the absence of either
kind of sheet, the peculiarities of the other would not be illumined by
the light of contrast. The presence in the single region under consid-
eration of sheets with the features of intrusions and extrusions there-
fore greatly increases the confidence that one may feel in the case, and
waxTants the acceptance of those sheets that we have called extrusive
as conformable and contemporaneous members of the Triassic series, by
means of which the dislocations of the formation can be detected.

The fullest statement of the method by which the extrusive trap
sheets can be thus employed is given in the article above referred to,^
by the senior author, in which the process of investigation followed by
the advanced section of the Harvard Summer School of Geology during
a week's work about Meriden is presented in detail. It is now our
design to continue the investigation in the district northwest of Hart-
ford, where a preliminary excursion has indicated a change in the course
of the faults from the uniform northeast trend that they possess in the
Meriden district. When the faults are mapped out over a considerable
area, comparison can be made between their course and the strike of the
schists on either side of the Triassic valley, on which the course of the
dislocations is thought to depend.

1 The Faults in the Triassic Formation near Meriden, Conn., Bull. Museum Comp
Zool., Geol. Series, I., 1889, pp. 61-87.

NOVEMBEK 16, 1889.

MUSEUM OF COMPARATIVE ZOOLOGY. 137

EXPLANATION OF PLATES.

PLATE I.

Fig. 1. Map of Triassic area in Connecticut from Long Island Sound to the north
bend of tlie Farmington River, based on Percival's map in his Geology
of Connecticut. The numbers in circles refer to localities on the sev-
eral trap ridges described in the text, and in most cases figured on
a larger scale in later plates. See page lOi.

Plates II. and III. contain outline maps traced from town maps in county
atlases, the trap ridges being sketched in black ; they cannot claim
much accuracy, but will probably serve as guides to the localities
that furnish exposures of critical contacts.

PLATE IL

Fig. 2. Adjacent ends of Saltonstall and Totoket Mountains. For locality 14,

see p. 110 ; locality 15, see p. 111.
Fig. 3. North end of Totoket Mountain. Locality 4, see p. 107 ; locality 16, see

p. 111.
Fig. 4. North end of Higby Mountain. Locality 6, see p. 107 ; locality 17 and

17', see p. Ill ; locality 24, see p. 115.
Fig. 5. Chauncy Peak, south end of Lamentation Mountain, and Quarry Ridge,

ISIeriden. Locality 7 and 7', see p. 108 ; locality 8 and 8', see p. 108 ;

locality 19 and 10', see pp. 112, 113.
Fig. 6. Notch Mountain and eastern ridges of the Hanging Hills. Locality 0

and 9', see p. 109 ; locaUty lU, see p. 109 ; locality 19 and 19', see

pp. 112, 113.

PLATE in.

Farmington Mountain and its anterior ridge. Locality 12, see p. 109.
Farmington River Gap, at Taritfville. Locality 13 and 13', see p. 110.
Rock Falls of Aramamit River. Locality 23, see p. 114.
North end of Lamentation Mountain. Locality 18, see p. 112.
Posterior ridges to Saltonstall Mountain. Locality 20, see p. 113 ; locality
21 and 21', see pp. 113, 114.

Fig.

7.

Fig.

8.

Fig.

9.

Fig.

10.

Fig.

11.

138 BULLETIN OF THE MUSEUM OF COMPAEATIVE ZOOLOGY.

PLATE IV.

Fig. 12. Overlying sandstone traversed by a small leader from the trap sheet of
Gay lord's Mountain at lioaring Brook, locality 3. See pp. 115 and
116.

Fig. 13. Angular fragments of trap imbedded in sandstone on the back of the
anterior ridge of Higby Mountain, half a mile southeast of East Meri-
den, locality 5. See p. 107.

Fig. 14. Drawing from a microphotograph of a section of vesicular trap from the
ridge posterior to Chauncy Peak at Highland Lake, locality 24. Tlie
trap is black, with white areas representing minute pseud-amygdules
and an occasional prism of plagioclase ; the large central space within
the trap is an amygdule, containing clastic material (dotted) at the
bottom, with the once horizontal lines of deposition now tilted parallel
to the general monocline of the region ; the upper part of the amyg-
dule is filled with calcite, of which part is stained with some ferrugi-
nous material (fine lines), and the rest is composite crystalline calcite
(blank). See p. 115.

Fig. 15. Drawing from photograph of sandstone in contact with vesicular upper
surface of trap, forming Lamentation Mountain, locality 18. The
black areas are the thin walls separating vesicles ; white spaces are
amygdules of calcite. See p. 112.

Fig. 16. Drawing from photograph of hand specimen of sand grains filling open
vesicles in trap. Falls of the Aramamit Kiver. Two vesicles have
lower bands of calcite, and the remaining space filled with clastic
material. Locality 23, see p. 114.

Fig. 17. Breccia from fault in a road-cut in the second posterior ridge to Salton-
stall Mountain, near Branford, locality 21. This fault is probably a
branch of the great fault by which the Triassic formation is limited on
the east. See p. 114.

PLATE V.

Fig. 18. The City Quarry at Meriden, looking northwest ; locality 19. a, a,
the -lower tiow in the southern part and the western alcove of the
quarry ; b, b, b, the upper flow, forming most of the mass here exposed ;
c, c, c, breccias of angular trap fragments and sandstone, traversing
the quarry. See pp. 112, 127. The northern extension of Cat-hole
Ridge is seen in the distance.

DAVIS & WHITTLE. TRIASSIC TRAP SHEETS.

PLATE I.

F/o. 1.

Photo Lith D_/ L S-Pupf frson & Son, Nevi Haven Ccnn

DAVIS & WHITTLE, TRIASSIC TRAP SHEETS

PLATE II

Photo Lith by L S Purderson &. Son. New Haven Conn

DAVIS i WHITTLE. TRIASSIC TRAP SHEETS

PLATE II

Photo Lith by L.S. Punderson t Son, New Haven, Conn.

DAVIS & WHITTLE, TRIASSIC TRAP SHEETS.

PLATE IV.

FIG. 12. "73

Photo. Lith b^ L S.Punderson & Son. New Haven. Conn.

PLATE V.

G. 18.

Photo Lith bv L- S Pundeison &. Son. New HavRn, Con"

No. 7. — The Topography of Florida, hy N. S. Shaler. With a
Note hy Alexander Agassiz.

[Published by Permission of the Director of the U. S. Geological Survey.]

Both in its general form and in the detail of its surface, Florida pre-
sents many interesting features. I propose in tiie following pages to
pass in general review the more important topographic elements of this
peninsula, and to consider the information which they give us as to the
general history of the continent.

The peninsula of Florida, as is readily seen by glancing at an ordinary
map, forms a salient on the coast hue which departs widely in its gen-
eral character from all the other great capes of the continent. The pre-
vailing trends of the eastern coast are from northeast to southwest.
This projection extends in a general northwest and southeast direction.
All the other greater peninsulas of the continent are distinctly moun-
tainous in their character. This of Florida is formed of low lands, rising
as a broad fold from the deep water on either side to a vast ridge, the
top of which is relatively very tiat, there being no indications of true
mountain folding in any part of the area. All the other great penin-
sulas of the continent, except that of Yucatan, which in certain ways
resembles Florida and may be causatively connected with it, are com-
posed of old rocks. The last named salient is made up altogether of
very recent strata.

The detailed topography of Florida is almost as anomalous as its gen-
eral configuration. Tlie region of the Everglades in the southern part
of the peninsula, and that known p.s the " Lake District " in the north-
ern, are both eminently peculiar in their configuration, having, so far as
I am informed, no likeness in any other part of this countrv.

The first question before us concerns the origin of the Florida uplift.
It will be observed that we have on the peninsula of Florida a very re-
markable ridge, which has grown up from the sea-floor to the altitude
of al)out five thousand feet ; and a somewhat similar elevation in the
archipelago of the Bahama Islands. Neither of these ridges has a

VOL. -XVI. — NO 7.

140 BULLETIN OF THE

mouutaiuous character. ludeed, it is at first sight difficult to find the
analoo'ues of these ureat anticlinal-like folds in the existing structures
of the land. They can hardly be classed with any of our known table
lands, for the reason that such elevations are in all cases more or less
associated with definite mountain folding. The only similar structure
which is known to me is that exhibited in the " Cincinnati anticlinal,"
that well known- ridge extending from near Columbus, Ohio, to Xorthern
Alabama. This elevation in length and breadth may be compared to
that of Florida, though it never had more than one half the height of
the Floridian peninsula.

In endeavoring to account for the Florida ridge, we must bear in
mind the processes of deposition which have evidently occurred in this
region. The geological histoiy of the sedimentation is about as follows.
To the west of Florida, in the Gulf of Mexico, we have a region which
from a remote time has been the seat of extensive accumulation of sedi-
ments. The Mexican Gulf in a more or less perfectly definite form is
one of the oldest topographic features of this continent. It is, as is
easily seen, the remainder of the continental trough which from an
early time has received a great share of detritus from the Appalachian
and Cordilleran fields. It has thei-efore been a region tending to sub-
sidence, through the well known influence of the weight produced by
sedimentary deposits on the surface of the earth. This subsidence has
naturally been attended by phenomena of counter-thrust elevation, the
characters of which I think are found in the curious uplifts of Yuca-
tan and Florida, which serve in part to bound this region of downward
movement. Besides the sediment contributed to the region of the
Mexican Gulf from the continental portions of North America, there has
been a considerable increase of such deposits from the island of Cuba.
This island, which probably came above the level of the sea in the
Mesozoic period, has evidentlv furnished a very large share of waste to
the neighboring sea-floors, as is shown by the extensive erosion indi-
cated in its highlands. A considerable portion of this detritus has
doubtless been accumulated in the region now occupied h\ the Straits
of Florida, and by the deep water between the Bahama Islands and the
greater land masses of Cuba and Hayti. I am disposed to consider the
extensive recent elevation on which rest the coral islands of the Baha-
mas as possibly due to the down-bearing of the crust caused by these
sediments.

The Caribbean Sea has likewise long been a seat of extensive sedimen-
tation. A number of great rivers draining from old mountain districts

MUSEUM OF COMPARATIVE ZOOLOGY. 141

subject to energetic erosion have contributed their detritus to this basin.
The line of the xVntilles appeal's to me to mark the phenomena of coun-
ter-thrust due to the accumulation of deposits off the coast of South
America, much as the peninsulas of Florida and Yucatan mark the
effects of sedimentation in that sea and in the Mexican Gulf. To the
down-thrust caused by sediments derived from the island of Cuba, and
deposited on the sea-floor to the northward, we may perhaps attribute
the sudden termination of the ITorida elevation on the south. The
general tendency to ccunter-thrust uplift produced by the growth of
strata in the Gulf of Mexico, and manifested in the peninsula of Flor-
ida, is here interrupted by the process of local sedimentation. It is
probable that, at the present time, the considerable energy with which
the Gulf Stream moves through the Strait of Florida may hinder the
process of deposition of sediment derived from the Cuban land mass ;
but, as I shall endeavor to show in the sequel of this paper, this limita-
tion of the Gulf Stream is probably a matter of very recent geologic
time.

Turning now again to the Cincinnati axis, let us note its relations to
the geography of the district at the time when it was formed, to see what
light it may throw upon the development of the curious elevations about
the Gulf of Mexico and the Caribbean Sea. The Cincinnati axis, as is
well known, is a singularly broad fold, which was developed on the floor
of the palseozoic sea at a distance of two hundred miles or more from the
then shore of the Appalachian Islands, and generally parallel with the
ancient land. At first, in the period of the Lower Trenton, this floor
appears to have been tolerably level. Before and during this period in
the history of the earth, a vast amount of detritus was borne from the
Appalachian land, and deposited on the sea-floor near its eastern shores.
Thus, along these old shores we had a vast thickness of sandstones of
the Okoee and Chilowee age, and above them a great thickness of rocks
belonging to the Knox group, which, though partly of organic origin, are
largely composed of inorganic waste from the old lands on the east.
There is no doubt that these last sediments were derived from the
Appalachian land, and they form an extremely massive system of sedi-
ments along the ancient shore. Following their onlaying, this portion
of the sea-floor which they occupied sank tf) a great depth, as is shown
by the peculiar character of the sediments and the organic forms in
the Trenton rocks of Eastern Tennessee. Apparently at this time the
Cincinnati anticlinal rose to near the surface of the waters, to a point
where it exposed the bottom of the sea to the action of currents siiffi-

142 BULLETIN OF THE

ciently powerful to break up and pack large shells in the manner
acconiplished by swift moving waters. It is possible that in part the
growth of the Cincinnati anticlinal took place at a later date, but the
greater part of its elevation was probably due to Silurian time. It
seems to me that it can best be explained in the manner above indi-
cated. It is a noticeable fact, that the Cincinnati'axis is most developed
along the line to the east of which this accumulation in the paleeozoic
seas derived from the bordering land was most extensive. Moreover,
the general form of the elevation is quite comparable to that of Florida,
though the axis of position is widely contrasted in the two cases.

If I am right in my supposition as to the oi'igiu of these curious
reliefs in the region about the Gulf of Mexico, if the anciently developed
axis of the Mississippi valley, the Floridian peninsula and that of Yuca-
tan, and the greater islands of the Caribbean, represent the results of
Cdunter-thrust arising from the imposition of sediments on the sea-floor,
it is evident that we have iii this part of tlie earth's sui'face a remark-
able exemplification of the effect of weight on the attitude of the crust.
It must be confessed that the matter is extremely speculative. I should
hesitate to give it note, were it not for the fact that the whole problem
as to the effect of weight of sediments is now much under discussion,
and it appears to me worth while to call attention to this district, where
there may be sometliing like critical evidence as to the verity of the
hypothesis.

The detailed topography of Florida is interesting from the light it
throws on two important jn'oblems, the growth of coral reefs, and recent
changes in the path of the Gulf Stream. As to the first of these ques-
tions I have little to add to the considerations which have been brought
forward by other writers. This little pertains to the distribution of the
livinir and the elevated reefs on the eastern shore between the southern
part of Key Biscaync and St. Augustine. Inside the living reef between
Key West and the southern part of Biscayne Bay the southern coast of
Florida is low. From the reports of others it appears likely that thei-e
are ridges, probably in their nature coral reefs, such as Long Key, at a
distance of a score or more miles from the swampy border of the land.
Xear the southern end of Biscayne Bay we find the first distinct reef
near the shore. This reef comes above tlie level of the sea about ten
miles north of Mangrove Point, just to the west of Old Rhodes Key.
It gradually rises, until at Cocoanut Grove, immediately west of Key
Biscayne, it has a heiglit of about twenty-two feet and a widtli of about
two miles. As the upitermost part of the deposit consists of character-

MUSEUM OF COMPARATIVE ZOOLOGY. 143

istic reef material, and is composed to a certain extent of corals, which
have evidently grown in place, this reef gives proof of a recent elevation
of the shore to the height of about tw§nty-five feet. There has been
a considerable loss by corrosive action, how great cannot be determined,
in the height of this reef. It is throughout honeycombed by subter-
ranean water passages, and the surface of the rock is much disrupted by
the overturning of trees in times of hurricanes, when the roots entangled
in the crevices of the rock break blocks away from their bedding.

The ]\Iiarai Reef, as I propose to term this interesting accumulation of
coral, extends northwardly with occasional interruptions for a great dis-
tance along the eastern shore of Florida. North of Dumbfoundling Bay
it appears gradually to lose its character as a true coral reef, and to take
ou the general nature of coquina. It appears to be the same elevation
that is traceable as far north as the northern part of Indian River, near
Titusville. I have only observed this ridge at certain points. North
of Lake Worth I have not observed any corals in the material. It ap-
pears that from Jupiter's Inlet to Titusville it is mainly composed of
raolluscan remains. It is therefore not certain that it is all of the same
age as the Miami Reef, but the fringe of beach material follows con-
tinuously on the line of that reef, fii-st in an almost meridional direction,
then turning at Lake Worth to the northwest. There can be no doubt
that the eastern shore of Florida from Miami to Titusville, and probably
all to the northward, has its position determined by the strong resist-
ance which this consolidated beach deposit has offered to the action
of the sea.

The steep escarpment which this barrier of old beach material pre-
sents, which is now elevated to a score of feet or more above its original
position, indicates that it long withstood the beating of the ocean waves.
The barrier of drifting sands now lying along the coast between this
escarpment and the open ocean has apparently been constructed in very
modern times, since the last elevation of the shore. A part of the cut-
ting which formed the escarpments of this reef evidently took place
during the process of elevation which brought the reef to its present
altitude. This is indicated by the fact that sea caves and other re-
entrants are formed in the cliffs at a considerable height above the
present plane of the sea.

The effect of this reef on the drainage of Florida is very great.
Although the rivers at many points have found their way across the
elevation, either by subterranean streams or through the low points of
the barrier, it serves to retain the land waters, and to bring into the

144 BULLETIN OF THE

condition of swamp a large part of the peninsula. The St. Johns River,
and the extensive swamps in which it heads, are in good part deter-
mined by the existence of this barrier. In a less complete way, the
waters of Lake Okochobee and of the Everglades to the south of it are
likewise prevented from finding a path to the sea by this natural wall.
Thus, at Cocoanut Grove, the waters of the Everglades at a distance
of only three miles from the shore in their time of lowest level lie
sixteen feet above high tide. In the rainy season they often rise to such
an altitude that they pour over the reef whenever it is less than twenty
feet in altitude. A sufficiently wide canal, having a depth of twenty
feet and a length of not over four miles, would drain the waters of the
Everglades into Biscayne Bay. The rivers which flow over this part
of the reef ccjnie down to the sea level over a series of rapids formed
upon the harder layers of the reef, and thus the full escape of the
Everglade waters is prevented. In the region more to the north, the
entanglement of the vegetation about the head-waters of the streams,
even where they have no rapids in their beds, likewise hinders the
escape of the marsh waters.

The superficial geology of the elevated reefs which constitute the keys,
as well as the section to the west of Biscayne Bay, aifords an interesting
subject of inquiry, which, owing to my brief sojourn in this region, was
but imperfectly followed. All the keys are evidently undergoing a
rapid corrosion by the action of the rain-water which falls upon their
surfaces, as well as a considerable marginal erosion by the mechanical
impact of the waves. On all the shores, it is also evident that the sea-
waters exercise a considerable solvent influence upon the limestone, but
this influence is much less manifest than in the case of the rain-water in
the interior portion of the key. In the strip immediately adjacent to
the shoi*e, where owing to the steepness of the slope the rain is quickly
shed from the surface, the solvent action of the fluvial waters is rela-
tively small ; but at a little distance back from the coast, where the
vegetation is more dense and the surface nearly level, the solutional
work is much more manifest, and is almost always distinctly traceable.
In time of heavy rain, the water gathers on the surface, being held there
in part by the dense mat of low growing vegetation. While so retained
on the surfoce, it doubtless obtains a considerable charge of carbonic
dioxide, which, as is well known, vastly increases the capacity of the
fluid for taking lime into solution. In most cases the water is conveyed
away through narrow crevices which penetrate the underlying rocks,
and discharge through small caverns communicating with the shore,

MUSEUM OF COMPARATIVE ZOOLOGY. 145

pouring foi'th theii' waters at about the level of mean tide. Wherever
these underground passages are formed (and they are very numerous
upon all the keys), the area about the neck of the orifices takes on the
shape of the sink-holes so well known in the great cavern districts of
Kentucky. Though these depressions in the Florida keys are never
so large nor so shapely as the characteristic forms of a similar nature
in older rocks, they are in many cases a conspicuous feature. Good
examples of such structures occur on Indian Key.

It is evident that the tendency of this corrosive action effected by
the rain-water is to lower the central portion of each island, for the
reason that in this interior field this water dwells longest upon the sur-
face, and becomes most charged with carbonic dioxide. If the process
were long enough continued, the effect would be to degrade the interior
region, leaving the marginal portion where the rain-waters are quickly
drained away in the form of a ring-like elevation. That this effect is
not more conspicuous in the case of the Florida keys is perhaps to bo
attributed to the fact that the sea cuts back the margin of the islands
so rapidly that time is not allowed for the development of such a topog-
raphy. If the shores of these reefs had been protected from the action
of the waves by the presence of the living coral, the islands would
probably exhibit a distinct internal depression.

The most interesting observations which I had an opportunity to
make concerning the erosion of the elevated reefs rest upon the sec-
tion from the western shore of Biscayne Bay into the Everglade district.
As before remarked, the western margin of this bay is formed by a
ridge of coralline material, which rises quickly from the surface of
the water to the height of about twenty-two feet above the surface,
attaining that elevation at an average distance of about half a mile
from the shore. This vcrsaut of the Miami reef, on account of the
steepness of its slope, provides a tolerably ready drainage for the rain-
water, which in no case rests for any time upon the surface. Owing to
the prevailing dryness of this well drained area, the low growing vege-
tation is small in amount, and the scattered pine blades afford little
woody material by the fall of its trunks, branches, and leaves. The
result is, that, owing to the absence of the decaying vegetable matter
and their brief sojourn on this sui'f\\ce, the rain-water obtains but little
carbonic dioxide.

When we pass from this relatively steep slope to tlie flat top of the
reef, where the watei's are less rapidly drained away, — where indeed
dui'ing the rainy season the surface is, as I am informed, very wet. —

VOL. XVI. — NO. 7. 10

146 BULLETIN OF THE

the growth of vegetation, and consequently the amount of humus upon
the soil, are considerably greater. The effect of the longer presence of
the rain-water, and the greater amount of the acids from decaying vege-
tation, is at once shown in the development of a great number of sink-
holes. Portions of the reef are so thickly set with these depressions,
that nearly all the rain-water appears to find its way by underground
channels to the sea, where we can note its emergence in great springs.
Descending from the summit of the reef towards the Everglades, I ob-
served that with each foot in height of descent the corrosive action of
the land water increased in amount. All portions of the reef which
were so situated as to be exposed to the waves of the lake which in the
rainy season covers this district, were very deeply corroded. Such sur-
faces often presented broad areas of rock so far eaten away by the
action of dissolving waters that, for the depth of a foot or more, the
remaining portion of the strata resembled the floor of a cavern covered
with stalagmitic materials. These decayed fragments of the rock often
assume curiously branched forms, and were so attenuated that the
pressure of the foot upon them would cause them to break down in
such a manner that it was impossible to walk over the surface. As
we approach the Everglades, the number of the sink-holes rapidly di-
minishes, probably for the reason that the elevation above the sea level
is not sufficient to impel the water to force a passage through the crev-
ices of the rock. Wherever the sink-holes occur, it is a noteworthy fact
tliat they frequently, if not generally, form the descending shaft which
gives exit to the waters in the central portion of some large coral. The
imbedded dome-shaped mass of the Meandrinas seem oftenest to be
chosen as the seat of these vertical shafts, which lead into the lower
lying caverns.

The quantity of material taken into solution by the swamp waters in
the rainy season, when the flooding of this area near the top of the
Miami Reef occurs, may be judged by tlie tliick coating of limy mud
which is deposited in the occasional closed sink-holes, from which the
waters have disappeared by desiccation. In these depressions the layer
of sediment, composed in large part of lime, often attains a thickness
of one fifth of an inch. As it necessarily i-eprcsents the amount of lime
in solution by the waters in a single season of rain, we may fairly take
it as a measure of the solutional work accomplished in one year. The
facts are not sufficient to permit a qnantitativc determination as to the
amount of this corrosion, but I am inclined to think that we are jus-
tified in assuming it to be a considerable fraction of an inch in each

MUSEUM OF COMPAKATIVE ZOOLOGY. 147

year. An inspection of the western face of the Miami Reef leaves in
my mind no doubt that the process of solution is rapidly extending
the general plane of the Everglades to the eastward. On the floor of
the more level and lower lying country which constitutes the eastern
margin of the Everglade district, we find a similar deposit of limy
matter, whicli has been laid down during the process of evaporation
of the swamp waters. At the distance of half a mile from the face of
the Miami Eeef, this layer was very much thinner than on the lower
portions of the reef itself. In the Everglade district the amount of
decavino- ve2:etable matter is srreat, and there can be no doubt that
running waters of this region become heavily charged with carbon di-
oxide, and are thus enabled to dissolve the limy matter with which
they come in contact. In the rainy season, as before remarked, these
waters rise to the height of from five to eight feet above their level
during the dry season, when I observed the district. As the waters ol
this swamp rise, they doubtless take a large quantity of the lime into
solution. After the rainy season passes, the water is drained away by
the numerous exits to the sea.

It is a noteworthy fact that Biscayne Bay, which receives through the
Miami and other streams discharging from the swamps of the Ever-
glades on its floor a vast amount of limestone mud. A portion of this
mud is composed of the remains of Foraminifera and other organisms ;
but microscopic examination of it shows that a large portion of the mass
does not exhibit evidence of having recently been in the organic con-
dition. It appears to be lime in the form whicli would be given it by
a precipitation from water. The quantity of this mud in the control
of the tidal and other currents which sweep through these embayed
waters is very remarkable. The volume of the material can best be
judged by the conditions exhibited by the deposits of limy matter at
the eastward end of the cliannel passing from Biscayne Bay to the sea,
at the point known as Caesar's Creek. The calcareous ooze moving out
from the Bay of Biscayne at this point is so large in amount, that it
forms a distinct delta, digitated at its seaward end in substantially the
same manner as the mouth of the Mississippi. It seems clear that this
great volume of mud comes from Biscayne Bay, and it would be difficult
to explain its origin by any action originating altogether in that basin.
I am disposed to believe that the surcharge of lime given to the sea by
the efiiueut water of the rivers which drain the Everglades leads to the
formation of a portion of this ooze by precipitation.

However it may be as to the origin of the limestone ooze, so plenti-

148 BULLETIN OF THE

fully formed along the coast where the Everglade waters discliarge, it
seems to me there can be no question as to the relatively rapid excava-
tion of the deposits of the Everglade district. There can be uo doubt
that this region shows in a very clear way how, under certain conditions,
the process of excavating the interior of a plateau of limestone deposits
may under favorable circumstances be cai'ried forward in an exceedingly
rapid manner. That a moi'e distinct reef has not been left around this
region of excavation is possibly due to the fact that this region has re-
cently been somewhat lowered above the height to which it formerly
attained. There is a good deal of evidence to the effect that the whole
peninsula of Florida has undergone a subsidence of ten or twenty feet
in altitude since the last period of elevation.

It is commonly supposed that the living coral reefs of Floiida cease
at Key Biscayne, and that no true reef exists to the northward of that
point. Although there is clearly no extensive development of reef de-
posits north of Cape Florida, my observations, though limited, are suf-
ficient to show that a distinct reef, essentially the continuation of the
main reef of Florida, that on which Fowey Eock Lighthouse stands, ex-
tends along the shore at least as far as Hillsborough River. In January,
1888, I was so unfortunate as to have my boat capsized on the edge of
the Gulf Stream, to the eastward of the mouth of tliat river. It was
necessary to bring the boat ashore bottom upward by swimming beside
it. Near the shore, there being a heavy sea on, we came upon a line
of breakers, beneath which the water was not more tlian six feet deep.
The effect of the surface of this reef on the bare feet of my party clearly
indicated that it was composed of firm coral rock. Subsequent inquiry
has shown that this reef is largely covered by living corals, including
many Gorgonias and actinoid corals, mostly of the common species of
Manacina, fragments of which are abundantly strewn along the beach
all the way from Cape Florida to Lake \Yorth. Between Hillsbor-
ough River and Jupiter Inlet the breakers show in times of storm the
continuation of a lower reef near the shore, and the fragments of
Manacina, often two feet in diameter, lying upon the beach, likewise
afford evidence of a living reef in this section. North of Jupiter
Inlet, my assistant, Mr. C. W. Coman, found fragments of Manacina
scattered along the shore for a distance of twelve miles. Beyond
this point, a careful search showed uo trace of stony corals. The
lessened development of the reef from Key Biscayne to Lake Worth
is doubtless in part to be explained by the fact that the Gulf Stream
depai-ts from the shore near Fowey Rocks. Its warm, life-giving waters

MUSEUM OF COMPARATIVE ZOOLOGY. 149

are only driveu in the form of a thin sheet against the shore in the
time of strong winds. At other times, when winds are blowing from
the westerly quadrant, the ciu-rent is separated from the shore by a
considerable interval of water which drifts from the north, and is much
cooler than the Gulf waters. Probably the decrease in the growth of
the reef north of Cape Florida is due also, in considerable measure, to
the southward movement of sands along the beach. A very large
amount of this sand is continually pouring around Cape Florida. The
history of this migrating detritus appears to be as follows. During the
glacial period, a very large amount of arenaceous material was con-
tributed to the sea in the region north of Cape Hatteras. The general
trend of the shore of this part of the continent is from the northeast to
the southwest, while the prevailing direction of the wind is from the east.
The result is, that so far as impelled by the waves, this sand works down
along the coast shelf to the southward. Wherever it comes upon the
beach and remains within control of the waves the southward movement
is quite rapid. Coming upon the coral reef, this sand tends to bury the
coral, and thus to limit its gi'owth. North of Cape Florida, the sea-fans,
or Gorgonias, which by their habit of growth ai"e in a measure protected
fi'om movements of detritus, are the principal representatives of the
polyps, the Manacinas occurring only as scattered clumps amid a
growth of the prevailing alcynoid polyps. North of Jupiter Inlet, the
Gulf Stream departs yet farther from the shore, and it is unlikely that
the temperatures are such as to favor the growth of a reef. The diffi-
culties incident to my shipwreck near Hillsborough River made it im-
possible for me to make more careful observations as to the condition
of this reef. I am indebted to iMr. C. W. Coman, who was formerly
keeper of the Lauderdale House of Eefuge of the Government Life-sav-
ing System, for a great part of the facts which are here given. He
has kindly followed my directions in observations on the extension of
this reef.

Imperfect as these observations are, they appear to me of interest
from two points of view. In the first place, they add nearly one third
to the known length of the living Florida Reef; and in the second place,
they show that while the reef may maintain itself for a certain distance
beyond the constant influence of the Gulf Stream, the polyps cannot
retain their full vitality when deprived of its current.

There is reason to believe that the marginal reef of Eastern Florida,
though it may now be extinct in the section north of Jupiter Inlet, has
recently been somewhat developed even as far north as IMosquito Inlet.

150 BULLETIN OF THE

In dredging for a canal now under construction on the eastern coast of
Florida, in the waters of Mosquito Inlet, near the point known as Oak
Hill, the engineers encountered a ridge of commingled shell and coral,
through which they were compelled to go for a quarter of a mile or
more in a north and south direction. The top of this ridge was some-
what below the level of the waters of the lagoon, and presumably below
the level of low tide in the neighboring sea. Some specimens of the
dredging shown me by Dr. John Westcott, the President of the canal
company, contained fragments of Manacina apparently the same as the
living species. It thus appears certain that at least one species of the
living reef-making coral has in recent times dwelt along the shore to
the north of Cape Canaveral.

The interior of the Floridian peninsula appears to be divisible into
three distinct districts. In the south, from the northern part of Lake
Okeechobee to Cape Sable, the surface is extremely level, formed proba-
bly in the main of organic waste accumulated behind the coral reefs,
upon which rests a thin and mterrupted coating of current borne sands
of inorganic origin. The only portion of this region which I have per-
sonally seen is the edge of the Everglades, about three miles west of
Cocoanut Grove. From the statements of Dr. Westcott and other ob-
servers as to the frequent occurrence of limy material in the Everglade
district, it seems to me most likely that the whole of this field above the
sea level is substantially composed of organic materials. The northern-
most part of the State, down into the base of the peninsula to a point
south of St. Augustine, probably consists of an older series of rocks,
mostly of Tertiary age, very uniformly covered by a deposit of detrital
sands brought to the region from tlie northward. Going southward
from the parallel of St. Augustine, we enter upon a region where the
surface is underlaid by the same sandy material as that found in the
northern part of the state, but the topography greatly changes its
character. In the northern section, the surface is in the main of the
gently undulating form belonging to the southern plain from Virginia
southwards. The deposits of sand are disposed so as to create gently
warped contours, the irregularities in height rarely exceeding ten or
fifteen feet within any one square mile. The form is that given by
slight marine currents where they act upon shifting sand. As we pro-
ceed southward, the irregularities of the surface become gradually more
and more accented, vuitil we gradually enter on a field known as the
Lake District, where the depressions without an outlet are so deep as
to enclose, not shallow morasses as they do in tiie more northern sec-

MUSEUM OF COMPARATIVE ZOOLOGY. 151

tions, but basins of water of such profundity that they have not become
closed by the swamp-building forces. The section of the lake district
extends from near Waldo to Lake Kissimmee, or perhaps yet farther
south. It has a length of at least two hundred miles and a width of
about eighty miles, though its limits in each direction are obscure ; the
area of open water basins gradually shades olf into the area of the
shallower depressions, now entirely occupied by swamps. By my rather
untrustworthy barometric observations, the highest point of the surface
in this lake district in the region about Apopka rises to near three hun-
dred feet above the sea. The number of basins contained in the area is
very great. If all those containing permanent open water were enumer-
ated, the total would probably amount to several thousand. In size
they vary, from the larger bodies, such as Lake Apopka, with a diameter
of ten miles or more, down to basins a few score feet across.

The most interesting feature in this district is the increase in the
measure of irregularity in the hills, as we rise above the sea level. On
either side, in passing from the shore, we cross a region which, though
occupied by sands, has, as before noted, a gently rolling aspect, remind-
ing one of the undulations of the sea when the waves of a great storm
have nearly sunk to rest. This is the condition of surface for a height
of from ten to thirty feet above the shore. For each fifty feet of as-
cent, careful observation shows a decided increase in the amount of the
irregularities, until they attain their maximum relief in the uppermost
portion of the country. So far as I have been able to ascertain, sub-
stantially all of these irregularities are moulded in recent sands. Only
occasionally are they affected by the form of the surface which existed
before the drifting sands came to this region. In certain cases the
underlying rocks are of a calcareous nature, and have been eroded by
subterranean waters. Where this has occurred, the pits formed in the
sands have occasional sink-holes in their bottoms. Some scores of such
openings w-ere seen in the course of four days' journeying between
Seville and Lakeland, in Polk County. It seems to me, however, that
these pits are not in any measure due to the causes which produced the
sink-holes. The great variety in their size, the lack of order in their
disposition on the surface, as well as the chance sections aflForded by
railways, all indicate that 'lie sink-holes are occasional concomitants of
these depressions, and in no sense their cause. My observations show,
moreover, that the sink-hole openings are often in eccentric positions in
relation to the pits, in some cases being actually above the lowest point
of the depression.

152 BULLETIN OF THE

The gradual increase in the measure of this irregularity of the super-
ficial sands, as we proceed from the shore towards the higher country,
clearly indicates that it is due to some cause the energy of which was
measured by the elevation of the surface above the sea. Moreover, the
fact that these ridges lie upon subjacent rocks of somewhat varied age
and composition, appears to indicate that they are not dependent on
any subterranean influences, such as the erosion or corrosion of the un-
derlying rocks.

"When I first came in sight of the lake district of Florida, the imme-
diate impression was that I had entered upon a kame district, a region
of pitted plain analogous to the kame belts along the New England
shore near the ancient frontal moraines, but on a far larger scale.
The surface has almost exactly the topography of the central part of
Nantucket, or the field of kame plain to the eastward of the Eliza-
beth Island moraine, where that lies at the base of Cape Cod between
Wood's Hole and Sandwich. It seems to me certain that anv geologist
familiar with this topography would, if taken blindfolded and ignorant
of his route to the lake district of Florida, at once come to the conclu-
sion that he was in a kame district of New England. Whatever were
the other circumstances under which our kames were formed, there can
be little doubt that they are the product of water flowing. I have else-
where argued that ordinary kames are in the main, if not altogether,
due to the tossing about of glacial waste under the influence of sti-ong
currents pouring from beneath the glacier into a water area where
mobile sediments were being laid down.

Although at first I endeavored to account for the peculiarities of the
surface in this lake district on the hypothesis that the warped surface
was produced by subterranean erosion, I was in the end forced to the
liypotliesis that these ridges represented the action of strong currents,
which served to move the sands, either in air or water. I then ad-
dressed myself to the task of determining which of these two agents of
transportation had given shape to the surface. I found myself quickly
driven from the hypothesis that these hills were due to the action of
the wind. In the first place, the gradual increase in the measure of
relief, as we go from the sea to the higher lands, is obviously against
the hypothesis of wind action. Next we note the fact, that on the ex-
isting coasts of Florida the dune building is slight in amount, though
the sands have in many places substantially the same character as those
which compose these hills. Furthermore, the shape of the hills is not
that presented by any of the extensive dune districts which I have

MUSEUM OF COMPARATIVE ZOOLOGY. 153

examined. Hills formed of blown sand are prevailingly sharper, and
are more in alignment, than are these ridges of Central Florida. In no
case known to me do they enclose such large basins as those of the
Florida district. Indeed, it is rarely the case that the deposits are suf-
ficiently dense to retain water. Last of all, the fact that these undu-
lations generally disappear as we go to the north, gradually passing into
the uniform southern plain, is against the hypothesis of their formation
by wind action. There are no atmospheric circumstances which w'ould
make this central part of Florida the seat of extensive duning, while
such action was absent from the northern part of the peninsula.

It appears to me that the most reasonable explanation of these tossed
sands is afforded by the supposition, which is apparently justified by
many facts, that the whole of Florida has recently been beneath the
level of the sea, and that during this period of submergence the Gulf
Stream swept across this portion of the peninsula, drifting the sands by
the action of its current into this complicated topography. The recent
submergence of the Floridian peninsula is indicated by the presence of
this large mass of detrital deposits of Pliocene or Post-Pliocene age. At
many points, as along the Indian River and elsewhere, these sands are
evidently overlying deposits containing altogether living species of ani-
mals. It is clear also that these sands have not been derived from the
erosion of sediments of an older date within the Florida district, but
have b'een imported from a distance. On this and other accounts we
may assume a recent submergence of the peninsula. Given this sub-
mergence without concomitant geographical changes which barred the
Gulf Stream from the Gulf of Mexico, we may suppose that the great
stream would have poured freely across the surface of the peninsula
within the region where we find this peculiar topography. Although
the Gulf Stream is confined at present within a narrow passage, where
it attains a speed of about four miles an hour, and possibly owes some-
thing of its rapidity of movement to the restriction of its exit, it would
doubtless, even with a larger opening, have a rate of movement suffi-
cient to exercise considerable energy on the bottom over which it flowed,
provided the floor was near the surface of the water.

Nantucket Shoals, near Cape Cod, and other similar regions of
shallow sea underlaid by sand where the ocean has moderate tidal cur-
rents, show us that a topography in a general way like this of the lake
district may be formed under water. The researches of the Coast
Survey have shown that rapid movements of submarine sand in this
district are taking place. As the currents in this district rarely have a

154 BULLETIN OF THE

speed greater than two or three miles au hour, it appears that the Gulf
Stream may have had sufficient velocity to bring about this arrangement
of the sands.

