Downloaded from http://jgslegacy.lyellcollection.

org/ by guest on May 29, 2012

Quarterly Journal of the Geological Society

On Nepheline-Rocks in Brazil, with Special

Reference to the Association of Phonolite
and Foyaite
Orville A. Derby

Quarterly Journal of the Geological Society 1887,

v.43; p457-473.
doi: 10.1144/GSL.JGS.1887.043.01-04.35

Email alerting click here to receive free

service e-mail alerts when new articles
cite this article
Permission click here to seek permission
request to re-use all or part of this
article
Subscribe click here to subscribe to
Quarterly Journal of the
Geological Society or the Lyell
Collection

Notes

© The Geological Society of

London 2012
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

01~ NEPH]~Ll-ITE-ROCKS I1~ BRAZIL. 457

33. On ~EP~Lr~r-RocKS in BRAzIr, w~th SP~cI~ REFrR~C~ tO the

ASSOCIiTI01r of PrrOlqOLITE and FOrAITE. By O~VXLLE A.
DERBY, Esq., F.G.S. (Read June 22, 1887.)

T~E nepheline-rocks heretofore recognized iu Brazilian territory

are the phonolites and associated basalts of Fernando Noronha, a
deep-sea island off the north-eastern shoulder of the continent, in
lat. 3 ~ 5' S., long. 32 ~ 24' 19" W., the volcanic nature of which
appears to have been first recognized by Darwin in the ' Voyage of
the Beagle'*. Recently a single small pebble from the little-
known island of Trinidade, in lat. 20 ~ 31' S., long. 29 ~ 19' W., has
come into my hands, showing that phonolite of somewhat different
character from that of Fernando Noronha occurs at that place
also.
Recent investigations have shown that nepheline-rocks of a
somewhat different character are also abundantly developed on the
mainland, and in such favourable conditions as to throw light on
the relations of the granitic type of foyaite or elmolite-syenite to
the other members of the group. The localities in which they have
thus far been recognized are situated in the Provinces of Rio de
Janeiro, SKo Paulo, and Minas Geraes, and their relative position
and relation to the main orographie lines of the region in which
they occur are shown in the accompanying sketch map (fig. 1). Three
of these localities, Campo Grande, Cabo Frio, and the peak of Tingua,
are in the immediate vicinity of Rio de Janeiro, the latter being in
the Serra do Mar range, the two former among its foot-hills.
:Further south another set of localities occurs in the same range
in the valley of the river Iguapd. In the Mantiqueira range the
peak of ItatiMa (3000 metres high, and the highest mountain of
eastern South America) and one or more other high peaks in the
neighbourhood are composed of these rocks, which occur also in the
Serra do Bocaina, a spur of the Serra do Mar range on the opposite
side of the :Parahiba valley. The other two localities are the Pocos
de Caldas (hot springs), on the southern margin of the great
westward expansion of the mountainous area which connects the
coast range through the Serra do Canastra with the central range
of Goyag, and Itambd in the Serra do Espinhaco range, a northward
extending branch of the Mantiqueira. As little more than a year
has elapsed since attention was first directed to these rocks, and as
the first knowledge of their existence in these different localities
was obtained almost casually, it may reasonably be supposed that

* A few of the •ernando-Noronha rocks are described by Renard (Bull. de

l'Acad, de Belgique, iii. 1882). A very complete collection made by Mr. g. C.
Branner for the Geological Commission of Brazil has been placed in the hands
of Prof. G. It. Williams for study.
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

458 "MR, 0o i. D E R B Y 0~1"

future investigation will give them a very considerable extension

beyond the limits above indicated *.
At the Cabo-Frio locality, a rocky island about three miles long
and 400 metres high is composed almost exclusively of foyaite of
two distinct types, the least abundant of which is referred by Prof.
Rosenbusch, who has kindly undertaken a microscopic examination
of these rocks, to nepheline-bearing augite-syenite, i single point

Fig. 1.--Sketch Map of parts of the Provinces of Rio de Janeiro,

$5o PauSe, and Minas Geraes.

of the island is occupied by a considerable mass of felspathic tuff.

The adjoining mainland is composed of gneiss cut by numerous dykes
of phonolite, basalt, amphibolite, diabase, and other rocks. A
detailed description of this important locality is deferred for a
future paper. At the Campo-Grande locality a network of dykes of
diabase, phonolite, trachytes, and various kinds of basalt, one of
which is limburgite, occur in gneiss in the railway-cuttings. Judging
* Since the above was written, I find the occurrence of nepheline basalt
reported by Poblman from Paraguay (' ~eues Jahrbuch,' 1886, vol. i. p. 244),
which makes it probable that the group of rocks here considered will be found
also in the Brazilian highlands bordering the Paraguay basin.
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

