General Geologic Features
The greater part of the Central Range consists of a mighty sequence of indurated metamorphosed argillaceous
Tertiary sediments. These rocks form the geologic subprovince immediately west of the pre-Tertiary
metamorphic complex. This subprovince covers chiefly the crest zone and the western flank of the Central
Range. It extends southward to the Hengchun Peninsula and encircles the southern end of the Tananao Schist
crystalline belt to the eastern flank of the Central Range. A narrow rim of slates and phyllites oh the
eastern edge of the Central Range from Taitung to Yuli also belongs to this geologic subprovince. For ease
of description, the rocks in this subprovince have been named the argillite-slate belt or argillits-slate
series of the Central Range (Ho, 1967 a and b).
The argillite-slate series extends from Santiaochio at the northeastern corner of the island southward to
Mutanshan on the Hengchun Peninsula for a length of approximately 350 kilometers. The maximum width reaches
about 50 kilometers. It forms all the high peaks along the crest zone of the Central Range and the two
highest ridges on the west, the Hsuehshan Range and the Yushan Range. In total this argillaceous belt
underlies a little less than half of the mountain areas of Taiwan. Although these Tertiary argillaceous
deposits .compose the sedimentary cover of the pre-Tertiary crystalline basement, they are well indurated
and deformed, and have furnished themselves clastic detritus to the younger formations in the basin further
to the west.
|Figure 4. Geologic subdivision and distribution of sandstorn facies|
in the argillite-slate series
LITHOLOGY AND GEOLOGIC SUBDIVISION
The various rock types in the argillite-slate sequence can be classified into two lithostratigraphic
associations, found in two distinct geographic regions. These form two belts in the Tertiary submetamorphic
terrain (Fig. 4). The western belt is called the Hsuehshan Range belt, named after Taiwan's second highest
ridge which is located approximately in the middle of the belt. The Hsuehshan Range belt is about 200
kilometers long, and averages 20-25 kilometers wide. It begins at Fulung on the northeastern coast,
extending southward through Wulai, Hsuehshan, Puli, and the Jihyuehtan (Sun-Moon Lake) to the Yushan Range,
which contains the highest peak in Taiwan, This belt continues southward to the upper course of the
Laonungchi stream in the Pingtung valley. To the east and south of the Hsuehshan Range belt is the Backbone
Range belt, covering all the crest ridges and the southern part of the Central Range. The Hsuehshan Range
belt is separated from the western foothills province on the west by the Chuchih boundary fault and from the
Backbone Range belt on the east by the Lishan boundary fault.
The great bulk of the metasedimentary rocks are dark gray, well-cleaved argillite, slate, and phyllite or,
more generally, indurated to metamorphosed argillaceous rocks. These argillaceous rocks contain abundant
small quartz veins. Argillite is more predominant in the western part, whereas slate and phyllite are more
abundant toward the east. The argillite is defined as hard mudstone which has suffered only slight
recrystallization. With the eastward increase of metamorphic grade, the hard shale grades into slate and
then phyllite gradually. The degree of metamorphism of the shaly rocks increases gradually from the western
margin of the belt toward the eastern pre-Tertiary basement found in the core zone of the Central Range.
There is no definite correlation between geologic age and degree of metamorphism which is controlled largely
by proximity of the argillaceous rocks to the center of deformation.
White and gray sandstone constitutes the other important litho-facies in this shaly sequence (see Fig. 4),
usually associated with thin and irregular lenses of impure coal or carbonaceous shale. The white sandstone
is medium- to coarse-grained but the gray sandstone is mostly fine-grained. The sandstone is either thick
and massive or contains interbeds of dark gray indurated shale and slate. Stratigraphic boundaries between
sandstone and shale are gradational. Interfingering relationships have been shown on large-scale maps. Limy
to marly lenses or nodules are scattered in the slates, mostly in the higher parts of the Central Range.
Thin and discontinuous conglomerate beds or lentils occur at several horizons in the eastern and southern
parts of the argillite-slate belt. The pebbles are derived partly from the metamorphic basement and partly
from the argillite and slate sequence. Disseminated small lenses or pods of volcanic rocks are scattered in
the argillaceous rocks. These are mainly basaltic pyroclastic rocks but there are also minor andesite,
dolerite, and other pyroclastic products of less mafic composition.