It is true that the an'angeraent of sands at and about Nantucket
Shoals is brought about by reciprocating currents caused by the succes-
sive movements of the tide, while the movements effected by such a
stream as that which flows from the Gulf of Mexico would be more con-
stant, or in one direction. Nevertheless, it is easy to see tliat A'ariations
in the wind cause even at present a considerable variation in the posi-
tion of the Gulf Stream off the coast of Florida. Strong winds trans-
verse to the surface of the current allect the flow of the superficial
waters, occasionally pressing them in against the shore, and again
causing a southward-setting current next the beach line. It is quite
possible that the variety of movements of the stream which may be
necessary to produce an irregular topogi-aphy have been brought about
by such variations in direction and force of the wind.

The gi'eatest difficulty I find in accounting for the topography of the
lake district is to explain the presence in the region of the large amount
of sand which has been shaped into these irregular ridges and hollows.
These sands have evidently come from the northward. It is not easy
to imagine the way in which they could have come into the control of
the Gulf Stream. One consideration, however, may aid us toward this
understanding. With the northern margin of the Gulf Stream crossing
Florida at the head of the present peninsula, its current would have
swept against the northern shore of the Gulf of Mexico. It thus might
have gained access to extensive deposits of sand, which had been accu-
mulated in the shallows along that shore. These sands it might have
borne onward until it brought them upon the Florida ridge. Tlie ex-
istence of a similar action is fdund in the movement of the sands against
the eastern coast since the last upheaval of the peninsula. The shore
from St. Augustine has received from the floor of the Atlantic an acces-
sion of detritus accumulated on the beach which separates the main
shore from the open sea. The amount of sandy matter appears equiva-
lent to more than one twentieth of that contained in the sand-hills of
the lake district. Within the limits of Florida this recently formed
sand barrier has a length of about four hundred miles, and an average
width of about three miles, and a probable average thickness of about
one hundred feet. It is therefore equivalent to a strip having a length
about twice as great as that of the lake district, and a rather greater
thickness of material. As the lake district averages not more than

MUSEUM OF COMPARATIVE ZOOLOGY. 155

sixty miles in width, we see that in a relatively short time, by the action
of ocean waves alone, this amount of detritus has been moved. If the
conditions had been such that the Gulf Stream had co-operated with
the wave action, it is not improbable that we should have had a mass
equivalent to all (if the sands of the interior of Florida removed from
the ocean floor, and brought to near the level of high tide.

If the surface of the lake district, and of Florida generally, was shaped
beneath ocean waters affected by strong currents, we are compelled
to believe that the elevation of the area above the sea level took place,
not in a gradual manner, but with extreme suddenness, at least for
all the height above the level of high tide. If the lake district had
emerged from the sea by a gradual upward movement, the ocean waves
would have produced a total change in the configuration of the surface ;
the incoherent sands of the hills would have been worn away, and moved
into the hollows. The aspect of the surface would be that of sand
beaches and plains extending towards the shores. If in the process of
elevation there had been pauses in the movement, during which the sea
even for a brief time, say for a few months, beat along a particular level,
then we should have had long beach lines with inclined aprons in their
fronts, such as now mark the elevated borders of our great lakes. So
far, I have not been able clearly to determine the existence of such fea-
tures in the Florida hilly country. At a few points indistinct signs of
such action are exhibited, but the wooded and swampy condition of the
country makes it difficult to trace these features. They will only be-
come apparent, if they exist at all, on careful study of the field. My
observations lead me to suppose that, if such features exist, they are
very imperfectly developed, and that, if we assume this surface to have
taken shape under water, we have likewise to assutne a tolerably rapid
elevation, which brought it above the level of the sea.

The problem in this field is substantially like that which we have in
the kame districts along the southern shore of New England. I have
elsewhere endeavored to show that these forms were clearly formed
beneath the surface of the sea, and came above it by a movement so
speedy that in the case of Nantucket the most delicate heaps of sand
were not disturbed by the ocean surges.

In the present state of my inquiries concerning the recent movements
on the eastern shore of North America, the evidence from the kame dis-
trict and that from the lake district of Florida appear alike to point to
the conclusion that a sudden elevation, or a series of such movements,
took place during the present geological period, a movement which must

156 BULLETIN OF THE

be deemed in its nature paroxysmal. It therefore becomes necessary to
make a very careful inquiry into the value of the evidence which this
case affords as to the speed with which these changes of level were
brought about. This problem should be considered by many field
geologists. It cannot adequately be solved by one student. On this
account, I venture to present the arguments in the case of the Florida
lake district, with the hope that they may be carefully reviewed by
other students.

MUSEUM OF COMPARATIVE ZOOLOGY. 157

NOTE.

The discovery by Professor Shaler of the northern extension of the great
Florida Reef beyond Key Biscayne, on the east shore of the southern extremity
of Florida, as far as Jupiter Inlet, throws a good deal of light on the probable
mode of formation of the Everglades. An examination of the map of South-
ern Florida in the Memoirs of the Museum of Comparative Zoology, Vol. VII.,
No. 1, Plate XXIII., or of the map (Plate VI.) in my Memoir on the Tortugas,
Memoirs of the American Academy, Vol. XI., 1SS3, or in the "Three Cruises
of the Blake," page 52, shows that in all probability the process of land-mak-
ing is simply more advanced in the Everglades than in the triangular stretch
of mud flats extending westward from the northern keys of Florida beyond
Cape Sable, and from that base in a general southwesterly direction to the Mar-
quesas. The presence of fossil reefs more or less concentric with the line of
keys induced Professor Agassiz,^ in his Report on the Florida Reefs, to look
upon the Everglades as holding to those reefs very much the same relation
which the mud ilats to the west of the main line of reefs hold to the latter.

Geologists 2 have, as a general rule, been opposed to this view, but they have
only examined the mainland nortli of the Everglades, and no geologist has as
yet penetrated farther into the Everglades than Professor Agassiz and his party.
A careful examination of the Everglades alone can determine whether their
fossil reefs are built upon a base consisting of the rocks which have been ex-
amined by Tuomey and others at Tampa Bay and Charlotte Harbor, or whether

1 Annual Keport of the Superintenaent of the Coast Survey, 1851. Report on
tlie Florida Reefs, by Louis Agassiz. Memoirs of the Museum of Comparative
Zoology, Vol. VII. No. 1, 1880, pp. 31, 57.

2 Report of Buckingham Smith ou the Drainage of the Everglades. Heilprin,
Trans. Wagner Free Institute of Science, Vol. I., May, 1887. Heilprin's explora-
tions were limited to the portions of the west coast of Florida included between
Cedar Keys and Punta Rassa, and did not touch the Everglade district or the great
Florida Reef. Likewise, the earlier researches of Conrad and Tuomey, and the .sub-
sequent ones of Smith, DalJ, and others, have all stopped short at the Everglades,
and the structure of the northern extremity of Florida has nothing whatever to do
with the formation of the coral reefs from Key Biscayne south. How far north this
reef structure extends is another point, and Shaler's interesting discovery goes far
towards giving us a clue to the mode of formation of the Everglades. That the
northern part of the peninsula of Florida is not made up of concentric coral reefs is
now very clearly demonstrated by geological and palaeontological evidence. What
is the southern extension of the formations which extend to the northern edge of
the Everglades, no one knows as yet.

158 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

the outcrops of these fossil reefs at the southern extremity of the Everglades
are only outlyers of the southern extension of these northern rocks. To the
damming up of the waters in the Everglades, and to the sudden outbursts of
gigantic masses of water charged with organic matter and lime, we may trace
the immense destruction of fishes which so frequently occurs on the shores of
the Florida keys and the waters surrounding them.

ALEXANDER AGASSIZ-

\

No. 8. — Contributions from the PctrograijJiical Laboratory of the
Harvard University Museum.

II.

On some Occurreiices of Ottrelite and Ilmenite Schist in Neio England.

By J. E. Wolff.

In the sei'ies of metamorphosed sediments which, in the many locali-
ties, represent nearly every geological horizon, a wide-spread type of rocks
are characterized by their fine grain, glistening micaceous aspect, and
perfection of cleavage, to which the names of phyllite, micaceous slate,
argillaceous mica schist, etc. have been applied. They represent
original fine-grained argillaceous sediments, in which the metamorphic
development of new minerals combined with the production of cleavage
has partially or totally changed the original character. These rocks
frequently attract attention by the presence of porphyritic, more or less
perfectly shaped crystals, scattered through the fine-grained micaceous
paste, which, unlike the analogous crystals of porphyritic eruptive rocks,
appear to have formed later than the " groundmass." Garnet, biotite,
andalusite (chiastolite), staurolite, albite, magnetite, ilmenite, and min-
erals of the ottrelite group, occur in this w^ay. In this paper some
notes are presented on schists or phyllites containing ottrelite or
ilmenite plates.

Ottrelite or chloritoid schists of Archean, Cambrian, Carboniferous,
and perhaps of later age, have been described from numerous localities
in Europe. The Cambrian phyllites of the Ardennes, among which the
classical ottrelite schists occur, have undergone a thorough chemical and
mici'oscopical investigation by M. Renard.-' One of these rocks is of par-
ticular interest in this connection, namely, the " Phyllade a ilmenite des
Forges de la Commune." The bluish gray rock contains numerous small
glittering metallic plates which can easily be mistaken for ottrelite ; in
the section they are transparent on the thin edge, with a brown color,

1 Hull. Mus. i;. Hist. Nat. Boly., Vol I pp. 212-'21'.). V(,l TI. pp 127-152,

\'.il. III. pp. 81, s.'>, •j:;0-2(!H

vOL. XVI. — NO. b

160 BULLETIN OF THE

and are bordered by a fringe of sericite. The optical examination com-
bined with chemical analysis led M. Renard to identify these plates
finally as ilmenite. In another rock (Phyllite ottrelitifere de Mon-
therme) they occur with ottrelite. These metallic plates had been
observed elsewhere by M. Eenard and others, but their true nature not
determined.

Minerals of the ottrelite family ("phyllite," chloritoid, masonite, etc.)
have been described from the rocks of New England by various miner-
alogists, and by T. Sterry Hunt from tlie palaeozoic schists of Canada.
The occurrence of this mineml in Maryland, in phyllite, has recently
been mentioned by G. H. Williams.-'

In the complex of gneisses, schists, and massive crystalline rocks
which cover the larger part of New England, there are certain areas of
partially altered sediments, the palaeozoic age of which has been estab-
lished by fossils or stratigraphic considerations. One of the most im-
portant of these is the strip forming the western edge of the Green
Mountains, which has been proved by the labors of Dana, Wing, Wal-
cott, and others to belong to the Cambrian and succeeding periods of
the Palaeozoic. These " Taconic rocks " consist of quartzites, crystal-
line limestone, phyllites of various kinds, and fine-grained gneisses, with
occasional conglomerates, especially near the base. That a large part
of the more highly crystalline rocks to the eastward, in Massachu-
setts at least, represent the same series still further metamorphosed,
appears to definitely result from the work of the United States
Geological Survey done under the direction of Professor Pumpelly, now
going to press.

Another important area of metamorphosed palaeozoic sediments oc-
curs in the eastern part of Rhode Island, on the shores of Narragan-
sett Bay, extending northward into Southerii Massachusetts ; it is of
Carboniferous age. The rocks are conglomerates, coal-beds, shales and
schists of various kinds, which like the Cambrian rocks of the Green
Mountains are intensely crumpled and metamorphosed.

There are two well-known localities for ottrelite in or near this
region : one that of the ^fasonite from Natic, R. I., described by
Jackson^ in garnetiferous mica schist which occurred as glacial boulders,
the other that of the ottrelite (Xewportite) from the vicinity of New-
port, R. I. ,Mr. T. N. Dale says of this occurrence, " Boulders and
pebbles of ottrelite schist abound about Newport, but I have failed to

1 Jolins Hopkins Univ. Circulars, September, 1889.
- Geology of Kliode Island, 1840, p. 47.

MUSEUM OF COMPARATIVE ZOOLOGY. 161

find any outcrop of it." ^ The carboniferous schists abound in little
black metallic plates which resemble ottrelite, so that the rock may
have been mistaken for ottrelite schist ; but since pebbles of this rock
ai'e associated with pebbles of the true ottrelite schist, there is little
reason to doubt that the latter occurs in place in the vicinity, and prob-
ably of Carboniferous age.^

Ottrelite Schist. — The ottrelite schist here described was collected by
Mr. Dale, occurring as pebbles on Easton's Beach, Newport. The rock
is a silvery-gi'ay, fine-grained mica schist, which has a well marked
schistosity (and cleavage), the plane of which bears no relation to the
distribution of the ottrelite. This mineral occurs in the well known
rliomboid or irregular plates, three or four millimeters in diameter,
with brilliant lustre and well marked cleavage surfaces. The latter are
pitted with little dull spots, which it is seen in the slide are grains of
quartz enclosed by the crystal.

Studied in the thin section, the rock is found to be composed of little
rounded grains of quartz, closely interlocking, when not separated by
the otlier constituents, and of minute scales of colorless mica with tlie
optical pi'operties of muscovite, which by their parallel ai-rangement
cause the schistosity of the rock. Certain wavy lines oblique to this
structure, which contain less mica and more quartz than the average,
may represent the original plane of deposition. A darker variety of the
rock contains occasional small plates of chlorite and bands of opaque
black substances, which are mixtures of graphite and titaniferous iron
ore (ilmenite'?) for the powdered rock gives a strong test for titanium
and also for graphite.

The ottrelite crystals and somewhat smaller black metallic plates are
seen to have no connection with either the plane of schistosity or
possible deposition plane. The former mineral occurs in plates of
irregular outline, appearing as lathe-shaped cross-sections, frequently
twinned several times, with composition parallel to the base, blue and
greenish plcochroism, and the other usual optical properties. They are
generally tilled with little grains of quartz of the same size and shape as
those composing the rock outside, which were evidently enclosed by the
crystal as it formed ; it is noticeable that the muscovite never accom-

1 A contribution to tlie Geology of Rhoile Island, Am. Journ. Sci., Vol. XXVII.
p. 222.

- Mr. Dale has found ottrelite seliist in place on Conanicut Island, opposite New-
port, but the rock has not been examined microscopically. Proceedings of Cana-
dian Institute, 1884-85, p. 21.

162 BULLETIN OF THE

parties the quartz in the ottrelite, but buts against the edge of the
crystal without altering its character or arrangement in proximity to it.
Sometimes the quartz grains fill the interior of the ottrelite in hour-glass
shape, but this form has no connection with twinning as in the case fig-
ured by Kosenbusch (Mik. Physiog., Vol. I. Plate XXII. Fig. 6), but is
evidently a case of crystal growth analogous to the forms so well known
in the augites of some eruptive rocks : a skeleton crystal of ottrelite
first formed, which did not enclose or else assimilated the quartz, while
a later growth, which filled out the double funnel-shaped cavity, was
able or obliged to enclose it. In the rock next to be described there
are skeleton crystals of ottrelite only partially filled up with the quartz-
bearing mineral.

The black plates in this rock are somewhat smaller than those of
ottrelite, with a jagged outline. They have sometimes a spindle-shaped
cruss-section, indicating then that they are discoid, but are generally
bounded by straight parallel lines; they are not transparent, but have
frequently a yellow leucoxene core, indicating titaniferous iron ore.
There is no doubt that they are ilmeiiite, as de+finnined by M. Renard
in the similar rocks of the Ardennes.-^

These ilmenite plates are generally bordered on both siues by a thin
sheet of chlorite, the base of which is parallel to the ilmenite. (The
similar ilmenite plates described by M. Renard are bordered by sericite.)
The plates are often entirely enclosed in the ottrelite crystals, some-
times one half in, the other half projecting out. The chlorite coating
disappears when they are found in the ottrelite, but they are then some-
times bordered liy a zone of ottrelite free from quartz inclusions, unlike
the rest of the crystal, of the same size and shape as the chlorite, sug-
gesting: that the latter was absorbed into the ottrelite when the crystalli-
zation took place. Small grains of titanite mixed with black ore are
scattered through the rock, and there are occasional prisms of tour-
maline.

Ottrelite Grammcke. -^This interesting rock was found by ]\Ir. Dale in
a glacial boulder at " Paradise," Newport, R. I.

The rock contains fragments of blue and white quartz, enclosed in a
dark gray micaceous cement, spangled with small plates of ottrelite.

The slides show that the rock has undergone intense dynarao-
metamorphic action ; the large fragments of clastic quartz in polarized
light exhibit all stages of change from mere straining to breaking and

1 He mentions the occurrence of these forms in PJiode Island ottrelite schist.

MUSEUM OF COMPARATIVE ZOOLOGY. 163

crushing, aud at the edges have yielded small broken quartz, -which is
mingled with the muscovite of the cement. The latter is made up of
fragments of detrital quartz, quartz derived from the crushing of the
large fragments, and perhaps some quartz formed chemically in situ,
with muscovite filling the interstices, and even filling the cracks made
in the large grains, and therefore evidently of metamorphic origin.
There are also larger fragments of quartzite, and rounded .aggregates
of quartz and muscovite, which represent decomposed clastic feldspar
grains.

The ottrelite occurs in this cement in plates of irregular shape, often
moulding itself around or enclosing the grains of quartz. It has all the
optical pi'operties of the ottrelite described above, and also encloses the
quartz grains of the cement, but not the muscovite, and very rarely
exhibits the least bending or straining ; hence it must have formed after
the crushing forces had ceased to act. There are sometimes skeleton
crystals of ottrelite, the hollow having the shape of an hour-glass, and
transitions to crystals in which the hollow is filled up by ottrelite
enclosing quartz. The cement also contains the black metallic plates,
small and imperfect, which are sometimes enclosed in the ottrelite.

We may conclude from the microscopic study of these rocks that the
ottrelite was the last mineral to form in them : it encloses the grains of
quartz of the cement, both when they are easily recognized as clastic
in the grauwacke and when of doubtful origin in the fine-grained schist.
The muscovite, which is evidently a metamorphic mineral in both
rocks, formed before the ottrelite, although not enclosed in it, for in
position and arrangement it is not affected by the latter, and it seems
necessary' to suppose a chemical solutioti of the muscovite -which filled
the space between the quartz grains at the time the ottrelite came to
fill that space. The ilmenite-chlorite plates also formed before the
ottrelite, since they are enclosed in it.

In the grauwacke the muscovite is found penetrating the crushed
pebbles of quartz along the cracks, and even penetrating into the sub-
stance of the quartz a minute distance where there is no visible break,
indicating a marked mobility for the solution from which this mineral
formed. The ottrelite, on the other hand, forms in comparatively large
unbroken areas enclosing the other minerals, somewhat analogous to a
concretionary formation. Such an ottrelite grauwacke illustrates anew
the position of ottrelite in tlie scale of metamorphism, occurring, as it
does often, in or associated with rocks that retain at least a part of
their original chai'acters. Its late formation in the rock, posterior to

164 BULLETIN OF THE

quartz, ilmenite, muscovite, etc., may indicate a higher degree of meta-
morphism than those minerals alone would do, or the presence of some
special geological or chemical conditions, to which we have as yet
no clue.

Graphite Schist with Ilmenite Plates. — This rock occurs as a boulder
on Miantonomah Hill, Newport, R. I., and contains plant impressions
(Dale).

It is a soft black graphite schist, containing irregular metallic plates
resembling ottrelite, which are two or three millimeters long and
0.12 mm. thick. These plates can easily be split off with a knife,
leaving a dull film of chlorite below them. They are imperceptibly
magnetic, are attacked with great difficulty by boiling hydrocliloric
acid, and the yellow solution gives a strong titanium test with tin-foil ;
they are therefore ilmenite. In the slides the rock is composed of small
grains of quartz, flakes of muscovite and chlorite, and specks of graphite
and iron ore (probably ilmenite). The large ilmenite plates have
frequently a spindle-shaped cross-section (i. e. discoid plates), and have
a kernel of leucoxene (titanite). Some are bordered on each side by a
thin plate of chlorite, some by brilliantly polarizing muscovite. The
rock is evidently of Carboniferous age.

Occurrences of minerals of the ottrelite group in the region of schists
and gneisses of Central and Western Massachusetts are mentioned in
mineralogies, but the writer has found no microscopical descriptions
of the rocks. A part of the so called ottrelite schists, such as the
" spangled mica-slate " of Hitchcock (Geology of Massachusetts) are
probably ilmenite schists.

In the Western or Green Mountain region, ottrelite and ilmenite
plates occur in schists or phyllites investigated by the writer for the
U. S. Geological Survey, both in specimens collected by Mr. T. N. Dale
from the Western Cambrian (or younger) rocks (Taconic region), and
also from the Cambrian series of Hoosac Mountain in the axis of the
Green Mountains, full descriptions of which will appear in the forth-
coming memoir.

The ottrelite schist of Hoosac Mountain occurs in several places in
the albite-phyllite series which overlies the basal Cambrian conglom-
erate. The ottrelite rock is a silvery greenish schist, containing crys-
tals of red garnet and small prisms of tourmaline, and spotted with
plates of ottrelite. In the slides the rock is composed of muscovite

MUSEUM OF COMPARATIVE ZOOLOGY. 165

(sericite) in the usual interwoven aggregates, irregular plates of chlo-
rite, grains of quartz, occasional crystals of albite and the ottrelite, in
small irregular plates with the usual pleochroism, etc., which sometimes
appear spindle-shaped in cross-section (discoid). Small, irregular black
metallic plates also occur in the rock.

The Taconic region of Greylock Mountain, the highest summit in
Massachusetts, lies immediately west of the Hoosac series, extending
west in turn to the Taconic range, which forms the boundary between
New York and Massachusetts. The rocks of this area are in large
part phyllites of many varieties and colors, often dotted with crystals
of albite like the similar rock of Hoosac Mountain, containing garnets,
tourmaline, etc. The black metallic plates are wide-spread in these
phyllites, exhibit the same properties, such as very feeble magnetism,
diiUcult solubility in hydrochloric acid, presence of titanium, etc., that
those from the Ehode Island graphite schists do ; they are therefore
ilmenite.

In the slides these rocks are composed of sericite, generally intimately
interwoven with chlorite, and small grains of quartz. Masses of black
ore, prisms of rutile, etc., are abundant. In some varieties the albitic
feldspar becomes an essential constituent. Tlie ilmenite plates are
commonly sandwiched between two plates of dark green chlorite, exactly
as in the Rhode Island rocks. In many of these rocks microscopic
plates of ottrelite, spindle-shaped in cross-section, exist enclosed in the
meshes of the mica.

April, 1890.

No. 9. — Contributions from the Pctrographical Laboratory of the
Harvard University Museum.

III.

On Keratophyre from Marhlehead Neck, Massachusetts.
By John H. Sears.

This interesting rock formation was first noticed by Prof. W. 0.
Crosby in the American Naturalist (Vol. XI. No. 10, 1877, p. 585),
where he says : " Near the middle of the southwest side of the harbor,
visible only at low tide, is a hard, whitish, fine-grained sandstone or
arenaceous slate. It overlies uncouformably the banded petrosilex
found on the shore." In the '* Occasional Papers of the Boston Society
of Natural History, III. Contributions to the Geology of Eastern Massa-
chusetts," Professor Crosby says again of Marblehead Neck (p. 263) :
" It is not generally known that this rocky peninsula, which may be re-
garded as lying on the extreme outskirts of the Boston Basin, includes
beds probably referable to the same horizon as the slate and conglomer-
ate on the south and west. Briefly stated the facts are as follows :
Near the middle of the northwest shore of the Neck, visible only at low
tide, is a hard, whitish, fine-grained sandstone or arenaceous slate ; it is
evidently largely feldspathic and turns yellowish on weathering. Por-
phyritically interspersed through the rock are clear, almost transparent,
rhomboidal crystals, from one eighth to one fourth of an inch long ;
these have been examined by Miss Hattie A. Walker and proved to be
orthoclase."

The next notice of this rock i? in the Proceedings of the Boston
Society of Natural History (Vol. XXI. Part 3, p. 288), "On the Tra-
chyte of Marblehead Neck," by Dr. M. E. Wadsworth, in which he
says, " Near Boden's Point, on the northwest shore of Marblehead
Neck, there is to be seen, exposed between high and low tide, the re-
mains of a trachytic overflow." On page 290, Dr. Wadsworth says :
" One of the feldspars, porphyritically enclosed in the groundmass, was
obtained in the section. This is clear, glassy, and contains only a slight

VOL. XVI. —NO. 9.

168 BULLET^' OF THE

amount of the ground mass and a few full fluid cavities. It is a simple
crystal of sanidin."

Dr. Wadsworth's field-work upon this formation was very thorough,
and but little is required in addition to his clear description of it. A
few notes, however, taken from his description and the observations of
the writer, may be of interest. This formation, now determined to be
keratophyre, can be seen at low tide near the residence of Mrs. Harding
on Boden's Point, Marblehead Neck. It appears as the much eroded
remains of a surface flow, and extends two hundred yards in a north-
easterly direction, with a width of sixty feet at the lowest point of ob-
servation. There are smaller masses of this rock three hundred yards
from this point in the same strike (noi'theast), which are exposed only
at extremely low tides. About five hundred yards south of Boden's
Point, near Flying Point, the eruptive granite cuts the metamorphic
slate of the Boston Basin series, and near this point also the granite is
cut by dikes of quartz-porphyry (felsite). Near the keratophyre, and
dipping under it, is a banded felsite. Both the granite and the felsite
are cut by diabase dikes. The felsite tends to the northeast, and forms
the larger portion of the bed rock of the Neck. The banding of this fel-
site dips towards the harbor nearly north, and lying upon it is the kerato-
phyre. Between the lowest points of observation and the banded felsite,
a conglomerate of varying thickness composed of fine felsitic debris,
holding rounded and angular fragments of the felsite, is found in several
places enclosed in the keratophyre. In some places the keratophyre
rests directly upon the felsite, while in others the conglomerate inter-
venes between them. The line of contact between the keratophyre
and the felsite debris is well marked ; specimens of the keratophyre
detached at this point show a basal surface very rough and pitted where
it conforms to the irregularities of the conglomerate. The keratophyre,
being exposed to the sun, rain, "^nd the action of the frost and the
ocean waves, is much decomposed ^a the surface ; but the least altered
specimens obtained are of a brownish or bluish gray color, having a
conchoidal fracture and a compact groundmass, holding, occasionally,
large glassy crystals of anorthoclase, some of which are one fourth of an
inch in length, and, rarely, plates of biotite. The groundmass in thin
section under the microscope is shown to be filled with lath-shaped feld-
spar crystals, which are somewhat decomposed. The base is an earthy
kaolinized mass, with irregular masses of quartz and earthy limonite.

Dr. Wadsworth described the rock from microscopical study as con-
sisting of a groundmass composed of ledge-formed crystals of feldspar.

MUSEUM OF COMPARATIVE ZOOLOGY. 1G9

cither in single crystals or simple twins, which had the optical prop-
erties of orthoclase (although some might be triclinic) and enclosed
between them varying amounts of a decomposed base, and of quartz
which he regarded as secondary. The porphyritic crystals were deter-
mined as orthoclase (sanidin). The rock varied considerably in fresh-
ness in the several specimens. This rock, occurring thus as a surface
How, was called " Trachyte " under the classification used by Dr.
Wadsworth, corresponding in this case to the " Quartzless Porphyry "
of Rosen busch.

During the season of 1889-90 eight sections of the keratophyre and
several sections of the detached anorthoclase crystals have been pre-
pared for microscopic study. Numerous crystals from the groundmass
have also been detached for the purpose of obtaining the specific gravity
and chemical analysis. IJiotite mica is often found in hand specimens,
and occasionally augite, although the latter has not as yet been detected
in any of the sections cut.

In the light of our present knowledge and with further investigation
it is possible to supplement Dr. Wadsworth's accurate descriptions, and
to determine the feldspar pheuocrysts as anorthoclase, and the rock as a
keratophijre. The pheuocrysts occur as crystals elongated parallel to
a, with a sqiiai-e cross-section owing to the presence of the base and
brachypinacoid ; in addition to the two cleavages there is a rough trans-
verse fissuring. The crystals are quite glassy when fresh. In the rock
slides, in polarized light, the different feldspar sections show marked
optical peculiarities ; there is often a very fine single, or double (micro-
cline) twinning ; sometimes the whole of one section of the mineral
consists of irregular areas not extinguishing in common, which resemble
the phenomena produced by mechanical causes ; these areas contain
very fine lines crossing each other at various angles in the difierent
areas ; in other cases there is a very fine zonal structure. Sections pre-
pared parallel to the base show this tine irregular double twinning, and
give an extinction l'^ to 2° tjblique to the line of the second cleavage
(oo P co), and sections parallel to the latter cleavage give an extinction
about 9*^ oblique to the line of the first cleavage, with an obtuse posi-
tive bisectrix about perpendicular t(i the face, the acute bisectrix a
making the angle of 9° with the basal cleavage. These sections also
show sometimes a very fine indistinct microperthite striation. The
angle between the two cleavages was determined in the roilecting goni-
ometer as approximately 89° 42', about thpt of microline. The specific

170

BULLETIN OF THE

gravity of fragments, determined by Westplial balance and Thoulet
solution, was between 2.570 and 2.572.

The following analyses of the feldspar (I.) and the rock (II.) were
made in the laboratory of the U. S. Geological Survey at Washington
by Dr. Thomas Chatard.

H»0 at 110° C.
H.,(J at red heat
Sibj . . .
TiO/ . .
P2O5 . .
Al.O^ . .

FeO .

MnO . .

CaO . . .

MgO . .

K2O . .
Nap

I.

Feldspar.

.0-1

.37

G5.GG

20.05

traces

traces

.13

.07

18

G.'JS

G.5G

IIL
Gmelin, No.

SiOo G5.90

10.4G

41

28

Al/)3

CaO

MgO

K,0

. . . . , . G.55

Ka.O G.U

1 '^

H^O

Specific gravity 2.587.

II.

Keratophyre.
.91

1.28

70.23

.03?

.OG

15.00

1.99

.24
.33

.38
4.99
4.98

IV.

No. 2.
G5.19
19.99

.63

.48

•7.03

7.08

.34

It is evident from the analysis and optical properties that this is a
triclinic soda-potash feldspar of remarkable purity, and veiy evenly
balanced percentages of Na and K, belonging to the anorthoclase gi'oup
of Itosenbusch. For comparison, analyses (III. and IV.) by Gmelin
are a])pended uf anorthoclase from the augite syenite of Norway (Brog-

* The TiOj wai not very ]iurr, aud its jircauiict, is not aLsolutuly ct-rtaiu.

MUSEUM OF COMPARATIVE ZOOLOGY. 171

ger, " Die Sil. Etagen 2 uiid 3," etc., p. 2G1). In the rock as a whole
the same even balance between ISTa and K is noticeable, and the insig-
nificant quantity of lime and magnesia. Allowing for the free quartz,
base, and decomposition products as causing a relative increase of silica
and iron and decrease of the alumina and alkalies, it is evident that the
feldspars of the gi'oundmass are closely allied chemically to the porphy-
ritic crystals, and ai'e probably also anorthoclasc. The rock is therefore
a very pure type of keratophyre.

The microscopical structure of the sections made are as follows : —

No. 21. Keratophyre witli anorthoclase crystal cut obliquely to an optic
axis. Groundniass made up of minute twinned lath-sbaped crystals of feld-
spar, somewhat kaolinized, some quartz, and an earthy fibrous kaolinized base.
In the centre of the porpbyritic feldspar crystal are numerous microliths and
a few ferritic masses, similar to and probably composed of the base, which
penetrates the edges of the crystal.

No. 21 A. Keratophyre and an aggregate of the porpbyritic crystals.
Groundniass nearly as in No. 21. One of the pheuocrysts shows twinning
after the Carlsbad type.

No. 21 B. Keratophyre with one porphyritically enclosed crystal. The
crystal is cut nearly parallel to the second cleavage, and gives an almost per-
fect interference figure of the positive bisectrix. The basal cleavage is well
developed, and the stria), or fine twinning, are well marked in polarized light
The groundmass is more generally composed of the minute lath-shaped feld-
spar crystals, some of which are clearly twinned anorthoclase of the same form
as the larger crystals. There are also small patches of quartz.

No. 21 C. Keratophyre with one large porpbyritic feldspar crystal cut
obliquely to the brachydiagonal, which in polarized light shows a niicroper-
thitic intergrowth and a very perfect example of fine and interruiited twinning.
Through the crystal are several fluid cavities and a few microliths of a reddish
color. The groundniass is more kaolinized, and the minute lath-shaped crys-
tals are less distinct. Small irregular masses of quartz and considerable
limonite and earthy matter pervade this section.

No. 21 D. Keratophyre section cut across a joint plane which is filled with
vein quartz ; numerous irregular patches of quartz are scattered all through
the section. On^mass is a basal section of original (?) quartz ; it gives the
uniaxial cross, and is shown to be positive by the mica plate. Some scales,
of biotite and numerous small grains of magnetite are seen in the groundmass,
which is composed of a fibrous feebly polarizing kaolinized mass of the decom-
posed minute lath-shaped feldspar crystals. One of the enclosed pheuocrysts
cut nearly parallel to the base shows nuniejous inicrolithic inclusions, and sev-
eral fluid cavities in which the Ijulible muvemeut i; seen The outer edge is
deeply penetrated l)y the LCioundiiuiss

172 BULLETIN OF THE MUSEUM OF COMPAKATIVE ZOOLOGY.

The occurrence of this keratophyre as a surface flow in close prox-
imity to the large intrusive masses of elfeolite- sodalite- zircou-syenite
of Salem Neck and the islands in Salem Harbor, and the augite- zircon-
syenite of Marblehead and the Beverly shore, is interesting, as showing
the various forms assumed here by the alkaline magmas under different
geological conditions or at diflerent periods.

June, 1S90.

;N'o. 10. — MdamorpMsm of Clastic Feldspar in Conglomerate
Schist. By J. E. Wolff.

In the complex of metamorphic rocks which occupy the region of the
Green Mountains in Western New England, two rocks are of importance
from their wide distribution in Vermont and Massachusetts, and their
striking appearance. These are the metamorphic conglomerate and the
albite schist.

Both rocks occur in typical development in Hoosac Mountain in
Western Massachusetts, exposed to perfection both in place and in the
great masses of fresh rock removed in the construction of the Hoosac
Tunnel. Here the conglomerate, representing the base of the Cam-
brian, rests on the underlying Archaean gneiss, with peculiar relations to
the latter, both as to mineralogical character and structure, whose im-
portance, as bearing on the origin of certain crystalline schists, has
recently been stated by Professor Pumpelly.^

This conglomerate attains a thickness of six to seven hundred feet,
and is then overlaid conformably by the second rock, the albite schist,
possessing a great but as yet undetermined thickness.

Detailed geological and petrographical descriptions of these rocks will
appear elsewhere, and are not presented here ; but the truly detrital
character of the conglomerate should be stated, containing as it does
true pebbles of quartz, feldspar, gneiss, or granite in a thoroughly crys-
talline matrix, and also the necessary detrital origin of the conformable
albite schist, now entirely crystalline. The latter rock is not confined to
the axis of the Green Mountains, but occurs abundantly in the fossiliferous
" Taconic " region immediately west, associated with limestones, quartz-
ites, and finer-grained schists or pnyllites. The albite occurs in irregu-
lar porphyritic grains of variable size, dotting the rock with its glassy
crystals, often twinned in two simple halves according to the albite law.
In thin sections it is strikingly clear and fresh, containing in the differ-
ent specimens inclusions of muscovite, biotite, or chlorite, grains of
quartz, grains or crystals of magnetite, epidote, rutile, etc., which are so

1 The Relation of Secular Hock-Disintegration to certain Transitional Crystalline
Schists. Bull. Geol. Soc. Am., Vol. II. pp. 209-224.

VOL. XVI. — NO. 10.

174 BULLETIN OF THE

similar iu character aud arrangement to the same minerals outside that we
have to suppose they were all formed together where we now find them.
This is the well known character of these pseudo-porphyritic feldspars
in phyllite-gueiss or feldspar-phyllite in general, aud in the present case
their character as albite has been proved by chemical and optical analy-
sis. That they should have some connection with the grains of true
detrital feldspar found in the less metamorphosed phyllites of this region
is easy to see, but difficult to prove.

It should be mentioned that the Hoosac conglomerate in its fine
schistose portions contains little glassy feldspars with albitic characters.

This conglomerate occurs at intervals for more than a hundred
miles to the north, well into the centre of Vermont, lying between the
Cambrian quartzite of the Taconic region and the underlying gneisses
on the east.

One of the most striking localities is on Bear Mountain in Walling-
ford, a few miles southeast of Rutland, Vermont, which was described
and figured in the " Geology of Vermont," Volume T. pages 32 and 34.
The conglomerate character is exhibited in wonderful perfection, the
pebbles varying from almost boulder size to that of the grains of the
finest sandstone. They are principally quartz (often blue) and feld-
spar ; the latter are generally of a dull white or reddish color, due
to kaolinization. One of the large feldspars, tested by the ordinary
methods, proved to be microcline. In the Hoosac conglomei'ate this
feldspar also occurs in the pebbles and in the underlying coarse gneiss
from which they were derived. In the finer grained layers of the Bear
Mountain rock the cement predominates, so that the rock is a crystal-
line schist containing little pea-sized angular grains of quartz and feld-
spar. This cement rock here and at other localities glitters with small
glassy feldspars, as does the albite schist of Hoosac Tunnel.

In the finer grained portions of this conglomerate schist the detrital
feldspars show interesting changes, whicli may throw some light on the
formation of albite schists. The descriptions which follow are based on
material from one hand specimen, so tli;it it will be understood that the
changes observed are ])rogressive only by comparison, feldspars side by
side having been aff'ected to a varying degree l)y the forces of metamor-
phism, obliterating partially or else completely the original characters.

In the hand specimen the rock is a silvery green schist, the color
due to large scales of a light green muscovite, which are arranged in
pai'allel layers, and enclose l)etwecn the meshes little elongated masses
of quartz, and little glassy feldspars, which are barely visible with the

MUSEUM OF COMPARATIVE ZOOLOGY. 175

leas. Grains of dull reddish or white feldspar and of blue quartz are
evenly distributed thx'ough the rock. The quartz grains pi'edominate,
having their longer axes parallel to the schistosity, and often forming
the cores of the lenticular bands of quartz. These quartz grains aver-
age the size of small shot, while the feldspars vazy from the size of a
large pea to that of the smallest shot.

The feldspars are often distinctly angular in shape. Even with the
unaided eye it is seen that they have a rim of clear glassy feldspar sur-
rounding the inner dull red or white core, the latter sometimes preserv-
ing its boundary, made by straight lines or sharp angles, the former
having a ragged edge which merges imperceptibly into the cement. In
other cases the clear rim, instead of representing only a small propor-
tion of the diameter of the feldspar, occupies half or more of the whole
grain, and little tongues of the clear substance then ramify into and
across the red core. The shape of these feldspars and their occurrence
with the larger feldspar pebbles of the coarse conglomerate make their
detrital character evident. The same is true of the grains of blue
quartz.

Small prisms of tourmaline, octahedra of magnetite, and rare grains
of apatite, occur as accessories.

In the slides these elements are easily recognized. The mica is
eutii'ely a greenish yellow muscovito in thick plates, which in the
thicker slides exhibits a pleochroism varying from yellowish green to
colorless, is free from inclusions excepting a rare grain of magnetite
and of titanite (?), and has the large axial angle of muscovite. The clas-
tic grains of quartz are i-ecognized by their large size, and by the fact
that in polarized light they are seen to have been strained, this effect
increasing until some grains pass into a peripheral cataclastic mosaic of
quartz grains produced from the original grain by crushing, between
which flakes of muscovite make their appearance. Quartz also occurs
abundantly hi the meshes of the mica, in aggregates of interlocking
grains, which sometimes enclose muscovite, and hence must have formed
by chemical action in situ.