N ~ P H I ~ L I N E - I ~ 0 C K S IIff B R A Z I L , 459

from analogy with other places examined, there should be some-

where in the neighbourhood some central mass from which these
dykes of phonolite, trachyte, and basalt radiate ; and an examination
of the hills in the vicinity will probably reveal such a centre. I t is
possible, however, that they should be referred to the eruptive mass
of the peak of Tingua, which is about 20 miles distant. Here only
the lower portion of a single spur has as yet been examined ;
this is composed of foyaite similar to that of the principal mass
of Cube Frio, resting upon gneiss. That it has not the character
of a dyke in the place examined, is proved conclusively by
a tunnel some 400 metres long, which has been cut through the
spur from side to side. Gneiss, cut by small dykes of basic rocks,
occurs throughout the tunnel and for a few metres above its mouth
at both ends, but the surface of the spur above the line of t h e
t u n n e l is occupied exclusively by foyaito. Small dykes of basalt
similar to those of Campo Grande and of a trachytie rock occur ; but
thus far they have only been seen in the gneiss. The only
phonolite seen is in a large boulder or projecting point in the midst
of a foyaite area ; it presents the appearance of a dyke about two
metres wide, with a sharp line of demarcation between it and the
foyaite, which adheres to both sides. The p h o n o l i ~ i s 4hickly
spotted with inclusions or segregations of foyaite similar to that o f
the sides, of all sizes up to an inch or more in diameter. The
cross sections of these inclusions show a tendency to geometrical
forms, appearing like sections of crystals *. The appearance is that
of a dyke which had caught up fragments of the enclosing rock; but
the regularity of the distribution of the inclusions and their similarity
of form is against this view, while, on the other hand,:the phenomena
to be described below from the Caldas locality make it seem
plausible to suppose that the phonolite is a portion of the original
magma t h a t has escaped complete crystallization, and that the
inclusions are crystallized segregations in the midst of it. A
petrographical study will doubtless determine which view is correct.
The great mountain mass of Itatiaia, rising about 2500 metres
above its base, is made up for the most part of a variety of foyaite
which has more of the granitic aspect than the prevailing rock at
Tingua and Cabo Frio, and which has only been met with in a
subordinate mass at the latter place, referred, as already stated, to
nepheline-bearing augite-syeni~e by Prof. ]~osenbusch t . Foyaite ~
of the ordinary type is also known to occur there, as likewise an
aphanitie rock, which may be considered a phonolite or a fine-
9 grained foyaite. As the excursion to this peak was made before
my attention was drawn to the group of rocks here considered,
m a n y other types doubtless passed unnoticed. The neighbouring
* A similar inclusion of foyaite in the phonolite of Yernando lqoronha was
found in a specimen from that place, when no large masses of foyaite were met
with in a careful examination of the island. Prof. Rosenbusch, to whom a chip
was submitted, regards it as an included fragment of an older rock.
t A specimen of this rock given to Mr. Henry :Bauer, of Iguap6, was sent by
him to the late Prof. Lasaulx, who described it in a recent number of the
' Sitzungsberieht der niederrheinischen Gesellschaft,' :Bonn. This is the first
published notice of the occurrence of these rocks in Brazil.
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

BO ~I~R. O. A. D:ERBY ON

peak of Pied is known, from specimens collected in the bed of a stream

flowing from it, to contain a variety of types of foyaite and other
nepheline-bearing rocks ; and another prominent peak, called Itajub~,
in the same vicinity and in the same range is, judging from its
topographical features, of similar structure. The Bocaina locality is
o n l y known to me by a specimen of foyaite brought from there
some years ago by the director of the National Museum when on a
botanical excursion.
The Iguap6 region is only known to me through specimens kindly
furnished at various times by a German engineer, Mr. Henry Bauer,
who has given considerable attention to the collection and study of
t h e rocks of his district. They include several peculiar types not
yet known from other localities; and suspecting that they were
associates of foyaite, I requested l~[r. Bauer * to search for that
rock, which he has recently found at a place called Jacfipiranga.
Nepheline-bearing rocks also occur in the vicinity of Xiririca
further up the valley, and there are reasons for supposing that a
number of other localities will be found in that region. The Itambd
locality is only known by a specimen in the National Museum of Rio
de Janeiro, sent many years ago by the German geologist Eschwege,
under the name of diorite, and which is pronounced by Prof. Rosen-
busch to be a fine example of nephelinite.
A cursory examination of some of the localities above mentioned
having shown an apparent and hitherto unsuspected relation
between foyaite, phonolite, trachyte, tuff, and certain types of basalt,
I determined to visit the Caldas region, from which a specimen
intermediate in character between foyaite and phonolite had come
into my hands, and where a railway under construction gave
unusual facilities for examining this series, while the proximity
of a sedimentary formation of palmozoic age gave hopes of ob-
taining some idea of its geological age. & splendid development
of foyaite, phonolite, and tuff was found associated with several
types that have not yet been met with in the other localities.
The tephritic basalts which characterize the other places are repre-
sented by leucite-basalt and by a peculiar rock having the external
aspect of a diabase, in which plagioclase is the predominant element,
and which I suspect will prove to b6 tesehenite. Traehytes, if
represented, only appear in dykes too much decomposed for accurate
determination.
The Mogyana railway, starting from Campinas in the province of
S~o Paulo, runs near the margin of the mountainous plateau of
southern Minas Geraes and the sedimentary plateau of S~o Paulo.
The former, composed of gneiss and metamorphic schists, has a
mean elevation of from 1000 to 1200 metres ; the latter composed of
horizontal strata of shale, sandstone, and limestone, cut by numerous
and large dykes of diabase, varies in elevation between 600 and 1000
9 Among the rocks sent me by Mr. Bauer is olivine-basalt (limburgite). In
the other two localities in which it is known (Campo Grande and Tingua) it
occurs also in connexion with the group of nepheline rocks here considered,
and it may be suspected that the relation is not purely accidental.
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