AGE AND FOSSILS
Foraminifers and less abundant mollusks are found in the argillite-slate sequence and calcareous
nannofossils have recently been found in the low-rank metamorphic shales. Other common organisms include
corals, echinoids, and algae. These fossils are concentrated in local zones, sparsely distributed in the
shaly rocks or in the limy lenses. Some fossils have been collected from the matrix of the conglomerate
lentils. Well-preserved organisms are scarce in the white quartzitic sandstone. Although low-grade
metamorphism has destroyed a part of the fossils, some are sufficiently well-preserved for paleontologic
identification. These fossils are distributed in a number of scattered exposures and offer the best evidence
for comparative age etermination. Complete faunal zonation is, however, hardly possible because no
continuous fossiliferous section is observed and barren stratigraphic sections are quite common.
In the earliest geologic maps and reports, the argillite-slate belt was considered pre-Tertiary because no
fossils or other means of dating the rocks were available. The early discovered microfossils are largely
foraminifers (Tan, 1937; Tomita and Tan, 1937). They are confined to the marly nodules or limy lenses and
the pebbles and matrix of the conglomerates exposed in the higher parts of the Central Range. These fossils
include Discocyclina, Nummulites, and Assilina and their age is mainly Eocene (Yabe and Hanzawa, 1930). The
undifferentiated argillite-slate sequence was thus inferred to be Eocene by early Japanese workers. After
the Restitution of Taiwan, L.S. Chang (1954) first reported the occurrence of Oligocene foraminifers at some
localities in northwestern Taiwan where the characteristic Eocene fossils are absent. Additional work
revealed more possible Oligocene localities within the argillite-slate sequence, which was then considered to
be Paleogene in general. Later extensive studies discovered that the foraminifers collected near the Lushan
hot spring in Wushe (Jenai-hsiang) of central Taiwan are of lower Miocene (Aquitanian) age (Chang, 1962b).
These Miocene foraminifers were subsequently found m many other areas in the argillite-slate belt, including
southern Taiwan, Hengchun Peninsula, northwestern Taiwan, and northeastern Taiwan (Chang, 1976). On the basis
of foraminiferal studies, the slate-argillite sequence to date has been assigned an age ranging from Eocene
to lower middle Miocene. However, the abundant fossil evidence gives only an improved knowledge of the
distribution of rocks of different ages in the argillaceous sequence; it provides little help in clarifying
different stratigraphic units that could be used in field mapping.
Two coral species, Astrocoena and Elephantaria, were discovered in one conglomerate bed intercalated in the
lower part of the slate sequence in northeastern Taiwan. The age probably latest Cretaceous or early
Paleocene. Based on this evidence, the slaty beds below the conglomerate were considered Cretaceous in age
by Yen and others (1956) and were named the "Pihou Formation." The conglomerate was named the E conglomerate
and considered a basal conglomerate separating the Eocene from the Cretaceous. However, the provenance of
the coral clasts in the conglomerate is still unknown. No positive Cretaceous faunas have yet been found in
the "Pihou Formation" to validate its age assignment. On the other hand, L.S. Chang (1966) discovered large
Eocene foraminifers in the E conglomerate and he believed that at least a part of the E conglomerate must be
intraformational conglomerate rather than a basal conglomerate. More recently Miocene planktonic foraminifers
have been discovered in the slates that were mapped as "Cretaceous Pihou Formation" in previous maps (L.S.
Chang, 1970). Miocene foraminifers were also discovered in the matrix of the E conglomerate (L.S. Chang,
1974). Paleontologic proof of "Cretaceous" in the slate sequence is thus still questionable, and convincing
evidence remains to be found. The E conglomerate will be discussed further in the section on orogenic events
and stratigraphic breaks in this chapter. In the more than 10 years since publication of the first edition
of this text, no fossils from the so-called "Cretaceous Pihou Formation" have been identified as definitely
Cretaceous; for this reason, no Cretaceous formation is shown on the present map.
In conclusion, the metamorphic shaly rocks have been assigned progressively younger ages with the progress
of paleontologic studies. This shaly sequence was first inferred to be pre-Tertiary on the basis of
structural features: later paleontologic work has identified Eocene. Oligocene, and Miocene ages in the
argillites and slates, but has still failed to corroborate a Cretaceous age.
The stratigraphic analysis of this thick argillite-slate sequence is still in progress because
lithostratigraphic subdivision of these undifferentiated argillaceous rocks is difficult. This is chiefly
because of the monotony in lithology, obscure stratigraphic sequence, restricted occurrence of fossils of
different ages, and lack of distinct structural breaks. Detailed stratigraphic subdivision is uncertain
without determinable lithologic or structural breaks. The complete stratigraphic succession and total
thickness of the argillite-slate sequence cannot be accurately determined because neither its top nor its
base can be clearly defined. It must be on the order of several thousand meters thick. Repetition of strata
by thrusting and isoclinal folding is common, and strata have been overturned over large areas.