Here and there little areas of clear feldspar occur, evidently the little
glassy crystals of the hand specimen. They often have a lenticular form,
flattened parallel to the schistosity of the rock, and the inclusions which
they commonly contain have their longer axes parallel to this direction,
and are distributed in rough parallelism to the same direction. The
inclusions consist of little flakes of muscovite and round or elongated
grains of quartz and of magnetite. One of these feldspars is represented

176 BULLETIN OF THE

in Figure 8 with black inclusions of magnetite and clear ones of quartz
and muscovite, and having the typical irregular flattened shape. The
feldspars of this type have a low even polarization tint, yellow of the
lower order even in thicker slides, where other feldspars show red and
blue. Often without twinning, or else twinned in single halves, less fre-
quently multiple twinned with few lamellae, they have the characters
of the albite of the schists. Some of these areas of glassy feldspar with
the same generp,l characters show the double twinning of microcline.

A variety of these feldspars is seen in Figure 7, which, having the
same flattened irregular shape, with inclusions of muscovite, quartz, etc.,
in parallel arrangement, the same even low polarization and fresh glassy
look, are clouded in the centre by an aggregate of dots which form a
central area with vague boundary, but having the same polarization
color as the outer area. With the high power these are seen to be fluid
inclusions with moving bubble, little flakes of colorless mica or kaolin, and
black opascite masses.

We now come to the clastic feldspars, which are generally larger than
those just described. Figures 1 and 2 represent typical cases.

In Figure 1 the enlargement is fifty diameters. The feldspar polar-
izes in one low color and has a homogeneous extinction. One cleavage
is well developed in the slide parallel to the short edge ; the other, indis-
tinctly parallel to the right hand edge. From the obtuse angle made
by the two cleavages, it is evident that the section is oblique to the zone
of either cleavage. The outer shape of the grain, as well as that of the
inner cloudy portion, is evidently determined by the two cleavage lines,
which is some evidence of clastic character when compared with the irreg-
ular shape of the albitic feldspars. The cloudy look of the central por-
tion is owing to streaks of opaque kaolinized (f) feldspar containing fluid
cavities, specks of black opascite, and stained by yellow limonitic pro-
ducts, which lie in the clear feldspar arranged parallel to the second
cleavage. These are evidently areas of decomposition. As seen in the
figure, these bands die out in the clear feldspar rim. The whole feld-
spar shows in polarized light indistinct multiple twinning parallel to the
second cleavage, which runs almost to the outer boundary of the clear
rim. There occur also, scattered through the central core, little bril-
liantly polarizing flakes of muscovite, in part arranged parallel to the
first cleavage. These become less abundant, but of larger size, in the
clear rim, where they are apt to arrange themselves parallel to the outer
boundary ; near the boundary they become still larger, and sometimes
connect with the mica outside. The outside boundary of the whole feld-

MUSEUM OF COMPARATIVE ZOOLOGY. 177

spar, as can be seen in the figure, is irregular, the quartz grains and
muscovite of tiie cement encroaching upon it. This is noticeable in the
upper left hand corner, where a flat grain, of quartz intrudes parallel to
the cleavage. It cannot be determined from this oblique section to what
variety of feldspar tlie grain belongs. The twinning excludes orthoclase ;
it may be microcline.

Figure 2 represents a section of another red clastic feldspar. This
also polarizes as one crystal, and, as the section happens to be normal to
an optic axis, it is easy to see by the position of the axial bar that
the cloudy centre and clear rim are in like crystallographic position.
No twinning is visible. Tlie cleavages are not visible in the figui-e,
although present in the section. The principal cleavage is parallel to
the long dimension, and the second cleavage parallel to the narrow side
of the grain; here again the outlines are determined by cleavage cracks,
or by the corresponding cr^'stallographic planes, unlike the albitic feld-
spars. The substance of the inner cloudy core is filled with little flakes
of kaolin or muscovite arranged parallel to the second cleavage, fluid
cavities, and opascitc specks. Here and there is a large, brilliantly
polarizing flake of muscovite. The boundary against the clear portion
is generally quite sharp, and parallel to a cleavage line. The two large
black cracks crossing the specimen obliquely seem to be secondary
weathering cracks filled with limonitic products. In the clear rim the
larger mica plates ai'e seen here and there, arranged parallel to the edge
of the crystal and to the general schistosity of the rock. The ragged
form of the outer edge is very marked in comparison to the even inner
boundary. The precise nature of this feldspar cannot be determined
from the section.

In the case of two other clastic feldspars (not figured) the following
method was employed. The rock was sawn through so as to cut the
feldspiir in two ; from one side a thin section was made, while from the
other a thin slice containing a section of the feldspar was sawn, the rock
surrounding the feldspar cut away with a penknife, and tlie specific
gravity of the fragment thus ol^tained determined, while a part of it
was crushed on a glass slide, and the cleavage sections thus developed
studied under the microscope.

The first had a Sp. Gr. of 2.585, and among the crushed material
cleavage pieces with the microcline double twinning are seen. In the
thin section the crystal has an elongate shape, but the general outline is
much less regular than in the previous c;ises (see Figs. 1 and 2). Tlie
central portion has the same opaque clouded appearance, owing to fluid

VOL XVI. — NO. 10. 12

178 BULLETIN OF THE

inclusions, plates of kaolin, and liraonitic masses, but the boinidary is
less distinct against the clear edge, which surrounds only part of the
crystal. The niicrocliuo twinning is seen in spots irregularly dis-
tributed among areas of single twinning or without visible twiiniing.
The twinning runs into the clear feldspar rim a short distance, then
dying out; this feldspar of the rim, however, does not stop at the
clouded edge, but extends into the clouded portion in irregular areas,
which extinguish together. Outside the entire feldspar grain, but often
touching it, there are areas of fresh glassy microcline, and also aggre-
gates of little clear feldspars in rounded grains of the albite type. Lit-
tle tongues of this fresh microcline' penetrate the clastic feldspar grain,
and little veins of the albitic feldspar also cross its corners. Quartz
grains and muscovite plates also occur isolated in the feldspar ; this
grain seems to be microcline associated with glassy microcline and albite,
and possessing a clear boundary, which cannot be separated from areas
w'ithiu the cloudy portion.

The second feldspar (No, A), studied in the same way, had a Sp. Gr. of
2.578 and indistinct microcline cleavage in the powder. In the slide it
shows a faint double (microcline) twinning ; as before, the central part
is cloudy with tlie little kaolin or muscovite plates arranged in parallel
lines, and also fluid cavities and irregular masses of brown limonite and
black iron oxide. The core is surrounded on the two long sides by a
clear glassy border of feldspar, about one eighth the width of the entire
grain. It has a single multiple twinning, which is parallel to and ex-
tinguishes with one of the sets of twins in the core ; the boundary be-
tween the rim and core runs in a general straight direction, although
jagged in detail. The rim contains flakes of muscovite, droplets of
quartz, and little singly twinned albitic crystals. At one end of the
grain, but outside, there occurs a large irregular area of doubly twinned
microcline, fresh and glassy, and distinct from the cloudy core which it
touches, the clear rim being absent. This fresh microcline, which is
evidently the same as that found in isolated areas in the rock, con-
tains mica and quartz inclusions. It extends as a vein half across the
clastic grain, quartz and mica mingling with the microcline. At the op-
posite end of the grain a little vein starts from the outside and entirely
crosses bntli the clear rim and the core ; it is at first composed of quartz
and muscovite, tlicn in the interior it changes to glassy microcline.

This grain seems to be microcline from the specific gravity and optical
tests ; it has freshly formed microcline adjoining it and crossing it in
veins ; it has a rim of clear feldspar in optical continuity with the

MUSEUM OF COMPAKATIVE ZOOLOGY. 179

whole grain, containing muscovite, quartz, and albitic feldspar, which
mav be itself microcline.

Figures 5 and G ilkistrate the case of .a clastic feldspar in which the
clear feldspar rim ramifies through the grain, entirely crossing it. Fig-
ure 6 repi-esents the left hand middle portion of the grain, which is
shown entire in Figure 5. The shape is roughly trapezoidal, with an
irregular edge bounded by the muscovite of the cement. The whole
feldspar polarizes as a unit, but in different tints. The clear rim and
the little connecting cross branches (see Fig. G) polarize in green, while
the cloudy portion polarizes in blue, the two parts passing gradually
into each other. The chnidy portion with high powers is seen to be
filled, as usual, with fluid inclusions, flakes of kaolin and limonitic
products, which, as seen in Figure G, are linearly arranged ; in polarized
light the cloudy material is seen to be arranged in spindle-shaped
masses. Two black lines which in both figures occupy the centre of
the clear tongues are aggregates of muscovite, which connect with that
outside the grain; several smaller tongues of muscovite also run in
a short distance from the outside. Here and there in the clear feld-
spar there are isolated large flakes or aggregates of the same mineral ;
minute colorless gi'ains with high single and double refraction occur,
which are probably calcite. The relations of the clear and cloudy fold-
spar are such that the latter occurs in little isolated areas encroached
upon by the clear mineral. In one patch only, the clear feldspar shows
twinning in a few isolated stripes. The cloudy portion shows none
whatever, and there is no means for determining its original character,
since only small residual patches remain.

In Figure 4 there is represented a small feldspar which polarizes with
a low even tint, is clear and glassy throughout, contains flakes of
brilliantly polarizing muscovite similar to that by which it is sur-
rounded, and has in general all the characters of the "albitic" feld-
spars with this exception, that the left hand portion is cloudy ; this is
due to the same cause as before, namely, fluid inclusions, flakes of kaolin,
and limonitic masses.

The feldspar of Figure 3 polarizes in the cloudy portion blue ; in the
outer clear glassy portion, a red of a higher order, the slide being thick.
The cloudy and clear portions have the usual characters ; the latter
shows here and there a single twin lamella.

In Figure 7 a small feldspar is represented of the albitc type ; that
is, it polarizes in an even low tint absolutely without twinning, and
contains comparatively large flakes of muscovite arranged parallel to

180 BULLETIN OF THE

the minerals outside. The centre, however, is cloudy, owing to fluid
inclusions and particles of limonitic material. Compared with the cases
pi'eviously described, the cloudy area is of less importance ; it not only
occupies a smaller portion of the whole grain, hut the inclusions are not
in such close aggregates ; there is also no difference in polarizing tint
between the clear and cloudy portions.

In Figure 8, already referred to, we have the type of the pure
" albite " variety of feldspar ; nameh', a homogeneous area of feldspar
without a linear boundary, having a somewhat longer dimension pai-allel
to the schistosity of the rock, glassy clear, and polarizing with a low
tint. In this case the black inclusions are magnetite, the others flakes
of muscovite. Muscovite bounds the grain on cither side.

In order to determine more accurately the nature of these feldspars,
a portion of the rock was powdered and a separation of the constituents
made by the Thoulet solution. From the powder thus obtained slides
were prepared by scattering a little of each powder in balsam on a glass
slide and cautiously grinding down to the required thinness ; the thin
sections thus obtained could be examined microscopically nearly as well as
in ordinary slides and with a predominance of cleavage sections. With
the first falling, the muscovite, magnetite, and tourmaline came down.

Between the specific gravity of anorthite, 2.76, and that of quartz,
2.65, a little material was obtained, which was found to be feldspar and
quartz weighted by mica or magnetite as impurities. At 2. 65, the bulk
of the powder came down, which was found to be quartz ; between this
and 2.60, the lower limit of plagioclase, a considerable quantity of pure
feldspar came dow^n, which in the slides exhibited the properties of what
has been described above as the albitic feldspar ; that is, the grains
are untwinned or simply twinned, clear and glassy, with occasional
muscovite or quartz inclusions. Sections could be found cut parallel to
the basal cleavage, twinned in two single halves and giving an extinc-
tion 4° oblique to the second cleavage, while other sections cut appar-
ently parallel to the second cleavage have an extinction 17° obhque to
the first cleavage, and show a bisectrix slightly oblique. Without
chemical analysis this is as complete a determination as was possible,
and by Sp. Gr. and optical properties indicates alhite. With this albite
there occur some grains of a multiple twinned plagioclase, and some
microcline apparently weighted by inclusions.

Another feldspathic portion of the powder was obtained between 2.60
and 2.56, and the larger portion of this was microcline. A part was of
the fresh glassy variety, free from inclusions, with distinct double twin-

MUSEUM OF COMPARATIVE ZOOLOGY. 181

ning iu many of tiie grains, and was evidently the glassy feldspar with
these properties described as occurring in the sections of the rock in
little areas associated witli tiie larger clastic feldspars, or in independent
flattened areas like those of the alliite. Another portion of the micro-
cline contained fluid inclusions, mica or kaolin flakes, and masses of iron
oxide, and seemed therefore to represent the clastic feldspar. Ortho-
clase was not identified.

The interpretation of these facts is not easy, and they do not seem to
the writer quite parallel to the cases of feldspar enlargement heretofore
described.

Van Hise^ described an enlargement of clastic feldspars in certain
Keweenawan sandstones in which the original orthoclase or plagioclase
grain, charaterized by its kaolinization and a border of ferrite following
the original rounded shape of the grain, was surrounded by a zone of
clear feldspar with ragged outer edge, which extinguished with the core,
and in which twinning bands were continued when present. The new
feldspar was therefore crystallographically co-ordinated with the old.

In the numerous cases of feldspar enlargement in eruptive rocks
described by several writers, the new feldspar sometimes extinguishes
with the old, sometimes does not, and appears then to be a more or less
distinct variety. Professor Judd has lately described feldspar enlarge-
ments'-^ in a " labradorite-andesite," which he believes to have been formed
after the consolidation of the rock and its alteration by weathering, the
new feldspar having formed through the alteration of the glassy base.
The original labradorite grain is surrounded by a clear feldspar fringe
across which the twinning planes of the core are prolonged, but in which
the optical constants have a dilferent orientation, appearing to belong
to a more acid feldspar. Tongues of this feldspar sometimes pene-
trate the old core, which is kaolinized.

Van Hise mentions the fact,^ that in the mica gneisses of the Elack
Hills the only microscopic clue to clastic origin of the rocks is found in
the presence of particles of iron oxide in the outer portions of the (en-
larged) quartz grains, but that this is evenly distributed through the
feldspar, which hence has entirely formed iu place.

The feldspars described in tlie present paper seem to represent both
this completed stage, and intermediate stages in which more or less
original clastic fi'ldspar remains.

1 Am. Journ. Sci.. Vol. XXVII. p. 309.

2 Quart. Journ. Geol Soc, Vol. XLV. pp. 175-186.

3 " Pre-Cambrian Rocks of the; Black Hills," Bull. Geol. Soc. America, Vol. I.
p. 227.

182 BULLETIN OF THE

It does not seem possible that the contrast of clouded kaolinized core
and clear rim could be due to selective decomposition acting on one
homogeneous grain, by which the centre was attacked, while the rim
was left clear. The clastic outline of the kaolinized portion, the in-
clusions of mica, quartz, etc. in the clear part, showing a difference of
origin, and the intricate manner in which the clear portion sometimes
ramifies through the core, seem to negative this supposition.

It seems necessary to regard the kaolinization as antecedent to the
formation of the clear feldspar.

What then is the I'elation between core and rim 1 In all the cases
described in this paper, the two parts extinguish together, aud, as in
the cases described by Van Hise, seem to be crystallograjijhically and
optically continuous. Twinning sometimes runs from one to the other,
sometimes dies out in the clear feldspar. In many cases, however, the
polarization tint of the core is different from that of the rim, and ob-
served with the highest powers this seems to be inherent in the feldspar
of the core and not due to the visible products of kaolinization ; in cases
such as that of Figures 5 and G the kaolinized portions are simply little
areas which grade imperceptibly into the clear feldspar, which perme-
ates them in every direction. It does not seem possible to explain all
these cases by mere outward growth of the feldspar grain by addition of
fresh feldspar of the same species to the core ; but rather by an actual
replacement of the detrital core by the feldspar of the enlargement,
which even in the least advanced stage in this rock has gone so far as
to leave but little beyond traces of the original feldspar, and the kaoim-
ization products. The whole series by which, if this exphmation is the
true one, even the kaolinization products are seen gradually to dis-
appear, could only be described and figured here by isolated cases,
although there is an intimate gradation in the slides. The smaller feld-
spars generally show this absorption most completely, until the per-
fectly clear type of Figure 8 is reached, which is undoubtedly albite.
The feldspar of the intermediate cases has entirely similar properties
aside from the presence of tlie fluid inclusions, kaolin, and iron oxide.
In some of the larger feldspars the core is microcline, and the rim is
perhaps the same (No. A) and this rim encloses small cr^'stals of albitic
feldspar and is adjacent to glassy microcline, which even penetrates the
whole grain ; therefore it is necessary to suppose a partial replacement of
the original grain by microcline (or regeneration if the clastic grain itself
was originally microcline), with enlargement and subsequent formation of
microcline outside the whole enlarged grain, which had been fissured.

MUSEUM OF COMPARATIVE ZOOLOGY. 183

Both microcline and probably albite have also formed independently
of any pre-existing nucleus, as far as their outline givea njy clue ; but in
other cases, if the facts are correctly interpreted above, the albite areas
owe their position to replacement and accompanying enlargement of
detrital feldspar cores of undeterminable species.

It has been stated that this conglomerate of Bear Mountain passes
into an albite schist in places, and that the Cambrian conglomerate of
Hoosac Mountain also passes into a similar rock conformably at the top
of the conglomerate series. The nature of the feldspar in the Hoosac
rock and in the similar schists westward in Greylock Mountain has been
proved by analysis.

In the slides of several specimens from the latter region the albites
show an apparent enlargement, not uncommon in the albites of albite-
phyllites in general. These albites are in large rounded grains, either
simple crystals or singly twinned. Each crystal polarizes homogene-
ously, having a very ragged outer edge. They contain a black mate-
rial disseminated through their substance, which is apparently black
oxide of iron mixed with graphite ; this substance is sometimes evenly
disseminated through the feldspar, or may be arranged in bands, some-
times wonderfully curved. These bands may be parallel in two adja-
cent crystals separated by the mica of the cement, showing a formation
in planes independent of the single albite crystals. Sometimes these
bands occur only in the core of the crystal, and are bounded by a zone
of clear albite, forming one crystal with the core. The outline of the
banded core is then sharp and bounded by straight lines, producing an
angular outline. In other cases the black material is irregularly dissem-
inated through the core, but bounds it against the clear rim by a more
or less continuous black line, which gives the core a rounded outline
bounded by gentle curved sides, simulating quite closely the well known
iron oxide bands which mark the limits of the original quartz grains in
quartzites, enlarged by new silica. These are therefore apparenly en-
largements of albite grains by new albite. Would it be possible instead
to regard them as replacements of original feldspar grains by albite
which grew beyond the limits of the original grain, did not entirely
resorb the iron products, and sometimes aft'ected their distribution by
any pressure and movement which may have accompanied the chemical
process by which the albite was formed?

Harvard Universitt, June, 1891.

DESCRIPTION OF PLATES.

Reproduced by the artotype process from microphotographs in polarized light.
The enlargement is given under each figure. The eight figures which are de-
scribed in the text represent a partial series exhibiting tlie supposed progressive
changes in original detrital feldspar grains until complete obliteration of the
original characters is attained.

Figures 1 and 2 represent the earlier stage in undoubted clastic grains.

Figures 3, 4, 5, and 6 represent an intermediate stage. Figure 6 is the left hand
central portion of Figure 6, more highly magnified, in order to show the rami-
fication of the clear feldspar in the cloudy feldspar.

Figure 7 represents the almost completed stage, and

Figure 8 the albitic type of feldspar in which there is no trace of clastic origin.

PLATE

FIG. 3 X 50.

PLATE

FIG. 7 X 46.

iG e X 4c

No. 11. — The Conditions of Erosion beneath deep Glaciers, hased
upon a Study of the Boiddcr Train from Iron Hill, Cumber-
land, R. I. By N. S. Shalek.

The conditions which determine the erosive action effected by conti-
nental ice sheets are as yet imperfectly understood. "We are not able
to penetrate beneath the existing accumulations of this type ; their
position makes even their superficial phenomena matters of difficult
inquiry. All that we can hope to ascertain concerning the work done
beneath these massive glaciers must be inferred from the eftects which
they have exercised upon the surfaces over which they once moved and
from which they have passed away.

Certain of these effects are so clearly indicated that there is no diffi-
culty in interpreting the actions to which they were due. Others as
yet remain extremely obscure. As the following inquiry was under-
taken with the hope of clearing up some of these obscure problems, I
shall preface this essay by a brief statement as to the present condition
of our knowledge concerning certain features of the mechanical work
done beneath the ice during the Glacial Period. The following described
points may be regarded as fairly well established, viz. : —

1. That the continental glaciers of the last Ice Period, though they
clearly moved forward in the direction of the ice front, did not always,
or even generally, accumulate large bodies of morainal matter at their
margins ; the frontal moraines being on the whole small in amount as
compared with the evident depth of the ice, and the distance over which
the materials were moved.

2. That wherever the rocks over which the ice moved were of a na-
ture to furnish hard fragments, these were plentifully removed, and the
consequent erosion of the surface went forward in a tolerably rapid
manner.

3. That the fragments loosened from the bed rock by the action of
the glacier were often borne to a considerable distance from their points
of origin, and that in this journeying the amount of erosion to which
they were subjected varied greatly, some of the erratics evidently re-
maining in contact with the bed rock, and serving to score or abrade its

VOL. XVI. — NO. 11. 1

186 BULLETIN OF THE

surface as they were urged forward ; others were lifted into the body of
the ice. Thus there wei-e two kinds of glacial wear : the one effected
by the plucking out of large fragments which separated from the matrix
along the lines of fracture formed by the bed or joint planes, and the
other effected by the rubbing of the bits of stone against the firm set
rock or against each other.

4. That there were in certain places beneath the ice considerable
streams of water moving between the overlying glacier and the firm set
earth, these sub-glacial currents being in many cases competent to move
large quantities of detrital materials.

The points above stated appear to me to be all which are of im-
portance, and which have already been established in a conclusive way.
The following problems as to the work done beneath glaciers seem to
be worthy of consideration, and to their elucidation the matters set forth
in this paper have been in the main dii"ected.

First. As to the rate of erosion on a surface of a given hardness
during the passage over it of a given length of the ice sheet.

Second. The relative amount of glacial erosion accomplished by the
dragging out of fragments, and by the grinding action of such fragments
upon one another and upon the bed rock.

Third. The relative distance the above named classes of detrital ma-
terials have been transported.

Fourth. The lateral dispersion of the debris on its journey from the
point of origin towards the ice front.

Fifth. The relative amount of wearing of pebbles of diverse hardness
in the course of their transportation.

Sixth. The effects of pressure melting on the lower lying parts of
deep glaciers, and the effect of such melting if it occurs in the process
of erosion.

This list by no means exhausts the catalogue of questions concerning
the nature and effect of glacial movement. It contains, however, a sug-
gestion of all the problems which it seems to me possible to deal with
from the facts which are considered in this paper. In searching for a
district in which to prosecute the above mentioned inquiries I sought to
find a field where there should be some small, sharply limited area of
rock, the material having a mineralogical character so distinct that it
would not be possible to confound its waste with that from any neigh-
boring locality. It was also necessary that the rock selected for study
should be of sufficient hardness to insure it against speedy destruction
in the rough conditions of glacial transportation. It is furthermore

MUSEUM OF COMPAKATIVE ZOOLOGY. 187

necessary that there should be a sufficient distance between the selected
locality and the sea to afford a chance for the display of a bowlder
trail of such length as to afford indications of value. After much pre-
liminary search of various localities in New England, I found that the
boulder train from the so called Iron Hill in Cumberland, R. I., pre-
sented by far the most satisfixctory basis for the proposed study.

General Description of the Train.

The accompanying map will afford the reader a sufficient idea as to
the geographical conditions of the district in which the trail from the
Iron Hill lies. It will be observed that the iron deposit is situated in the
town of Cumberland, which lies on the eastern margin of the valley of
the Blackstone River, at a point about fifteen miles north of Providence,
R. I. Although the contour map which accompanies this report, and
-which is reproduced from the plates prepared by the U. S. Gee -gical
Survey, will afford an excellent idea of the topography, the scaie on
which it is printed is somewhat smaller than is desirable for the com-
plete display of the matter with which we have to deal. It would have
been possible, from the original maps of the Survey, to have made this
illustration on a more considerable scale, but the chart would have then
been too large for convenient use.

It will be observed that, except for the occurrence of a few somewhat
isolated hills, the reliefs of this district have no great height. They
rarely, indeed, exceed one hundred feet of elevation above the neigh-
boring valleys. For about thirteen miles from its point of origin, this
glacial trail lies upon a surface of bed rock of varied hardness, which
still retains in good part the topography given to it by erosive agents
which operated before the advent of the last Ice Age. The boulder
clay or ground moraine of which the train forms a part is, except where
it crosses these ancient valleys, usually not more than ten feet thick-
It is likely that in this district there may be many deep valleys in the
bed rock which have been entirely effaced by the thick deposits of drift
which mantle this part of the coast between Valley Falls and Aquidneck
Island. On that island the bed rock is again near the surface, the drift
being rather more than fifteen feet in average depth, and except next the
shores, where some waslied drift occurs, consisting altogether of boulder
clay. In this part of its course the movement of the glacial stream
which bore the material deposited in the train may have been somewhat
guided by the long ridge which forms Aquidneck Island, and by the

188 BULLETIN OF THE

deep trough which constitutes the eastern part of the fiord known as
Narragausett Bay. At present there is about three hundred feet differ-
ence in the altitude of the bottom of the channel and the top of its
eastern wall. It is likely, however, that the debris which has accumu-
lated since the Ice Period has diminished the depth of this trough to
the amount of a hundred feet or more. The axis of the depression lies
approximately in the path of the ice stream. It was, indeed, partly
excavated by the same glacial movement which conveyed the fragments
of the boulder trail, and therefore it naturally has the same general
direction as that followed by the glacier.

An important effect of the channel which forms the main or western
part of Narragansett Bay has been to hide a great part of the surface of
the train beneath the waters of the sea. As will be seen from the map,
the train appears on the shore upon both sides of the Narragansett main
channel, and upon the islands in the bay wherever they exhibit deposits
of till. It is evident, however, from the distribution as shown on the
map, that the ice stream, while pretty closely following the path of the
depression, inclined to surmount the eastern border of the channel.
This tendency becomes more marked as we approach the seaward end of
the bay, for at that point we find the train inclining away from the
mainland, or western shore of the embayment, and crossing Aquidneck
Island to its eastern side. It will be observed that this eastward ten-
dency is probably increased in the part of the train which lies below the
level of the sea. This is indicated by the fact that the area occupied
by the boulders touches the southwestern end of Martha's Vineyard. A
glance at the map will show that this requires a very sudden turn of the
train to the eastward on the area beneath the sea between the mouth of
the bay and the district occupied by Martha's Vineyard.

At first sight, it seems improbable that such a sharp flexure in the
path of the :ice as is above noted would have been likely to have
occurred. There is evidently no barrier in the region beyond the mouth
of Narragansett Bay which could have served to bring it about. While
still in a state of doubt concerning this feature I had occasion to ob-
serve a similar sharp eastward turn in the ice which escapes from the
southern end of the fiords at Mt. Desert, Maine, where we have valleys
comparable to that of Narragansett Bay as regards their direction,
debouching into the open sea. In the Narragansett Basin we are not
able to find very clear evidence as to the path of the ice after it escapes
from the groove between Aquidneck and Conanicut Islands ; but on ^It.
Desert we can see on the islands which lie to the southeast of Somes

MUSEUM OF COMPARATIVE ZOOLOGY. 189

Sound distiuct proof, afforded by the numerous glacial scratches, that
the ice turned sharply to the east through about the same angle as
would be necessary to lead the train we are considering to the western
end of Martha's Vineyard. The same eastward turning is observed in
the direction of the glacial scratches at the west end of the mountain
range of Mt. Desert, where a change of over fifty degrees of arc in
the path followed by the ice evidently took place within a north and
south distance of about one thousand feet.*

A glance at the map which accompanies this report will make it evi-
dent that by far the larger part of this boulder train lies beneath the
level of the sea. At least nine tenths of its area is so hidden from view.
Fortunately for the inquirer, a considerable, and perhaps the most impor-
tant, part of the evidence which it affords remains open to inspection.

Description of the Source of the Train at Iron Hill.

The conditions of the deposit whence the boulder train was derived
deserve special attention. They are eminently peculiar, and singularly
well suited to be the subject of an inquiry such as is proposed in this
writing. The material of which the Iron Hill is formed is a peridotite,
a phase of olivine gahbro. The percentage of metallic iron in the ore
varies between 25 and 45 per cent. The remainder of the mass con-
sists in the main of olivine and felspar : the olivine is in small rounded
grains, but the felspar generally occurs in the shape of considerable
crystals. These are irregularly scattered through the mass, appearing
in the greater part of the deposit : they constitute a very striking
feature in the rocks, and enable the observer with great certainty to
recognize a fragment derived from it. Under the influence of the
weather, the surface of the boulders derived from this locality assumes
a dark rusty brown color, while the felspar crystals retain a whitish hue.
Pebbles of this material, on account of their conspicuous features, are
often visible at a surprising distance, and present such peculiarities
that the trained eye will never mistake them for those formed from
other deposits.

The area occupied by the peridotite deposit of Iron Hill is singularly
limited. It appears as a unique boss of rock rising from the tolerably
level country about it to the height of sixty feet above the base. (See
Plates I., II., and III.) The greatest length of the mass is about 1,200

* See Report on the Geology of Mt. Desert, Eighth Annual Report of the Di-
rector of the U. S. Geological. Survey, p. 1004.

190 BULLETIN OF THE

feet, its greatest width about 500. The precise position of its flanks
is more or less hidden by deposits of drift, but a careful survey with the
dipping needle indicates that the deposit has a rudely elliptical form of
the above named dimensions, and with the major axis in an approxi-
mately north and south position. The rocks which border this deposit
are not disclosed at the contact. There are no outcrops within some
hundreds of feet of its base. Where found this " country rock " con-
sists of syenitic and granitic gneiss and mica schist. These are, indeed,
the only materials which have been observed in the immediate neighbor-
hood of the hill.

All the evidence which has been obtained tends to show that this
mass of ilmenite is in its nature a dike. The considerations which
lead to this conclusion cannot well be discussed in this memoir. They
are in general as follows. The rocks of this section of country are
evidently steeply tilted. All the exposures which have been observed
show dips exceeding thirty degrees of declivity. Therefore, if this iron
ore were a bedded deposit, it would probably appear along a much
longer line than that on which it occurs. We should expect, even
allowing for possible faulting, to find the bed repeated at other points in
the field. It is true that bedded iron ores sometimes occur in thick
pockets, but in no case known to me do these accumulations have the
solitary character or the proportions indicated in this deposit. The
studies made by M. E. Wads worth appear to afford good petrographic
arguments in favor of the hypothesis that this ore has been injected
into its present position.*

A very careful search with the dipping needle, over the fields for a
distance of some miles from Iron Hill has failed to show anything which
could excite suspicion that similar deposits exist beneath the thick
mantle of drift which covers the greater part of this country. A yet
more careful exploration for boulders has shown that only a very few
small pebbles of this ilmenite exist in the district north, east, or west of
the hill, and these are all of a size and shape which make it eminently
probable that they have been conveyed to their present site by the
aborigines or by the white occupants of the country.

Pebbles of the Iron Hill peridotite have been more or less employed for
various useful purposes, and so have secured a certain measure of arti-
ficial distribution. They serve remarkably well for weights, especially
those used for fastening horses. They make excellent ballast for boats.
They are adapted for heating water, where the heat is communicated to

* This Bulletin, Vol. VIL (Geol. Series, Vol. I.), p. 183 et seq.

MUSEUxM OF COMPARATIVE ZOOLOGY. 191

the fluid by hot stones, — a method much practised by the Indians and
other primitive people, and still in use by our farmers. Therefore it is
not surprising that occasional specimens are found beyond the limits of
the boulder train, but in many weeks of search I have never discovered
a fragment away from the trail which did not by its present use or its
association with other objects indicate that it had been artificially trans-
ported. Only three boulders have been found to the north of the hill.
These were rounded bits, such as may have been used by Indians or
whites in various simple arts. As will be seen from the description of
tlie glacial trail, the pebbles within its belt are extremely numerous,
constituting for some miles of distance a considerable percentage of these
drift materials. Therefore tliese sporadic fragments, always of small
size, but not the smallest, and of a form likely to commend themselves
for the before mentioned uses, need not excite suspicion that there are
other and nnlooked for sources of this material which may serve to
confuse the indications which the train affords. The form of the train,
so far as it is traceable, and the condition of the fragments at equal dis-
tances from the apparent source, alike indicate that all the materials
which it contains have been derived from one locality. Throughout the
observable portions of its extension, the train steadfastly and uniformly
widens, and the fragments of peridotite grow smaller with the increase
of distance from the source of the material.

At no other point in New England has an ore which could be con-
founded with this from Iron Hill been discovered ; nor, so far as I can
learn, has a rock of like aspect been obsei-ved in any other part of this
country. It seems therefore safe to assume that this boulder train
affords excellent evidence as to the conditions which brought about the
transportation of its fragments.

The rock of which the Iron Hill is composed has been a good deal
used in making pig-iron. From time to time the material has been
quarried and exported to furnaces in New York and Pennsylvania. The
large amount of ilmenite which it contains makes it an unsatisfactory
ore, but the absence of phosphorus and the small amount of sulphur
tempt manufacturers to essay its use. Tlie result has been a sufficient
amount of mining operations in the form of open pits somewhat to change
the original aspect of the hill. Moreover, a large part of the boulders
near the source of the train, and at various points in its length, have
been gathered and shipped to distant parts. The illustrations (Plates
I., II and III.) give the general form of the mass as it now appears.
In its original aspect it had the regular arched shape which is common

192 BULLETIN OF THE

to isolated masses of hard, homogeneous rock which has been worn
by glacial action. At the close of the Glacial Period this region prob-
ably was for a time below the level of the sea, and during the re-
elevation of the land the surface of the hill appears to have been exposed
to the washing action of the waves, by which the covering of drift mate-
rials was in good part removed from its summit and accumulated about
the base of the elevation. The slightly undercut cliffs on the eastern
face of the mass may likewise be due to wave action. The expansive
effect of frost acting in the frequent open joints, which plentifull}' occur
wherever the rock is not covered by the drift, has somewhat disrupted
the supei'ficial portions of the mass. Notwithstanding these natural
and artificial changes, the elevation still retains the general aspect which
it had at the close of the Glacial Period. It has probably not, on the
average, lost more than six inches in depth of material since the ice
passed away.

In the details of its structure the iron ore which constitutes the Cum-
berland Hdl is remarkably uniform. There are relatively few joints,
and these, except in the upper parts of the hill, are usually latent ;
that is, they are generally developed only wdiere the rock is subjected to
a considerable rending force. At one place on the western side of the
hill, near its northern end, there was until recently a large surface of
rock exposed by the artificial removal of the drift, preparatory to quarry-
ing work. This surface, containing an area of about five hundred square
feet, was in precisely the condition in which it was left at the close of
the Glacial Period. The finest scratches made by the moving ice were
not in the least effaced (see Plate IV.). From this and the other small
and imperfectly preserved glaciated exposures which the hill affords, I
have sought to determine the nature and measure of the ice action on its
mass. Of the several possible problems, two were selected for discus-
sion ; first, as to the direction of movement of ice over the surface of
the hill ; and secondly, the proportionate amount of wearing done by
the plucking out of fragments and by the rubbing down of the surface.

The data for determining the general direction of the ice movement in
the region to the north and west of this district are imperfect, for the
reason that the rock is not of a nature to have formed or retained glacial
8tri£e, and is generally covered with glacial detritus. The scattered ob-
servations which are obtainable, serve to show that the average move-
ment was the same as that which is indicated by the central line of the
boulder train. When, however, we study the scratches exhibited on this
hill, especially those which occur on the large artificially exposed area

MUSEUM OF COMPARATIVE ZOOLOGY. 193

on its western side, we see that the general motion was accompanied by
local and subordinate movements which were of a very complicated na-
ture. In the lower part of the large exposure just referred to, the
scratches, which are peculiarly well preserved over the entire surface,
have the general direction proper to the normal movement of the ice in
this held ; but within the space of a few feet they turn abruptly in the
manner indicated in Plate IV. The change in the direction of the
movement of the ice, which is indicated by the scratches in the distance
of about five feet, is as much as eighty degrees of arc. Unfortunately,
at the time of the observation the mass was broken away at the point
wliere the turn was made, but from the remains of the surface it is clear
tliat the total change in direction was accomplished within the distance
of ten and probably within less than five feet. It is difficult to account
for this sudden change in the course of the ice movement. The geueral
inclination of the surface on which it occurred is slight, the hill rising
on the eastward with a slope of not more than fifteen degrees, and to a
height of only about thirty feet above the place where the turn occurs.
The decline of the surface toward the valley on the westward appears to
have been equally gradual.

It may be remarked, in passing, that sudden turnings of the ice where
it came in contact with slight obstacles appear to be much more com-
mon in Southern New England than in the more northern parts of the
continent. This may possibly be explained by the fact that the glacier
during the last stages of its movement in this southern district con-
tinued to advance even after it became relatively thin, and that in this
condition it was more influenced by local and slight irregularities than
was the case in the regions where it had a profounder depth. It is
not at all likely that the steadfast flow which conveyed the materials
of the great boulder train belongs altogether to the same part of the Ice
Period as these, the last formed scratches on Iron Hill.

The surfaces of the hill which remain as they were left by the ice are
generally smooth. They exhibit, as is the case usually with massive rocks,
a breadth of level or gently curved faces on which are incised the shallow
depressions of the glacial scratches. Here and there we find cavities out
of which large fragments have been plucked by the dragging action of
the ice. \yherever these depressions left by the violently separated
fragments exist, we observe that the angles of the depression have been
more or less rounded by erosion subsequent to their removal. They
occur in all states of obliteration, some retaining their originally sharp
outlines, others being so far effaced as to be almost indistinguishable.

194 BULLETIN OF THE

About one fifth of the surface is occupied by the pits which were left by
the removal of these larger fragments, the remainder being, except for
the post-glacial wear to which it has been subjected, in the planed down
condition resultinsr from the attrition of the fragments embedded in the
moving ice. We have therefore to conclude that at least four fifths of
the down-wearing of this mass during the later stages of the glacier's
action was accomplished by the scoring and scratching action of the
glacier, and not by the plucking out of large masses such as constitute
the boulders in the great train. The portion of the eroded matter re-
moved from the grooves was, at the time of its formation, broken into
the state of sand, the grains of which, like the larger fragments, were
borne on by the glacial movement.