NEPHELINE-ROCKS 1 ~ BRAZIL. 461

9netres, the highest of the denudation-ridges in the part here con-

:sidered rising to a little over 700 metres. The geological age of
this plateau has only in part been determined. The lower beds are
:soft shales and sandstone with flaggy siliceous limestones, which
last are remarkably persistent, having been found across nearly the
whole width of the Provinces of S~o Paulo and Parana, a distance
of about 300 miles. The limestone has afforded fossil reptiles, wood,
and a few unsatisfactory shells, all of which indicate Upper Palaeozoic
.(Carboniferous or Permian) age. The most satisfactory fossils are
the silieified woods, which include Lepidodendroids, conifers of the
Dadoxylon-type (conifers with a single row of pores also occur), and
ferns of the type of Psaronius. Above the limestone come heavy
beds of sandstone with intercalations of a melaphyre-like rock, often
porphyritic and amygdaloidal, which has not afforded fossils, but is
presumably Permian or Triassic. The railway for some distance
out from Campinas is at times on the sedimentary series, at times
among gneiss hills of the mountainous zone ; but after crossing the
river ~r the latter disappear and the main line to Casa
Branca follows a flat-topped sedimentary ridge between that river
sad the Rio Pardo, leaving the Caldas group of mountains, which
lies between the head-waters of these two rivers, to the eastward.
The Caldas branch runs over the sedimentary series to within about
10 miles of the foot of the mountain, where it disappears, giving
way to the gneiss foot-hills. Sedimentary rocks, however, appear
in a narrow belt along the foot of the mountain, and it is probable
t h a t the interruption above noted is due to the fact that the railway
follows the bottom of a valley, and that on the heights on either
side the sedimentary strata extend continuously to the mountain.
At all events there are no reasons for supposing that the beds at the
foot of the mountain belong to a different series from those further
:away, whose geological age is fixed by the occurrence near Casa
Branca of limestones with characteristic fossils.
The railway ascends the mountain on one side of the valley of the
Corrego (brook) do Quarf,el, rising in a distance of 18 kilometres
from 820 metres at the Prata station to 1270 metres at the summit.
The eruptive series begins to appear about 2 kilometres from the foot
of the serra in a cutting near the Prata bridge, where a greenish
spotted phon01ite appears associated with gneiss. Then follow a
few low cuttings of sandstone of no special interest. The ascent
proper commences with the passage of a narrow gap between
magnificent cliffs of sandstone rising about 50 metres above the
road-bed. This gap is cut through a narrow ridge set like a wall
across the end of the rather broad valley o f the Quarrel. The
sandstone is a very hard fine-grained white rock, broken by joints
into small angular fragments, and is very similar in appearance to
some of the beds of quartzite (itacohmite) of the metamorphic series,
and quite unlike the ordinary sandstones of the sedimentary group
above described. That its relations are with this group rather than
with the older one is, however, proved by thin beds of soft clay-
shales intercalated in it near the base along with a thin layer of
Q. J. G.S. No. 171. 2i
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

462 ~R. o. A. D~BY o~r

chert, this last being a very characteristic feature of the Carboniferous:

series of S~o Paulo. No fossils could be found, and its complete
identity with the Carboniferous series could not be satisfactorily
established, since it is possible, though not very probable, that two ~
distinct formations in this region may be characterized by cherty
layers. I f it is distinct I am very strongly of opinion that it
will prove to be older rather than newer. The beds are inclined at
an angle of 15 ~ to the eastward (away from the mountain), strike
N. 20 ~ W. So high an inclination is unusual in the series to which
this rock is referred, which, whenever it has been examined, is essen-
tially horizontal or with only slight local inclination. As, however,
the eruptive activity of this region continued after the sandstone
was deposited, the disturbance may be regarded as local. Passing
the gap the road bends round and follows the base of the ridge~
cuttings in sandstone appearing for a distance of two or three kilo-
metres. In some of these the rock is seen to pass into a tuffaceous
conglomerate containing pebbles of quartzite, and pebbles and
boulders of eruptive rocks. A peculiar feature of this eonglomerate~
is the presence of abundant and often large masses of brown mica,
which, from its occurrence in masses of considerable thickness and of
almost perfect crystalline form, not in detached flakes, must have
been formed in place after the coarser material of the rock w a s
deposited.
In the cuttings both the sandstone and the conglomerate are con-
siderably decomposed, and the contact is not perfectly clear. In one,
sandstone is seen above, conglomerate below, with no apparen~ line
of demarcation between them, although the passage from coarse con-
glomerate to fine-grained and homogeneous sandstone is an abrupt
one. That an actual passage occurs seems to be confirmed by t h e
fact that in the coarser parts of the conglomerate, and near the
supposed contact, the cement is extremely quartzose, whereas further
away and in the finer portions the eruptive elements predominate
in the cement, and the same rock with its characteristic crystals of
mica presents the aspect of an ordinary tuff. The predominant
types among the boulders of the conglomerate are two rocks which
were not found satisfactorily exposed in situ. One has the external
aspect of a moderately coarse-grained diabase, which, under the
microscope, shows remarkably fresh plagioclase as a predominant
element, the other elements being altered beyond recognition by me,
and in a manner which I have never seen in any diabase. From
its behaviour with acids it appears to contain nepheline, in which
case it is probably a teschenite. The other rock common in the
boulders seems to me to be a somewhat altered phonolite, but, if so,
it is much richer in iron than any undoubted phonolite that I have
examimed. Undoubted specimens of phonolite also appear in the
conglomerate.
In one of the cuttings in sandstone a dyke of phonolite occurs,
also a dyke of a rock too much decomposed for positive recognition,
but which appears to be identical with the diabase-like rock of the
boulders. In another cutting in conglomerate there is an inclined ~
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

N E P H E L I N E - R O C K S IN BRAZIL. 463

dyke, about 9 metres wide, which is of extreme interest(fig. 2). Owing

to decomposition, only detached fragments, still in situ, of perfectly
sound rock could be obtained, those near the margins being typical
phonolite, those near the centre foyaite equally typical. Another
case of the peripheral development of phonolite as a phase of
foyaite will be mentioned hereafter.