Many different stratigraphic names have been proposed and used for all or parts of the argillite-slate
sequence. Generally a single unifying rock-unit name was proposed for the whole sequence due to the
continuity of the "undifferentiated" beds, their monotony, and the lack of known stratigraphic breaks. In
the early Japanese reports, this entire argillite-slate sequence was named the "Slate Series" and was
divided into upper and lower parts. Later, more detailed geologic mapping named the argillites in
northwestern Taiwan the Wulai (Urai) Series (Ichikawa, 1929) and the slates in eastern Taiwan the Suao (Suo)
Series (Ogasawara, 1933). Both series were further subdivided into groups and then into individual
formations. The Suao Series is considered older mostly on inferences drawn 1'rom their lithologic,
metamorphic, and structural features. In central Taiwan the slate series has been called the Shuichangliu
(Suiyoryu) Formation (Hayasaka and others, 1936) and the Puli (Hori) Formation (Oinouye and others, 1928) in
separate areas. It has been named the Chaochow Formation (Rokaku and Makiyama, 1934) on the Hengchun
Peninsula and the Changshan Formation (Tsan, 1964) in Kaohsiung-hsien. Most of these stratigraphic names
cover part or all of a mighty sequence of argillaceous rocks that is probably of different ages in different
places. Details of physical stratigraphy of these units, such as type sections, the nature of top and lower
boundaries, total thickness, and stratigraphic correlation, are generally uncertain. No further attempt will
be made to review fully the original definitions and usages of the names that have been used for these
Some geologists attempted to classify the rocks merely on the basis of contained fossils or inferred ages.
Such stratigraphic units suffer the common drawback of the lack of distinct mapping boundaries and of
well-represented lithologic sections. The fossil-dated rocks are also difficult to restore to complete and
undisturbed sequences so that stratigraphic relationships still cannot be unraveled.
Although dark gray argillite and slate (phyllite) are the predominant rock types in both the Hsuehshan and
the Backbone Range belts, the detailed lithologic composition of the two belts differs significantly and
serves as the basis for lithostratigraphic classification. The variation of lithologic sequence in these two
belts represents differences in depositional environments and conditions of formation. The rocks in these
two belts are therefore represented by two separate systems of stratigraphic nomenclature and map pattern in
the map legend. The Hsuehshan Range belt has more carbonaceous units and thick beds of quartzitic sandstone,
and nearly no limy lenses. Except a small part of the seemingly non-metamorphosed shales, the shaly
sediments are mostly indurated into argillites or mildly metamorphosed into slates. Conglomerate is rare and
pyroclastic effusives are more abundant in the northern and central parts of this belt. In the Backbone
Range belt, on the other hand, the rocks are comparatively highly metamorphosed. Slate to phyllite is the
predominant rock types, containing intercalations of marly or limy nodules, siltstone, sandstone, and
conglomerate. In places the slate is intimately interbedded with thin to medium beds of quartzitic sandstone,
but thick beds of white quartzite are not present. Pyroclastic rocks are exposed mostly in the middle and
southern parts of the Backbone Range belt.
Two distinct lithofacies can be distinguished in the Hsuehshan Range belt, an argillite-slate facies and a
carbonaceous sandstone facies (Fig. 4). The carbonaceous sandstone facies is represented by thick- to
medium-bedded white or gray sandstone including thin coaly lenses or carbonaceous layers. These thick
sandstone beds provide the best lithologic markers for stratigraphic subdivision in the "undifferentiated"
argillaceous series. The boundary between the argillite or slate and the sandstone is reasonably
satisfactory for mapping and accordingly forms a suitable boundary between different rock-stratigraphic
units. On this basis, a number of lithostratigraphic units can be differentiated in the Hsuehshan Range belt
for field mapping and stratigraphic analysis. Due to facies change from the north to the south, the
stratigraphic systems and nomenclature differ between the northern and central parts of the Hsuehshan Range
belt, as fully discussed in the section on "general stratigraphy."