For reasons which will appear hereafter it is important to consider
whether or no this relative excess in the quantity of the fine-grained
material removed by the ice was limited to the closing stages of the
Glacial Period. At first sight it seems likely that the thicker the ice
moving over a rock surface the greater would be its tendency to rend
the rock over which it flowed. The pressure of an ice sheet upon its
base is directly as its depth, and up to a certain point the abrading
power of a glacier must increase with its vertical section. When, how-
ever, the ice has attained a certain thickness, it must attain the maxi-
mum effect which it can exercise upon the surface over which it moves.
After that, the shearing action of the upper upon the lower parts of the
ice must cause the superior part to flow over the lower without propor-
tionately increasing the erosive action.

In corroboration of the view that there was probably no great diminu-
tion in the plucking action of the glacier in the later stages of its work,
we find that the boulders of the train which lie nearest its source are the
largest which appear anywhere in its path, and that from its origin to
its extremity the fragments in the train gradually, and rather uniformly,
diminish in size through the process of rending by attrition to which
erratics are commonly subjected in the process of glacial carriage. We
furthermore note the fact that almost everywhere within the glaciated
district where we can examine a large surface of any of our mass-
ive rocks, we find, as here, that the proportion of the scored away
or ground down to the rent surfoces is generally very great. From
these considerations I am disposed to assume that the material removed
from Iron Hill in the form of boulders was very much less than that
which was carried away in the form of sand or yet more finely divided
matter.

MUSEUM OF COMPARATIVE ZOOLOGY. 195

The reason for this relatively large wear of the rock surfaces by the
process of scratching and grooving is easily seen : every part of the
base of the ice was armed with bits of stone, which, held in the moving
glacier, were dragged over the subjacent surface. When one fragment
was worn out in the rude work, another was pressed down in its place.
Each bit of rock worn from these grooves in turn became a cutting
point : even the finest served to polish, and in so doing to remove, a
portion of the rocky matter over which the ice moved. Thus the scor-
ing and grooving work continuously went on, but the plucking out of
fragments took place intermittingly. In order to have a mass thus
plucked out from its bed it was necessary to have the strain which the
moving ice applied to it so disposed as to lift it from its bed place, and
in most cases this movement required that the detached rock should be
overturned in a manner which required it to rise up into the lower part
of the glacier. It is evident that the disruption of blocks of rock from
their bed places would be more easily accomplished on a slope which
inclined away from the course of the ice than in other conditions of
exposure to the glacial flow, for in the first named position the blocks
would have less support, and would need to rise to a less distance into
the ice, than where the slope faced a current. To these conditions we
may fairly attribute the ruder character of the surface observable on the
southern slopes of many rocky hills as compared with the northern
aspects of the same elevations. Unfortunately, the southern end of
Iron Hill is so much covered with glacial waste that it is not possible
accurately to determine the relative amount of plucking which went
on there. The exposed surfaces, though limited in area, appear to indi-
cate, however, that the amount of degradation due to this cause was not
materially greater than upon the northern and more visible part of
the hill.

Although in its present condition the surface of Iron Hill exhibits
but few boulders of its own material, the history of the place makes it
clear that in its natural state it was plentifully scattered over with
these erratics, the greater portion of which have been removed to be
used in the manufacture of iron. Within the period of my own memory
of the locality a large part of the erratic material has thus disappeared,
and at the same time a large number of boulders from the southern end
of the train have likewise been taken away. There still remain upon
the surface of the elevation a number of considerable erratics, which
being from parts of the deposit where the ore was of low grade have not
been deemed worthy of export. The evidence goes to show that the pro-

196 BULLETIN OF THE

cess of plucking out boulders from this exceedingly firm set rock contin-
ued down to the very time when the ice ceased to act upon the surface.
I have found these facts very difficult to reconcile with tlie common
view as to the circumstances under which the glacier passed away from
this part of the country. The opinion is, in effect, that the ice
gradually lost its energy of movement, and slowly came to an inactive
or stagnant state.

The evidence as to the suddenness with which glacial action ceased
at this point is paralleled at many other places in New England which
lie at elevations within a few hundred feet of the shore line. It seems
to me that this feature may be accounted for on the supposition that
this region was considerably depressed beneath the sea at the time
when the ice lay over it, and that the glacial mass was not removed by
gradual melting, but floated away in the form of icebergs. The transi-
tion from the conditions of the ice sheet between the time when it
rested on and moved over the surface to that in which it disappeared
in the form of icebergs, may well have been brought about suddenly
by the progressive melting of the mass. I cannot discuss this hypoth-
esis as to the conditions under which the glaciers of this district ceased
to act, but I may be permitted to remark that this view seems more
consonant with the evidence than that which holds the glaciers in this
part of the country to have gradually stagnated and passed away by
sub-aerial decay.

Detailed Description of the Iron Hill Boulder Train.

The general course of this boulder train, and its relations to the sur-
face on which it lies, are fairly well shown by the map which accom-
panies this paper. The most striking feature of the train is the pro-
gressive widening of the belt of country which it occupies from its
source to the sea. At the point of origin the train is not over 900 feet
in width, and saving a few small scattered blocks which may owe their
dispersion to the action of sea waves during a post-glacial submergence
or to the interference of man, the width of the trail at its source does
not exceed 700 feet ; yet near Providence, R. I., a point about 75,000
feet south of Iron Hill, it has widened so that the distance between its
borders slightly exceeds 20,000 feet, and at the sea the belt occupied
by the Frratics is probably more than 40,000 feet in width.

Throughout its extent this boulder train is in its central part most
distinctly marked by the erratics. From this middle line the boulders

MUSEUM OF COMPARATIVE ZOOLOGY. 197

progressively dimiuish iu numbers towards the boundaries of the area
over which they are scattered. Tims in the more definitely indicated
portions of the train, say at a point five miles south of its origin, it is
possible, in favorable positions, to find a dozen or more fragments on
the surface of a square acre, while on the margins the average number
may not be one to ten acres of area. As indicated on the map, the
boundaries of the train denote the limits within which fragments have
been actually found. It is likely that occasional bits of the rock occur
at considerable distances to the east and west of the points where they
have been actually observed. From Iron Hill south to the city of
Providence, field observations indicate that the margin of the train has
been tolerably well defined. South of that point the boundaries become
constantly more and more obscure.

The obscuration of the margins of the train between Providence
and the open sea is due in part to the fact tliat in the more southern
district the boulders are to a great extent worn out by the attrition to
which they have been subjected, but it is mainly attributable to the
fact that a large part of the trail is covered by the waters of Narragan-
sett Bay, or by the deposits of terrace gravels which have been formed
since the ice moved over this district. The only part of the boulder
train which is distinctly traceable in the region south of Providence
are the marginal portions of its field. The central axis of the accumu-
lation lies in the main chaiuiel of the bay. Not only is this southern
portion of the train to a great extent hidden beneath the sea, but the
portions of it which are exposed to view have been more exposed to
the interference of man than the section north of Providence. Along
the shores of Narragansett Bay the forests have been mostly cleared
away and the land brought under tillage. These conditions have been
favorable for gathering the boulders of iron ore, which conld thence
readily be shipped by water to the furnaces along the Hudson and else-
where. Moreover, next tlie shores, the pebbles of convenient size have
been much sought for boat ballast, for which use their great weight
admirably fits them.

Trusting the indications given by occasional fragments found in
fields on either side of Narragansett Bay, the trail in the parallel of
East Greenw^ich, Pt. I., is about 27,000 feet wide, and near the mouth of
Narragansett Bay it has a width, as before remarked, of about 40,000
feet. Thus, in a distance of thirty-five miles, the average widening of
the path occupied by these boulders is abont 1,000 feet for each mile of
distance from the source, or in other words the lateral or excursive

198 BULLETIN OF THE

movement of the rocks from the central path of the train is about one
tenth of the direct forward movement.

At the source of the train the peridotite boulders are plentifully in-
termingled with, and to a certain extent overlaid by, materials brought
from the region which lies to the northward of Iron Hill. It is evident
that, these schistose rocks yielded an abundance of erratics, and it is
therefore not surprising to find that even at the outset the train has
not more than one tenth of its mass made up of the fragments of ore.
The remarkable fact is that the boulders of peridotite rather plentifully
occur on the surface of the drift in the section immediately to the
south of the place whence they were derived. This thorough com-
mingling of the debris from different parts of the erosion field clearly
indicates that there was some feature of movement in the ice which
tended to comingle the detritus which it bore, and to effect the com-
bination in a very rapid manner. In other words, the vertical scatter-
ing of the fragments in the body of the glacier evidently went on
simultaneously with the horizontal dispersion which brought about the
widening of the belt over which they were dispersed. At the distance
of not more than half a mile from the source of supply the intermixture
of the peridotite boulders with those derived from the bed rock appears
to be complete. The occasional artificial sections in road cuttings and
wells show this feature in a clear manner.

It should be noted that the average distance between the pebbles of
ore rapidly increases as we depart from the source of the ti-ain, and that
this increase is accompanied by a somewhat proportionate diminution
in the size of the fragments. This is approximately represented in the
following table, which assembles the observations made on many thou-
sand erratics in the section of the train between Iron Hill and Provi-
dence. South of that point, as before remarked, the greater part of
the trail is covered by the waters of Narragansett Bay.

Average Diameter. Largest Fragments obserred.

3 feet. 5 feet.

2 " 6 in. 4 "

1 «' 6 " 3 "

1 foot. 2 "

10 in. 1 foot 6 in.

9 " 1 " 3 "

8 " 1 " 2 "
7 " 12 "

It should be observed that this table must not be taken as repre-
senting accurately determined data at the several points on the train

Place.

At source.

2 miles

south,

4 "

ti

6 "

(C

8 "

u

10 "

«

12 "

(<

14 "

<I

MUSEUM OF COMPARATIVE ZOOLOGY. 199

designated by the distance from the source. It indicates, however, in a
diagrammatic way, the average of the observations, and may be taken
as a fair approximation to tlie facts. In this connection it is well to
note that one of the three fragments of ilmenite which have been found
on the western extremity of Martha's Vineyard had a mean diameter of
about five inches.

The greater part of the fragments of peridotite which constitute the
Iron Hill boulder train are contained in the till deposits which occur in
its path. All those of large size are found in the drift of this nature.
A few smaller bits, however, and these genei-ally much rounded, have
been discovered in the washed gravels along the shores of Narragansett
Bay. The infrequency of the occurrence of this material in the stratified
deposits of glacial waste is doubtless due to the fact, which has been well
observed in this part of Xew^ England, that these water-borne sands and
gravels have always been conveyed from a much greater average distance
than the matter contained in the till or ground moraine. Observations
which I have made in this field indicate that the average distance to
which the ordinary rocks in the till have been transported is probably
not more than three miles, while the mean carriage of the stratified
materials is at least four times as great.

A careful inspection of the fragments observed in the trail shows that
the diminution in size of the erratics in their southward journey has
been mainly due to attrition. Not one per cent of the fragments
indicate by their form that they have been subjected to division along
joint planes since they parted from the original bedding. As they de-
crease in diameter they become more and moi-e spherical or subovate in
form, until they attain a size no greater than an ordinary billiard ball.
It is a noteworthy fact that none of these bits have been found hav-
ing a diameter of less than an inch, and the number of those which
approach this size is remarkably small. As a deliberate and extended
search has been made for these smaller fragments, it seems likely thnt
their appareiit absence is not due to their inconspicuousness. I am in-
clined to explain it by the supposition that the smaller the erratic the
more likely it is to be crushed into fragments by the rude strains which
have served to round the larger bits. It is evident that the resistance
to pressure arising from being squeezed between other boulders, or be-
tween erratics and the bed rock, is in a measure proportional to the
diameter of the pebble. A strain which would be withstood by a frag-
ment six inches in diameter might crush to the state of powder one
which contained only two or three cubic inches of material. It may in

VOL. XVI. — NO 11. 2

200 BULLETIN OF THE

sreucral be assumed that the resistance which rocks of this intensely
hard nature oppose to crushing strains is inversely as their size, and in
this way we may account for the rapid manner in wliich most of the
observed boulder trains disappear, so that they are not traceable for more
than a few miles from their origin.

After observing the deficiency of small pebbles of material derived
from Iron Hill, I endeavored to trace the comminuted material within
and near the evident path of the trail by determining the quantity of
magnetic sand which the glacial deposits afforded. Examinations for
this purpose were made at several hundred points by carefully separating
the grains of magnetite attracted by a horseshoe magnet from the other
debris with which it was mingled. At first it seemed likely that this
method of inquiry would give important results, for it was at once made
evident that the till material of the district contained from two to six
per cent by weight of magnetic sand. It soon, howevei', became certain
that this material could not be accepted as an indication of the trail, for
an extended search showed that something like this amount of magnetic
sand commonly exists in the glacial waste derived from the metamor-
phosed schistose rocks of Southern New England. The fact is that these
ancient strata throughout contain a notable percentage of crystalline
magnetite. An effort to discriminate the fine-grained material from Iron
Hill b}' the amount of titanium it contained also proved fruitless, for the
reason that this substance very commonly occurs as an ingredient in the
ferruginous sands of the drift. Thus the result of about a hundred
assays made by my assistant, Mr. Robert Robertson, was purely nega-
tive, so f;ir as the study of the boulder train was concerned, though it
served to throw a good deal of light on the mineralogical constitution of
our glacial deposits. It proved that this heavy and little oxidizable iron
ore is in a measure concentrated by the actions which have brought
about the formation of our glacial deposits.

I have now set forth the most important features concerning this train
which are not made sufficiently evident by the delineation of its path on
the map. I next propose to make these facts the basis for some consid-
erations as to the nature of the actions which distributed the material
over the surface between Iron Hill and Martha's Vinej^ard.

Cause of the Fanning out of the Train.

I have already noted the fact that the Iron Hill boulder train widens
from its source to the sea, or for a di.stance of about thirty-five miles, at

MUSEUM OF COMPARATIVE ZOOLOGY. 201

the rate of about a thousand feet for the mile of distance. It is evi-
dently a difficult matter to determine the cause of this remarkable
feature. At first siglit it seemed to me possible that the result was due
to the action of floating ice, operating during the retreat of the glacier,
and more or less aided by the action of the ocean waves during the
submergence which apparently continued in this field throughout the
closing stages of the Ice Period ; but the fact that the fragments with
rare exceptions are found in the till renders it difficult, if not impos-
sible to accept this explanation. Moreover, the gradual widening of
the trail at the uniform rate above stated seems of itself to make this
hypothesis quite untenable.

There is another apparently possible cause of dispersion to be found
in the successive advances and recessions of the ice during the closing
stages of the Glacial Period. "VVe might conceive that the successive
forward movements varied somewhat in direction, and that the waste
from Iron Hill might thus have been shoved about so as to widen the
field which it occupied. In the region in which this boulder train lies
there is no distinct evidence of such successive movements of advance
and retreat of the ice, but in a region about twenty miles to the east of
this line, in the valley of the Taunton River, we find from the sections
along the line of the Old Colony Piailway abundant proof that there
Avere many successive, though slight movements of this nature. In a
portion of the valley of the above named river corresponding in length
and position to that extending from Iron Hill to Providence, there is
good evidence of at least six of these successive movements of advance
and retreat. These oscillations were slight and temporary, as is shown
by the fact that the ice in each southward going did not clear awaj' the
previously formed incoherent deposits, nor were there any distinct
frontal moraines formed at the margin of the ice field. The facts indi-
cate that these variations in the position of the ice front amounted to
only a few hundred feet of distance in the axis of the motion. In these
successive advances and recessions of the glacial margin there may pos-
sibly have been some alterations in the direction of the ice flow. It is
a well observed fact that the margin of a glacier, if the ice sheet have
a considerable front, is apt from time to time to put forth lobes which
push forward in directions somewhat independent of the general course
of the ice field. Moreover, we often find, in a district where the glacial
scratches are well preserved, that the last scorings inflicted on the rock
lie at a considerable ansjle to those which were antecedentlv formed.

It will he easily understood that a uniform shifting in the course of

202 BULLETIN OF THE

the ice over the whole trail would not necessarily alter its width. It is
necessary to suppose, if we are to account for the widening of the field
in this way, that the movements were at once so numerous and so or-
ganized in relation to each other that they resulted in a dispersing
action on the boulders, which was the greater the ftirther they were
away from their source. In other words, the difficulty with this
hypothesis is that it will not account for the gradual and essentially
uniform widening of the train from its source to the sea. This feature
seems to indicate some dispersing agent which has served in a constant
manner to increase the distance between the frajrments during their long
southward journey, without at the same time deflecting the path from a
tolerably direct line. I shall now submit the hypothesis which it seems
reasonable to frame to meet these conditions.

"Where a continental glacier rests on and moves over a surfixce of rock
it is clear that its bed is subjected to powerful al^rading forces. The
energy of position of the glacier breaks up the rock, grinds the greater
part of it into small fragments, and thereby converts a certain amount
of energy into heat. The experiments made by Robert Mallet, which
are set forth in his paper in the Philosophical Transactions in 1873,
show that the energy required to crush certain of the harder kinds of
rock is sufficient to elevate the temperature of the material to about
200° Fo-hrenheit. This amount of heat appears to be sufficient to melt
somewhere near an equal bulk of ice, provided the ice was at the tem-
perature of 32° when the melting began.

The effect of this rock crushing, which is accomplished at or near the
base of the glacier, is necessarily to bring about the melting of a certain
amount of the ice. Another and doubtless a more important supply of
lieat in this deeper part of the ice arises from the shearing movement
of the particles of the glacier over each other. There can be no doubt
that the principal part of the energy of position of the glacier is ex-
pended in the work of impelling the particles of ice by each other in
the unceasing movement to which they are subjected in the lower por-
tion of the sheet. Some part of its energ}' is doubtless expended in the
differential movement which takes place in the upper part of the section,
but by far the greater portion of it must be spent in bringing about the
tortuous shearing motions which we know from the glacial scratches
occurred in the section immediately adjacent to tlie i)ed rock. The
value of this heat-giving action cannot be well estimated, but there is
reason to believe that in amount it is many times as great as that
derived from the disruption of the rock which occurs at the base of the
glacial mass.

MUSEUM OF COMPARATIVE ZOOLOGY. 203

To the above mentioned sources of heat we have clearly to add a
third element of action ; that which arises from the constant escape of
temperature from the earth's interior. Hhe amount of this is not well
ascertained. It is probably not sufficient to melt more than a few
millimeters of ice each year, but it has to be reckoned on in estimating
the influences which are at work to bring about melting at the base of
the glacier. From these three direct heat-giving actions, we are entitled
to assume that a certain amonnt of the ice at the base of a moving mass
of the material will inevitably be melted. In addition to these direct
influences we must also take into account the probability, indeed, we
may say the certainty, that the pressure of the superincumbent glacier
indirectly favors the liquefaction of the ice at the lower parts of the
section. As long ago as 1849, James Thomson theoretically estab-
lished the conclusion that the eff"ect of pressure was to lower the freezing
point of water.* Shortly after, Wm. Thomson experimentally proved
the verity of these theoretical considerations, and it now appears that
where a glacial sheet has a thickness of three thousand feet, the freezing
point of water is lowered to about 30°, or approximately two degrees
below that at which it congeals in ordinary atmospheric conditions.

It may be regarded as fairly well established that the ice in several
points in Eastern New England attained a thickness of more than three
thousand feet ; it is possible, indeed, that when the glacial envelope was
most completely developed its depth much exceeded this amount. As
we do not yet know the temperature at which ice enters a glacial mass
such as recently covered the northern portion of this continent, we can-
not make sure that this pressure alone would be sufficient to bring
about melting at the base of the mass, even if it had the thickness of
a mile. There can be no doubt, however, that the effect arising from
the lowering of the melting point would be greatly to favor the liquefac-
tion arising from the sources of heat before mentioned. It seems to me,
therefore, that we are justified in assuming, at least as a working hy-
pothesis, the existence of large amounts of molten water next the bed
rock on which a deep glacier lies. This hypothesis seems to be ade-
quately supported by the physical conditions of our drift deposits.
Wherever these have been observed, we note the existence of many
accumulations of washed sand and gravel, commonly known as kames or
osars, which distinctly map out the position of subglacial streams of
fluid water, which often coursed for scores if not hundreds of miles
under the ice arches before they escaped at the margin of the glacier.

* See Transactions of the Royal Society of Edinburgh, Vol. XVI., Part 5.

204 BULLETIN OF THE

111 some cases, notably in the somewhat continuous kaaie belt which
extends up the Genesee valley from its mouth to its source, we are
compelled to believe that the stream flowed beneath ice which probably
had, at least over a part of the path followed by the current, a depth of
half a mile or more. It is impossible to believe that the water supply
of this stream could have been furnished by the descent of the fluid
from the surface through crevices to the bed rock, for owing to the ease
with which ice moves under pressure the weight of the superincumbent
materials would make the maintenance of such passages for the descent
of the waters impossible. We are therefore driven to the conclusion
that these subglacial channels occupied by the rivers which flowed upon
the bed rock were fed from the melting which occurred at the base of
the glacier.

The foregoing considerations make it appear likely that the water
lying at the base of a continental glacier would be subjected to the very
great pressure of the overlying ice. With ice having a depth of three
thousand feet this pressure would amount to near ninety tons to the
square foot.

It appears tolerably evident that, notwithstanding the pressure of
the superincumbent glacier, the water which flowed beneath it passed
through continuous arches leading from the interior portions of the ice
field towards its margin. We may, therefore conceive that this water,
for doubtless it had to wrestle with the ice for passage, moved irregu-
larly towards the margin of the glacier, making its way in a violent
manner by the obstructions which it encountered. The position of
the kames of serpent iform outline whicli mark the position of these
ancient subglacial streams clearly indicates that the course of the ice-
covered waters was liable to much change; as is shown by the eskers,
their paths shifted in a somewhat sudden manner from one side of the
valley to the other. Xear the margin of the ice where the depth of
the sheet would permit the formation of crevasses which for a time at
least would remain open, the erratic movements would doubtless be
greater in amount than in the sections where the ice was so deep as to
prevent the formation of fractures. We may fairly conceive that the
motion of these subglacial streams, urged as they were by a pressure of
many tons to the square foot, would, at least near the margin of the ice
sheet, take place with paroxysmal violence. Quantities of the debris
which came in the path of these vigorous currents would doubtless be
urged laterally away from the axis of motion of the glacier, or borne up
from near the bed rock into higher lying parts of the ice. Each ecu-

MUSEUM OF COMPARATIVE ZOOLOGY. 205

sideruble disruption of the glacier, by thus favoring the movements of
the imprisoned waters, -would be likely to bring about the transverse
scattering of the rock debris.

When first laid down, after transportation by these currents of water,
tlie detrital materials would naturally have the washed and bedded
character proper to deposits such as occur in kames, but we know by
observation that it often happens that such accunuilations were soon dis-
rupted by the motion of the glacier, the fragments taken into the mass
of the ice to be redeposited with the aspect of ordinary till. Much of
the drift material in Southeastern New England evidently consists of
debris which has recently been in the form of washed and stratified
gravels. A careful study of the drift in this section of the country has
convinced me that by far the greater part of its mass has been at least
once, and probably again and again, assorted by water before it was
finally taken into the ice for the last time, to be laid down in the shape
in which we now find it. It therefore seems to me that we are justiiied
in supposing the horizontal dispersion of the materials contained in the
boulder train from Iron Hill to have been mainly brought about by the
violent movements of subglacial water.

Attention has already been called to the fixct that the fine debris
derived from the scoring and polishing of Iron Hill, and from the com-
minution of the boulders which are plucked from it, is not distinctly
recognizable in the path of the boulder train. The evidence of wearing
afforded by the hill itself clearly shows that at least three fourths of
the erosion which took place upon its surface delivered the iron ore
to the glacier in the form of fine sand, such as is ground out from gla-
cial striations or worn from the polished surfaces between the grooves.
Moreover, by far the greater part of the mass of the erratics which were
plucked from the rock was reduced to a similar state of division b}' the
attrition to which the fragments were subjected. If this iron sand
had been transported in substantially the same manner as the larger
boulders, we should be entitled to expect evidence of the material in
the path of the trail ; but, as before noted, this comminuted magnetite
is scarcely, if at all, more abundant in the field occupied by the boid-
ders of the substance than in the other parts of the country to the
north, east, and west of the train. The only way in which I can ac
count for the disappearance of the fine debris is by supposing that it
was borne away to a considerable distance by the subglacial ciu'rents
of free water.

Although there is considerable difference in the measure of wear to

206 BULLETIN OF THE

which all glacial material is subjected, the debris from Iron Hill affords
the best indication of such variation which I have ever found. It is a
noticeable fact, that within half a mile of the source of the material we
occasionally find boulders which have been completely rounded com-
mingled with others which have been subjected to such slight attrition
that their original form and size has hardly been altered at all. Pro-
ceeding down the train, we note the fact that gradually all the fragments
become more and more rounded, but even at thirty-five miles away some
of the bits appear to have retained some share of their original outline.
It is true that the angularity of these fragments may in part be at-
tributed to the successive fracturing to which they have been sub-
jected, but for some miles from the source it is evident that many of
the erratics have been in a manner preserved from attrition with the
bed rock, or against other moving fragments. This has probably been
brought about by a process which uplifted the well preserved erratics
into the bodv of the ice.

Before passing from this part of our inquiry, it is woi-th while to note
the striking contrast exhibited by this train of peridotite boulders as
compared with certain other trains of softer material which are traceable
in this section. In the region to the southwest of Iron Hill in the town
of Smithfield, R. 1., there are several outcrops of a crystalline limestone
which are sufficiently limited in area to aff'ord distinct boulder trains.
Although this highly metamorphosed limestone has the hardness of
ordinary marble, and by its structure afl'ords erratics which are on the
average larger than those plucked out from the ice at Iron Hill, the
trains which are formed of it cannot be traced for more than five or six
miles to the southward of the outcrops. We thus perceive the measure
in which the singular hardness of the rock from Iron Hill fixvors the
preservation of the boulders derived from that locality, which has been
able to journey more than ten times as far as the hard marbles of the
Smithfield district.

I may allude, in passing, to the fact that the relative hardness of the
bed rock of any district, as compared with that of the fragments borne
over it by glaciers from other fields, is of much importance when we
seek to explain the distribution of glacial drift. Where the bed rock of
any locality is hard, and the rocks lying just above it in the path of the
glacial flow are soft, we generally find the surface of these hard ma-
terials occupied by little coarse detrital Avaste, and this for the reason
that the fragments are readily ground out against the nether millstone.
If the conditions are reversed, and the rocks from which the glacier came

MUSEUM OF COMPARATIVE ZOOLOGY. 207

are very resistant to erosion, while the given field is underlaid by soft
deposits, we usually find the area thickly covered with boulders. This
consideration will explain the very distaut carriage of the erratics which
we find in Southern Ohio, many of which have doubtless come from the
region north of Lake Erie, or from points more than two hundred and
fifty miles away from their present site, the whole of their course having
been over rather incoherent strata. The reverse of these conditions
exist in Nevr England, where the firm set I'ocks are usually hard, and all
the debris which comes in contact with them in the process of glacial
transportation is apt to be worn to the state of minute fragments. In
this part of the country we can rarely trace a glacial train for more
than a score of miles.

The Kate of Glacial Wearing at Iron Hill.

One of the most interesting series of observations which it proved
possible to undertake in this inquiry concerned the rate of erosion which
took place at Iron Hill during the time it was suV)jected to glaciation.
Although, as will shortly be seen, the results of this research are of a
rather discrepant nature, certain general conclusions which appear to be
of value were obtained. As will be seen from the map, the part of the
trail which lies to the north of Pi"ovidence has the form of an acute-
angled triangle, of which the base may be assumed to have a length of
about 20,000 feet and the sides an extent of about 75,000 feet. The
area included in this field is about 750,000,000 square feet. Our aim
is now to ascertain the amount of the debris from the hill which lies
upon this surface. From a careful inspection of the ground, I came to
the conclusion that it is reasonable to estimate the quantity of detritus
from Iron Hill which lies within this area as equivalent to an inch in
depth over its whole surface, or say 60,000,000 cubic feet. Now the
visible part of the Iron Hill rock does not exceed 000,000 squai'e feet.
It may be prolonged to the southward beneath the detrital deposits for
the distance of a few hundred feet, and may extend somewhat to the
westward under the drift materials ; but if we assign to the mass an
area of 1,000,000 feet, we give it all the area which careful inquiry with
the dipping needle shows to be probable. On this basis we may say
that the train north of Providence contains the equivalent of a mass
sixty feet in height which has been removed from the hill.

Attention has already been called to the fact, that not more than one
fifth of the eroded matter from Iron Hill went away in the form of

208 BULLETIN OF THE

boulders, — at least during the closing stages of the Ice Period. By far
the greater part of the material was removed in the condition of fine
sand, which has been to a great extent swept away from tlie path of the
trail. Accepting this estimate, we have to reckon the erosion of this
surface during the period when the trail between the hill and Provi-
dence was formed as amounting to about three hundred feet in depth.
As the distance between Iron Hill and Providence is about seventy-five
thousand feet, the question arises. Can we assume that, during the pas-
sage of the ice along this length of its course, anything like this gi'eat
amount of wearing was brought about at the source of the trail 1 It is,
liowever, by no means certain that the distance traversed by the ice
during the formation of the trail did not exceed the length of the field
occupied by the debris which it conveyed. As before remai'ked, there
are reasons to suspect thiit the ice advanced and retreated several times
while it lay on this part of the shore-land, and these advances and re-
treats may have materially prolonged the time during which the ice
continued to move over the surface of the hill. If the ice long retained
a stationary front at any point between Providence and Iron Hill, or if
its margin were subjected to successive oscillations, at no time falling
back to the north of Iron Hill, then the boulders on this field represent
the wearing effected by the passage of a much greater length of ice
sheet than is indicated by the longitudinal extent of this part of the
trail.

After careful examination, I am inclined to doubt whether any con-
siderable irregularities of movement such as have just been suggested
ever occurred in this part of the glacier while the train was forming. I
can find no trace of frontal moraines, such as would have been caused
by any considerable pause in the retreat of the ice or the re-advance of
its frontal wall. Therefore, while granting the probability of a certain
amount of oscillation in a glacial margin, I am not disposed to think
that these accidents could have been of such magnitude as entirely to
invalidate the computations as to the rate of erosion which we have just
made.

I have before noted the probability that the ice went off" from this
district, not by the gradual retreat of its front to the northward, nor by
stagnation followed by a slow process of melting, but by the floating
away of the thinned glacier in the waters of the sea, which at the close
of the ice time stood at a higher level than at present. If the ice sheet
thus departed in the form of bergs, we may the more readily account for
the prevailing absence of small frontal moraines which we might expect
to mark the stages of its retreat.

MUSEUM OF COMPARATIVE ZOOLOGY. 209

There is yet another way of approachiug this question of the rate
of erosion brought about by the passage of a given amount of ice
over the surface of the bed rock, — a method which is applicable in the
study of many glaciated rock surfaces. Tliis may be set forth as follows.
The indentations on the surface which has been ex'oded by the glaciers
are divisible into four classes : the pits which were left where disjointed
masses of the rock were plucked out and borne away by the moving ice ;
the grooves, or more or less distinct relatively broad channels, which
have been carved in those parts of the rock made particularly accessible
to erosion by the local softness of the material, or by the form of the
surface, which led to local intensifications of the erosive work done dur-
ing the passage of many successive cutting points composed of bits of
hard rock held down upon the bed by the moving ice ; the scratches,
which are distinguished from the grooves by the fact that they have
been formed by the incisive action of a single point of hard material
urged forward by the ice; and, lastly, the general polishing of the sur-
face accomplished by the attrition of very small powdery fragments,
which were not large enough to be fixed in the ice or suflSciently hard
to make pei'ceptible grooves, but which served to smooth the rock much
as a polishing powder acts when rubbed upon a surface of metal by the
human hand.

For our present purpose we shall limit ourselves to that form of gla-
cial wear which is effected through the action of the distinct scratches
or indentations which are produced by the movement of a point of hard
rock over the glaciated surface. On many rocks which are thus eroded
it is possible to measure the length and breadth of these indentations,
and to determine the relative amount of wear which is in this manner
brought about. It is rarely the case that the evidence to this effect is
so cleai"ly indicated as on the unweathered portions of Iron Hill. By
cai'efully examining the glaciated surface shown in Plate IV. we find that
we may estimate the depth of these scorings at an average of one twen-
tieth of an inch, and we may reckon the channels as covering one fifth
of the surface, the intermediate spaces being occupied by parts of the
rock which have been polished in the manner above described. The
average length of these grooves appears to be about eight feet. It thus
appears probable that while these rock fragments which made the incis-
ions moved for the distance of fifty feet, they eroded somewhere about
one twentieth of an inch from the surface of the rock which is the sub-
ject of this computation. At this rate, while the cutting fragments were
moving for the distance of a mile, the aggregate erosion accomplished

210 BULLETIN OF THE

through the formation of scratches would amount to about one hundred
twentieths of an inch, or to near five inches, and during a journey of
these catting fragments to the distance of fifteen miles, to about
seventy-five inches. Allowing, as before estimated, that the plucked
out material carried away in the form of boulders amounts to as much
as one fifth of that removed in the other forms of erosion, the aggre-
gate wear may be estimated during the time when thirteen miles of ice
was passing the point as somewhere about seven or eight feet.

There is a manifest source of error in the computation last given,
which arises from the fact that no account is taken of the form of
erosion which occurs between the scratches, which results in the produc-
tion of a smooth surface, or in scorings which are so delicate as not to
make an impression on the eye. As at least three fifths of the surface
is of this charactei", it appears to me that we must reckon the abrasion
due to the rubbing of the rock by very fine particles of deti'itus to be
about twice as effective in removing material as that which produces tlie
scratches. Were tliis not the case, the discernible indentations would
occupy a lai'ger part of the field. AVe therefore see that to the com-
bined scratching and polishing actions we may perhaps reckon a total
lowering of the surface during the passage of the ice over the distance
between Iron Hill and the town of Providence at about twenty feet.

The large proportional share taken in erosion eflected by the so called
polishing woi'k done by the glacier lias been generally overlooked. It
appeai-s to me that on most hard rocks it has been the efficient means
"by which they were worn away during the passage of the glacier over
their surfaces. A careful examination of any completely smoothed ma-
terials, such as our harder granites or the denser clay slates, will make it
evident to the eye that the most of the waste which was removed from
the surface, at least during the last stages of the glacial erosion, was
taken away in the form of a very fine powder, the so called glacial flour,
which we know by many observations is likely to be carried to an indef-
inite distance by the sti'eams which flow from the glacial area. In tins
way, we can account for the fact that tliis fine detritus has generally
disappeared from the districts about the ice front.

There is yet another evident doubt which serves to diminish the value
of the above given rude approximations to the rate of glacial wearing.
This is due to the fact that we do not cleaHy know that the hard points
which effected the incision of the scratches were firmly held in the ice
as they were urged over the surface of the bed rock. It is well known
tliat a wire suspending a weight and resting upon a block of ice, the

MUSEUM OF COMPARATIVE ZOOLOGY. 211

mass of wliich is at or near the freezing point, will, throngli the effect
of pressure in promoting melting, gradually work down through the
block without leaving any crevice behind it. It therefore seems not
improbable that where the bit of rock which made ifhe incision in the
bed was small, the ice which held it moved more rapidly than the tool
itself, and that the rate of movement of the cutting points was an un-
known fraction of that at which the ice moved towards the margin of
the glacial field.

The time required for the passage of a length of the glacier equal to
the distance from Iron Hill to Providence, is a matter of almost as
much doubt as the amount of erosion which it accomplished. Our only
possible source of information is found in the rate of movement of exist-
ing ice streams. ^Ve are, it is true, tolerably well informed as to the
speed attained at the extremities, and at various points on the surface,
of valley glaciers of the Alpine type. Such observations as have been
made on the larger ice streams in Greenland and Alaska show very
clearly that the glaciers of these countries move far more swiftly than
the better known streams of Switzerland and Norway. From the ob-
servations which have been made on the arctic fields, it seems not un-
reasonable to compute the motion of the Xew England ice at not less
than twenty feet per diem, or say at the rate of about a mile a year, or
from thirteen to fifteen years for the journey over the part of the train
which we have endeavored to subject to computation.

Although no kind of final value can be assigned to the results of the
computations above given, it seems to me that they serve to indicate
that the erosion accomplished by the ice while it lay upon the surface of
this part of the continent was probably effected with great rapidity.
The impression left upon the mind of the student who attentively con-
siders and carefully reckons the more computable form of wearing which
is brought about by scratching and polishing is to the effect that the
surface must have worn downward at an annual rate which is certainly to
be measured by inches, if not by feet. If after inspecting this evidence
he will follow the course of the boulder train which we have been con-
sidering, he will find that the quantity of the debris from the hill which
it contains forces him to a similarly high reckoning as to the rate of
the glacial wear. Even if from the data he obtains he should conclude
that the estimate of the peridotite in the train which I have made is five
or ten times too great, he will still be compelled to believe that the
down-wearing took place in an exceedingly rapid manner. jNIinimizing
the estimates in every possible way, in a manner which need not here

212 BULLETIN OF THE

be sot forth, I have not succeeded in making the amount of the down-
wearing less than an inch per annum.

The more accurate our knowledge as to the genesis of the topography
within the ice-worn region becomes, the more clearly is it proved that the
essential features of the surface are not due, as was former!}' supposed,
to the erosion effected during the Glacial Period, but are to be ascribed
to the ordinary agents of erosion which operated on this district during
the pre-glacial ages. Nowhere is this fact more evident than in tlic dis-
trict about Iron Hill. The surface of that field still discloses a drainage
system which in its main features is clearly very ancient. The valleys
have the normally digitated character which is characteristic of the work
done by fluid water, and though these depressions are everj'where more or
less modilied, and sometimes very greatly changed, by the erosive work
of tlie ice, the type of the topography is truly fluviatile, in this regard
differing from such characteristically glaciated districts as Labrador,
Scotland, or Scandinavia. Only the smaller tril)utaries of the streams,
those occupied 'by the lesser permanent brooks, have lost their valleys
by the process of glacial ei'osion. Although I have made numerous
efforts to secure some basis for a quantitative estimate, however imper-
fect, concerning the amount in depth of the material which was removed
from this district during the Glacial Period, I have not succeeded in
obtaining any data deserving consideration here. I can only state the
general impression made by a I'eview of the topography, which is to the
effect that the wearing brought about by moving ice cannot have
amounted to as much as an average of one hundred feet over the region
within a radius of thirty miles from Iron Hill. It is difficult indeed to
reconcile the hj'pothesis of even this amount of erosion with the remark-
ably well preserved details of the river work in this region.