:Fig. 2.--Section of Decomposed DyTce in cutting, Cortege do Quarrel.

9m.

a. Decomposition-nodulesof phonolite in decomposed dyke.

b. Decomposition-nodulesof foyaite in decomposed dyke.
e. Decomposed micaeeousconglomerate.

Leaving the sandstone and conglomerate area at an elevation

of about 900 metres, which is presumably about its highest level, the
road, while following in general a nearly straight line for about 10
kilometres to the Cascara station, near the summit winds about for
a considerable space marked off by lateral valleys, iu which the road
makes long and sharp bends to the eastward. The extreme points
of these bends are marked by the Pinhalzinho culvert, the tunnel
and viaduct, the intervening distances being 3~ and 2 kilometres.
In the U-shaped bends of Pinhalzinho and of the tunnel and viaduct,
where the road crossing the lateral valleys enters most into the
material of the mountain, nepheline-rocks alone are found, while
along the projecting portion of the lower space a dark-coloured
basic rock with prominent crystals of pyroxene is exposed in several
considerable cuttings. This is in general totally decomposed, showing
only rarely, in spots, stony nuclei of a mottled bluish and brownish
colour and heavily charged with white zeolites; but in all cases the
outlines of the pyroxenc-crystals are sufficiently well preserved to
serve for its ready identification. In one cutting only, between the
tunnel and Pinhalzinho, is the rock perfectly preserved. Two
types are here presented : one is a jet-black basalt, which, according
to a note kindly furnished me by Prof. Rosenbusch, is a ]eucitite ;
the other is a bluish-black tuffaceous rock containing pebbles up to
the size of the first of the leueitite, but as it also contains prominent
crystals of pyroxene, the two rocks cannot be distinguished in the
decomposed masses. The relations of these basic rocks to the
nepheline-series is shown in two cuttings between the tunnel and
the viaduct, represented in the annexed sketches (fig. 3). Both rocks
are much decomposed, but in spots are sufficiently preserved to make
their identification perfectly certain, while owing to the marked
difference in colour of the decomposition-products the line of contact
is remarkably sharp. I could not make out satisfactorily whether
the leucite-rock was originally a basalt or a tuff. The appearance
212
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

464 ~ . o. A. n~BY o~

is suggestive of denuded hills of leucite rock buried beneath a flow

of phonolite. The former rock occurs in the little valley between the
two cuttings, the relative position of which is very much as repre-
sented in the figure, so that it is probable that the two masses are
connected underneath a capping of phonolite. The peculiar-shaped
detached mass in the left-hand figure can hardly be a dyke, and is

Fig. 3.wSections on Railway near Pinhalzinho.

a. Decomposed leucite rock.

b. Decomposed phonolite.

perhaps a fallen boulder enveloped in the phonolite. I t is 4 miles

wide in the widest part and lies about 20 metres away from the
main mass. In the cuttings in leucite rock below Pinhalzinho
several dykes, from 2 to 4 metres wide, occur, some of which are
evidently of decomposed phonolite, while one, which is better pre-
served, although altered to some extent, is either a traehytic rock
or a more felspathic phonolite than any elsewhere observed.
In the bends of the tunnel and of Pinhalzinho there are considerable
cuttings in dark blue phonolite and in a peculiar red rock intimately
associated with it. The latter is best preserved at the tunnel;
but even there, although the rock is apparently perfectly sound, its
brisk effervescence with acid shows that a part of its original con-
stituents have been transformed into carbonates. Under the micro-
scope, I could make nothing out of it beyond the occurrence of
minute dark microlites in a very finely granular ground-mass. In
places, dark red glassy crystals of hexagonal outline and irregular
whitish spots occur sparingly; both appear to be of secondary
origin. Generally the rock appears very homqgeneous, but in
places thin undulating streaks of lighter and darker red, giving an
appearance Of fluxion-structure, are seen. In other places' there are
patches and streaks of bluish and greenish phonolite, which appear
to shade off into the red rock without well-defined outlines, such
as would be expected if they were foreign inclusions. Patches
of included ])ebb]es and boulders with well-defined rounded out-
lines are a l ~ seen; and two or three large cuttings near Pinhal-
zinbo are exclusively through a coarse boulder-conglomerate, which
is, however, so much decomposed that only on the closest scrutiny
can it be distinguished from the ordinary red rock. This con-
glomerate is well exposed in a ridge just above the tunnel, inter-
calated between two closely adjacent ridges of the red rock, and
passed by a cutting about 80 metres long and 15 metres high, in
which hundreds of broken boulders with perfectly fresh fracture are
seen. They are all well rounded and of all sizes up to a cubic
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