The Backbone Range belt is the least known geologic terrain in Taiwan. Only widely spaced route traverse
mapping has been carried out in this belt because of rugged' terrain and difficult accessibility. No
carbonaceous sandstones are exposed in the Backbone Range belt, so the stratigraphic scheme of the Hsuehshan
Range belt cannot be extended eastward into this belt. An extensive area of the slate sequence is overturned
in the northern part of this belt, increasing the difficulty of determining the undisturbed sequence.
Although this belt covers more than half of the Tertiary submetamorphic terrain, only two broad stratigraphic
units are distinguished: the Miocene Lushan Formation and the Eocene Pilushan Formation. Both formations
were proposed for the first time during the compilation of this new geologic map and are composed mainly of
slate and phyllite with local intercalations of indurated sandstone. Sandstone interbeds are more abundant
in the Pilushan Formation, especially its lower part. Disseminated marly, limy, and conglomeratic lenses are
found in widely scattered places in the Pilushan Formation. The stratigraphic and structural relationships
of the argillites and slates between the Hsuehshan Range belt and the Backbone Range belt are not clear and
are still controversial.
No distinct lithologic break has been found between the Lushan Formation and the Pilushan Formation. The
essentially similar argillaceous lithology of these two rock units makes the boundary arbitrarily defined at
most places. Distinction between these two formations has to be made largely on fossil evidence and this
faunal boundary cannot be applied in field mapping. The distribution of these two formations on the geologic
maps is, therefore, quite uncertain and constant revision is needed as additional faunal data become
available. No Oligocene faunas have been found in the Backbone Range belt. Questionable late Oligocene rocks
have been reported only in the southern part, (C.T. Lee, 1977; T.C. Huang, 1980b) and are of limited
distribution. The Miocene and Eocene slate formations could be separated by a stratigraphic gap representing
an unconformity. However, no good evidence of an angular discordance has ever been discovered. The presence
of conglomerate between these two formations in southern Taiwan was reported by L.S. Chang (1972) and was
named the N conglomerate by him. A paleontologic break between Oligocene and Eocene rocks on the south
cross-island highway was recently reported by T.C. Huang (1980b). This break could represent an unconformity,
possibly correlative with the unconformable surface suggested by Chang's N conglomerate.
TIME-STRATIGRAPHIC (STAGE) SUBDIVISION
Stage is a stratigraphic unit characterized by the fauna the unit contains. Because no faunal units at the
stage level are of global extent, different sets of stages are defined for different geologic provinces. The
stage name is usually created by adding the suffix "ian" to the name of a locality with good fossiliferous
sections. Stages are used loosely as time-stratigraphic units although strictly a "stage" is a
biostratigraphic unit, because it is defined entirely by the faunas it contains. A stage is not a rock unit
and cannot be used as a cartographic unit for field mapping. Based on the study of foraminifers, L.S. Chang
(1962a and 1963a) defined four stages in the argillite-slate belt of the Central Range and summarized below.
Pilushanian stage (Eocene)
Rocks of this stage are distributed in the crest zone, the western flank, and the southeastern flank of the
Central Range. The chief lithology is represented by well-cleaved slate or phyllite and gray quartzitic
sandstone. The Eocene age is proved by the occurrence of Nummulites sp., Assilina formosensis Hanzawa,
Discocyclina sp., and Asterocyclina sp., in the rocks.
Hsuehshan stage (Eocene).
This stage is represented by several different rock units: the Szeleng Sandstone, the Hsitsun Formation, and
the Hsinkao Formation. The age of this stage is reported to be early Eocene or older Paleogene; however,
diagnostic Eocene fossils have been found only in the Hsinkao Formation. The Hsuehshan stage is considered to
be older than the Pilushanian stage which is chiefly middle or late Eocene.
Shihtsaoan stage (Oligocene).
This stage is represented by all the argillite-slate units overlying the Szeleng Sandstone and including the
uppermost Aoti Formation. The characteristic smaller foraminifers include Gaudryina hayasakai Chang,
Globigerina ampliapertura Bolli, and others.
Lushanian stage (Miocene).
A 14-km wide slate belt near Lushan of Nantou-hsien on the immediate western slope below the crest of the
Central Range forms the type section of this stage. The rocks are mostly dark gray slates intercalated with
dark gray compact sandstone and small marly nodules. The characteristic fossils include Orbulina suturalis
Bronnimann, Globigerinoides bisphericus Todd, and others. These fossils are mostly of early Miocene
Aquitanian age or slightly younger. Recent fossil studies indicate that the Lushanian faunas could be mainly
middle Miocene, and extend partly into early Miocene. The Lushanian stage is also represented by the slaty
beds on the Hengchun Peninsula.