The sliirhtness of the wearing which seems to have occurred in North-
ern Rhode Island is paralleled at many other points wliich are much
farther within the boundaries of the great North American glacier. I
shall note but two instances of the many which I could cite for the
purpose of showing that the erosive work accomplished during the last
Glacial Period was at certain points even less than I think it was in
the neighborhood of Iron Hill. In the region about Pittsfield, Mass., the
considerable areas of limestone rock there exposed retain the sink-holes,
or shallow pits, which are normally foi-mcd wliere calcific limestones
ai-e exposed to long continued weathering. Tliese depressions have
been filled with glacial waste, but the pits have evidently lost but little
of their original depth. In the same region the decayed mica schists

MUSEUM OF COMPARATIVE ZOOLOGY. 21

o

and other related metamorphosed rocks have not been planed away by
the glacier, but remain with the peculiar aspect which is commonly sup-
posed to be limited to the district south of the glaciated field. Again, in
the region immediately north of Kingston in Canada, a place situated
in what is supposed to have been the very heart of the great glacier,
the horizontal rocks of the Silurian Age retain their delicately incised
valleys, which were formed before the Ice Period, in a state of preserva-
tion almost as perfect as those formed in rocks of the same age and
character in Central Kentucky, in a district a hundred miles south of
the ice front. Here and there these valleys of the Kingston field are
somewhat embarrassed by accumulations of glacial waste, and at other
points the streams have made slightly deeper excavations in their old
paths, but on the whole the topography is substantially that which
existed before the advent of the glacier.

It is evident that, if Ave assume the rate of glacial wearing to have
been rapid, and yet at the same time the amount of effective work to
have been small, we are at once compelled to believe that the duration
of the cutting action was but brief. Along the margin of the ice the
condition of the frontal accumulations of debris at a number of points on
this continent leads us to the conclusion that the southernmost part of
the field occupied by the ice was tenanted by the glacier for but a short
time. Thus m the central parts of New Jersey the morainal accumula-
tions are generally slight, while the margin of the field occupied by
the ice in the northernmost point in Kentucky, though the indications
which point to the presence of the sheet are unmistakable, shows no
frontal moraine whatever. If these peculiar instances of slight wear-
ing were limited to the margin of the glacier, we could sufficientlv
account for the facts by supposing that a sudden forward movement of
the glacier had occurred, during which the fringe of the ice sheet occu-
pied for a very brief time an area which the climatal conditions did not
permit it to remain in. Such temporary excursions of the ice, though
on a smaller scale, have been frequently observed at the lower extremity
of the Swiss glaciers. Owing to the existence of such slightly worn
areas as we have noted in the interior portions of the American gla-
ciated field, we cannot account for the facts in the manner just indicated.
It appears necessary to suppose the existence of some conditions which
would permit the glacier to rest over a surface, and at the same time
prevent its abrasive action on the bed rocks.

Having been for some years engaged in preparing a series of maps
and reports on the surface geology of New England, I have been led to

214 BULLETIN OF THE

study a dozen or more parts of the field in which the evidence as to
the small amount of glacial wearing is particularly clear. These areas
are widely scattered, and from the additions which are constantly being
made to the list it is evident that they are numerous. While this
essay was in preparation, my assistant, Mr. J. B. Woodworth, discov-
ered a characteristic field of this nature in the region of hill land, on
the western border of Rhode Island, where the decayed schistose rocks,
the decomposition of which evidently occurred in pre-glacial times, have
not been removed by the action of the ice. At first I was disposed to
attribute the absence of erosion in these districts to some local arrest of
the ice movement, but a careful inspection of the localities has generally
disclosed the existence of areas of hard rock, which bore the normal
marks of glacial wearing, showing clearly that the ice moved in the
ordinary manner over the area. I therefoi-e felt compelled to frame
another hypothesis to account for the arrest of glacial wearing during
the greater part of the time when the ice sheet lay over the areas in
question. While I am still in much doubt as to the value of these sug-
gestions which I have to offer, I may say that they have withstood my
own criticisms and those suggested by several of my fellow students of
the phenomena for a period of ten years, and it therefore seems well to
offer them for more extended debate.

Hypothesis concerning the Conditions of Continental Glaciers.

We have already had occasion in the preceding pages incidentally to
note the effect of pressure in lowering the freezing point of ice, but it
appears to me that we have by no means exhausted the considerations
as to the conditions of deep glaciers which are open to us by the im-
portant discoveries as to the effects of pressure on ice which were made
by the brothers Thomson about forty years ago. I propose, therefore,
to review the matter, with the hope of discovering some explanation of
the arrest in the wear of the bed rocks which seems to have occurred
during the time when a thick ice sheet occupied the northern portion of
this continent. There can be no doubt that pressure melting operates
in an effective though slight manner even in the superficial portions of
an ice mass. The phenomena of regelation exhibited when two bits of
frozen water are pressed together, clearly shows the way, as has often
been observed, in which the conditions operate, and many other simple
experiments serve to indicate an action of the same nature. There now
appears to be little doubt in the minds of those who have inquired into

MUSEUM OF COMPARATIVE ZOOLOGY. 215

the facts of the ice movement in the Swiss glaciers, that pressure melt-
ing plays a considerable part in determining the movement of those
relatively small ice streams. So far as I am aware, however, no in-
quirers have endeavored to ascertain the efiect of pressure melting on
wide-spread and deep sheets of ice.

It is now tolerably clear that during the last glacial epoch a large
part of the field occupied by the continental glacier was buried to the
depth of about a mile beneath the accumulations of frozen water. If it
■were necessary for our purpose, it could readily be shown that the thick-
ness of the sheet was probably much greater than six thousand feet, but
the pressure which a mass no more than a mile in depth would bring
upon the surface of the earth would be sufficient to lower the freezing
point to about 30° Fahrenheit. We cannot ascertain at what tempera-
ture the accumulations of snow were built into the mass of the glacier.
There is, however, reason to believe that the initial heat. was not much
below the freezing point of water. It would not, however, militate
against the hypothesis to suppose that the mean annual temperature of
the surface of the glacier, and consequently that of the accumulating ice
sheet,- was as low as 25° or even 20°. AVe have next to note, that, with
the progressive deposition of snow, the layers formed each year would be
brought nearer to the bed rock, which process would lead to a constant
increment in the pressure which they sustained from the superincumbent
material. Thus the melting point of the ice would be progressively
lowered.

Not only does the progressive descent of the ice towards the bed rock
serve, through the influence of pressure, to bring the material ever nearer
the melting point, but with each stage of the down-going the particles
come nearer to that portion of the mass where several different causes
act together to produce a positive increase in temperature. There is
little doubt that the shearing movement of the ice due to the friction
of its mass upon the surface of the earth progressively, and at last very
rapidly, increases as we approach the base of the glacier. This inter-
stitial motion is necessarily attended by the conversion of a great part
of the energy of position of the mass into heat, which is^commuuicated
to the neighboring ice, and on account of the slight conductivity of the
material escapes towards the surface in a very slow manner. Next
the bed rock the actual friction of the ice upon the base over which it
moves, and the abrasion of the rock, convert yet more of the force
which leads to the motion of the glacier into heat. To these sources of
temperature we must add the slight but not unimportant -effect of the

VOL. XVI. — NO. 11. 3

216 BULLETIN OF THE

contribution of heat poured forth into the ice from the earth's interior.
All these actions tend to promote the liquefaction of the lower part
of the glacial envelope.

Although it is easy to perceive the existence of a number of efficient
causes tending to bring about melting in the lower portion of a conti-
nental glacier, it is difficult to form an adequate conception as to the
precise way in which these influences would operate. The facts, how-
ever, justify us in supposing that the temperature induced in the lower
portion of the ice would to a great extent be retained in the deeper
parts of the glacier. It is a well known fact that ice is a poor con-
ductor of heat, and therefore we may fairly assume that a considerable
increment of warmth would be likely to be brought about in the lower
part of the section, while the upper portion remained substantially unaf-
fected by the condition of the lower parts of the mass. The hypothesis
is, in effect, tha.t at a cei'tain stage in the development in thickness of au
ice sheet the portion of the mass next the bottom, while still remaining
below the temperature of 32° Fahrenheit, becomes converted into water,
or into very much softened ice, which cannot escape vertically or hori-
zontally froipa the field in which the melting was developed, but remains
as a fluid or semi-fluid sheet interveniug between the solid ice and the
surface of the earth.

The conditions of a mass of water at the base of a glacier, owing its
essential fluidity to the combined influence of pressure melting and
positive contributions of heat, is so peculiar, that it is necessary for us
somewhat carefully to examine into its state. At first sight it may seem
likely that such a mass of fluid would inevitably be urged by the pres-
sure of the superincumbent ice away from the field in which it was
formed, moving in the direction of least resistance, which would gen-
erally be towards the margin of the glacial field. It is evident that the
motion could not be in an upward direction, for in a deep glacier the
yielding nature of the material must prevent the formation of fissures
at any great distance below the surfjice. Even if such crevasses should
be made, they would quickly be closed by the pressure of molten water,
which would instantly freeze when it entered their free spaces. It is
more difficult to account for the hindrance to the movement of the
fluid toward the margin of the glacier. We must, however, conceive
that as soon as such a movement took place, and for the reason that it
did take place, the pressure molten water, having attained to a position
where less weight was imposed upon it, would quickly refreeze. It
would be aided in making a certain excursion towards the marj'in of the

MUSEUM OF COMPARATIVE ZOOLOGY. 217

field by the fact tluat it was urged forward by the pressure of the over-
lyiug ice, and this energy of movement would, to a certain extent, be
converted into heat through the frictions which the liquid encountered
on its journey.

The migration of pressure molten water towards the margin of the
ice would doubtless be somewhat restrained by the plastic condition
which occurs in ice when pressure melting begins. In the familiar
experiment which is made by subjecting a column of ice to com-
pression, we observe that the melting does not occur simultaneously
throughout the mass, but it begins along the planes of junction of its
crystalline or fragmental elements, films of water developing along these
planes, and gradually extending in width until the whole mass becomes
softened to the point where it loses its rigidity without becoming
generally fluid. It seems reasonable to conceive that the passage from
a sub-glacial area, where the water was melted at a temperature below
32° Fahrenheit, to a thinner part of the glacier, where the solid ice
rested on the ground, would be through a belt where the ice was in a
semi-fluid condition, which would serve through the frictions which
would there be engendered somewhat to restrain the flow. With these
preliminary suggestions as to the probable state of the bottom of a
deep glacier, we may now proceed to examine into certain corollaries
which may fairly be drawn from the main propositions.

As long as a glacier rests upon the bed rock in the form of ice, its
foundation seems necessarily subjected to intense erosive action, but as
soon as the ice next the bed rock is converted into pressure molten
■water, this wearing must cease, and the area would probably at once
become more perfectly insured from any form of erosion than any other
portion of the earth's surface. This exemption from change would
continue until, by a process of thinning of the glacier, its base was
permitted to return to the frozen state. It therefore seems possible
that where a deep glacier is developed upon any area we are likely
to have at first active erosion ; then a state in which wearing rather
suddenly ceases, because the ice thickens and becomes warmed, and
therefore melts in the manner before described ; and, last of all, with
the passing away of the ice, the thinned sheet may come again to
move over the bottom, and for a time to repeat the erosive work which
was discontinued while the ice retained a great depth.

In case pressure molten water were extensively developed at the
base of a great glacier, such as occupied the northern part of this con-
tinent, we should have to conceive the bottom of the ice, as regards its

218 BULLETIN OF THE

relations to the bed rock, divided into three zones. Next the maririn
there would be a belt occupied by completely frozen water, which lay
upon the bottom ; within this belt, a section where the pressures were
sufficient to pi'oduce only a partial melting or softening of the ice ;
and in the central part of the held, an area in which the ice rested ou
pressure molten water, or on ice which was made by the combined
action of pressure and heat so soft that it could not exercise any erosive
effect. I am not inclined to believe that this body of water, reduced
from the state of ice to the fluid or serai-fluid condition, would ever be
likely to become of any great depth. As soon as the measure of liquefac-
tion was brought about which would prevent the ice from holding firmly
to the bed rock, the heat due to the shearing motion of the glacier and
to the grinding up of mineral matter would no longer be produced. At
that stage I conceive that the motion in the inner parts of the field
which conveyed the annual rainfall towards the margin would in part
be affected by the gradual working out of the pressure molten water,
and in pai't by the squeezing of the softened ice near the base towards
the glacial front. Neither of these actions would serve to convert any
considerable part of the energy of position of the mass into heat.

It is commonly supposed that the immediate application of pressure
will serve to melt a mass of ice, even if its temperature be a degree or
two below the freezing point. Some experiments made under my direc-
tion by Mr. E. W. "Wood while a student in Harvard College have shown
that this is not the case.* If to such a mass of ice even a great pressure
is suddenly applied, only a small amount of water becomes melted : tliis
pressure molten fluid abstracts heat fi'ora the remainder of the ice in
such a measure that, if the pressure be rapidly accumulated so that
the ice has no chance to gain in temperature from without, we have a
result which apparently contradicts the hypothesis which is here pre-
sented. I see no reason to doubt that, if we could at once impose upon
a surface a glacier having the thickness of a mile and a temperature of
31° Fahrenheit, we should have but little indication of pressure melting
at the base of the ice ; but here, as elsewhere, the element of time and
the continuity of slight actions have to be taken into account. Reck-
oning with these, we perceive that the friction of such a hypothetical
glacier on the bottom, the grinding of the debris which it will produce,
and the vast amount of shearing action which would take place in the
particles of ice as they struggled over the surface, and by each other
for a great distance above the bottom, together with the heat poured

* See American Journal of Science, 1801, Vol XLI. p. 30.

MUSEUM OF COMPARATIVE ZOOLOGY. 219

forth from the earth's interior, would gradually bring a certain thick-
ness of the section into a state of more or less perfect fluidity, — into
a condition in which the mass would flow, though with much less ease
than fluid water, with such facility that in a slow movement it would in
no wise aff"ect the condition of its bed.

The experiments which we are able to make on the surface, either by
compressing a mass of ice to the point where a good deal of water
appears between its units of structure, or by mingling snow with water,
seem to indicate that we may have semi-fluid masses formed containing
enough ice to move with a certain speed, and yet, as far as erosion is
concerned, behaving like liquids. It appears to me likely that, while
in some of the deeper valleys below a continental glacier we might
have considerable masses of water in a state of perfect fluidity, the
greater part of the material, the cohesion of which was etFected by
pressure and heat, would, although the water would be water and the
ice ice, have as a mass the essential properties of a fluid. As this ma-
terial, ranging in its rigidity between water and ice, moved toward the
zones of diminished pressure, it seems to me that it would, through the
reduction of pressure, gradually acquire the normal resistance of un-
compressed ice.

The reader has doubtless already perceived the objection which I
find suggests itself as an iusuperaiile obstacle to the acceptance of the
hypothesis of the central part of the field of ice resting upon water
made more or less completely molten by pressure. He will ask how it
is possible that this fluid material is not at once driven forward in the
direction of the ice front to the point where, on account of the dimin-
ished pressure, it would become refrozen. To meet this point, we should
attend to certain considerations already presented, though in a some-
what preliminary way, concerning the conditions under which this
pressure molten semi-fluid is compelled to advance. It should not be
supposed that the central portions of the ice field rest upon a deep
sheet of pressure molten water, which would be eff"ectively urged
towards the margin of the glacier by the weight of superincumbent
material. We have to assume the depth of the ice in the neigh-
boring portions of the glacier which rested upon the bed rock not to
differ considerably from that which rested on the fluid material. A
very slight difference in the depth of the section would be sufficient to
bring about the change from the rigid to the mobile state. The con-
ditions would probably be such as to maintain these two parts of the
ice field in a delicate adjustment of their depths. As the central area

220 LULLETIX OF THE

thickened, thereby iucreasing the amount of pressure molten water, a
portion of the fluid would be squeezed under the ice of the peripheral
zone, thei-eby augmenting the thickness of its section, and at the same
time thinning the ice in the central area. In this manner we cani con-
ceive the creation of a balance in the impulses and resistances affecting
the movement of the ice which was softened by pressure melting, so
that the drift of the material toward the margin of the continental
glacier would be slow and uniform.

The conditions of our hypothesis require us to suppose that the effects
of pressure melting would first be felt in the deeper parts of the glacier,
those portions of its mass which lay in the valleys, and that the soften-
ing of the ice might there be completely effected while the frozen water
was still in contact witli the earth at the higher levels of the surface.
It thus might well happen that the considerable elevations of the coun-
try, those hard parts which had survived under tiie conditions of ordi-
nary land erosion, would be much more effectively worn down than the
rock beneath the river valleys. We can thus account for the destruction
of such a prominence as Iron Hill, which w^as probably a sharp peak of
considerable altitude when the glacier began its work, while the neigh-
boring valleys were but little worn by the action of the glacier.

So long as a glacier is receiving a considerable annual contribution of
snow which is built into its mass at a low temperature, it may well be
that the accumulations of heat due to the work done near the base of the
ice would not affect any considerable portion of the central section of the
mass. If now for a time the annual snowfall diminished to the point
where, by the thinning of the glacier, pressure melting ceased to take
place, the vrhole section of the ice might gradually acquire a relatively
high temperature, so that any sudden increase in pressure might bring
about very extensive melting. If in this condition of the deposit the
amount of snowfall should, for a number of years, be greatly increased,
the result might be a great development of pressure molten water, which
would be pushed forward towards the margin of the glacier to the point
■where, owing to the diminution in the thickness of the ice, it could be-
come refrozen. In this way we may perhaps account for those sudden
and temporary advances in the margin of the glacier which are so clearly
indicated at various points in this country.

No direct verification of the hypothesis above deduced is to be ob-
tained by observation or experiment. The only approach to proof which
we can hope to secure is by an inspection of the facts exhibited in the
records of glacial action with a view to ascertaining how far they may

MUSEUxM OF COMPAKATIVE ZOOLOGY. 221

be explained by the suggestions which we have been considering. Be-
ginning this comparison with the Greenland glacier, the only field where
we can find conditions approaching those which existed in the greater
ice field of the American continent, we note the following facts. In
the marginal portions of the Greenland ice field the slope of the surface
towards the sea is tolerably steep, and is rent by numerous crevasses.
Gradually, as we pass from the frontal portions of the ice to the interior
of the field, these crevasses disappear, and the slope of the glacier be-
comes slight and unbroken. On the crest there is a wide field where
the glacier has the character of a great plain with a slope so slight that
we cannot well conceive the movement towards the margin as taking
place over the surface of the underlying rocks in the manner in which it
occurs near the borders of the sea. These conditions are reconcilable
with the assumption that the central part of this great glacier rests upon
ice which has been softened by pressure to the point where it no longer
behaves with its normal rigidity, but acts substantially as a fluid, while
in the peripheral section, that which is beset with crevasses, we have to
suppose that the glacier rests upon the bed I'ock.

Turning now to the conditions of the area on the mainland of this
continent, so far as they were effected by the ice of the last Glacial
Epoch, we may briefly review the features which are explicable by the
hypothesis which we are considering. We note at the outset the fact,
to which the reader's attention has already been directed, that the ero-
sion accomplished by the ice in the interior of the glaciated field is often
very small. We may now extend this statement by saying that the
wearing which has occurred in the central portions of the area occupied
by the ice bears no kind of proportion to the depth to which the sheet
evidently attained, or to the length of time which it must have remained
on the surface. If space permitted, it would be possible to bring up an
extended array of instances, such as that cited from the region north of
Kingston, Ontario, where in districts in which the glacier must have
been very deep and long enduring the erosive work was less than in the
marginal parts of the field. These facts do not seem to be explicable on
the supposition that the glacier wore the surface over which it lay in a
measure at ail proportionate to its depth or the continuity of its action.
If, however, we suppose that only the marginal zone of the ice prevail-
ingly rested on the surface of the earth, and that a great part of the field
lay upon a fluid or semi-fluid stratum of water, the difficulties which we
encovmter are cleared away.

It seems impossible to explain the motion of a continental glacier on

222 BULLETIN OF THE

the supposition that the ice throughout the field rests upon the bed rock.
Under these conditions it appears necessary for its surface to have a
slope towards the margin of some degi'ees of declivity in order that the
sheet may be impelled downward with sufficient energy to overcome the
great resistance due to its friction on the bed rock. A slope sufficient to
accomplish this purpose would require an inconceivable thickness of ice
in the central part of the Xorth American glacier. The hypothesis of
pressure melting shows us a way out of this difficulty. We have only
to conceive the central parts of the area of the glacier to be freed from
the basal friction, to avoid the need of hypothecating a considerable
slope of the surface except near the margin of the ice. In this view,
the element of friction on the bed rock is substantially reduced to a belt
of limited width into which the ice is fed from the areas where pressure
melting occurs.

The sudden advances and recessions in the position of the glacial
front can be better accounted for on this hypothesis than in any other
way. A slight increase in the pressure in the central portions of the
field, such as might be brought about by an increased snowfall extending
over a term of a few years, would probably lead to the discharge of
water rendered more or less fluid by compression into the marginal
portions of the area. This would naturally be attended by a sudden
outward march of the ice. In this way we may explain the prevailingly
wide fringe of territory in the Mississippi Valley which lies to the south-
ward of the southernmost distinct moraine, and which appears to have
been temporarily occupied by the ice sheet. This district is covered
by a layer of glacial waste, but at its outer margin we find none of
those accumulations of detritus wfiich indicate the permanent occupation
of a line by a "glacial front.

It appears to me that we may by the hypothesis of pressure melting
explain the formation of those very thick deposits of till which occur in
certain parts of the glaciated area, and this in the following manner.
Until a glacial sheet has accumulated to such a depth as to bring
about pressure melting, the combined erosion of the bed rock and the
irregular movement of the ice near the surface over which it moves
bring about the admixture of rocky material with the frozen water to
the depth, it may be, of some hundred feet above the earth. If now
pressure melting begins, the debris will gradually drop upon the surface,
and this action will continue until perhaps all the detritus previously
intermingled with the ice has become separated from it. If from time
to time the glacier became so far thinned that its solid parts again

MUSEUM OF COMPARATIVE ZOOLOGY. 223

rested on the bed, the result might be the formation of those striated
pavements which have been observed in till deposits. It appears to me
not improbable that in the end we may be able to account for the for-
mation of drumlius, those most puzzling of all glacial deposits, by the
action of pressure upon ice, the compressive action operating in the
following manner.

Let us suppose that a glacier such as covered the drumlin field of
.Southeastern New England had acquired in the process of its movement
a great store of rock detritus, distributed through several hundred feet
of the ice which lay next the earth. Let us further suppose that,
through the thickening of. the sheet combined with the development of
heat near its base, this debris-laden part had been brought to the critical
point where very slight increments of pressure would bring the impris-
oned water to the fluid state, and lead to the precipitation of the mineral
matter, the result would be the rapid formation of a till sheet. Wher-
ever, through the existence of irregularities on the surface of the earth,
projections existed of sufficient height to rise into the glacier a little
above the level at which complete pressure melting occurred, the ice in
its motion would be subjected to a certain amount of strain as it moved
over the elevations. As, according to the supposition, the water of the
glacier was very near the point of fluidity, we may well conceive that a
very trifling resistance, in amount insufficient to exercise any distinct
erosive effect on the mass of till, might cause still further melting, and
thus bring about an increase in the deposit of debris. In this way the
growing drumlin would rise up into the ice to the point where detritus
ceases to be supplied, or perhaps to the level where the resistance of the
glacial material was sufficient to bring about erosion. Even if the mass
did not at first have the shapely lenticular form proper to these eleva-
tions, it would, during the subsequent thinning of the ice which probably
everywhere preluded the disappearance of the envelope, be eroded to the
arched shape which characterizes the deposits.

As I propose in this essay only to indicate in a general way the pos-
sible value of the hypothesis above set forth, I sliall not undertake
further to discuss the explanatory value of this view. Enough has been
set forth to show that, if it proves tenable, it may serve to rationalize
our views as to the mode of action of continental glaciers, by extending
our conceptions as to the conditions under which they do their singu-
larly important work. As the considerations which have been adduced
are to a certain extent novel and somewhat difficult to grasp it seems to
me well in closing to submit them to a brief review. Leaving out of

224 BULLETIN OF THE

account the minor propositions, we may make the following condensed
statement.

In the growth of a glacial mass, the snow is built into it at a tem-
perature below the freezing point, and each annual contribution is ever
brought nearer to the surface of the earth, and tends to become molten
by pressure. Effective melting near the base of the ice is probably
secured by the conditions which make for the development of heat at
that level. It is highly probable that, when the ice has attained a depth
of a mile or two, its lower part is either converted into water or so far
softened that it ceases to be an eroding agent, and may be forced to
move in essentially the manner of a fluid towards the zone of less re-
sistance. Arriving at a point where, owing to the thinning of the ice,
the pressure is sufficiently diminished, this water gradually refreezes and
is rebuilt into the firm glacier, and as such pursues the remainder of its
journey. We have thus to conceive a deep glacial envelope, such as that
which now covers Greenland, to be divided into two realms ; a central,
in which the ice does not come in contact with the surface of the earth,
and a peripheral, in which it exercises the familiar erosive action on the
bed rock.

During the development of a continental glacier, until the sheet had
attained a thickness at which the pressure melting action would begin,
the whole of the mass would rest upon the surface of the earth. As
the inner parts of the field attained the depth which would cause the
ice next the ground to become softened or melted, the erosive work
would be limited to the peripheral zone. With the further increase in
the profundity of the glacier there would be a tendency, rapidly to push
outward the peripheral parts of the accumulation where the glacier
rested on the bed rock. When, in the closing stages of the period, the
ice sheet thinned, this zone of erosion would gradually be withdrawn to-
wards the centre of the field, or towards the point where the glacial
conditions lingered longest. In this way we can account for a long con-
tinued sojourn of the ice in the fields which we know it occupied, with-
out being required to suppose that the aggregate erosion was very large.
If the width of the peripheral zone were, say one hundred miles, and
the distance from the centre to the farthest point to which the ice ex-
tended one thousand miles, the time during which the eroding zone
occupied any part of the surface may have been but a small portion of
the duration of the Glacial Period.

The hypothesis of pressure melting enables us to account for various
peculiarities of glacial movement which cannot otherwise be readily

MUSEUxM OF COMPARATIVE ZOOLOGY. 225

explained. It seems likely to solve the euigtnus presented by the very
sudden variations of a temporary nature in the position of the ice front.
It appears to explain the way in which the ice journeyed for gi-eat dis-
tances over surfaces of slight inclination in the direction of glacial flow,
or which sloped towards the centre whence the glacial movement radi-
ated, for it limits the friction to the probably nan'ow zone where the
glacier rested upon the earth. The hypothesis will clearly account for
tlie small amount of erosion which is often traceable in the regions which
lay in the central parts of the glaciated district, and therefore beneath
the deeper parts of the accumulation ; for in that part of the field
pressure melting was probably first established, and must have continued
for the longest time. It furthermore bids fair to explain the very puz-
zling phenomena exhibited by drumlins or lenticular hills, by showing a
way in which, through the thickening of the ice, the rocky matter which
it had taken up from the bed rock might be rapidly deposited in the
form in which we now find it.

The only justification for presenting such highly speculative consid-
erations as are offered in this writing is that they may serve to explain
phenomena which, in the present state of our knowledge, cannot be
otherwise rationalized. It seems to me, therefore, that this view may
fairly be submitted to debate. I am by no means sure that it can
withstand the criticism which it merits, but it seems to me worthy of
in(piiry on the part of those who are well trained in the interpretation
of physical phenomena.

Q

<

-1

■X

M

D
U

^ s

•-' z

o

X

MAi'oi'Ai'Ainoi'' Massachusetts axd Rhode Island, siiowim. posiiiox ornii-;

Toiio',;rnj)ln fronv U S Cteolo;^icnl Sur\'cr 100 H Coiiloiu >[a]»

4

h

r 51:; : .,ii? mr'm, i

- ' : ■■ - .' ' • .* * ■ P \r iIdl lUV .1 ill J

•\^

n-*

I

^#

^-^ ^5r »^.^

^^ ^J'uM

c -.,.. ''\

•i 3?i;
*" f- ,-.

»• 1 :.'r •

1

N^

1

-! r.

PC

Q

Z

D

<

CC

J

<:

cc

!?

UJ

<

S

W

ID

CT

u

2

1— 1

id

J
J

O
O
►J

n:

z

o

d:

Xo. 12. — Contributions from the Pefrograpliiccd Laloratory of the

Harvard University Musenm.

Acmite Trachyte from the Crazy Mountains, Montana. By J. E. Wolff

AND R. S. Tarr.

In the prorji-ess of the field investigation of the eruptive rocks of the
Crazy Mountains, Montana, by one of the writers/ in 1889, the occur-
rence of a group of eruptive rocks was noted, which were distinguislied
from the theralites and associated rocks by their gray or greenish gray
color and somewhat greasy lustre, and of which about a dozen specimens
from as many localities were collected for further study.

Field Occurrence. — They were only found in the northern half of the
range, in general associated with theralite, and occurring like the lattei-,
and so far as known all the other eruptives,^ solely as intrusive rocks.
The field types can convenient!}' be classed under three heads : 1st. Dikes
cutting the Cretacous (Laramie T) shales and sandstones ; 2d. Small in-
trusive sheets parallel to the bedding; 3d. Thick, bulging, laccolitic
sheets, which may send apophyses into the adjacent shales. In tlie last
form, sheets of the rock have been observed which are a hundred feet
thick at the bulge and a mile long, and, conforming with the tilted
position of the strata, produce long high ridges with sloping back and
steep front. As with the theralite, diorite, and other eruptives of the
range, the rock is coarse, almost granitic in the thick sheets, fine-grained
and porphyritic in the smaller sheets, dikes, and apophyses.

Acmite-trachyte Type. — AVhen occui'ring in the latter forms, the rock
is of a fresh green to grayish green color, with a somewhat greasy lustre
and a conchoidal fractui'e. Glassy feldspar ciystals 5 or 6 mm. long-
give it a porphyritic character, and smaller augite phenocrysts are com-

* No. IV. Metamorpliism of Clastic Feldspar in Conglomerate Schist, by .J. K.
Wolff.

1 J. E. Wolff.

•■2 Ibid., "Geology of tlic Crazy Mountains," Bull. G. S. A., Vol. III. p. 44-3.

VOL X\I — NO. VI.

228 BULLETIN OF THE

nion ; in some specimens the rock is dotted with small white crystals of
sodalite. In one instance the rock holds large phenocrysts of flesh-
colored feldspar in tabular prisms with tei'minal planes, and is porous
owing to the presence of small irregular cavities, whose rough walls are
lined witli limonite. Plates of Liotite are rarely seen. There is a dis-
tinct fluidal structure in the dikes and sheets which is visible in the
hand specimens by the parallel aiTangement of the phenocrysts and a
slight fissility parallel to the plane of flow.

Characters in. Thin Sections. — Feldsjmr. — In the thin sections of this
type the feldsjiar phenocrysts appear in long rectangular sections, or in
broader and less regular crystals. Carlsbad twins are as frequent as
single crystals. A certain number of these feldsjiars polarize homoge-
neously, and extinguish parallel to the length of the rectangular sec-
tions, or nearly so, like sanidine ; but in many sections a very fine
multiple twinning is present parallel to the brachj^pinacoid (albite law),
and sometimes there is present with this a faint double twinning (mi-
crocline structure), or segments are multiply twinned in one direction,
whicli is not the same in adjoining segments of the feldspar, ("leavage
sections were jirepared from one of tliese phenocrysts with the following
result : —

Basal cleavage (0 P) extinction 2i° to 3° oblique to the trace of the

2d cleavage (qq P ^ ).
Second cleavage (oo 1' oo ) extinction about 9° oblique to the trace of

the basal cleavage.
Specific gravity of the piece 2.583.

This specimen is therefore a triclinic feldspar intermediate between
microcline and albite (soda-rnicrocline — anorthoclase) and the anomalous
polarization of the feldspar sections in the rock is thus explained. These
phenocrysts contain as inclusions apatite, sodalite, augite, fcgirine, and
biotite. The pegirine needles appear to have formed toward the close
of the feldspar period, as they are commonly included in the peripheral
portion of the crystal.

Sodalite is present in some specimens, occurring either in crj'stals
large enough to dot the rock macroscopically, or only visible in tlie
section. It appears in tlie slides in the usual apparent hexagons or
squares, sometimes rounded or corroded by the groundmass. The min-
eral is colorless, with a low refractive index and isotropic, unless zeo-
litization has occiu'red. It sometimes contains secondary calcite. Tlie
mineral gelatinizes easily witli acid, and the micro-chemical tests sliow

MUSEUM OF COMPAKATIVE ZOOLOGY. 22'J

the presence of CI, Xa, SO3, and a little Ca ; so that there is a mixture
of the hauynite molecule. The speciHc gravity of an isolated piece was
2.19, which is low for sodalite (2.28), and must be attributed to zeoliti-
zation. W. Lindgreu has described,^ under the name " analcite-basalts,"
basic rocks from tlie Highwoods (a range lying north of the Crazy
Mountains and a locus of similar magmas), composed of olivine, mag-
netite, a little biotite, and a mmeral determined as analcito, but no feld-
spar or nepheline. The analcite appears in the slides in hexagonal or
octagonal sections, clear and isotropic, and was determined to be such
by specific gravity and chemical analysis. Lindgren considers this pri-
mary. Judging from the descriptions, there is some resemblance to the
mineral identified by us in the present rock, without quantitative anal-
ysis, as sodalite ; but as the latter has the form and other properties of
sodalite, and is identical with that mineral as occurring in the granular
and porphyritic theralite, this seems the correct determination.

Aurjite. — Tlie porpl)yritic augite has planes developed in the pris-
matic zone, sometimes terminal, and the common orthopinacoidal twin-
ning. In sections the pale green of the centre of the crystals gradually
increases in deiitli towards tlie edge, which is formed of deep green
feo-irine, as in the theralites, havinii- the characteristic small obliquity of
extinction in the prismatic zone and axis of elasticit}', n, near the vertical
axis. The augite is generally fresh, and includes crystals of magnetite
and sodalite.

The slides contain occasional plates of biotite and grains of magnetite.

Groiindmass of the Acmite-trachi/le Type. — This is composed essen-
tially of slender lathe-shaped feldspars and acicular crystals of a^girine,
whifli lie either between the feldspars or included in them, and by
preference in the outer poi'tion. These minerals appear in sections
transverse to the flow with perfect fluidal arrangement ; but when cut
parallel, many of the feldspar sections are broad and rounded, and contain
the a'girine needles in a network without parallel arrangement. The
feldspars are therefore somewhat tabular in the plane of flow. Single
crystals, Carlsbad twinns, or multiply twinned crystals occur, with ex-
tinctions parallel or but slightly olilique to their length, and are doubt-
less also anortlioclase in part. With the green a'girine a few brown
needles of acmite occui\ Apatite and magnetite are found in the usual
form. The combination of nearly parallel feldspar and ajgirine needles
gives the former a peculiar feathery look in polarized light with low

1 " Eruptive Kocks from Montana," Proc. Cal. Acad. Sci., Vul III. p. 51.

230 BULLETIN OF THE

There is a variable amount of interstitial matter between the feld-
spars of the groundmass which fills the triangular spaces left by their
divergence, or appears as irregular ai'eas in sections parallel to the flow.
This substance is colorless, has a low index of refraction, is sometimes
fresh and glassy, sometimes clouded by a fibrous substance, showing
rarely traces of a rectangular cleavage, and containing scattering regi-
rine needles. In some cases it polarizes so strongly as to be evidently
a tabular feldspar section (or a third generation of feldspar), but gen-
erally polarizes feebly, or is completely isotropic, and then gelatinizes
with acid. The feebly polarizing part is probably nepheline, and the
isotropic clear areas analcime, derived by alteration from the nepheline.
The a;girine needles occurring as inclusions in the nepheline or analcime
can hardly be regarded as secondary, since they are identical in size
and parallel or network arrangement with the wgirine needles so abun-
dant in and between the feldspars, and evidently a primary constituent.
Brogger ^ describes undoubted cases of secondary segirine in analcime,
and J. Francis Williams - a3girine needles in the analcime of the " gray
granite " of Arkansas as secondary, similar in occurrence to those of the
Montana rocks. The interstitial element varies greatly in quantity, and
may become so considerable as to give the rock a phonolitic character.

All the specimens have the t3-pical tracliytic structure.

(El<eolite) — SyenUe Type. — This coarse variet}' occurs in tlie thick
sheets. The rock has a gray color, passing into wliite as the decompo-
sition of the feldspars increases, and has a tendency to porphyritic struc-
ture. The feldspar phenocrysts, unlike tliis mineral in the groundmass,
are in part fresh and glassy, — a fact which assists the optical determi-
nation,— and have a fine striation on the basal cleavage. The minerals
have an indistinct parallel arrangement, due to flow. The rock is al)out
half as coarse as the "gray granite" (ela3olite syenite) of Fourche
Mountain, Arkansas, which it resembles.

Character in Thin Sections. — In thin sections the structure is panidi-
omorphous, the angular spaces between tlie feldspars being occupied
by nepheline. The large feldspars are glassy clear, having the peculiar-
ities of twinning previously described. The outer zones are sometimes
opaque, owing to decomposition, and filled with regirine needles. The
extinction angles on basal cleavage sections (0 P) (Specimen No. 145)
were 2° to 4° to the trace of the second cleavage ; and on second cleav-
age sections (go 1* oo ) fi'c^i ^° to d],° oblique to tlie lirst cleavage. The

1 Mincralion i1. Syciiitpofimatitganfre. p. '■]?,().
- Igneous Ivocks nf Arkansas, pp. (iS ami 70.

MUSEUM OF COMPAllATIVE ZOOLOGY. 231

specific gravity in most determinations was 2.621, but went as high as
2.623. A sufficient quantity of this feldspar was carefully selected,
freed from visible impurities, and ana\j'zed in the laboratory of the
United States Geological Survey at Washington by Dr. W. F. Hille-

following result : —

SiO.^

G2.31

AI0O3 (containing a very little iron)

22. G3

CaO

.63

SrO

.57

BaO

.77

KoO

4.79

NagO

7.68

H2O at 100° C.

.16

HP above 100° C.

.72

100.26

The optical characters show the triclinic nature of this feldspar, and
an apparent homogeneity, even with high powers, excludes its reference
directly to a microscopic mixture of microcline and all)ite (microcline-
microperthite of Brogger) ; it appears to belong in the anorthoclase
group of Rosenbusch (soda-microcline of Brogger). The per cent of
strontia and baryta is \inusual, and only comparable to the baryta and
strontia sanidin from the nephelinite from Meiches, analyzed by Knop.^

The smaller feldspars occur in long lathe forms, and are more decom-
posed than the others ; in decomposing they become opaqne and fibrous.
The augite crystals are similar to those of the fine-grained rock, and
have the same segirine border. Independent acicular crystals of segirine,
and sometimes of acmite, also occur. The angular spaces between these
minerals are occupied generally by a feebly polarizing substance, which
gelatinizes with acid, and is evidently nepheline. In decomposing, it
breaks up into strongly polarizing fibrous zeolitic aggregates. Sodalite
is rare in the coarse rock, except in the apophyses, or near the contacts
of tlie sheets, where it occurs in small crystals between the feldspars.
The coarse rock under these conditions assumes the acmite-trachyte
character of the dikes and smaller sheets.

The following analyses of these rocks are presented here, but the
discussion of their relations to the other alkaline rocks of the Crazy
Mountains is deferred to the monograph in preparation. Nos. 65, 131,
and 297, represent the Acmite-trachyte type, and No. 145 the Elseolite-

1 N. J. Min., 1865, p. 688.

232 BULLETIN OF THE

syenite type. These analyses were made by Dr. W. H. Melville in tlie
laboratory of the United States Geological Survey at Washington. For
comparison, Analyses I. of Theralite/ II. EliBolite-syenite ^ (" grnv
granite," Fourche Mountain, Arkansas), and III. Acmite-trachvte ^
(from the Klihlsbrnnnen Siebengebirge, Germany), are introduced for
comparison : —

65.