NEPHELINE-I~0CK$ I N .BRAZIL. 465

metre, mingled together in the greatest confusion, and loosely

cemented by a paste of pebbles and minute rounded grains of
the same nature as the boulders. W i t h the exception of a few
masses which are fragments of a preexisting and more firmly com-
pacted conglomerate of the same aspect as that of which ~hey now
form a part, all the boulders and pebbles seen are of t h e same
character and, to m y eye at least, undistinguishable either macro-
scopically or microscopically (except by a slight difference in colour,
9a light shade of red or a leaden colour being predominant) from the
red rock of the adjacent cuttings. At one point a small mass of the
red rock, decomposed, but evidently in situ, rests upon the con-
glomerate, which is also cut by a smM1 dyke which, in its decom-
posed state, also resembles the red rock. A relation is thus estab-
lished between this large and almost detached mass of conglomerate
and the smaller patches, which are clearly i n e h d e d in the red rock,
and the latter is thus seen to pass on the one hand into a fragmental
rock, and on the other into a compact phonolite
On the same spur, between the tunnel and :Pinhahinho, occurs
the largest exposure yet known in the region of foyaite, which
evidently forms a considerable portion of the mass of the spur, and
appears in connexion with the red rock throughout a distance of
about 2 kilometres. The rock is in general of rather coarse grain,
but of even texture, and weathers into large rounded boulders,
which, if the massive rock was not seen in the cuLtings, might be
taken for erratics. I n one place, at the side of a small ravine, the
texture is porphyritic, with large and small polygonal patches of
coarse-grained whitish rock and large and perfect felspar-crystals
scattered through a bluish finely granular ground-mass, in which,
however, the granitic texture is still apparent. At the tunnel the
relations of the foyaite to the red rock are very well exposed.
High up on the side of the peak above the tunnel a considerable
mass of foyaite is seen close alongside of a considerable exposure of
the red rock. The tunnel is excavated in a large irregular dyke-
like mass of foyaite that cuts the red rock, and is most probably
continuous with that of the top of the peak some 300 metres above
it. This mass of foyaite is separated by an intervening mass of red
rock from another about 100 metres further up the ravine, in which
a quarry has been opened. The annexed sketches (fig. 4) of the two
openings of the tunnel and of parts of its sides near the upper end
show the relations of the two rocks. At the upper end, the lower
part of the arch is of the red rock, rising highest on the left or
upper hill-side ; the upper part is of ibyaite. A road cut on the
right side, on a level with the floor of the tunnel, shows the foyaite
cutting out the small patch of red rock of the right-hand side of the
mouth, but giving way to it again a little further round the hill.
This appearance can only be explained by regarding the foyaite as
an irregular dyke-like mass, some 10 metres or more thick, cutting
* Judging from the reference in Rosenbuseh's 'lVfikroskopische Physio-
graphie,' vol. ii. p. 299, a comparison might be made between this rock and
that from Tenerifi~ denominated eutaxite by Fritsch and Reiss.
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

466 m ~ . Oo A. DERBY ON

the red rock at a low angle. The tunnel, which describes a strong
curve, soon enters wholly into the foyaite, which appears in the
floor, roof, and sides ; but a few metres beyond, the red rock appears
again irregularly, still rising highest on the left side when the
exposure is continuous (3), while in front it is divided into three
distinct masses, the foyaite sinking between to below the floor of the
tunnel (4). The road-bed again rises above the level of the lower

Fig. 4.--Sections across and on the sides of the tunnel near Pinhalzinho.

t~ t~ b o'
k
1 2

3
a. Foyaite. b. Red rock.
1. Upper mouth of tunnel
2. Lower mouth of tunnel.
3. Left side of tunnel near upper end.
4. Right side of tunnel near upper end.

contact of the foyaite and the red rock, and continues in the former
to the lower opening, where the latter again appears in very small
patches on each side, which only rise very slightly above the floors.
The foyaite forms quite a regular arch over the lower mouth of the
tunnel (2) ; this comes out at the upper surface of the mass, which is
covered completely by the red rock. The latter is here so broken into
small fragments as to resemble an immense heap of chestnuts, and a
land-slide of this incoherent material had, at the time of my visit,
revealed a considerable surface of the foyaite on the slope over the
mouth of the tunnel. This contact-surface was irregularly undu-
lated, and inclined at an angle of 15~ ~. The rock-openings and
a part of the interior contacts have been concealed by masonry, but
a portion of the latter are still exposed.
Both rocks near the contact are generally decomposed, and the
red rock is everywhere too much so to reveal any modifications that
it may have suffered. I n places, however, the foyaite shows an in-
teresting contact-phenomenon. At about a metre away the rock
becomes finer-grained, an,t passes rapidly into foyaite porphyry and
finally into true phonolite, the phonolitic facies extending for 10-15
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

N E P H E L I N ~ R O C K S I N BRAZIL~ 467

.centimetres from the immediate contact. Unfortunately I was

unable to ascertain whether or not the same phenomenon is presented
along the upper contact, as the places where it had been exposed
were either decomposed or covered up by slides of the overlying
r e d rock.
Above the tunnel, for a distance of 5 or 6 kilometres, the cuttings
(with the exception of those hL the red rock and leucite-rock already
mentioned, and situated close to the tunnel) are mainly in phonolite,
generally much decomposed, and of no special interest. I n one a
dyke of phone:life is seen, cutting a mass of decomposed foyaite,
showing that if a part of the phonolite (as in the case of that
enclosed in the red rock and the conglomerate near the foot of the
mountain) is older, or (as in the case of the peripheral portions of
t h e mass at the tunnel, of the dyke in conglomerate above mentioned
and of the rock described below) contemporaneous with the foyaite,
a part also is newer. A cutting about 2 kilometres below the
Cascara station shows inclusions of foyaite in bluish phonolite, some
of which are of considerable size. These present sharply defined
outlines, and are either circular or show a tendency to mimic poly-
gonal crystalline forms. The next cutting above, which is through
a broad low ridge of ibyaite, exhibits exactly the reverse inclusions,
t h a t is to say, of phonolite in foyaite.
This rock, which is apparently identical with that of two quarries
.off from the line of railway near }he Caseata station, and almost in
a straight line with the cutting, the furthest being at least a mile
away, presents several interesting characteristics. The rock in the
main resembles quite closely that of the tunnel, but contains a
glassy honey-yellow ingredient, which has not been noticed else-
where. I t is also marked by indistinct circular spots slightly
darker in colour than the generality of the rock, as if drops of oil
had been sprinkled over it. In places also the nepheline, which is
generally bluish, takes on a rather brilliant red colour, and appears
to present more distinctly marked crystalline forms. Small points
:and thin irregular lines and, in one case, a pear-shaped inclusion
two inches long, of an amethystine colour also appear. The most
interesting of its peculiarities, however, are the inclusions. Some
of these are irregular masses more coarsely crystalline than the
enclosing rock, which, owing to the.predominance of large crystals of
felspar, have something of the aspect of pegmatite. Others are
pebble-like masses of finer-grained and darker foyaite, while others
again, and these are the most common, are of phonolite. These are
of ali sizes up to that of a man's head, with sharply defined and
generally angular outlines, though without t h e mimicry of crystal-
.line form presented by t,he reverse inclusions of foyaite in phonolite.
The smaller inclusions, often no larger than the end of the thumb,
are perfectly homogeneous in appearance ; but in some of the
larger ones there is a more or less distinct mixture of granitic and
felsitic material. Two of the largest inclusions seen are hero
xepresented (figs. 5 and 6).
Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