1.31

297.

145.

I.

II.

III.

SiO.,

58.70

G2.17

64.33

59.60

43 17

.59.70

64.21

AI2O3

19.2G

18 58

17.52

16.97

15.24

18.85

1G.98

Fe.Oa

3.37

2 15

3.06

3.18

7.61

4.85

6.69

FeO

0.58

1.05

0.94

1.15

2.07

MnO

0.10

tr.

0.35

0.19

CaO

1.41

1.57

0.56

2.32

10.63

1.34

040

MgO

0.76

0.73

0.34

0.80

5.81

0.08

0.18

NaaO

8.55

7.56

7.30

8.38

5.68

6.29

5.13

K2O

4.53

3.88

4.28

4 17

4.07

5.97

4.41

TiO.,

tr.

tr.

tr.

tr.

P.O.

0.10

0.11

tr.

0.14

Dss at 105°

2.57
0.07

1.63
0.07

0.95
0.04

2.53
0.07

3.57
SO,
0.94

1.88

1.00

100.00 99.50 99.67 99.56 99.39 99.56 99.09

Comparison with other Aanite-trachijtes. — In chemical and mineralogi-
cal composition and habitus, the fine-grained rocks are almost identical
with the classical acmito-trachyte from the Kiihlsbriinnen in the Sie-
bengebirge. The microscopic characters of the German rock are given
in Rosenbusch, (Mik. Physiog., Vol. II. p. 599,) where it is stated that
the rock, when weathered, is filled with peculiar round pores, which do
not exist in the fresh rock, but are there represented hj areas of a
brownish yellow isotropic, or partly cryptocrystalline substance, which
is occasionally developed in radially built spherulites of positive charac-
ter, a single spherulite occupying the space of a subsequent cavity. In
slides from a vei'y fresh specimen of the German rock collected by one
of the writers, the yellowish brown color of these areas is very faint or

1 J. E. Wolff, Petrography of tlie Crazj- Mountains, 1885.

2 Williams, he. cit., p. 81.

•' G. Bischof, in Von Declicn's Geogn. Besclir. d. Siebengebirges. Ver. d. Preuss.
Uli. und West., VA IX p 310.

MUSEUM OF COMPARATIVE ZOOLOGY. 2^^j3

lacking ; they arc in some cases isotropic, have often a polygonal shape,
and the acmite needles, which are abundant in the rock, arrange them-
selves parallel to their sides when in proximity ; they also gelatinize
strongly with acid, and thus resemble the small irregular sodalite ci'ys-
tals of the Montana rocks. The feldspar phenocrysts of the Siebenge-
birge rock have the triclinic twinning described above, and thus the two
rocks are nearly identical, at least for the American variety with little
nepheline.

The Sijenitic type resembles in appearance and structure the elreolite-
syenite ("gray granite") from Arkansas, described by J. F, Williams,
in chemical and mineralogical character. It is closely allied to the Mon-
tana rock, excepting that it has more nepheline, and that the feldspar
was referred by Williams to the microcline-microperthite of Brcigger (a
microscopic interlamination of microcline and albite), while the rea-
sons are given above for considering the feldspar of the Montana rock
a microscopically homogeneous triclinic soda-potash feldspar. Lind-
gren {loc. cit.) has described as '' augite-trachytes " rocks from the High-
Avood Mountains closely resembling these.

The previous descriptions illustrate the dependence of rock structure
on physical conditions of cooling, which is so striking a feature of the
eruptive I'ocks of this range, the syenitic or trachytic character of the
I'ock depending on the variation in the thickness of the rock mass.

Camisridge, Mass., January, 1893.

No. 13. — Reports on the Dredging Operations off the West Coast of
Central America to the Galapagos, to the West Coast of Mexico,
and in the Gxdf of California, in charge of Alexander Agassiz,
carried on hy the U. S. Fish Commission Steamer "Albatross,"
Lieut. Commander Z. L. Tanxek, U. S. N., Commanding.

[Published by permission of Marshall McDonald, U. S. rish Commissioner.]

V.

Ueport upon Rocks collected from the Galapagos Islands. By George P.

Merrill.

So far as the present writer is aware, the volcanic roclcs of the
Galapagos Islands have been the subject of but little investigation.
Darwin in his "Voyage of the Beagle " (pp. 372 et seq.) describes the
islands as volcanic, and the character of the material as black basaltic
lava with scoria and tuffs. Naturally, his studies did not at that time
include an examination of thin sections under the microscope. A more
detailed description of the rocks of these islands is given by Dr. F. A.
Gooch, in Tscherraak's " Mineralogische Mittheilungen" for 1876 (pp. 133
et seq.). x\ccording to the latter, the volcanic materials examined by
him are all of a basaltic character, and in part closely correspond to
those at present under consideration. Dr. Gooch's specimens were how-
ever from the islands of Bindloe, Abington, Hood, and Cliarles, none
of which localities are represented in the series received from the Fish
Commission.

A reference to this paper is however of importance. How well the
speicmens now examined represent the sources from which they were
taken, the present writer has no means of deciding. It is assumed that
they are fairly typical.

VOL. XVI. — NO. 13.

236 BULLETIN OF THE

Of the numbers given below, the first is that of the specimens as
now entered upon the Museum Catalogue, and the second, enclosed in
parentheses, is that by which the specimen is known in the records of
the Fish Commission.

67526 (8110). Basalt. Wreck Bay, Chatham Island : near haci-
enda of Mr. Cobos. This is an ordinary basalt, consisting of a dark
gray vesicular mass, which in thin sections is seen to be made up of faint
wine-colored augites in very irregular forms, lath-shaped plagioclases,
small opaque particles of iron ore, and numerous large, clear, faintly
greenish blebs of olivine. 'No. 67527 (8111), from the same locality,
offers no difference worthy of note.

67528 (8113). Basalt. Landing on northwest face of Malpelo Island.
This rock much resembles the last in mineral composition and structure,
with the exception that extensive alteration has badly obscured the
groundmass, and given rise to abundant chlorite, epidote, calcite, and
sundry ferruginous decomposition products. The presence or absence
of a glassy base cannot with certainty be determined. In the hand
specimen this is a finer grained, more dense rock than 67526 or 67527,
and shows small amygdaloidal cavities filled with a white zeolite. The
general aspect is that of an older rock than the others (perhaps a mela-
phyr) though obviously it will not do to speak too positively from an
examination of the section alone.

67531 (8116). Andesite (1). Near Chatham Bay, on Cocos Island.
In the hand specimen this is a light brownish gray, fine-grained, in
some cases vesicular rock, studded with small (1-2 mm.) white specks
indicative of feldspar, but in which none of the constituents are of such
size as to be accurately determined by the eye alone, or even when
aided by the pocket lens. The thin section under the microscope
shows a dense aggregate of stout feldspars which are in part sanidins
and in part a soda-lime variety, obscurely striated, and occurring in
stout lath-shaped forms. The structure is indistinctly granular, and
occasionally slightlj' porphyritic through the development of the larger
feldspar above noted. So far as observed there is no interstitial glass.
Abundant red ferruginous and opaque hornblendes, granules of iron ore,
minute colorless apatites, and ferruginous decomposition products com-
})lete the list of determinable constituents.

MUSEUM OF COMPAKATIVE ZOOLOGY. 237

A rough analysis yielded results as below. This indicates that the
rock is more nearly related to the andesites than trachytes, though
abnormally rich in potash. The rock vjas found to contain a notable
amount of free sodium chloride, derived presumably from ocean spray,
which may in part account for the high pei'centage of soda.

Per Cent.

SiOj

56.50

AI2O3, Fe^Og

28.20

CaO

2.83

MgO

.98

K,0

4.25

NagO by difierence

6.68

99.44
Washington, D. C, October, 1892.

Xo. 14. — Contributions from the Petrograpliical Laboratory of the
Harvard University Museum.

VI.

A Basic Dike in the Connecticut Triassic. ^ By L. S. Griswold.

Occurrence.

Loose boulders of the rock to be described in this paper have been
known for some time to Mr. S. "Ward Loper, but the rock was first found
in place by Mr. J. C. Graham of Wesleyan University while at work for
the United States Geological Survey. It occurs as a dike, exposed on
the outlet of Beseck Lake, about a quarter of a mile west of the Air
Line Eailroad at Baileyville. The dike intersects the shale just under-
lying the "posterior" or uppermost trap flow of the Triassic series of
Connecticut. A small fault of about six inches was noted in the dike.

Physical Characters.

The rock is dark colored, almost black, and dense. The specific
gravity (average of three determinations) is 3.036. It attracts attention
by the gi'eat number and large size of the black phenocrysts contained,
also to some extent by the small spots of a white mineral. The rock is
very fresh immediately below the weathered surface.

Examined with a magnifying glass the phenocrysts are found to be
of augite, hornblende, and an occasional large biotite. The maximum
diameter of the phenocrysts is perhaps f inch (19 mm.). The spots
of white mineral give a brisk effervescence with cold hydrochloric acid,
so they represent secondary fillings of cavities by calcite. The calcite
frequently occurs in the phenocrysts as a decomposition product.

Microscopical Examixatiox.

With the microscope two divisions of tlie components of the rock can
be made: the phenocrysts, comprising perhaps a third of the total mass,

1 Published with the permission of the Director of the United States Geological
Survey.

VOL. XVl — NO 14.

240 BULLETIN OF THE

and a fine groundmass. Among the phenocrysts angite greatly exceeds
hornblende, biotite is rare. The hornblende crystals are brown and
strongly plechroic ; they have lost their crystal angles and show rounded
resorption outlines. The same is the case with the few pieces of biotite
seen. With the angite, however, it is different ; this mineral appears
commonly with complete or partial crystal outlines, though some show
the rounded resorption form. It is evident that most of the angite
crystals were formerly rounded and have gained the angular form by
a later growth in the magma, for irregularly rounded cores can be dis-
tinguished surrounded by portions having different extinction angles.
The structure thus given closely resembles zonal structure ; zonal
structure does sometimes occur in these secondary borders. Tlie older
portions of the crystals are pale green in color, the borders are pinkish;
in one case the pink border is pleochroic, pink to greenish yellow.

The phenocrysts of hornblende and angite may each contain rounded
inclusions of the other mineral, inclusions of the angite in the hornblende
being much more common. Thus there seem to have been five stages
in the crystallization of this magma: first, a time when hornblende and
angite formed in good-sized crystals ; second, these crystals were resorbed
until they became rounded grains ; third, another separation of large
crystals of angite and hornblende, which often enclosed indiscriminately
and without crystallographic relation the grains above mentioned; fourth,
this second generation of angite and hornblende underwent resorption
sufficient to destroy the ci-ystal outlines ; fifth, a final separation of
angite, renewing tlie crystalline form of the augite phenocrysts, and
pi'obably taking place at the same time as the crystallization of tlie
groundmass. The inclusion of hornblende crystals in augite, as well as
the converse, has been noted before,-^ but the peculiar association above
noted is perhaps new. The accompanying plate shows the essentiul
points above mentioned. Both varieties of phenocryst also contain
inclusions of caleite ; the frequent occurrence of iron oxide with tliis
calcite may indicate tliat the caleite fills cavities left by the solution of
the iron oxide.

The groundmass is composed chiefly of minute augite crystals of a
pale pink or green color, closely compacted together. ^Magnetite in fine
crystals is abundant, perhaps composing a third u{ the bulk of the
groundmass. Small ciystals of brown hornblende are common, but
compose no considerable percentage of the mass.

1 J. F. Kemp and V. F. Marsters, Amer. Geo!., August, 1889. Also J. F. Kemp,
Amer. Geol., ^Lirch, 1890.

MUSEUM OF COMPARATIVE ZOOLOGY. 241

Occasional small structureless areas are seen which with crossed nicols
polarize feebly or are isotropic ; one instance was noted where there was
the faint appearance of a few minute lath-lil^e forms in radial arrange-
ment, as though plagioclase crystals had separated from a glassy magma.
This mineral gives no gelatinization with hydrochloric acid. On sepa-
rating the rock powder by gravity solutions, grains of this mineral settled
between 2.80 and 2.60, and more came down between 2. GO and 2.51.
Microchemical tests with liydrofluosilicic acid on these grains gave pretty
abundant cubes of potassium and some prisms of sodium, the tliorn-like
forms of calcium were also noted. This substance would thus appear
to be of a feldspathic nature and not nepheline since it did not
gelatinize. This determination would be of importance were the quantity
of the mineral large, but it perhaps does not compose more than one
per cent of the rock mass.

Classification.

Since the study of this rock has begun, a complete chemical analysis
has been found necessary to determine its position definitely ; until this
has been made, only the possibilities can be given. The rock has much
in common with the group of Fourchites of Dr. J. Francis Williams.-'
If it is regarded as belonging to this group, and is named according to
the predominating minerals, it would be called an augite ampliibole
fourchite. The occurrence of a feldspathic constituent rich in potassium
in the rock would tend to exclude it from this group, however, since
these rocks properly contain a lime-soda feldspar, nepheline or leucite ;
for this reason it may seem better, since the amount of this constituent
is very small, to associate this rock with those of the pyroxene group
(pyroxenite) although these have been regarded as containing no feld-
spathic constituent. In this latter case this would be the first dike rock
of the group.

This dike is interesting as being the first of the group of basic dikes
found in the Eastern United States which lias a geological age determin-
ably later than the Carboniferous ; the rocks which it intersects being
above the middle of the Connecticut Triassic. Of course the idea that it
is of later age than the Triassic is not excluded, though the fact that it
is broken by a small fault might be brought forward as an argument
for the intrusion of the rock before the time of deformation, which is
conceived to have followed closely the Triassic deposition. The wide
difference in character between this rock and the Triassic effusives may

1 Arkansas Geol. Survey, Ann. Rep., 1890, Vol. II p. 107.

242 BULLETIN OF THE

indicate something with regard to relative ages. The eflfnsives are fine-
grained olivine diabases, in which the augite crystals are much decom-
posed. The freshness of the dike rock would indicate that it is much
younger.

Most of the basic dikes of the Eastern United States previously 'de-
scribed, and having a near relation to the above, have been Campton-
ite ; in Northern New Jersey, however, some dikes have been noted by
Prof. J. F, Kemp,^ which he decides are practically identical with the
Ouachitite of Arkansas, and with the Ouachitite the relation of the Con-
necticut dike is close. If this Triassic dike rock is finally placed in the
Fourchite group, it will be the second occurrence outside Arkansas of
the rocks of the group.

1 Arkansas GeoL Survey, Ann. Rep., 1890, VoL II. p. 403. Also Am. Jour. ScL,
April, 1893.

Petrographical Laboratory, Harvard University,
June, 1893.

DESCRIPTION OF THE PLATE.

Reproduced by the artotype process from a micro-photograpli taken with polar-
izer only. The enlargement is 23 diameters.

The plate shows an angite phenocryst having a border of secondary augite and
enclosing a resorbed grain of hornlilende of the first generation. A portion of a
resorbed hornblende cr^ystal of the second generation is also shown. Some of the
larger augite crystals of the groundmass are distinguishable.

No. 15. — Notes on the Geology of the Island of Cuha, based vpon
a Eeconnoissance made for Alexander Agassiz. By Egbert
T. Hill.

PAGE

Letter of Transmittal 243

Introduction 243

I. Elementary Geologic Structure . . 24G
The Pre-Tertiar}', Metamorphic,

and Igneous Foundation . 246

The Pre-Tertiary Sedimentaries . 247

The Tertiary Subsidence ... 248

The Post-Tertiary Folding . . 253

The Post-Tertiary Formations . 254
II. Geologic History recorded by the

Topography 258

General Topograph}" .... 258

Topography of the Inland . . . 259

The Topography of the Coasts . 264

Terraces and Benches .... 267
The Soboruco or Elevated Reef

Level 268

Beach and Cliff Terraces . . 268

The Cuchillas of Baracoa . . 269
Older Higher Levels of the

Yunque Type .... 269
The Havana Levels .... 270
The ]\Ioro and Cantera Lev-
els ... 271

The Matanzas Levels ... 271

The Santiago Levels .... 274
Comparison and Correlation of

various Levels 275

Two Great Periods of Pleisto-
cene Regional Elevation . 275
The older Yunque Elevation
and succeeding Period
of Base Levelling . . 276

PAGE

The Modern or Tripartite Bench
Elevation and the Elevated

Reef 276

Inquiry into Evidences of Alter-
nating Subsidence .... 277
Absence of Upland Deposition 278
The Testimony of the Harbors 279
Absence of Fiords and Estu-
aries 280

Absence of Terraced and Re-
vived Stream Valleys . . 280
III. Resume^ and Conclusion .... 282
History of the Island reviewed 282
Obscureness of Pre-Tertiary

History 282

Evidence of Jeep Marine

Waters to North and

South of Cuba during the

two Cretaceous Epochs . 283

Tertiary Subsidence regional . 284

The Continental Shore during

this Epoch 284

The Orogenic Movement at
Close of Tertiary Time not
indicated in North Amer-
ican History 285

Reflections on the Character of
the pleistocene Regional

Uplifts 285

Absence of known Evidence of
any former Union of the
Island with Continents . . 285
Description of Plates 287

LETTER OP TRANSMITTAL.

Washington, D. C, May 1, 1894.

Dear Sir : — I beg to submit the following report upon my recent
studies in Cuha, undertaken at your expense and under your direction.

I left Washington, D. C, Sunday, February 11, arriving at Havana
on the following Wednesday. A day was spent in Havana to ascer-

VOI.. XVI. — NO. 15.

244 BULLETIN OF THE

tain the most feasible methods of working. I then made a reconnois-
sance east to Villa Clara, from which point a reconuoissance was made
north and south across the island. This trip was made primarily to
obtain some knowledge of the interior of Cuba. I was accompanied by
some American engineers who were familiar with the island, and who
were of great assistance to me in my subsequent operations. Returning
from Villa Clara to Havana, I stopped at several points, including
Matanzas and neighboring places. There were many opportunities to
study the basal contacts and structure of the limestones in relation
to the older nucleal area of Cuba upon which they were deposited. At
Matanzas I made a thorough study of the cut of the Yumuri River
of Matanzas, and of the limestone formations composing the Pan de
Matanzas and the country to the interioi', as set forth in the following
pages. Returning to Havana from Matanzas, I spent a few days studying
the geology of the site of that city and its environs, and in making a
north and south section across the island from Havana to Batabanos.

On February 28, a coasting steamer was taken from Havana for
Baracoa, which stopped for four or five hours at various points on the
way. I arrived at Baracoa on the 4th of March. With Baracoa as a
base, I made a study of the country west of Yunque mountain and east
to Cape Maysi, taking a trip up the Yumuri River of the east.

I sailed from Baracoa on the 13th of March, arriving at Philadelphia
on the 18th.

I wish to acknowledge my obligations to Mr. E. Sherman Gould, of
New York, Consulting Engineeer of the Havana Water Works ; Rev.
Lorenzo J. Arrubla, of Real Colegio de Belen, Havana ; Seiior S. Fortun,
Chief Engineer of the United Railways of Havana ; Hon. A. B. Dickey,
Consular Agept of Baracoa, and many others throughout the island, for
valuable data furnished ; also to Prof W. H. Dall of Washington, for
his valuable assistance in paleontologic determinations of the age of the
rocks discussed herein.

Looking back at my expedition, I now see many points which I would
like to examine again, but I hope that, upon the whole, you will not be
dissatisfied with the results.

Very truly yours,

Robert T. Hill.

To Alex. Agassiz, Esq., Cambridge, Mass.

MUSEUM OF COMPAKATIVE ZOOLOGY. 245

INTRODUCTION.

Paleontology, stratigraphy, and physical geography are the three
co-dependent criteria by which geologic history is interpreted. A proper
study of the paleontology of Cuba requires years of residence and patient
labor by an expert in Cenozoic and Mesozoic fossil forms. The struc-
tural geology, involving vast petrographic research and tedious travel,
likewise demands long study; the thick residual soils, the dense vege-
tation, and the absence of higliways and general lack of exposures every-
where, retard the worker in these branches of geology. While geologic
research in these fields has already given us criteria without which the
present paper would have been impossible, the topographic forms reveal
a story equally interesting and more intelligible, and in the following
pages I Iiave endeavored to interpret them, with such assistance as
could be derived from the co-ordinate branches of paleontology and
structure.

Literature contains many descriptions of the general geography and
geology of the island, — especially the works ^ of Humboldt, Salterain,
Suess, Crosby, Ramon de la Sagra, Don Manuel Fernandez de Castro,
and the various pul)lications of Mr. A. Agassiz. Don Manuel Fernandez
de Castro's brief pamphlet, accompanied by a geologic map, is an excel-
lent resume of the stratigraphy and paleontology of Cuba, and should
be consulted by any one contemplating the perusal of the present paper.
Mr. Agassiz's investigations have made known to science the wonder-
ful topography of the surrounding ocean floor. M. Elisee Rectus has
recently compiled the general physical and political geography of the
island as ascertained by previous investigators, and its relation to the
surrounding seas and the West Indian archipelago. Professor Suess
has compiled a chapter on the Antilles, setting forth the present state
of knowledge concealing the geology of the island of Cuba. These,
together with Salterain's description of the geology of Havana, have
explained in a preliminary way the geography and geology in a manner
to prepare for an intelligible discussion of the topographic evolution of
the island, by which its history may in part be finally interpreted.

1 The titles of these publications are given in notes accompanying the references

to tlit'iu

246 BULLETIN OF THE

I preferred to examine the phenomena entirely de novo, so as to be
influenced by no preconceived hypotheses, and hence made no study
of these writei-s until my report was formulated. After outlining this
paper I examined this literature, and I hope that the present discussion
of the topographic and diastrophic phenomena will further advance the
■work so well begun hy others. I am also glad to affirm, with a few
minor exceptions, the views of the structure set forth by these earlier
writers, and I would earnestly request those who read this paper to
consult them.

I. ELEMENTARY GEOLOGIC STRUCTURE.

The Pre-Tertiary, MetamorpJiic, and Igneous Foundation. — This is a
floor of ancient (certainly Pre-Tertiary) rocks, consisting of serpentine
and igneous material, mostly basic in aspect. These are exposed by
erosion at various points throughout the island, but, except in the
Santiago region, they seldom, if ever, form the rocks of the immediate
coast, although often found quite near it. Thej'^ now underlie most
of the island at no great de[)th, and are exposed in many drainage
cuts beneath the limestones. Wherever I have seen these rocks — at
Villa C'lara, Havana, and near Baracoa — they had once been covered
by the Tertiary limestones, but I cannot speak with certainty concern-
ing the province of Santiago de Cuba, where, according to Kimball,
traces of limestone as high as 2,300 feet are pi'eserved on the south side.
On the north side, opposite Santiago, they are certainly overlapped far
interiorward by the limestones. Some of the igneous and metamorphic
rocks of the Santiago region may be of later origin and intrusive
through the limestone, but generally throughout the island they have
been completely covered by the latter. No Post-Tertiary eruptive
sheets were seen hv me except one small dike which intrudes into the
greatly folded limestone near the water-works back of Havana. (See
Plate I. Fig. 3.)

These older rocks consist of diorites, serpentines, schists, and rarely
granites, as reported from Santiago ; of serpentines, gi'eenstonc, porphyry,
and basic igneous rocks brought down by the rivers of the north side of
the east end of the island ; of serpentine and metamorphic rocks with little
quartz, as seen underneath tlie limestone in the vicinity of Villa Clara ;
and of serpentine, tuffs, and old volcanic material, as back of Havana.

The rare presence of eruptives and of sedimentaries older than C're-

MUSEUM OF COMPARATIVE ZOOLOGY. 247

taceous, reported by De Castro aiid occasionally seen by me in this old
complex, testifies tliat in Pre-Tertiary time the old metamorphic floor
protruded above the level of the sea as a land area, during a period of
active vulcauism. From their composition and occurrence there can
be little doubt that they once formed an ancient land area of unknown
size and relations, but certainly almost as large as the present island,
which was partly or completely submerged during early Tertiary time.

The Pre-Tertiary Sedimentaries. — Resting upon this metamorphic and
igneous foundation at various places there is a formation of stratified,
non-fossiliferous, sedimentary clays. These are older than the Tertiary
limestone, and apparently immediately preceded them in origin. They
are not of great thickness, and are void of determinable fossils wherever
I have observed them. In the Havana section, in the southern suburbs
of the city, only a few feet are exposed beneath the old limestone in
contact with the underlying tuffs. They are here green in color, and
somewhat unctuous.

In travelling overland toward Villa Clara, I found that the limestones
extend beyond Colon, but between tiiat place and San Pedro they are
eroded through down to the underlying clay formation, which extends
from thei'e continuously east to Villa Clara, being best exposed at
Esperanza. Here the railway has cut across alow anticline of clays
which show well defined stratification planes and alternate strata of
softer and harder beds. In general they consist of (1) an upper division
of light colored, laminated, thinly banded clay, with persistent bedding,
of which twenty feet are here exposed, and (2) a lower series of coarser
beds, the harder persistent strata being loosely cemented and having a
mealy consistency, with siliceous pebbles, while the alternate beds are
laminated. About one hundred feet of these are exposed. These clays
are folded and slightly faulted in places.

I could find in tliese beds no fossil remains except one poorly preserved
plant impression, a monocotyledon, which, with the general character of
the material, gave the impression that these clays were deposited when
the conditions of sedimentation avound Cuba or in Cuba were far differ-
ent from those of the present, and, to a certain extent, they indicate a
previous land. I cannot say positively that the Esperanza clays are
identical with the clays of Havana, but both occupy the same relative
position between the Tertiai-y limestones and the metamorphic for-
mation, and both are exposed by the erosion of the limestone from
above them,

De Castro refers these clays of Esperanza to the Cretaceous period.

248 BULLETIN OF THE

upon what ground, except stratigraphic position, I cannot say. He
reports other Mesozoic sedimentaries at both ends of the island, together
with Ammonites and RadioUtes, which would clearly indicate the occur-
rence of Cretaceous deposits in Ciiba.^

The Esperanza clays with plants disclose a Pre-Tertiary land accom-
panied by erosion, but its area and extent I cannot interpret. There
are indications that these clays have been metamorphosed into schists
in places, as may he seen nine miles north of Villa Clara, where the
anticlinal rocks overlying a vast intrusive bed of asphaltiim show many
evidences of transition. This idea, however, is at present merely a
suggestion.

Formations of the Tertiary Subsidence. — In strong contrast with the
Esperanza clays and the older metamorphic floor are the Tertiary and
later limestones which cover them, and which are the predominant geo-
logic feature of Cuba. These certainly once extended over all the island,
with the possible exception of a small portion of the high mountainous
region before mentioned, in the vicinity of Santiago de Cuba. They
still occupy by far the greater portion of the whole area. The study
and classification of the limestone complex is difhcult, owing to the
folding, induration, and erosion, the thicl< covering in places of residual
soil and vegetation, the universal alteration they have undergone through
solution and internal changes, and the general concealment by tufaceous
incrustation of well defined stratification planes and partings. The lime-
stones are clearly divisible into the two general categories of the newer
and the older, or basal. The former consist mostly of unfolded rocks of
undoubted coral reef origin, and occur on the lower levels adjacent to
the coast, while the latter, if of coral reef origin, have- lost all character-
istic features, of rocks of such origin, are undulated and folded, and
constitute the uplands and high coastal scarps against and around which
the later coral rock grew.

The more ancient limestones nearly everywhere constitute the upland
of the island, and by alteration and underground decay have lost their
coralline structure, if they ever possessed it. These, so far as my
observations extended, constitute all the limestones of the island above
an altitude of one hundred feet. These older limestones are diverse in
texture and composition. Where good exposures are obtainable, they
usually exhibit well defined stratification and separation planes, never
seen in the undoubted reef rock, and sometimes alternate with more

1 Pruebas Paleontologicas de que la Isla de Cuba, etc., por Don Manuel Fernan-
dez de Castro, Madrid, 1884, p. 6.

MUSEUM OF COMPARATIVE ZOOLOGY. 249

marly or very slightly siliceo-argillaceous beds. The limestone beds
have beeu well described lithologically by Sagra, as follows : —

" It is white in color, or light yellow, with a fracture sometimes smooth,
sometimes conchoiJul, containing some concretions, — very often casts with
petrifactions. The fossil substances enclosed in the limestones are very abun-
dant. . . . The porous beds of the middle part of this locality, as near Batabano,
resemble those spongy and calcareous banks of the Jurassic of Francone, near
Dondorf, Pegnitz, and Tumbach. These yellow cavernous beds, which show
cavities from four to five inches in diameter, alternate with others entirely com-
pact and less charged with petrifactions. The line of hills which border the
valley of Los Guiues toward the north, and which unite the hills of Camoa
and the Tetas de ]\Ianagua are of this last variety, the color of which is a rosy
white, sometimes almost lithographic, like the Jurassic limestone of Pappenheim.
The compact and cavernous beds contain small ferruginous masses, and are the
same formation that Humboldt designated the Calcarie de los Guines, which is
exposed on the southeast near Trinidad, on the hills of San Juan, already
referred to, and on the north coast, near Matanzas. In these different localities
it exhibits grand subterranean cavities, wdiere rain water accumulates, and in
which many considerable streams submerge." ^

I miglit add to this description the remark that these rocks bear a
striking lithologic and structural resemblance to the Neocomiau and
Middle Cretaceous rocks of Texas.

Although distinctly stratified, the limestone is irregular in texture.
While it is, in general, of a cellular structure, a cubic foot of it in any
locality exhibits great irregularities in hardness and compactness. There
are spots so hard and crystalline that it is difficult to break them with
a hammer ; other spots are firmly crystalline and banded ; still others
are rounded indurations ; and again there are soft, pulverulent spots.
All of this irregularity of texture is secondary, or in a condition of
alteration produced by aqueous solution. In some places the cellular
cavities are many feet deep, while the remaining portions are indurated
into sharp edges of coarse, sometimes crystalline limestone. So com-
pletel}' has the work of solution and intei'stitial change gone on that it
is doubtful whether the ori!j,inal nature of the rock is anvwhere well
preserved.

This weathering and induration is very similar to that which I have
often noticed in the chalky Lower Cretaceous limestones of Texas. On
the resisting summit points the rock is hardened and worn into the
peculiar Liliputian ridges known in the Alps as " Karrenfelder," while

^ Histoire Physique de Cuba, Tom. L p. 109.

250 BULLETIN OF THE

the steeper bluffs are thickly coated with" tepetate," or tiifaceotis deposits.
Great caverns abound in these rocks in many parts of the island.

These limestones have been so greatly altered since their original
deposition that, from macroscopic examination, it is difficult to tell their
original character or the conditions under wliich they were deposited.
They certainly do not anywhere exhibit the enormous proportions or
abundance of coral remains so apparent in the reef rock, nor do they
show, except occasionally, an abundance of casts and moulds of molluscan
shells, and I seriously doul)t whether, as alleged by Crosby ^ and Kimball,'^
and formerly by A. Agassiz,^ they are coralline in origin, as in the modern
reef rock. They sometimes contain traces of coral, but I do not think this
proves that they were reef rock, for all corals are not reef building, and
the organic remains are far more abundantly molluscan than coralline.
Neither can they be called chalks, although very foramiuiferous in places,
for they are too coarsely crystalline, clastic, and molluscan, and lacking
in that fineness and uniformity of texture seen in the chalky limestones,
which I have had considerable experience in studying. In places at their
basal contact they are certainly detrital, showing (as at the reservoir
south of Havana, where they are in contact witli the older series of clays
and serpentines) a distinct conglomeratic structure, and being compdsed
largely of shell fragments and beach wash. Near Villa Clara they contain
very small fragments of igneous material derived from the older rocks
which they buried. In many places they are distinctly sedimentarv, as
seen in the Castillo Principe Plateau west of Havana, where tliev con-
tain alternations of stratified, slightly yellow argillaceous layers, while
the several hundred feet exposed in the caiion of the Rio Armendaris,
south of Havana, exhibit far more molluscan remains than coral, although
some corals are present. Likewise at Matanzas the older limestones ex-
hibit every character of sedimentaries with molluscan remains, rather
than coral reef structure. At Baracoa, Nucvitas, and elsewhere on the
west coast, tlie limestones not only appear to be sedimentary, but they
alternate with beds of a yellow argillaceous and arenaceous material,
clearly sedimentary, and containing great numbers of molluscan fossils.
In fact, I do not believe that any of the limestones below No. 2 of the
Matanzas section (Plate II. Fig. 4) are of reef rock origin, but am of
the opinion that they are mostly organically and chemically derived sedi-

^ On the Elevatefl Cora] 'Reefs of Cuba. Proceeilings of the Boston Society of
Natural History, Vol XXW. pp. 121-129.

2 American Journal nf Science, Tlccemher. 1SS4.

8 Bull. Mus. Comp Zo<iI., Vol. XXVI. No. 1, December, 1894.

MUSEUM OF COMPARATIVE ZOOLOGY. 251

merits of marine lime mixed with the calcareous debris of the life of the
ocean's margin, with, in places, an almost imperceptible proportion of
the finer physical sediments of the nucleal island.

While these limestones and alternating beds have a great areal extent,
it would be a mistake to assign to them a proportional thickness, for
accurate measurements will not make their thickness anywhere greater
than one thousand feet. 1 estimated from the dips in the Rio Armen-
daris section tliat tliey were from eight hundred to one thousand feet ;
the incomplete section in the canon of the Yumuri of Matanzas reveals
eight hundred feet ; the canon of the Yumuri of Baracoa shows six hun-
dred feet ; the summit of Ynnque displays less than one thousand feet ;
while the section from fourteen kilometers south of Havana to Bata-
bano is not over one tliousand feet. (Plate 11. Fig. 1.) In fact, they
may be said to constitute a comparatively thin veneering over the old
metaraorpliic tioor.

The old limestone formations occur from end to end of the island, and
extend in many places completely across it down to water level. Their
continuity is interrupted only by erosion along the central axial region,
and only the low portion adjacent to sea level is covered by later deposits.
De Castro's geologic map of Cuba ^ shows in an excellent manner their
general disposition. In places, as between Mata and Yumm-i, they form
the north wall of the coast. They cap the highest eminences of the
island seen by me, overlooking all other rocks, being overreached only
b}' the Sierra Maestro, the geology of which is unknown. Their close
proximit}' to the north coast and their abrupt protubei'ance above the
newer formations have an important bearing on the history of the island
as a whole. So extensive is this old limestone formation, and so abruptly
does it rise above the coast, that, if all the coastal formations were
stripped away, or if the island should subside for one hundred feet, its
superficial extent would hardly be perceptibly diminished or its outline
materially altered.

The greater part of these limestones seen by me are of Eocene and
Miocene or of Pliocene age, as alleged by De Castro. In the Armendaris
section, near Havana, they arc both Eocene and Miocene, as has been
asserted by De Castro and others, and as is shown by my collections.^

1 Croquis Geologica de la Isia de Cuba, por D. Manuel Fernandez de Castro,
ampliado por T>. Pedro Sallerain y Legarra. 18G9-8o. Printed in Vol. IX. of the
Congreso Internacional de Aniericanistas.

- The determinations of age in tliis paper are hased upon the palcontologic
determinations of l^r. William H. Dall, of the U. S. Geological Survey, who kindly
examined the material collecicd.

252 BULLETIN OF THE

At Baracoa the upper layers are Miocene. Humboldt, De Castro, La
Sagra, and others have recognized the Tertiary age of these limestones,
and their distinctness from the modern reef rock, or soboruco. Concern-
ing them De Castro says : —

" The Tertiary terrane in the island of Cuba is more important in view of the
great extent it occupies, the abundance of its fossils, and various circumstances
which are peculiar to it, and which" would supply matter for a long discussion.
I shall have to confine myself, however, to saying that at one time it must
have covered nearly the whole surface of the island, judging by what still
remains of it, notwithstanding the denudation which it has undoubtedly suf-
fered. A glance at the sketch will serve for a description or enumeration of
the localities where it is found,^ although it is probable that, when the whole
territory of the island is studied as has been the immediate vicinity of Havana,
Matanzas, Cienfuegos, and Santiago de Cuba, part of the color representing the
Tertiary terrace will have to be replaced by colors indicating older formations,
which, like the Cretaceous, have not yet beeu recognized owing to lack of data.
" The presence of Carcharadon me^aiorft; /i, belonging exclusively to the Mio-
cene period in Europe, although found in America also in the Eocene ; the
abundance of Orbitoides mantelli, a foraminifer which in the United States
characterizes a bed belonging to the Upper Eocene ; the occurrence of Orbitoi-
des at many points, as in the vicinity of Pinar del Rio, at the western end of
the island of Cuba, and at localities on the eastern part of the island of Santo
Domingo, forming an extensive horizon, would permit the exact determination
of the age of the different beds above and below those containing this forami-
nifera. For the present I will confine myself to saying that in Cuba there
undoubtedly exist deposits of the three periods into which the Tertiary is
divided, because among seventy genera and more than two hundred species
of fossils thus far found, there are, besides those of the Eocene and Miocene,
a great number which belong to the Pliocene.

" The Eocene is perfectly represented, and there are many fossils which, if
they are not identical with those referred in Europe and India to the Num-
mulitic, greatly remind one of them. ... It may be said that in Cuba the exist-
ence of the Miocene and Pliocene has more evidence, in view of the abundance
of fossils distinguishing those ages."^

It would be desirable to make a complete study of these old tertiary
deposits, but it would require years of careful paleontologic and strati-
graphic investigation. The observations made by me at various localities
were not sufficiently numerous to enable the construction of a generalized
section, or to permit deductions concerning the permanency or continuity

1 Found also by me in the Armendaris section near Havana.

2 Pruebas Paleontologicas, p. 7.

MUSEUM OF COMPARATIVE ZOOLOGY. 253

of the horizons and the alteraations of lithologic material shown in the
local sections, such as the great beds of fine siliceous and argillaceous
mixtures with the lime, as noted at Matanzas and seen from there east
to Baracoa, forming thick strata of yellow material, containing, at least
at Baracoa, Miocene Mollusca and corals, as determined for me by Dr.
Dall and Mr. T. Way laud Vaughan. (Plate I. Fig. 5. 3.)

The slightly arenaceous yellow beds outcrop at Nuevitas, Gibara, and
many other places along the coast, and are included between thicker strata
of limestone, and I think they are underlain by several hundred feet of
that material, and belong near the limestone capping Yunque and the
Yumuri bluti's. These yellow bluffs underlie the soboruco reef at Bara-
coa, and are capped by a thick stratum of old limestone back of the city.
The harbor is largely formed by their undermining. They are also well
developed beneath the old reef points of Mata Bay.

A peculiar rock material in the old limestone series at Baracoa, and not
seen elsewhere, is a hill (Plate I. Fig. 5. 4) of almost vertically stratified
siliceous material, which at fii'st sight resembles gray chalk, but has the
light specific gravity of some of the diatomaceous earths. Under the
microscope this material is found to be composed largely of siliceous re-
mains of minute organisms, mostly of Radiolaria, with sponge spicules and
echinoid fragments, but containing no diatoms, so far as I have studied it.
This material is distinctly stratified, and contains oocasional thin separa-
tion layers of a gray-blue clay and some flint-like siliceous nodules. It has
clearly undergone great disturbance, as is shown by the vertical arrange-
ment of its beds, and apparently lies below the yellow beds, which are
Miocene, as determined by Dr. W. H. Dall. This material has always
been a source of great perplexity to the people of Baracoa, who could not
classify it or understand its qualities. The reservoir for the village
water-works is located upon the single hill where it outcrops, on the
southwest side of the harbor. The beds are over five hundred feet in
thickness, and I think they overlie the oldest of the limestones, but this
I could not ascertain with certainty. Neither this material nor the
yellow beds which together constitute at least five hundred feet of the
tertiary sequence can be classified as of coralline origin.

The Post-Tertiary Folding. — The chief feature which separates the older
limestones into a distinct system from the modern reef rock is the strati-
graphic unconformity between them, and the fact that the former have
undergone great folding and disturbance prior to the deposition of the
latter, which are always subhorizontal. In no locality have I seen the
newer reef rock folded or greatly pitched, but the older limestone is

254 BULLETIN OF THE

frequently tilted at an angle of forty-five degrees, as at Baracoa, and
sometimes intensely folded into anticlines, as back of Havana, and pre-
sents every degree of folding and disturbance in the numerous railway
cuts between Havana and Matanzas, at Villa Clara, Yumuri, and else-
where. In fact, it is seldom if ever subhorizontal on the north coast,
and the later deposits are entirely unconformable with it. While the
details of this disturbance could not be wrought out during the brief
time which I spent upon the island, its character can be seen in the
various sections and illustrations given on Plate I.

The general lay of the old limestone is that of a low anticline whose
axis corresponds with that of the island, with folds more greatly devel-
oped along the northern coast. This folding took place clearly near the
close of Tertiary time, and prior to the deposition of the Post-Tertiary
formations and elevations to be described, and indicates one of tlie most
important epochs in the geological history of Cuba, representing, as it
does, an orogenic folding not elsewhere traceable in North American
history. This folded condition of the limestone, however, has in no
manner influenced the later topographic detail of the surface, and was
mostly, if not entirely, antecedent to the great regional elevations to be
described later.

The Post-Tertiary Formations. — In strong contrast to the older Ter-
tiary limestones is a more modern group of limestones of undoubted
coral-reef origin, which border the coast in most places, or form small
coral islets adjacent thereto, and are locally known as soboruco.

M. Ramon de la Sa!2:ra has defined this formation as follows : —

" L'autre forni.ition de calcaire modenie, qui a recu dans le pays le noin de
sohnruco, se tronve de long de la cote dans plusieurs endroits de I'ile ; elle est
tellement recent, que son agglomeration continue menie aiijourd'hui, et c'est
a elle que I'on doit les cayes, les recifs et tons les bas-fonds de coraux. Les
parties superieures s'elevent parfnis a partir d'une profondeur de vingt a trente
brasses. Toutes les inegalites de cette roche sent recouvertes d'une couche
calcaire agglouieree avec des restes d'animaux, des coquilles, de coraux, et de
madrepores." i

The elevated reef rock can always be recognized by the perfection and
abundance of well preserved remains of reef-making corals, which form
the greater proportion of the mass and unmistakably show its origin, and
by the absence of subdivisions into lamination and bedding planes. The
surface is practically the old level of the submerged reef, the sharper

1 Histoire Physii^ue, etc; de I'lle de Cuba, Tom. I. p. 110.

MUSEUM OF COMPARATIVE ZOOLOGY. 255

irreo-ularities having been levelled by solution. The formation averages
about thirty feet in thickness, and usually extends inland only a short
distance, often only a few yards, as on the northwest end of Moro Point,
or not over an eighth of a mile, as at Baracoa, and is especially well
exposed along the narrow points of the numerous small harbors, as shown
on Plate II.

The soboruco is a topographic as well as a stratigraphic feature, for its
surface is a bench gently sloping to the sea ; it has neither been covered
by later deposits nor greatly denuded. It usually forms a cliff at the
surf line, about fifteen feet in height, against which the surf beats with
great force, wearing deep indentations. The spray breaks over the
summit, with the aid of the sun producing the surface induration which
is visible wherever rain or other moisture falls upon the hot limestones,
or w'earing the surface into cavernous Karrenfelder. This solution and
induration at Baracoa, for instance, has converted the limestones in
spots into a coarse saccharoidal marble, and has aided in the segregation
of small lumps of iron ore direct from the coral.

Where I was able to examine the base of the elevated reef rock,
mostly at the mouths of rivers, it seems to have been deposited rather
abruptly upon a semi-argillaceous terrane of silt, and occasionally very
fine pebbles, which have been brought down and deposited by the rivers.
(Plate II.) I did not find it growing upon the larger gravel which is
deposited immediately at the river's mouth, as is seen off the Yumuri of
Baracoa, where the river empties into the sea, and not into a bay. Further-
more, the present submerged fringing reefs do not grow immediately
where the rivers send their fresh waters into the sea, but are interrupted
there by a barren area simulating a submarine channel, as is shown in
the accompanying illustrations of the harbors. This fact has an impor-
tant bearing upon the origin of the present circular harbors, and upon
tlie theories of alleged subsidence, both of which subjects are more fully
discussed in later pages.

Is is impossible to describe all the localities at which the soboruco was
observed.^ Sometimes, as along the Havana coast, it occupies a narrow
coastal strip extending from the])oint of one harbor to another. Again,
as on Moro peninsula, opposite Havana, it occurs only as a small patch
in a slight indentation in the old headland composed of folded Miocene
rocks. (Plate I. Fig. 2.)

At Tanamo and other places on the north coast the soboruco not only

1 See A. Agassiz, Bull. Mus. Conip. ZooL, Vol. XXVI. No. 1, Plates XLIV.-
XLVII.

256 BULLETIX OF THE

forms the border of the mainland, but constitutes many bordering islets
of great areal extent. Generally these are low, standing only a few feet
above the water. There is a vast elongated archipelago of these elevated
reefs bordering the coast all the way from a point east of ]\Iatanzas to
Nuevitas. I passed most of this region in the night, and I can say
little concerning it. At Xuevitas, in the harbor, there are three peculiar
islands, known as Los Ballantos, which have very great resemblance to
the Keys of the Bahamas, presenting a bold, rounded escarpment at the
north point, composed of yellow friable material that may have been
pither coral sand or the yellow Miocene clays. It was impossible to get
ashore to these to examine them, although this was the only locality
seen by me where there wae a suspicion of wind-blown formation. The
greatest areal development of the Hat soboruco was found along the
outlet of this liarbor.

Nowhere have I seen the elevated reef rock folded or otherwise dis-
turbed except by the gently coastward inclined elevation it has under-
gone. The interior margin I have never observed at a height of over
forty or fifty feet. In general, there is only one massive layer of this
old reef rock exposed, but at Matanzas there is undoubted evidence of
two older underlying reefs, the inner edges of which have been elevated
with the modern reef so that they do not form distinct terraces. It may
be that the apparently continuous reef around Cuba represents more
than one of these layers. Whether one or several alternations of reefs,
the soboruco as a whole certainly represents a recent and uniform ele-
vation of the whole periphery of the island at a verj- recent period of
geologic time, but sufficiently long ago to have permitted considerable
alteration and erosion. It is found from Cape San Antonio to Cape
Maysi on the north side of the island, and at many places on the south
side, e.specially near Santiago, as described by Kimball.^

Cantera is a term used tliroughout Spanish countries for any stone
that is soft enough to be hewn or sawed with ordinary carpenter's tools,
as distinguished from a stone requiring mason's implements. Much of
the cantera in Cuba is composed of a soft molluscan or coralline lime-
stone, which has not the irregularity of composition and density and the
varying hardness of the older limestone, nor the unaltered coral structure
of the elevated coast reef. It is intermediate between the two, and may
represent a stage in alteration between them. It is finely cellular, or
porous, not usually saccharoidal, and is filled with small cavities some-
times lined with botrvoidal lime coating.

^ American Journal of Science, December, 1884.

MUSEUM OF COMPARATIVE ZOOLOGY. 257

Quarries of cantera are found adjacent to most of the cities on the
north coast of the west half of the island. They occupy a slightly
higher altitude than the coast reef, and usually constitute the first
bench of the island above it at a lower level than the erosion planes in
the older limestones. (See Havana and Matanzas sections, Plate I. Fig.
1. 2, and Fig. 4. 1.)

I did not observe any break between the cantera and the older lime-
stone, owing to obscurement, except in the Matanzas section, where it
clearly appeal's that the cantera is mostly old reef rock which has no
topographic integrity, antl which was unconformably deposited on the
older limestone after the latter had been considerably elevated. In other
words, it there represents the oldest of the recognizable fringing reefs.

At Havana, in the convict quarry, northwest and at the foot of the
Castillo Principe Plateau, which is made up of the older limestone, there
is a great cantera that seems more molluscan than coralline. The same
deposit is also- worked in the banks of the Piio Armendaris, two or three
miles southwest. Topographically it here underlies an erosion level
intermediate in heipht between the level of the modern reef and that of
the Moro Plateau.^ Xo cantera was observed east of Matanzas in the
ports of Segua, Gibara, Nuevitas, or Baracoa. It is not here proposed
to establish the cantera as a persistent geologic unit, for tliere may be
other cantera beds in the old limestone. The coralline cantera of Ma-
tanzas and the molluscan cantera of Havana are not found at an alti-
tude of more than one hundred feet, and they are always near the coast.
At the furmer locality the cantera is the oldest of the rocks of prob-
ably coralline origin, and at the latter it is intermediate in position
between the modern reef rock and the older limestone.

Throughout Spanish America the term " playa," meaning literally
a flat beach, lacustral, or shore deposit, is applied to alluvial flats or
mud plains composed of gravel, sand, and clay. In Cuba I found that
the term was generally used for an alluvial deposit sometimes lining the
inner margin of the circular harbors, as at Havana and Baracoa. These
are small in area, and are usually adjacent to the zone at which tlie
rivers come out of the highland into the harbors ; they represent delta
deposits that have undergone slight elevation coincident with that of the
modern elevated reef Their origin is more fully discussed under the
head of Harbors.

^ Specimens of this cantera received from Havana since tliis report was written
soniewliat eonfirni tlie impression that they represent a late Post-Pliocene deposit laid
down aijainst the ohier Tertiaries, and prior to the elevation of the soboruco.
VOL. XVI. — NO. 15. 17

258 BULLETIN OF THE

Flat, marshy alluvial deposits (cienagas) occur in many places, but
mostly on the south coast. At Batabanos, opposite Havana, the coast
for a mile or more inland is composed of ancient alluvial material, ap-
parently similar to a calcareous mud now depositing and forming the
bottom of the adjacent sea for a mile out from laud. These cienagas
and cienaga deposits are reported to have considerable extent at various
places, especially on the south coast. The elevated portion is synchronous
with the soboruco elevation on the north coast.

A striking' peculiarity both of the older structures and the coast
deposits of Cuba is the scarcity — almost total absence — of arenaceous
or sandy deposits. Nowhere is found the fine quartz sand such as
accumulates arovuid the northern littoral of the Gulf of Mexico, and the
presence of pieces of quartz gravel is very rare, even in the delta deposits.
This is owing to two reasons : (1) the formations of the island, both the
older metamorphic foundation and the limestones, possess Yevy little free
quartz, and (2) the littoral sands or physical sediments of the peripheral
drainage of the Gulf, derived from the continental Americas, are not
transported as far as Cuba, as Professor A. Agassiz has alread}' pointed
out. Even the building sand of Havana and other places is calcareous
beach debris.

Beefs. — No description of the geology of Cuba would be complete
which stopped at the ocean level, without allusion to the adjacent sub-
merged coral reefs that fringe its shores or lie a short distance away,
which, with the adjacent submarine topography, have been so ably
described by Mr. Agassiz.^ So fully has he described these phenomena
that it is not necessary to discuss them further, except to call atten-
tion to the fact that there are often considerable deptlis between the
barrier reefs and the near-V)y land. This has important bearing upon
the topography of parts of the coast.

II. GEOLOGIC HISTORY RECORDED BY THE

TOPOGRAPHY.

General Topograph^/. — Having reviewed the fundamental rock struc-
ture upon which the sculpture of the land is dependent, we can now pass
to a more intelligible discussion of the general topography and its evo-
lution. It is neither necessary nor possible to give a minute detailed
description of the geography of Cuba, but only so much as may relate

1 Bull. Mus. Conip. Zool., Vol. XXVI. No. 1, December, 1894.

MUSEUM OF COMPARATIVE ZOOLOGY. 259

to the genesis of the island. Its sliape and outline have been described
by various writers, notably Humboldt, Sagra, and Reclus, and I shall
touch upon only those details or generalities that will help to elucidate
its geologic history.

The outline of the island might be compared to that of a hammer-
headed shark, the head of which is the culminating range along the
straight Santiago coast, from which extends westward the longer, lower,
and more sinuous mass of the islaiid, while the fins are represented by
pendant coral reef archipelagos. The Santiago coast is excessively
mountainous, and is supposed to be in some way genetically connected
witli the mountain systems of the other Antilles. Concerning the com-
position, age, and topography of these ranges, we know little excejit
what has been told by other writers, to the effect that they extend
aproximately in an east and west direction, towering far above the
levels of the remaining purtion of the island, and occupy a very narrow
strip close to the ocean's margin. Extending away to the north and
west from this nucleul elevation is the main body of Cuba, which is
primarily a great limestone plateau intensely eroded and terraced, with-
out any well defined axis of higher elevation except as indicated by the
headwater drainage that diverges from it and flows into tlie opposing
seas, the latter feature corresponding to the longitudinal axis of the
island. The highest elevations do not occur in a continuous ridge, but
are irregularly dispersed, as if they were remnaiits of a dissected elevated
plateau, diversified by plains and irregular chains of hills, often nearei
the margin than the centre, and seldom over two thousand feet in
altitude.

The topographic forms belong to two categories, those of the inland
and tliose of the coast. The former depend upon uplift and erosion, and
are mostly the product of superficial agencies. The coastal featiires,
on the other hand, are the product not only of uplift, but are largely
influenced by the sea, — its life, its sediments, and its surf erosion.
The latter are of little areal extent, especially on the north side, where
they form the merest fringe around the island. The uplands extend
close to the coast, where they are terminated by abruptly terraced
cliff's, or series of successively lower levels.

The Inland Topograph}!. — This includes the whole surface of the
island except the narrow coastal plains and elevated reef, and is pecu-
liarly and strikingly diff"erent from any topography to which we are
accustomed in tlie United States. Under the influence of excessive
humidity, the solvent limestone material of which it is mostly composed

260 BULLETIN OF THE

has assumed the aspect of an ancient and much carved land surface,
while the geology, on the other hand, tells us that it is comparatively
new. The peculiar surface configuration of the island is solely due to
the decay and erosion of these older limestones. Where they predomi-
nate, the interior country remote from the somewhat abrupt coast is
marked by extensive flats or Llanuras, upon which rest, without any
apparent systematic arrangement, sliort ranges of abrupt hills or knobs.
Botli the plains and the mountains are composed of the same old
limestone. (Plate III.)

Although the rocks dip at various angles in gentle waves or folds, and
in some cases with numerous short anticlines, it was impossible to make
out any system of disturbance which influenced the minor topography.
This sub-arrangement of the underlying limestone structure does not
influence the present surface topography, as the strata are of insuffi-
cient induration to create topographic irregularities.

The limestones decay so rapidly at the surface that they can seldom
be seen in a freshly exposed condition except in railway cuts and quar-
ries. Everywhere they disintegrate into a rich red residual soil, some-
times of great depth, which contains a large percentage of iron derived
from the limestone. This sul>aerial decay is beautifully shown in the
cuts of the Havana and Matanzas railway near the latter citj', an illus-
tration of which is shown on Plate I. Fig. 7. This iron has great
economic value in the vicinit}' of Santiago de Cuba, under the conditions
shown by Kimball. The limestone is also melting away by deep un-
derground solution, and is very cavernous. This deca}', proceeding
irregularly, being much greater in some spots than in others, produces
large sinks, as in the limestone regions of Kentucky, only of a vastly
greater area. Thus it is evident that the deep, sloping indentations
are produced not solely by surface corrosion, but partly 1)\' underground
decay, the streams carrying away tlie lime in solution, ratlioi- than as
physical debris. A most remarkable illustration of the effect of solution
upon the topography is the valley of the Yumuri de Matanzas, a plan of
which is shown on Plate II. Fig. 9.

When this rapid disintegration and degradation of the old limestone
is considered, it is evident that the ancient limestone summit of the
island must necessarily have been greatly lowered in outline. In other
words, the present highest levels of the limestones do not represent tlie
altitudes wdiich they formerly attained, but are merely planed off rem-
nants. It is impossible to determine how great a portion of the top of
Cuba has been thus removed, except so far as to say that, judging from

MUSEUM OF COMPARATIVE ZOOLOGY. 261

the jiresent rate of denudation, it must have been a large amount, for in
places it has laid bare the old metamorphic floor. "Wherever I have seen
the latter, it unmistakaMj shows that it was once covered by the lime-
stone. This is well shown in the accompanying section across the island,
through Havana, where the older foundation rocks always appear in the
valleys of erosion beneath the escarpments of stratification formed by
the adjacent limestones, the latter being so tilted around their periphery
that, clearly, they once extended over them, as at Villa Clara. It is
easy to conceive tliat, if erosion proceeds in the future as in the past,
without further elevation of fringing coast deposits, the island will ulti-
mately be planed down to its original core of serpentine and allied rocks,
without material alteration of its coastal outline.

The most ancient part of the longitudinal limestone arch, as tack of
Havana, has been removed down to the older metamorphic rocks, and
a strip of the older limestone formations running parallel to the coast
remains between this valley and the sea. (See Plate I. Figs. 1 and 2.)
This, in turn, by a cross erosion of the streams, is serrated into frag-
mental reumants of the limestone, like the Moro and Castillo Principe
Plateaus at Havana. The Pan de Matanzas, near Matanzas, and the
peculiar mountains of Moa and Yunque, near Baracoa, are remnants of
older and higher levels which have been presrrved in this manner. In
fact, the erosion has been so great that tlie limestone is almost removed,
e.Kcept where preserved in ■':;olated mountain buttes like the Sierra Yun-
que and the Pan de Matanzas, and headwater ei'osion is constantly de-
stroying the remnants of the original limestone plain by deepening the
cols down to tlie metamorphic floor. Granting that the older limestones
once extended over most of the islaml in the contour of a low dome,
and that this arch has un ergone several periods of intermittent eleva-
tion with corresponding intervals of rest, accompanied by base levelling,
the topography can be more easilv explained.

"Where the limestone is the prevalent formation, as in the sugar
country of central Cuba, the surface is marked by extensive level tracts,
covered with the deep residual tierra colnrnda, one of which plains is
well shown in the accompanying photograph. (Plate VI.) The continu-
ity of these plains is broken by abrupt hills, either single or in groups,
some of which seem to have no persistent axis of direction, and are clearly
remnants of the higher level below which the land has been degraded.
The plains show very little slope to the eye, and project abruptly to the
foot of these limestone hills. They varv in size from many square miles
to a few arrcs. Even the small {)Iaius, when entirely encircled by

262 BULLETIN OF THE

mountains, are very flat, and abruptly terminate against short lines of
hills, and there is no reason to doubt that they and tlie limestone hills
are the product of the miequal resistance of the different limestones
above described, the mountains representing the remnants below which
the plains developed on another harder plane. (Plate Y.)

The eminences of Cuba called mountains, with the exception of the
Sierra Maestra and kindred ranges of the Santiago coast, can now be
easily classified. (Plate I. Fig. 8.) They are all either (1) the direct
remnants of the old limestone covering carved out by circumscribing
erosion, or (2) inequalities of the ancient metamorphic floor from which
the limestone has been denuded.

The mountains of the former kind may be placed in two general
classes, according to their altitude and degree of erosion. First are the
high limestone peaks, mesas, and riilges, having an altitude of from one
thousand to two thousand feet. The Sierra Yuuque of Baracoa, the Pan
de Matanzas, and the Tetas de Managua, are examples of isolated peaks,
standing close to the north shore of the island. Each of these is
surrounded by deep drainage valleys cut almost down to sea level.
They are many miles away from any masses of land of similar altitude,
and form conspicuous landmarks along the coast. Their summits are of
the sub-horizontal strata of old limestone, while the base of at least one,
the Sierra Yunque, consists of the older metamorphic rocks. The high
ridges, like the Sierra San Juan, and the high summits of the central
portion of the island, ai"e remnants of the same old level, and differ from
the more isolated peaks in having been less dissected. Xot owing their
outline to any structural folding, but being entirely the product of the
drainage, these have no regularity of arrangement or trend, but are
found in irregular patches throughout the island.

The Spanish language, to which our geographic nomenclature is
already so much indebted, has provided an appropriate name for moun-
tains of this class, which have lower altitudes, ranging from four hun-
dred to seven hundred feet. These are the cuchillas, or knives, so called
because of the numerous sharp salients marking their slopes, caused by
the deep incision of the old plain or general level of which they are the
fast fading remnant. These are the hills forming the sharp background
to the coasts, especially at the east end of the island. The cuchillas are
generally composed of the old limestone, which dips at many angles and
degrees, but sometimes they consist of a complex of limestones, yellow
beds, radiolarian beds, and the old mctamorj)hic floor. At the Yumuri
River of the east and around Cape Maysi they consist of a more massive

MUSEUM OF COMPARATIVE ZOOLOGY. 263

and unbroken wall of the old limestone, but as we go westward they
become more dissected, as in the line of high hills along the coast, and
in tlie background against which the little harbors are cut out as far
west as Xiievitas. Still farther westward the contour recedes slightly
inland. On the south or Santiago coast, the same level of the cuchillas
summits is preserved in diorite and syenite.

In addition to the limestone mountains of erosion described, there are
many low hills in the central part of the island adjacent to Villa Clara
(Plate IV.) and Puerto Principe which are clearly structural remnants
of the older metamorphic floor, from which the folded limestones have
been eroded, the latter being often preserved on top of the higher ele-
vations, or sharply inclined around their edges. The series of sharply
rounded hills between Havana and Matauzas is also the result of the
wearing away of the limestone covering down to a floor of tuff's and ser-
pentines, which, owing to its softer nature, is more deeply and sharply
sculptured than the limestone regions proper.

Concerning the geology of the Sierra Maestra of Santiago, Kimball
says that the old limestones preserved on their slopes show that at
least twenty-three hundred feet of their elevation are Post-Tertiary, and
there is no recorded evidence of any Post-Tertiary eruptives or flows.
I incline to believe that these ranges belong in the same class with those
of the Villa Clara type. Although the close of the Tertiary was marked
by much folding, recognizable mountains simulating Post-Tertiary struc-
tural folds, or evidences of Post-Tertiary extensive volcanic action, are
certainly rare, if they exist at all. The present irregularities are all
the result of erosion. I made every possible reconnoissance over the
island to study the upland topography, and I think my conclusions are
founded on abundant evidence, proving beyond doubt that the higher
limestone elevations are solely the remnant of the former area of the
older limestone mass. For instance, the sharp lines of limestone sum-
mits on the high divide of the island between Havana and Batabano are
clearly the old scarps of the Armendaris drainage cut out of a former
plateau. The mountains on the road from Havana to Villa Clara and
back of Matauzas are either of similar character, or are the perimeters
(knobs) of vast basins, like the sink-holes of other limestone regions,
only much larger, owing to the more solvent nature of the substructure.

The isolated mountains of high elevation along the north coast, like
the Pan de Matanzas and the Vunque, are fragments of the older areal
summits, which have been separated by circumscribing erosion from
the main body of the upland, and stand as solitary remnants of the

264 BULLETIN OF THE

ancient and older plateaus, which, when their limestone cap is finally
removed down to the metaniorphic base, will assume the Villa Clara
type.

The drainage system of Cuba is extensive. In general the streams
flow from the central axis toward the opposite coasts, and are of the type
which modern geographers would term simple consequent or autogenous
streams. They have light-colored blue water like that of the limestone
springs of Florida and Texas, and in many cases obtain their supply
from the undeiground waters of the limestone region. Where seen
throughout the interior upland plateaus of Cuba they are small in
volume, and flow in slightly indented channels in wide valleys, and are
remarkably free from the incisions of lateral drainage. A typical stream-
way is shown on Plate IX. They do not possess deep barrancas or
canons until they begin to cut across the edge of the CuchiHa plateaus
near the coast. In many cases these rivers are intermittent, disappear-
ing into and reappearing from the cavernous limestones. As they ap-
proach the escarpments of the coastal platforms, they reach the sea either
by sinking into tlie limestone, by tumbling cascades, or by cutting deep
vertical caiions. They are all slightly tidal at their mouths, the salt
water extending at high tide a short distance up them, but never reach-
ing far inland of the soboruco. Most of them bring down to the coast
the metamorphic and igneous rock of the old nucleal foundation, Init in
no case have I observed limestone fragments, although the rivers must
degrade and transport in solution far more lime than any other material.
The vertical canons in some cases, like that of the Yumuri of the east,
extend to the sea, and testify to the rapid rising of the land, confirming
the story of the cliffs and base levels as will be described later. (See
Plate I. Fig. 6.)

Perhaps the most important factor in the evolution of the topographic
conditions of Cuba is the superficial and underground destruction and
alteration of the limestones by solution, lieretofore mentioned. Owing
to the porosity of the limestone rocks, the drainage of Cuba is largely
underground in the limestone regions, and flowing surface streams and
lateral drainage channels, such as the dendritic headwater ramifications
so common elsewhere, are notably scarce in the higher region inland
from the coast, and the upland limestone region seems to be dissolving,
rather than oorrading, into a sink-hole topography of vast fjroportions.

The Coastal Topography. — None of the topographic features of Culia
are so peculiar as the innumerable subcircular harbors^ which indent its

1 See Agassiz, Bull. Mus. Comp. Zoul., Vol. XXVI. No. 1, Tlates XIII., XIV.

MUSEUM OF COMPARATIVE ZOOLOGY. 265

northern coast, examples of which are common on all the pilot charts.
Most of these are constructed upon the same fundamental type, consist-
ing of a subcircular or reniform bay outletting through a narrow neck
or strait into the sea. Into the back of the bay usually flow one or
more of the small rivers of the country. Generally the landward side
of these harbors is or has been the elevated, broken Cuchilla highland,
while the points of the narrow necks enclosing the outlets to the sea
are sub-level plains composed at the sea margin of soboruco or recent
elevated reef rock.

On the landward side of some harbors at the foot of the cuchillas,
those of Havana and Baracoa, for instance, there is sometimes a playa,
or alluvial plain of small area, composed of ancient sediments of the
river, which has participated in the general elevation of the coast.
The accompanying plate (Plate II.) enables us to discuss more intelli-
gently these phenomena, and their bearing upon the elevation of the
island.

There are two possible hypotheses concerning the origin of these
harbors. The first is that of subsidence and superimposition, as set
forth by Crosby,^, given more fully in the portion of this paper treating
of evidences of subsidence. This implies that the elevated reef rock
once extended across the area now occupied by the neck or outlets, and
at a former epoch of elevation was eroded through by the rivers, and
that by subsequent subsidence -the waters of the sea encroached upon
the land through the channel tlius worn, producing an estuary. A
second hypothesis is that they are the result of the growth of fringe and
barrier coral reefs adjacent to or opposite the mouths of rivers, which
were subsequently elevated and unequally eroded. In my opinion, the
harbors were evolved from the simple type of rivers now emptying
directly into a fringed reefed sea, like that of the Rio Yumuri of Baracoa
and the Limones. (Plate II, Figs. 1 and 2.) The rivers all originally
emptied directly into the sea, as do the Yumuri and the Limones of
to-day, and the coast line was the precipitous bluff of the Cuchilla high-
land, now forming the background of these harbors, in front of which
was a basal shelving beach. Delta material was discharged off their
months into a deeper area between tlie shore and an outlying barrier
reef, as now seen in the harbor of Jaragua, or into reefless submarine
areas ])roduced in the following manney. The entrance of the fresh
water iut > tlie sea prevented the growth of reefs immediately opposite
the mouth of the river as for out as the freshening influence of the

1 Op. cit.

266 BULLETIN OF THE

river water was felt, which includes the delta region of coarser gravel
deposit. But it can readily be seen that certain physical sediments,
like tine gravel, will receive enough impulse from current and surf to he
carried into the margins of the salt water, so that unusually opportune
conditions are created for coral growth immediately where the water be-
comes sufficiently salt and food abundant. Thus it is that fringe reefs
do not usually form immediately at the mouth of rivers, but leave non-
coralline gaps in the reef simulating an extended submarine channel of
the river. This is clearly shown in the chart of Limones (Plate II. Figs.
1 and 2), and other rivers, where the present submerged fringe reefs
make a projected channel into the sea.

It is the rule, whether the land is subsiding, rising, or stationary, that
the sea always indents the mouths of rivers after they have once reached
its level, and tends to wear away the angular points bordering its mouth.
This wearing is produced by the diurnal change of tidal level, and the
resulting constant corrasion, however small, of the bottoms, whether by
fresh or tidal current ; so the level of the sea, even in a delta-making
stream of perceptible age, will constantly encroach inland and cause
small estuarine deposits in the indented mouth at high tide, to be
moved outward with the ebb. Thus it is that the steep rivers of Cuba,
which are all very old and permanent, have slightly indented base level
with deposits of gravel extending inward coincident with the fluctuation
of the tide.

The playa deposits found along the interior border of the harbors of
Havana and Baracoa represent the coarser gravel and silt thus formerly
given lip by the rivers upon reaching tide level, before the latest eleva-
tion. In the present Yumuri of the east the flood tide extends a mile
or less up the river. At the time of the general elevation of the coast
reef, the older delta deposits similarly formed were elevated correspond-
ingly, and are now found surmounting the lowest terrace.

Snch an elevation as has taken place, and has produced the elevated
coast reef, would raise the present growing reefs above the water, so that
they would form indurated points at each side of the river's moutli, and
if there were a barrier reef its elevation would convert the old inside
deep into a land-locked harbor, while the old indented gravel would
form playas at the back of the harbor. Probably this is what has taken
place. Furthermore, the sides, of the narrow necks and sea fronts of the
harbors are composed of a harder and more durable stratum of reef rock
than the country back of them. The beach-like sides of the harbor
within the reef-like points are subsequently widened by undermining, as

MUSEUM OF COMPARATIVE ZOOLOGY. 267

now seen at Baracoa, where the surf line, which enters the narrow bay
with gathered force, breaks against the unconsolidated miocene yellow
deposit constituting the sides of the harbor back of the elevated reef,
and huge blocks of the latter constantly topple over into the bay.

There is general remark throucrhout Cuba that the harbors are becom-
ing shallower. Captain Mclntyre, a trustworthy mariner, who has been
trading at Baracoa for over forty years, assured me that within his
memory the anchorage area has steadily decreased, and that ships which
formerly discharged their cargoes at a pier, are now dependent upon
lighters. While it is very probable that the silt from the rivers is a
partial cause of this, it may be probable that steady elevation now going
on, as it has certainly gone on in very recent time, may be productive of
the shallowing.

The ovoid harbors thus developed from the simple type of river
emptying directly into the sea and undermining the contiguous reefs
attain a third stage (Plate II. Figs. 7 and 4) in which the regular margins
become denticulated and irregularly indented by erosion, as seen in the
harbors of Havana and Escondido. In the harbor of Havana the ex-
cessive irregularity of the interior margin is increased by the fact that
the limestone background has been cut through down to the tuffs,
serpentines, and clays underlying it, which degrade into more irregular
topography than that of the limestones.

Terraces • and Benches.

The most striking feature in the topography of Cuba consists of the
well defined terraces and benches which mark its jcoasts in many places.
These are often so distinct, especially at the east end of the island, that
their continuity is traceable for many miles, as they rise abruptly from
the water's level, one above the other, in a series of cliffs. On the west
end of the island they are not so distinctly visible from any single point
of view, for the flat benches are much wider, but they are nevertheless
traceable. In other places denudation has destroyed them.

Besides these benches and terraces, whose integritv is distinctly pre-
served, remnants of older and more denuded levels can be traced, and
for convenience they may be classified as follows : —

4. The Sobornco, or elevated reef level, j

3. Elevated beach and cliff lines and ? 1. Later (Lower) levels,
the Havana base levels. '

2. The Cnrhilla level.
1. Tlie Yuiiiiue level.

\ 2. Older (Higher) levels.

268 BULLETIN OF THE

The Sohoruco, or Elevated Reef Level. — The general extent and topo-
graphic character of the soboriico is explained on a previous page. It
forms the lowest bench immediately adjacent to the entire north coast
and along the Santiago front, and is topographically and geologically an
elevated coral reef. Synchronous with this level are the elevated playa
deposits in the harbors, and the elevated cienaga or mud deposit on the
south side of the island at Batabano.

The Beach and Cliff Terraces. — Ou the east end of the island, the north
coast is marked by three distinct and abrupt cliffs and terraces cut out
of the steep slope of the old six-hundred-foot plain, or the Cuchilla level,
which forms the highland. Between Cape Maysi and Baracoa the coast
is practically inaccessible. The three terraces seen in this region ai'e so
clear and distinct that they are readily visible at one view, and their
continuity is clearly traceable for miles. They can be best understood
from the accompanying figure (Plate I. Fig. 6), and a description of the
coast adjacent to the mouth of the Yumuri of the east. Here the river
empties directly into the sea through a precipitous canon affording a
fine cross-section of the benches, so that their architecture and origin
can be seen. The coastal scarp consists of three narrow sub-level
benches, each surmounted by a vertical cliff. Bench No. 1 is the first
sub-level strip above the sea. This in general represents the level of
the elevated reef, which nearly everywhere forms the low-lying coastal
plain and breaks off at the sea in a surf wall some ten feet in length.
Its interior margin against the base of the first great clift" is forty feet
high, and it nowhere exceeds one hundred yards in width. Imme-
diately off the mouth of the Yumuri River, however, a gravel delta fan
spreads out in brackish water, a hundred yards or so on each side.
The present -submerged fringe reef does not grow immediately where
this delta fan is being deposited across the river mouth, but appears on
each side. If the present bottom, constituted as above, should be ele-
vated forty feet, it would produce a beach exactly similar to the elevated
one now seen ; that is, it would be composed of alluvial gi-avel imme-
diately where the rivar once emptied, and of coral limestone a hundred
yards or more on each side of it.

This lowest terrace (Xo. 1), which is usually formed of elevated reef
rock, is composed of alluvial gravel immediately off the cut of the river,
and of elevated reef rock a quarter of a mile away seaward. This lowest
bench consists of several small levels, the uppermost of which is the
specially well defined alluvial gravel plain.

This old beach abuts against a cliff (No. 2) about one hundred and

MUSEUM OF COMPARATIVE ZOOLOGY. 269

twenty feet high (one hundred and seventy feet above the sea) worn out
of the lower part of the old white limestone. Its sides are vertical in
most places; and inaccessible. This clitf is in turn surmounted by another
bench (Xo. 2), which was likewise formerly an old beach level, from
which any renmant of the deposit that may have once existed has been
eroded. It is covered by a dense growth of vegetation. The river
canon which cuts across these cliffs and benches shows that they are
not elevated built-up coral reefs, but are clearly cut sea terraces in the
old limestone. The second bench is about a hundred feet in width, and
abuts against a second vertical cliff, the summit of which is nearly as
high as that of the first one, or about three hundi'ed and fifty feet above
the sea. The level bench (No. 3) mounting this cliff is similar in appear-
ance to No. 2.

This last bench in turn abuts against a third and uppermost escarp-
ment of the highland, which terminates at a height of from five hundred
to six hundred feet in the irregular upland plain forming the fourth
level above the sea.

The Cuchilla Level. — This fourth level is the genei'al upland plain as
it appears from the sea, and represents the old land from which was
carved the group of cliffs above described. This highest escarpment
forms a comparatively unbroken plateau at the eastern end of the
island, overlooking the sea, but westward the increasing drainage cuts
it more and more into numerous serrated hills known as the Cuchillas,
or " Knives," whose summits have a general culmination of from five
hundred to six hundred feet, and are clearly remnants of the Yumuri
Plateau. These Cuchillas form a very conspicuous coast feature from
Nuevitas to the east end of the island.

The Tunqve, or Higher Level. — A single glance, at the peculiar iso-
lated mountain known as the Yunque, or Anvil, situated six miles west
of Baracoa, is sufficient to show that its sub-level summit is the remnant
of an ancient higher level than that represented by the Cuchillas.^ This
is a magnificent butte, whose summit is put upon the pilot chart at
eighteen hundred feet, and so estimated by Crosby.* The summit is an
ovoid mesa, which is apparently level, but which really shows deeply
carved drainage ways and ancient topography indicating long exposure.
The upper portion is composed of a mass of the older tertiary rocks one
thousand feet in thickness, whose perimeter is an almost inaccessible
cliff. This rests upon a base composed of the metamorphic rocks of the

1 A. Agassiz, Bull, Mus. Comp. Zool., Vol. XXVI. No. 1, Plate XLI.

2 Op. at.

270 BULLETIN OF THE

old Pre-Tertiary nucleus. From this summit oue can look down upon
the Cuchillas, the Yumuri terraces, the elevated reefs, and the wide
expanse of the ocean ; and inland toward a country sliowiug its own
level, overreached by still higher mountains of the Sierra Maestra to the
south. On every side the drainage has cut deep below this peculiar
mountain, carving the low-lying country into an intaglio of serrated
hills.

No one can view this summit without being impressed with the won-
derful story it tells of the great erosion that has taken place around it,
as well as the fact that the difference in elevation between its plateau
and that of the lower-lying Cuchilla level represents a vast hiatus in the
history of the island's elevation, — a long period during which land
stripping and degradation ensued, reducing the surrounding areas to
the old Cuchilla erosion level. The relation of this peculiar remnant
to similar phenomena in other parts of the island will be discussed later.
The accompanying illustration (Plate I. Fig. 5) of the harbor of Baracoa
gives a clear presentation of the various levels seen in the east end
of the island. Let us now examine the kindred phenomena in other
localities.

The Havana Levels. — The western half of the island also presents an
interesting series of topographic surfaces that, for convenience, we will
denominate the Havana levels, which, although varying in expression,
have genetic relations to the cliff phenomena of the eastern end of the
island. In the Havana region the wider area of the terraces makes
the cliffs less conspicuous than in the east. The highest of the dis-
tinctly preserved levels coming under my observation is the one that
encloses the harbor of Havana as seen on the summit of the Moro penin-
sula to the east (Plate I. Fig. 2), and its continuation west of the city,
known as the Castillo Principe Plateau. These have an average altitude
of fifty meters (or one hundred and sixty-five feet) as determined by ^Mr.
Gould, who has made a contour survey of the region. From the still
higher eminences back of Havana, or from nearly any point of view, the
contour of this plateau can be easily recognized, and it clearly represents
an old level of erosion, — either a very wide beach level, or a base level
of erosion. The Castillo Principe peninsula, which represents a portion
of the Moro Plateau, is a narrow flat divide of the land extending north
and south, lying between Rio Armendaris and Havana Bay. Its struc-
ture consists of the gently disturbed older limestone series surrounded
by lower beds of cantera and soboruco, constituting the lower levels
upon which the main portion of the city of Havana and the suburb of

MUSEUM OF COMPARATIVE ZOOLOGY. 271

Carmelo are built. It terminates inland in a limestone escarpment
of erosion, which has worn down to the underlying floor of igneous
tuffs.