468 ~[R. O. A. D E R B Y ON

Fig. 5.--Phonolite-inclusion in Fig. 6.--Crystalline inclusion in

Foyaite, one sixth natural Foyaite, reduced about one
size. half.

The larger one (fig. 5), which is 9 inches long and 4 inches wide,.
is a blue phonolit,e, with tolerably abundant, crystalline inclusions in
the lef%-hand port,ion, which become rarer towards the right. On
the left side there is also a long, curved, ribbon-like inclusion of
foyaite, which shades off at the lower end into the including rock
through a group of scattered crystals, such as are common in phono-
lite. The other one (fig. 6), which is about 4 inches long, is repre-
sented above as it, appears on an irregularly fractured surface. The
dark-shaded portion is felsit,ic, and may, I think, be considered as
phonolite, notwithstanding its abundant, crystalline inclusions. I t
forms a distinct, sheath, sharply defined against the enclosing foyaite,
about, the whole inclusion, and also about, the three principal
crystalline masses imbedded in it in the lower part of the mass.
It appears, however, to shade into these, and into the smaller and
less-defined inclusions of the upper part. These masses differ con-
siderably in aspect, from the enclosing foyaite, and are flecked with
small dark patches apparently related to the felsitic mass. As
the specimen has been placed in the hands of Prof. t~osenbusch, no~
further description of it will be attempted here, as, for my present
purpose, it, is sufficient to signalize the double nature of the inclusion,
that is to say, of a rock of granitic texture in one of felsitic character,
which is itself enclosed in a rock of granitic type. The appearance
of this and other inclusions, coupled with the facts already stated,
as to t,he occurrence of phonolite as a peripheral facies of foyaite, lead
me to regard the inclusions, whether in t,he one or the other of the
rocks, as parts of the same original magma. A petrographical exam-
ination will doubtless determine whether this view is correct or not.
From the cutting in foyaite, above mentioned, to the Cascara
station the road winds for abou~ a kilomet,re around a prominent
spur, some 400 metres wide, of bluish-black and greenish tuff, which
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

I~EPHELINE-I~OCES IN B E A Z I L 469

in general appears quite homogeneous, but in many places shows

layers and patches of pebbles up to the size of a hen's egg. The
pebbles could not be certainly recognized, but the most abundant
appear to be similar to the diabase-like rock of the conglomerate
at the foot of the mountain, only finer-grained and somewhat
altered, so that only lath-shaped felspars could be made out with
certainty. This tuff further resembles that associated with the
conglomerate by the presence of rare and small flakes of brown
mica. I t is cut by small dykes of phonolite, and by dykes about
20 centimetres wide, too much decomposed for recognition, but
which appear tobe of a basic rock. One mass which I had noted
in a field as a dyke, one metre wide, of a basic rock, proves on exami-
nation to be composed mainly of granular quartz and magnetite, and
is probably not a dyke, though it certainly presents the appearance
of one. The tuff is also traversed by horizontal vein-like masses,
from one to two inches thick, of a highly felspathic rock of granitic
or coarse porphyritic texture, and by vertical dykes, from 3 inches to
2 feet wide, of a rock that appears to be a more crystalline variety
of that of the horizontal dykes or sheets. In the largest of these
dykes, which is much decomposed, the felspar crystals attain the
diameter of an inch. In aspect this rock resembles the felspathic
veins of granite and gneiss much more than it does the foyaitic
rocks o f the region in which it occurs. Under the microscope it
differs markedly from any rock known to me ; but I suspect that it
will prove to be an augite-syenite, or perhaps a liparite. Whatever
it may be, it, with the phono]ite dykes, serves to connect the tuff
with the crystalline rocks of the region.
Close by the Cascara station there is a small cutting in decomposed
quartzite intercalated between two cuttings in tuff. The rock is so
decomposed and broken by joints that its position could not be satis-
factorily determined, but it appears to dip to the eastward at an
angle of about 20 ~ I t is cut by small dykes of phonolite and by
very thin irregular veins of pegmatite, showing large quartz-grains
and kaolin in the form of felspar, which I take to be different
from the felspathie dykes and layers in the adjacent mass of tuff.
In appearance the rock is not very unlike the sandstone at the foot of
the mountain ; but its occurrence at a much higher level, the presence
of granitic masses apparently distinct from the eruptive group that
characterizes the region, and, above all, the occurrence in similar
conditions at another point, to be described below, of undoubted
itaeolumite, lead me to refer this exposure to a series much older
than that represented at the foot of the mountain.
At Cascara the road leaves the wooded slope of the valley of the
Quartel and enters the Campo region of the mountainous plateau of
Caldas. This plateau extends northwards some 15 or 20 miles to
the Rio Pardo, which, where I crossed it, flows at an elevation of
875 metres. The mean elevation of the plateau is about 1200 metres,
the undulating surface presenting differences of level of from 100
to 200 metres. I t is bounded on the west and north by an approxi-
mately semicircular arc of ridges rising abruptly from 200 to 400
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