There can be no doubt that this peninsula was once an eastward con-
tinuation across Havana Bay of the plateau west of the Eio Armendaris
and the Moro Plateau, and that it has been disconnected from them by
the streams which flow in the intervening valleys. The plateau in
which Moro Castle is situated is similar in surface and structural features
to that of the Castillo Principe, and liliewise terminated inland in the
hilly region north of Picgla by an escarpment of stratification which
destroys the possibility of tracing its former extent inland. In area it
represents an elongated east and west narrow platform forming a vertical
coast line as far as Guanos Point, and extending toward Matanzas, where
I think its level is represented by some of the terraces in that vicinity,
and projecting, in places, as at Moro Point, fully to the ocean's edge.
This general level of the Moro Plateau is not an elevated coral reef, but
is an ancient level of erosion representing a period in the history of the
island when the area it now occupies was approximately near sea level,
and which has been subsequently elevated. Its surface in no manner
represents a deposition plain or the surface of an old reef growth, but is
produced solely by base-levelling erosion, and this in spite of the irregu-
lar sinuosities seen in its substructure, the old tertiary limestone.

The Caniera Elevation. — Around the base of the Castillo Principe
Plateau may be traced the remnants of another level, approximately
twenty-five meters (eighty-five feet) in height, which, for convenience,
I will term the Cantera level. This, too, represents another and later
epoch of levelling, and has likewise been greatly destroyed by later
erosion. Below it, and adjacent to the sea, is the soboruco, or elevated
reef level.

Older Levels. — Back of Havana there is a line of still liigher, greatly
eroded hills, which overlook these levels, and have an altitude of about
five hundred feet, merging southward into a plateau constituting the
divide between the north and south shores of the island. This lime-
stone plateau gently slopes away to the south coast, and undoubtedl}'
once covered the hilly area back of Havana. The highest point on the
railway, which goes through a saddle, is 101 meters (332 feet), the
country rising to about two hundred feet above this.

The Matanzas Levels. — Matanzas Bay diff'ers from the general type of
the sac-like harbors of the north coast only in that it is rhoraboidal
in form, and seems more deeply cut into the high background which sur-

272 BULLETIN OF THE

rounds it. The topography and geology of this locality alone is so inter-
esting and complicated that it would require a lengthy paper to describe
it, and it can best be explained in brief by reference to tlie accompanying
topographic sketch and section. (Plate I. Fig. 4, Plate 11. Fig. 9.)

Two diminutive rivers flow into the harbor, the Yumuri of Matanzas
and the San Juan, both emerging suddenly from the highland. The
highland or sky line surrounding the harbor on the two sides is about
one. hundred meters (three hundred and fifty feet) in altitude, as deter-
mined by aneroid at Mount Serat, and constitutes a flat-topped mesa or
plateau north of the Yumuri, and a poorly defined bench against a still
higher hilly region south and east of that river. Out of this plainly
marked level are carved the sloping and narrow lowlands immediately
surrounding the harbor, upon which the city is built. Between the
level of the city and the highlands the narrow remnants of a few ter-
races or pausation planes are faintly traceable. One of these is about
one hundred and fifty feet above the sea, and the other, upon which is
located the railway station back of the city, and which constitutes the
bench back of the Versailles church north of the I'umuri River, is about
fifty feet.

The Yumuri enters the harbor valley through a deep precipitous
canon cut athwart the high level above described. Viewed from the
city, this cafiou appears to be a chasm in a mountainous background.^
Upon ascending it for half a mile it is seen to open out into a wide and
beautiful amphitheatre, some four leagues in circumference, bordered by
steeply sloping walls, and with a wide sub-level bottom. The bottom
of this valley is only a little above sea level, and if submerged a few
feet would become a circular harbor from the inflow of the sea.

Upon climbing to the summit of the canon to the Church of the
Hermit, upon the high level, a grand view of tliis peculiar amphitheatre
is seen. It is clearly carved out of a vast sub-level plateau having the
general altitude of the Mount Serat eminences, whose remnants consti-
tute the plateau lying between the Yumuri River and the sea on the
west side of the harbor to the north and east of the amphitheatre.
Traces of this plateau ^ also surround the south raai-gin of the amphi-
theatre, forming a bench from which I'ises a line of higher hills, more
serrated, — the same which are crossed and seen between Havana and
Matanzas, and which do not exceed six hundred feet in altitude. This
remarkable valley and the more remarkable canon which connects it

1 Soc A. Agassiz, Rail. Mus. Comp. Zool., Vol. XXVI. No. 1, Plate XLII.

2 Ibid., Plate XLIII.

MUSEUM OF COMPARATIVE ZOOLOGY. 273

with the sea are subjects conceruing which I am at a loss to offer
explanation.

The geologic section (Plate I. Fig. 4) of the caxion shows that there
are at least three alternations of old reefs and thin gravel beds, which in
turn rest upon a great thickness, not less than eight hundred feet, of
the older Tertiary limestones, out of which the amphitheatre proper is
carved. The beds are all tilted to an angle of fifteen degrees, but the
highest elevation of the undoubted coral laid down against the old
limestone is less than one hundred feet.

That this amphitheatre was once an indented harbor and the Yumuri
River canon its outlet, is a hypothesis which may he suggested. The
denuded floor shows no trace of evidence that would convey this impres-
sion, but around the walls of the amphitheatre are traces of terraces
corresponding in height to the hundred-and-fifty-foot bench outside
the harbor, and these may represent the former floor of the amphi-
theatre when at sea level. If they do, then the Versailles, or highest
elevated reef rocks, were formed off the point of an old outlet through
the Yumuri. In the canon itself, however, there is no distinct evidence
of planation terraces, such as would indicate pausation periods followed
by renewed epochs of cutting, although just out of it on the west side of
the harbor, back of Versailles church, old river gravels are preserved
about twenty feet above sea level.

Between Havana and Matanzas the interior is a very broken country.
The railway runs back of the interior of the escarpment of the old coast
limestones, and sub-parallel to them, for thirty-six kilometers from
Havana, upon a floor of underlying metamorphics, constituting a very
hilly country. At thirty-seven kilometers the railway again cuts the
bottom of the limestone at an altitude of two hundred feet, enterinjr a
level limestone plain at Aguacate, separated by a deep eroded valley
from a range of limestone hills two kilometei's to the north. At sixty-
two kilometers the road cuts through this range of tertiary limestone
hills, which have an altitude of six hundred feet. At Serba Mocha the
peculiar limestone hills known as the Pan de Matanzas are seen to the
north across an eroded valley. These summits are to the western half
of the island what the Yunque is to the eastern, — remarkable isolated
remnants of the nearly destroyed older levels which once surmounted
tlie island. The Pan de Matanzas is alleged to be twelve hundred feet
high. It consists of a double eminence, the intervening valley present-
ing precipitous walls. The summits are of limestone, and are clearly
remnants of the old limestone mass of the interior, from which they have

VOL. XVI. — NO. 15. 18

274 BULLETIN OF THE

been disconnected by erosion, while the intervening valleys are cut down
to the metamorphic floor. The Tetas de Managua are of similar nature
and origin.

The Santiago Levels. — Mr. James P. Kimball has published a paper
entitled " Geological Relations and Genesis of the Specidar Iron Ores of
Santiago de Cuba," ^ which gives valuable details concerning the occur-
rences of terraces on the south coast of the east end of the island in the
vicinity of Santiago and Guautanamo. Concerning these he speaks as
follows : —

" The immediate coast presents a remarkable development of coral rock, or
coral limestone, in three terraces, of which the upper is about 350 feet above
the sea. The second terrace is at an altitude of about 175 feet, and the present
shore, a plateau of comparatively recent elevation, about fourteen feet above
tide. These terraces mark successive elevations of the Sierra Maestra ranse.
These stages of elevation were in direct, but probably remote, succession with
other elevations which I shall show to be indicated by traces of more ancient
corallines (coral formations) about two miles still farther back from the present
coast.

" The last terrace, or that of the present shore, falls away vertically into deep
water soundings, at the mouth of the Carpintero, 150 feet off shore, giving a
depth of 165 feet. It retains to a remarkable degree the structure of solid reef,
studded with distinct forms of coral, and is strewn with fragments of coral
rounded by the waves, but in good preservation, and numbering a large vari-
ety of species."

He also shows that traces of the old limestones are found in the high
flanks of the Sierra Maestra. Of these he says : —

" The several terraces of recent coralline mark, as already- indicated, succes-
sive and in chronological order the later uplifts of the Sierra, in vertical range
not less than five hundred feet. These, together with the series of corallines
of the second line of foot-hills, as recognized by the bodies of hematite and
marble, are proofs of a sum of uplifts of not less than thirteen hundred feet.
Obscure traces upon the first range of foot-hills of still more ancient corallines,
to which I shall again refer, point to a still more remote succession of uplifts
whose vertical range — referred to the latest indicated level of coral formations,
some one hundred feet below the present shore — may be estimated at about
twenty-three hundred feet. From the syenite hills may have disappeared by
subaerial erosion intervening corallines, between those of the present coast and
the line of ancient and now metamorphosed corallines traced along the contact
or southern margin of tlie diorite mantle."

1 American Journal of Science, December, 1884.

MUSEUM OF COMPARATIVE ZOOLOGY. 275

Comparison and Correlation of Various Levels.

Let us now compare the various data presented concerning the coast
and inhmd topography at the various localities mentioned, and inquire
into their relation to each other. Figure 8 of Plate I. will aid in under-
standing the presentation to follow.

That the soboruco or elevated reef represents the same general level
around the north and south coasts of Cuba is indisputable, and can be
interpreted in no other way than that there has been in recent time a
uniform elevation throughout the nine hundred miles represented in the
length of the island. It is the same formation topographically and geo-
logically, wherever seen, and establishes the fact that the elevation of
the island, at least during one epoch, was general, and not local or spas-
modic. If such a uniform movement has beyond doubt taken place at
a modern epoch, it establishes the principle that similar elevations were
not impossible in the past.

The levels represented in the three terraces of the Yumuri of the east
have remarkable identity with the levels of the west end of the island,
as at Havana and Matanzas, where my detailed studies were made. The
only ditference is, that the latter are wider than the former, owing to
the lower and more rounded character of the country out of which they
were cut. The correspondence in altitude is such that no one can doubt
that they represent synchronous and identical regional movements and
pausations, and that they were once continuous throughout the length
of the north coasts of the island, and around Cape Maysi to the Santiago
coast.

The Cuchilla, or dissected peneplain of the east, presents a remarkable
analogy to the higher dissected summits back of Matanzas, constituting
the upland divide of the west end of the island in the latitude of Havana.
Hei-e the old levels represented by these summits are less distinct than
in the east, probably owing to the fact that this end of the island had
not previously been so highly elevated as the east.

The oldest and highest limestone summits, appi'oximating from fifteen
hundred to two thousand feet, as typified in Yunque, the Sierra del
Moa, the Pan de Matanzas, the table land of Mariel, and the Managua
Paps of the west half of the island which follow near the north coast, the
highest limestone at Santiago and other places, represent the remnant
of the oldest and highest level or levels, which have been so completely
dissected and planed down that their extent can only be estimated.

276 BULLETIN OF THE

These elevations may be only the remnauts of an aboriginal uneven
surface, but collectively they generally represent a higher land than
existed before the Cuchilla plains were developed. Whether the high
summits of the Sierra Maestra adjacent to the Santiago coast contain or
preserve traces of still older levels is an interesting problem for the future.

These phouomena may now be grouped into three distinct age cate-
gories, one of which is still further divisible into many subdivisions.
These are (1) the modern or well preserved tripartite group of lower
lying levels, cliffs, or terraces, including the modern soboruco, tlie highest
level of which approximates three hundred feet ; (2) the dissected and
greatly denuded remnants of the old Cuchilla level, five hundred to seven
hundred feet above the sea, the remnant of an old general height whose
integrity is almost destroj'ed, and which is less easily traceable than the
first ; (3) remnants of the almost destroyed more ancient upland, as
preserved in the isolated buttes of the Yunque and Pan de Matanzas
type and the higher limestones of Santiago and Cienfuegos, which
demonstrate that there was once an old surface at least two thousand
feet above the modern sea level.

The obvious history of these levels is as follows : —

(1) In a period near the close of the Tertiary, to be ascertained, long
previous to the emergence of the present elevated reef and cautera and
the erosion of the Cuchilla plain, there was a great upward movement
of the island to the height of at least two thousand feet, which as yet
has revealed no history of its details further than that, from the absence
of later deposits and from the character of its ancient and much sculp-
tured topography, we may fairly infer that it has not since subsided
beneath the sea, but has remained mostly dry laud, and that its area
and outline were very neai-ly as great as those of the island of to-day.
This includes those portions of the island above the dissected Cuchilla
plain.

(2) The Cuchillas at five-hundred-foot level constitute a plain or
plains produced by base levelling in the epoch immediately following
this oldest period of elevation, and represent the time interval between
it and the later movement recorded in the first or lower group. The
country was planed down by erosion to near sea level. The Cuchillas
summits indicate a long pausation period between the old Yunque and the
renewed modern elevation recorded in the Yumuri cliffs cut around them.

(3) The tripartite group of modern cliffs and base levels below and
against the Cuchilla escarpment are the product of a renewed and mod-
ern upward movement, which elevated the old Cuchilla coastal plain to

MUSEUM OF COMPARATIVE ZOOLOGY. 277

a plateau, and subjected it to the erosion whicli has since dissected it
into its present rugged outlines. The Yumuri cliffs were carved from
it where it formed a sharp coastal scarp, and the Havana and Matanzas
lienches represent synchronous levels with the latter in the west end of
the island, where the Cuchilla level was of less extent. That this mod-
ern group of elevations was intermittent, as shown by its alternate cliflFs
and terraces, is evident, the modern soboruco representing the latest
uplift.

The elevated benches and terraces which border the coast of Cuba,
with the single exception of the soboruco, or modern coast reef, are not
ancient coral reefs either topographically or lithologically, as has been
asserted, but, on the other hand, are beach and erosion plains, produced
by rapid elevation of the island in Post-Tertiary time, and carved from
various formations, principally the older limestones, regardless of struc-
tural arrangement and composition. Even though the old limestones
may be remotely of coral origin, which I do not think, and which idea
I have discussed on a previous page, these old terraces can in no wise be
interpreted topographically as elevated reefs, for none of the original
reef topography is preserved. On the other hand, 1 can give numerous
instances where the same benches are carved out of the varying compo-
nent material, which was much folded or disturbed prior to their erosion,
as is shown in most of the figui'es.

The series of terraces around Cape Maysi and Yumuri are carved out
of a massive matrix of old limestone of undulatory structure, as shown
in the figures. The terraces or base levels at Matanzas are cut out of a
series of beds widely divergent in lithologic composition, all dipping at
angles of from ten to twenty-five degrees. The Moro and Principe Pla-
teaus at Havana form a planation surface upon a floor of folded limestone,
in which distinct anticlinal structure can be traced. The terrace i;pon
which the Military Hospital at Baracoa is situated is carved across tb.e
almost vertically inclined edges of the older Miocene limestones. The
summit of the Yunque, instead of being a coral reef, is a greatly degraded
peneplain. The soboruco alone of all the levels is topographically an
elevated reef, and this, as before stated, dees not rise anywhere over
fifty feet.

Lack of Evidence of Subsidence.

I must confess my inability to distinguish any positive evidences of
subsidence since the beginning of Tertiary time or accompanying these
elevations, although it would be rational to think that tlie movements

278 BULLETIN OF THE

must have been oscillatory. It is easy to imagine evidences of subsi-
dence, but to prove them is difficult. Geikie has said : —

" It is more difficult to trace a downward movement of the land, for the evi-
dence of each successive sea margin is carried down and washed away or covered
up. . . . The student will talie care to guard himself against being misled by
mere proofs of the advance of the sea on land. In a great majority of cases
where such an advance is taking place, it is due not to subsidence of the land,
but to erosion of the shores. . , . The encroachment of the sea upon the land
may involve the disappearance of successive fields, roads, houses, and villages,
and even whole parishes, without any actual change of level of the land." ^

I failed to find any traces in the upland areas of recent deposits which
would indicate any submergence whatever. The soils, especially the
" Tierra Colorada," are everywhere residual, and nowhere did I observe
any that could be attributed to transported material or overplacement
of alluvial matter, and particular care was taken to look for such evi-
dence. De Castro reports extensive upland alluvial deposits in the
region of Puerto Principe, but gives no evidence whereby we may deter-
mine whether they were produced by inland deposition or by submer-
gence of the land to sea level. Neither do the rivers show any revival
or other evidence of such subsidence, but all have continuous down-
ward cutting sections.

Whether there has been recent subsidence immediately preceding the
deposition of the elevated coral reef of soboruco, whereby the circular
harbors were produced, as Crosby alleges,^ is also a point which is diffi-
cult to determine. In support of his position he points out the structure
of the circular harbors and the great thickness of the older limestones,
which he believed to be ancient reef rock. 1 have endeavored to show
on a previous page that there is no evidence to support the theory that
the older and elevated limestones were coralline reefs in origin, and
hence it is not necessary here to discuss this testimony further.

Concerning the circular harbors, however, I must confess that Mr.
Crosby has some reason for his argument, although the evidence may
be more strongly interpreted to indicate elevation. Concerning these
harbors, he says : — ■

" The coast of Cuba is probably not rising now, at least not at all points.
On the beach near Baracoa the erect stumps of large trees may be seen,
standing where they grew, near the low-tide mark. The numerous harbors
of Cuba are nearly all formed on one plan, of which Baracoa is a good ex-

1 Text-Book of Geology, p. 261. 2 Qp, dt.

MUSEUM OF COMPARATIVE ZOOLOGY. 279

ample. It is an approximately circular, almost completely landlockerl basin,
communicating with the sea through a narrow but deep passage between broken
walls of the coral rock. The larger harbors departed from this plan chiefly in
their more irregular outlines, all agreeing in having deep, narrow mouths-
Every harbor is at the mouth of one or more rivers, and their inlets, as I con-
ceive, are the work, not of the sea, but of rivers at a time when tlie land was
higher than now. While the main body of the harbor, in each case, is simply
tlie broader and older portion of the river valley behind the barrier reef, which
has been invaded liy the rising sea, the circular form of many of the smaller
harbors is largely due to the fact that the sand brought down by the rivers is
thrown up by the sea into curved bars, cutting off the inequalities of the shore.

" During the formation of the most recent of the elevated reefs, which, as
already stated, forms a level about thirty feet above the sea, the mouths of the
smaller streams were behind the reef, discharging into irregular channels or
basins between the reef and the shore. On account of the turbidity and fresh-
ness of the water, the reef, especially on its inner border, grew less rapidly at
these points than elsewhere, the basins behind the reef becoming filled with
debris from the land. "When the reef was finally raised to something above its
present level, each river scoured out a large part of the sand and gravel which
it had deposited, and cut a narrow channel through the reef itself. During
this period of elevation, Cuba, like most rising lands, had few harbors, but
when subsidence began the sea occupied the channels and basins which had
been excavated and cleared out by the rivers, and thus a large number of har-
bors came into existence.

" Opposite the mouths of larger rivers, such as the Toar and the Molasses
in the vicinity of Baracoa, the reef in question w\as interrupted, and these
streams discharged into broad, open bays, while the lower portion of their
valleys show equally with the harbors that the land is sinking. They are half-
drowned valleys filled to a considerable depth with land detritus, conditions
■which could not exist if the land was rising or had risen." ^

The interpretation of Ihe evidence of the harbors depends upon the
correctness or incorrectness of Mr. Crosby's hypothesis that the narrow
outlets through the reef rock represent a channel cut by the scouring of
the rivers themselves. It may be that they are channels representing
originally areas of non-coralline growth, such as ai'e now known to exist
in submerged reefs remote from areas with developed land streams and
in atolls, and such as biological laws tell us should exist opposite the
mouths of rivers, — such channels as now exist off shore around the coast
of Cuba, where reefs are growing. Mr. Crosby admits that the reefs
grew less rapidly at these points than elsewhere, on account of the
turbidity of the waters.

The channel of the Havana harbor and the canon of the Yumuri

1 Op. cit.

280 BULLETIN OF THE

River of Matanzas are certainly eroded in the manner Mr. Crosby alleges,
not through coral reefs, but through older rocks which have been ele-
vated across the tracks of the rivers, though most of the harbor necks
in the east end of the island are certainly old submarine reef valleys,
resulting merely from the fact that the coral has grown up around them.
The harbor of Havana is a much better example of supposed subsidence
than is that of Baracoa, but even here the channel cut out of the old
Tertiary walls of the harbor does not necessarily imply that the land was
formerly higher than now, for the heavy surf may be seen cutting many
similar indentations into the limestone sea front, which action, with the
aid of that of the rivers, could have easily made these indentations.

Concerning the mouths of the rivers themselves, their alluvial de-
posits and the evidence of their valleys may bo interpreted to mean
elevation more positively than Mr. Crosby interprets them to mean sub-
sidence, nor can I understand why he calls them " half drowned." There
is a singular absence of fiord-like valleys or indentations, or of ancient
estuarine deposits, around the coast of Cuba, such as ordinarily indicate
subsidence. (Plate I. Fig. 6.) In fact, the rivers in nearly all cases,
like the Yumuri of the east, run directly to sea level through almost
vertical chasms cut straight across the line of terraces, and are void of
any terraces within their canons, showing unmistakably that they have
been cut down to sea level since the terraces and their own deltas were
elevated, and that there is no superimposition indicative of subsidence
previous to the reef-making epoch.

That some of these rivers do at present reach tide level a shoi't dis-
tance from the beach is true, but so short is this distance that vessels
can always obtain fresh water from them by sending light boats up
stream less than a mile. I think that this slight indentation of tide
level up these rivers is indicative, not of " drowning," or of an ancient
subsidence, but that, on the contrary, it means merely that the rivers
and surf are doing their normal work of degrading the land. If they
were really drowned rivers they would be navigable some distance inland,
but in the three largest streams, the Armendaris of Havana and the two
Yumuris of Matanzas and Baracoa, I found it impossible to go inland
over a mile in the shallowest row boat, being soon retarded by rapids.

On the other hand, these streams are new forming delta deposits in
places outside their mouths, which is more indicative of present eleva-
tion than of subsidence. Furthermore, at tlie mouth of the Yumuri of
the east these deltas were also formed iuiinediately preceding the eleva-
tion of the coast I'cef, which may be accepted as evidence of elevation at

MUSEUM OF COMPARATIVE ZOOLOGY. 281

that time. At any rate, if there had been any long epochs of subsi-
dence, they would be recorded in great fiord-like valleys or low passages
acr'oss the central axis of the island, sucl^ as do uot exist, and to which
the oval harbors may not be compared, for the origin of these is due
entirely to the pre-existing fringe reefs.

It might be alleged that all the ancient topography showing subsi-
dence is still beneath the ocean level, and that the angular edges of
Cuba are indicative of the fact that the present outline merely represents
an ancient summit which is re-emerging. The submarine topography,
however, is not within the province of this paper, but I agree with Pro-
fessor Agassiz that its irregularities were indicated long before the period
of history herein recorded. The three alternations of gravel and reef in
the Matanzas section may also have indicated slight alternations of sub-
sidence and elevation.

AVithout committing myself to an emphatic negation as yet, I must
confess that no evidences of great subsidence are apparent at present,
and although I hold my conclusions upon this subject in abeyance to
future observations, I seriously doubt its existence.

It is now possible, with the aid of the stratigraphic and paleontologic
data previously given, to make a few conclusions concerning the Ceno-
zoic history of Cuba. It has been shown by the stratigraphy that the
topographic levels are not old reef levels, but that, with the excep-
tion of the modern reef, they have all been carved out of the old
Tertiary limestones, which had previously been folded and disturbed by
movements that could not have been generally uniform, but which were
erogenic, and hence the present bench topography of Cuba is subsequent
to this period of Post-Tertiary wrinkling, and represents a different kind
of movement, which was regional or epeirogenic. Since the old folding
or orogenic movement occupied at least a small portion of Post-Tertiary
time, we may reasonably conclude that the two great periods of uniform
uplifting recorded in the old levels must have taken place at least since
the beginning of the Pleistocene. In other words, they are compara-
tively modern in geologic time, — some of them absolutely recent.

It is not maintained in this paper that these epochs of regional
elevation were continuous and uninterrupted, or unaccompanied by
pauses or even alternating epochs of subsidenco, but that their general
progress has been periodically upward, and that, if there were epochs of
subsidence they are difficult to distinguish, and were of short duration,
and insignificant in comparison with the ereat uplifting movement that
has generally taken place.

282 BULLETIN OF THE

III. RESUME AND CONCLUSION.

The known geologic history of Cuba may be stated as follows : —

1. In Pre-Tertiary times an old land existed, almost as extensive in
area as the present island. Whether this old land was insular, multi-
insular, or connected with other Antillean areas or the mainland, I will
not speculate. The submarine topography indicates that it was not.
Its composition ami structure, however, show that it was an area of
active vulcanism accompanied by great metamorphism and eruptive
flows. If there are preserved in it any traces of Pre-Tertiary sedi-
mentation they are largely overwhelmed and almost obliterated by the
vulcanism, metamorphism, and later erosion. Paleozoic, Triassic, Jurassic,
and Cretaceous sediments have been reported by De Castro ^ in localities,
but their physical history is unknown.

2. It is also certain that during Tertiary times, embracing the Eocene
and Xeocene periods, this ancient nncleal land, with all of its geographic
outlines, completely subsided beneath sea level, and that it was covered
with limestone sediments, which were organically derived from the sea,
not the island itself, for there is no semblance of limestone material in
the rocks of the Pi'e-Tertiary land which could have furnished material
for the Tertiary rocks. That this subsidence was profound we may
reasonably conclude from the thickness of the older nucleal region, now
visibly covered by the limestone beds, which have been horizontallv ele-
vated to a heiglit of at least two thousand feet. In other words, the
Pre-Tertiary subsidence may have been at least to an equal depth.
Durin": this epoch of Tertiary subsidence a thousand feet of Tertiary
limestone were accumulated over the old nucleal island.

3. After the close of Tertiary times the Tertiary sediments were
greatly warped and folded, concurrently with an emergence of the land
from the sea. This movement was orogenic.

4. Following this began the epoch of epeirogenic or regional elevation.
During' Pleistocene time the island underwent the first of these upward
impulses to its present height, with the exception of about six hundred
feet represented in still later movements. This older Pleistocene or
Yunque elevation raised the main area to a height of at least two thou-
sand feet in its eastern half, and fifteen hundred feet in its western half.
How much higher it extended we cannot tell, so great has been the
erosion. This elevation was so rapiii and general thi-oughout the island

1 Op. cit.

MUSEUM OF COMPARATIVE ZOOLOGY. 28

o

that no coastal accumulations are preserved around its perimeter. This
elevation likewise developed the present outline of the island almost in
its entirety, and perhaps in greater area, jvhich has since been destroyed
by erosion.

5. Following this older and greater Post-Tertiary elevation, and inter-
vening between it and the time of the Cuchilla, or five-hundred-foot
level, there was a long period of erosion, cutting down the country to the
Cuchilla plain, which was at that time marine base level.

6. Renewed and general elevation of the island commenced in recent
time, after the period of rest recorded in the Cuchilla level. The later
terraces, sea cliffs, base levels, and modern coral reefs and savanna
deposits of the south coast were tlien elevated. It is also evident that
in this later period elevation was intermittent, accompanied by slight
pauses. It is difficult to exactly fix the time of this latest elevation.
It was certainly very recent, and a considerable period later than the
old Yunque elevation. It cannot be older than late Pliocene, and it
may or may not be in pi'ogress at present.

It is not the province of this paper to discuss the history and origin
of the Antillean sub-continent to which Cuba belongs, but I cannot
refrain from presenting a few thoughts which may be of service to those
who may consider this subject.

The old metamorphic floor represents, beyond reasonable doubt, a land
that existed probably in Cretaceous time, and much of its metamorphism
and igneous extrusion took place in that period. Similar phenomena
have been recorded in Santo Domingo and Jamaica. In fact, it is not
proved that any rocks older than Mesozoic existed upon any of these
islands. Similar disturbances and excessive vulcanism and metamor-
phism are known to have been extensive on the Cordilleran region of
the North American continent, including all of Mexico and the Cordil-
leran region of the United States, which have been described by the
writer and others, and lately most aptly termed by Lawson the ]\Ieso-
zoic revolution.^

As I have shown in a previous paper,^ the marine waters extended
across the isthmian region of the American continent, at least during
the earlier half of Cretaceous time (the Comanche epoch). Whether
this was by union of the two oceans, or by an eastward indentation of
the Pacific, or vice versa, I am not prepared to say. It is certain, how-

1 See tlie Journal of Geology, Vol. I. No. 6, September- October, 189-3.

2 The Cretaccotis Formations of Mexico and their Relations to North American
Geographic Development, American Journal of Science, April, 1893.

284 BULLETIN OF THE

ever, that during the Cretaceous, at the close of the Comanche epoch,
great erogenic forces were active, and that the strike of their corruga-
tions constantly bent eastward until in the latitude of Southern Mexico
they were in the direction of this old Antilleau axis, and that the latter
may have been part of the protuberances marking this line of great
erogenic movement, which in general was peripheral or concentric to
the old Appalachian land.

How great an area was involved in the upper Cretaceous and Tertiary
subsidences it is difficult to say. No attempt has been made to trace
the former event in the Antilles. The latter certainly included all the
great Antilles, — a region fifteen hundred miles in length from east to
west, — and the Atlantic and Gulf margins of the North and South
American continents, and probably all the isthmian region, which was
possibly land in Upper Cretaceous time, again connecting the Gulf of
Mexico.

One of the misty epochs in Cuban history is that of the folding and
disturbance at the close of the Tertiary, and I can only suggest that it
belongs with the orogenic phenomena wliich enveloped or overlapped
the periphery of the older Mesozoic Cordilleran region, in Central Amer-
ica, and in northern South America. This involved the Tertiary forma-
tions of the other Antilles, but there is no trace of it along the northern
periphery of the Gulf of Mexico.

I do not mean to say that these vast and apparently uniform regional
elevations which have taken place since tlie earlier folding of the Mio-
cene limestones were unaccompanied by faults or warping, but these are
nowhere prominently apparent, and their importance is secondary to the
former, which were not local, but general or epeirogenic in character,
and involved the uplifting of the whole island approximately, uniformly,
and synchronously.

That this uplifting was confined to Cuba alone of the Antilles, it
would be prepostorous to suppose, and we can in no way avoid the con-
clusion that it represents only a small portion of a great regional uplift,
including much of the surrounding area of the Mexican and Caribbean
gulfs. The adjacent islands must have been involved in these great
regional movements, the periphery of which must have been some dis-
tance from the present island, but I do not allege that the islands were
thereby connected.

Whether the movements can ultimately be correlated with those of
the surrounding American coasts, or the topographic irregularities of
the surrounding ocean floor, is a question which I shall not attempt to

MUSEUM OF COMPARATIVE ZOOLOGY. 285

answer. We have recorded evidence that similar terrace phenomena
occnr in Nicaragua, Yucatan, Jamaica, and San Domingo, and the coasts
of South America have participated in these regional uplifts of Pleisto-
cene and recent time, to which the slight elevation of the Gulf coast of
the United States is insignificant.

In these studies I have found no evidence that Cuba, since its earliest
historj' (the Mesozoic) has had land connection with the United States.
Unless there was some profound subsidence in Post-Tertiary time, such
as I have been unable to detect, no possible deduction can make such a
connection. In fact, I know of no positive evidence that it has been
connected with our continent at all, and have only hypothetical evidence
that the Pre-Tertiary land may have once extended toward the Yucatan
peninsula, and that it was only then, if ever, that the Antillean and
Cordilleran islands were united. Neither can we avoid conceiving that
the subsequent elevations have brought the isthmian region up with it,
making the present land connection between tlie continents.

The axial direction of the general Antillean Post-Tertiary elevation is
approximately east and west, and hence it is presumable that the sub-
marine ridges wei-e more likely to have been extended in that direction,
and that to the north and south of this axis, which must theoretically
be the remnant of a great east and west swell or fold, there must have
existed corresponding sloping sides and synclinal troughs. It is but
natural, then, that evidence of the continuation of the Cuban dias-
trophism must be looked for in east and west lines rather than in lines
north and south.

We can also reasonably conclude that the orogenic development of
Cuba, begun in some unknown period of antiquity, was practically com-
pleted at the commencement of the Pleistocene, — that is, the develop-
ment accompanied by displacement, folding, and vulcanism, — and that
the stage of elevation then began, bringing iip the old Pre-Pleistocene
architecture and carving the mass into its terraces and present outlines.
The group of regional elevations which I have described, although
marking a wide interval of time, all occurred in a comparatively recent
geologic period. To fix this time exactly would be impossible wuth the
scant data at hand, but we can make some approximations.

The oldest of the elevations, now represented by the Yunqne level,
certainly followed the period of folding wliich the Tertiary limestones
underwent after their deposition. This folding, we may safely say, was
Post-Tertiary, and took place in late Pliocene or early Pleistocene time,
approximately, and marks the beginning of the re-emergence of modern

286 BULLETIN OF THE

Cuba, and the teiTaces are all of later age. Before this period, which for
convenience we will call early Pleistocene (properly late Tertiary) it
must be acknowledged that the area of Cuba, crests and coasts, was at
least two thousand feet lower in altitude than at present. We cannot
imagine that such a depression was locally limited to the island of Cuba
or the Great Antilles, or that it would have abruptly terminated along
the east and west axial line, and hence it is not difficult to infer, espe-
cially in the light of existing geologic evidence, that it involved the
isthmian portion of the continent south of the great escarpment of the
Mexican plateau, and that oceanic connection then existed between the
Atlantic and the Pacific, as has been already indicated by the paleon-
tology and by the living forms. -^

^ See A. Agassiz, The Origin of the West India Fauna, Mem. Mus. Comp. Zool.,
Vol. X. No. 1, p. 79, 1883; also, Three Cruises of the "Blake," Bull. Mus. Comp.
Zool., Vol. XIV., 1888.

MUSEUM OF COMPARATIVE ZOOLOGY. 287

EXPLANATION OF THE PLATES.

PLATE I.

Fig. 1. Geologic Section across the Island of Cuba from Havana to Batabano.
Scale, 5 inches to 1 mile. (1) Pre-Tertiary Formations. (2) Tertiary-
Limestones. (3) Soboruco Reef. (4) Mud Deposit of Batabanos.

Fig. 2. Detail of Moro Plateau, North End of above Section. Figures have same
reference.

Fig. 3. Dike near Water- Works, South Edge of Havana. (1) Dike Material.
(2) The same, more weathered. (-3) Supposed Cretaceous Clays.
(4) Surface showing Tertiary Limestone on right.

Fig. 4. Geological Section of the Canon of the Rio Yumuri of Matanzas. (1) Mas-
sive Coralline Cantera, Reef Rock, 85 feet. (2) More Arenaceous Lime-
stone, with Molluscan Remains, 15 feet. (3) Stratified Calcareous Clay,
with MoUuscan Remains, 10 feet. (4) Same, with great number of small
Pebbles. (5) A very white Lime Material, with Bands of Clay. (6) At
Base. (7) Calcareous Matrix, with Pebble, 10^ feet. (8, 9, 10, 11) Mio-
cene Limestone with Molluscan Remains, becoming arenaceous at
Base (12).

Fig. 5. Section at Baracoa. (a) Sea Level ; (b) Elevated Reef Level ; (c) Military
Hospital Level; {d) Cuchilla High Lands (1827), Yunque Level ; (e) Ra-
diolarian Hill. (1) Soboruco. (2) Miocene {1) Limestone. (3) Y'ellow
Clays with Miocene Mollusca. (4) Hill of Radiolarian Earth.

Fig. 6. The Canon and Terraces of tlie River Yumuri of the East. Vertical height,
eOO feet.

Fig. 7. Section near Aguacate, showing Decay of Limestone into Red Residual
Soil.

Fig. 8. Ideal Illustration of the Epochs of Elevation in Cuba. (1) Soboruco or
Elevated Reef. (2) Cliffs of the Coast. (3) The Cuchilla Level.
(4) The Yunque Level. (5) The Sierra Maestra.

PLATE II.
The Evolution of the Circular Harbors of the North Coast of Cuba.

Figs. 1, 2. Mouths of simple Rivers, with Fringing Reefs growing off their Points.
(1) The Y'umuri of tlie East. (2) The Limones.

Figs. 5, 6, 8. The Development of the Circular Bay, by Erosion of the softer Ma-
terial back of the harder Points of elevated Reef Rock. (5 and 8, Mata
Bay ; 6, Baracoa.)

288 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

Fig. 7. Example of Irregular Outline resulting from still more advanced Erosion.

(Harbor of Escondido.)
Figs. 3, 4. Double-mouthed Harbors produced by Elevation of Barrier Reefs in

combination with the Fringing Reef. (3, Harbor of Jaragua; 4, Ya-

maniguey.)
Fig. 9. Matanzas Harbor, showing Yumuri Valley, Canon, and adjacent Levels.

Note. — The artist has transposed the east and west sides in Figs. 3, 4, 5, 7, and 8.

PLATE III.

Limestone Mountains south of Matanzas.

PLATE IV.
Villa Clara, Metamorphic Mountains.

PLATE V.

Contact of Upland Plain with Limestone Hills.

PLATE VL
Typical Plain, Central Cuba.

PLATE VIL
Yumuri River, Matanzas Bay.

PLATE VIII.

Yumuri Valley and High Levels.

PLATE IX.
Yumuri River.

R.T. hill-Notes ok the Geoloov of Cuba.

•^M^SSML

_^,.^- T'-;':?*-

,3SE:^:jr:i^:;^.

'^^^^^W^^^^I^^i^^

(2)

??3i^^J0-^^^'<^^^

"/-'-

O.

A.

r

(8

"(fX-,

V y

(6)

i;i

(7)

\ \

tl Ciisflnrt.h;ft.Ke"Hftv^

.HILL - Notes on the Geology of Cuba

PL. II

''/i^^i'^fC^

ms^=^^'^m<^^^

'^//"A '1,1 }i'Y "' V,

V/^^y///,//////y ■■///,

<^ TERTIAHr KMESTCNi, CTC

'///''//'A'^elevated nEer^ or soborucc

:\SOBMS/}GED SffOtV/A/G ffEEFi,

/t/, ':m

Ski .■ •^'littS^^I v' ■. •>•

J? ■•

<:

<

Eh

Ct4

O

Eh

D

o

oT

I — I
<

D

O

o

t— t

<
O

o

I — I

Pi
O

Eh

<

-<
J
O

<

<

CO

UJ

z

o

h

CO

W

<

a!

Q
2
<

(X
D

Ll,

o

h
u
<
h

Z

O
o

<

m

D
U

<

h
z

UJ

u

(X

<

o

>-

>

<

CD

<
N
Z
<

<

>

D

CO
J
u
>

X

o

Q

z

<

UJ

<
>

IX.

>

X