470 ~R. o. A. ~)~.Rmr o~

metres above the general level. Similar but shorter detached ridges
(the Serra do Caraeol and Serra de Caldas) to the southward and
,eastward appear to complete an approximately circular or elliptical
enclosure (fig. 7). This circular arrangement of the higher ridges
is peculiar, and, taken in connexion with the character of the pre-
dominant rocks of the plateau and of the bounding ridges, as far as
,examined, is probably not without significance.
Fig. 7.mSketch Map of the Caldas region.
4" "~

>

12'00B
,seuL'~ \
:,~,z (,o~,,,.)

Serra d~ Caracol

In the 18 kilometres of railway from Cascata to the village of

I'o~os the cuttings are mainly in decomposed rock, which is, however,
sufficiently preserved in patches to show its original character, either
in loose masses or in the traces of structure still visible in the clay
resulting from decomposition. The predominant rock is phonolite.
Foyaite appears rarely in a few cuttings , always totally decomposed.
No evidence of the existence of other rocks was met with, and the
absence of dykes is noticeable. About 4 kilometres from Po~os two
cuttings in decomposed phonolite show an abundance of decomposed
analcime, some of the crystals measuring 3 inches in diameter.
lqear Cascata two quarries, already referred to, have been opened
to the right of the road ia hills of foyaite with abundant inclu-
sions. Close to one of these there is a hill of coarsely porphyritic
phonolite unlike any variety seen on the railway. The crystalline
inclusions are of all sizes up to an inch in diameter, with ill-defined
limits and generally stained red, the red colour extending at times
to the ground-mass, which is normally bluish, the inclusions being
white. This reddening of the rock without visible signs of decay
has extended to nearly the whole quarry, so that masses free from
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

NEPltELIN~-ROCKS I1~".BRAZIL, 471

i t are Somewhat difficult to find. I t is seen, however, to be most

pronounced near the surface and in the neighbourhood of fissures,
and is undoubtedly due to weathering or to infiltrating waters.
This phonolite differs further from all other varieties seen in the
region, except that of the Prata bridge, in being fit for a b u i l d i n g -
stone on account of its freedom from the fine joints and splintery
fractures t h a t generally characterize the Caldas phonolites. About
t h e village of Po~os the phonolite is generally distinctly granular,
and might perhaps be regarded as a fine-grained foyaite. The
hot sulphur springs (temp. 45 q C.) that give name and importance
to the place issue from the midst of the phonolite. The hill close
to the village (1600 metres high, the highest of the semicircular
ridge) was examined for a distance of over a mile, and found to con-
sist exclusively of phonolite. This rock also characterizes the road
from Po~os to the Rio-Pardo margin of the plateau, occurring also
with foyaite along the descent, but, apparently, not extending
beyond the river.
The R i o das Auras, the principal stream of the plateau, just
before it breaks through the ridge to descend to the Rio Pardo,
traverses for several hundred metres a considerable patch of
quartzite. I n places this shows the fiaggy structure and other c h a -
racteristics that identify it with the itacolumites so abundant in the
province of ~ i n a s . The geological age of the series to which it
belongs has not been determined for Brazil, but it is certainly very
old. Rocks similar to it and to its associates are called Huronian
b y many North-Americ~m geologists. The strike is N. 20 ~ W. ~ ;
dip 20 ~ S.W. A small exposure of phonolite occurs close to the
quartzite, and two quarries in foyaite have been opened in the imme-
diate vicinity, but the relations of these rocks to the quartzite could
not be seen. A little above a cascade formed by the quartzite there
is another, in which the rock, at first sight, appears to be similar;
but on closer inspection it is seen to be a greenish tuff enclosing
fragments of eruptive rocks. This continues for a considerable dis-
tance up stream, then isolated masses of quartzite begin to appear
associated with it, though in no case could an actual contact be
discovered, nor could pebbles of quartzite be detected in the tuff,
though grains of quartz and fragments of gneiss are not uncommon.
Apparently, however, the tuff forms a laver on an irregular surface
of quartzite, points of which occasionally appear through it. T h e
tuff is traversed by joints running lg. 70 ~ ~E., and dipping 70 ~ IN.,
which in places produce a sort of flaggy structure.
The Rio Pardo appears to form the northern limit of this eruptive
group as the Rio de Prata forms the southern. Going northward
from the Rio Pardo only gneiss, mica-schist, and granite were met
with between that river and the Rio Grande. 1North of the latter
* It may here be remarked that I found a north-westerly strike predomi-
nant between Caldas and the Serra do Canastra, as well as in that range, whereas
in Eastern Minas and along the Serra do Espinha~o the strike is almost univer-
sally north-westerly. It is for this reason that I have ventured to suggest (by
the dotted line of the map, fig. 1) the derivation of the Serra do Oanastra as
a branch of the Mantiqueira range.
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

472 ~. o. A. I m R , ~ O~

the quartzite (itacolumite) and metamorphic-schist series (repre-

sented in the Caldas region by the quartzite of the Rio das Antas
and Cascara) appears in the well-defined mountain-range known as
the Serra do Canastra, which forms the western wall of the S~o
Francisco basin and the watershed between that river and the upper
Parana. The sandstones and bedded melaphyres of the upper par~
of the S~o Paulo plateau abut against the western side of this range,
their eastern margin cutting across the rivers Pardo and Grande,
some distance to the westward of the Caldas mountains. Gneiss
is said to occur on the mountainous plateau near the city of Caldas,
and fine and large crystals of zircon come from the same locality,
leading to the suspicion that zircon-syenite or a zireoniferous variety
of foyaite may occur in that vicinity. The limited time at my dis-
posal did not permit me to follow up this interesting suggestion 4.
Pending the detailed microscopic examination of the Caldas rocks
which Prof. Rosenbusch has kindly undertaken, it does not seem
desirable to attempt to go very far in drawing conclusions from the
observations recorded in this interesting region. I t is hoped, how-
ever, that this imperfect exposition may prove as convincing to the
minds of others as the phenomena observed in the field were to my
own mind on the following points : - -
1st. T h e substantial identity as regards mode of occurrence and
geological age of the Caldas phonolites and foyaites.
2nd. The connexion of the latter through the phono]ites with a
typical volcanic series containing both deep-seated and aerial types
of deposits.
3rd. The equal, if not greater, antiquity of the leueite-rocks as
compared with the nepheline-rocks, whether fclsitic, as phonolite, or
granitic, as foyaite.
4th. The probable Palmoz0ic age of the whole eruptive series.
The interruption in the section of the sedimentary series near the
foot of the Caldas mountains renders the last conclusion less positive
than could be desired. The conglomerate at the base of the sand-
stone at the foot of the mountain proves that the eruptions com-
menced before or during the deposition of the sandstone, while the
dykes cutting the latter show that they continued after the deposi-
tion. I t is certain that at the time when the present sedimentary
plateau of S~o Paulo was occupied by a late Palaeozoic or early
Mesozoic sea, the Caldas mountains, with an elevation of at least 200
metres, rose either at the margin, or not more than 10 m.~jes distant
from the margin of that sea, and this, in accordance with the law of
the relation of volcanoes to the coast-line, may be taken as an argu-
ment in favour of the great age of the eruptions, since no late
Mesozoic or Tertiary marine deposits are known nearer than the
present Atlantic coast between Rio de Janeiro and Bahia, or t h ~
Pampa region of the Paraguay basin. :Freshwater Tertiary deposits
are known about the city of S~o Paulo and between the Serras do
* Zircon has not been recognized in any of the Brazilian foyaites thus far
examined. Sphene is a characteristic and abundant microscopic accessory in
that of Cabo Frio, Tingua, and Itatiaia, but has not been observed in that of
Caldas, which also differs in other respects from that of the other localities.
 Downloaded from http://jgslegacy.lyellcollection.org/ by guest on May 29, 2012

N-EPHELINE-ROCKS I ~ BRAZIL. 473

~[ar and ~Iantiqueira in the valley of the Parahiba ; but these are
.certainly newer than the eruptions of similar material of Itatiaia.
Unfortunately the other localities mentioned in this paper afford no
prospect of throwing light on the geological age of the nepheline-
rocks, as there are no sedimentary beds in their vicinity interme-
diate between the gneiss and the recent deposits. They promise,
however, to give important results on the relations of various types
of eruptive rocks(and it is hoped that an opportunity will soon be
afforded for examining and describing them more fully.

Dlscusslo~.
The P~SIDE~T said it was seldom that a paper containing such
important facts was presented to the Society. I t was reserved to
Mr. Derby to have proved that plutonic rocks containing nepheline
(foyaite) passed into volcanic masses which were true phonolites.
This Mr. Derby had clearly established by observations in the field.
He had also shown that leucite existed in rocks of palveozoic age,
thus rendering untenable the last stronghold of those who insisted
on making geological age a primary factor in petrographical classi-
fication. He alluded also to the value of the independent determi-
nations of Profi Rosenbusch.
Mr. BAV~R~A~ had been over portions of the ground with the
Author, and was glad to add his testimony to the value of the
paper. He spoke of the importance, in a geological sense, of these
generalizations. I t was remarkable how highly crystalline masses
of rock pass over into a sort of phonolite. These were associated
with paheozoie masses, which were pre-Permian or at least pre-
Triassic. He alluded to the difficulty of investigating Fernando
Noronha, and also to the difficulties attendant upon the investigation
o f rocks in Brazil, which were subject to such an enormous amount
of local alteration.
Profi B o ~ w z also expressed his sense of the value of the paper.
He alluded to the comparative rarity of nepheline- and leucite-rocks,
and to the confusion in the nomenclature. H e was reminded of the
nepheline-rocks near ~ontrea], where dolerite was broken through
by nepheline-syenite, associated with tephrites and phonolites.
Although there might be a doubt here, these rocks were most pro-
bably of Silurian age; but the evidence in Brazil was still clearer
as to the .~al~eozoic age, and he believed that, in the case of some
other masses, the evidence had satisfied the Canadian geologists.
He alluded also to the nepheline-roeks in the Katzen-Buckel, where
there was a similar passage from coarse-grained to fine-grained.
Dr. ]=[ATc~ said that in this case leucite was clearly shown to be of
almozoic age, and he regarded the paper as a step towards the better
assification of this group of rocks.
Profi Sw~n~r asked for evidence as to the identification of the
leucite.
The P ~ S I D ~ T thought there was no possibility of mistake in
this respect. As regards the rocks of the Katzen-Buekel, none
were truly holocrystalline, and hence they could not be Compared
with foyaite or elzeolite-syenite.