RESEARCH PAPER
First fossil records of the tholichthys larval stage of butterfly fishes
(Perciformes, Chaetodontidae), from the Oligocene of Europe
Norbert Richard Micklich ? James C. Tyler ?
G. David Johnson ? Ewa S?widnicka ?
Alexandre F. Bannikov
Received: 28 May 2009 / Accepted: 17 June 2009

 Springer-Verlag 2009
Abstract The first fossils of the tholichthys larval stage
of a chaetodontid from the lower Oligocene (Rupelian, Fish
Shales; about 30.1 MYA) of Frauenweiler (Baden-Wu?rt-
temberg, S Germany) are described, along with two less
well-preserved probable tholichthys from the lower Oli-
gocene Menilite-Formation (IPM3, 30?29 MYA and
IPM4A, 29?28 MYA) of Przemys?l (Outer Carpathians, SE
Poland). The fossils are compared with tholichthys larvae
of the extant genus of Chaetodontidae to which they are
most similar, namely Chaetodon. The German specimens
are identified as tholichthys larvae by the plate-like
expansions of the supracleithrum and posttemporal that
extend posteriorly well behind the head and by an expan-
ded preopercle with a large posterior spine; these bones,
and the frontals and supraoccipital, are distinctively rugose.
The Polish specimen from IPM3 has similar features so it is
safe to assume that it also is a chaetodontid tholichthys,
whereas the specimen from IPM4A is so poorly preserved
that it can only very tentatively be referred to the Chae-
todontidae. The occurrence of tholichthys larvae prompts
reconsideration of the paleoenvironmental situation at their
respective fossil sites.
Keywords Morphology, paleoecology 

Tholichthys, Chaetodontidae 
 Lower Oligocene 

Frauenweiler, S Germany 
 Przemys?l, SE Poland
Kurzfassung Die ersten fossilen Nachweise des Tho-
lichthys-Postlarvalstadiums von Schmetterlingsfischen
(Chaetodontidae) werden beschrieben. Zwei komplett er-
haltene Exemplare stammen aus dem Unter-Oligoza?n
(Rupelton, Fischschiefer; ca. 30,1 Mio. J.) der Tongrube
Frauenweiler bei Heidelberg (Baden-Wu?rttemberg, Su?d-
Deutschland). Zwei weitere, weniger gut u?berlieferte
Funde stammen aus der unter-oligoza?nen Menilit-Forma-
tion (IPM3, 30-29 Mio. J. und IPM4A, 29-28 Mio. J.) von
Przemys?l (A?u?ere Karpathen, Su?dost-Polen). Die Fossilien
werden mit den Tholichthys-Larven heutiger Chae-
todontidae verglichen, und zwar insbesondere mit solchen
der Gattung Chaetodon, denen sie am meisten a?hneln. Die
Funde aus Frauenweiler ko?nnen aufgrund plattena?hnlicher
Auswu?chse am Supracleithrum und am Posttemporale die
deutlich hinter das Kopfende hinausreichen, sowie ein
breites Pra?operculum mit einem gro?en, nach hinten ge-
richteten Stachel als Tholichthys-Larven bestimmt werden.
Ebenso typisch ist die rugose Oberfla?chenskupltur mancher
Scha?delelemente, insbesondere diejenige der Frontalia und
des Supraoccipitales. U?bereinstimmungen mit dem polni-
schen Exemplar aus der IPM3 Zone legen nahe, dass es
N. R. Micklich (&)
Natural History Department, Hessisches Landesmuseum
Darmstadt, Friedensplatz 1, 64283 Darmstadt, Germany
e-mail: micklich@hlmd.de
J. C. Tyler 
 G. D. Johnson
National Museum of Natural History, Smithsonian Institution,
MRC-159, P.O. Box 37012, Washington, DC 20013-7012, USA
e-mail: tylerj@si.edu
G. D. Johnson
e-mail: johnsond@si.edu
E. S?widnicka
Department of Palaeozoology, Zoological Institute,
University of Wroc?aw, Sienkiewicza 21,
50-335 Wroc?aw, Poland
e-mail: gama@biol.uni.wroc.pl
A. F. Bannikov
Borisyak Paleontological Institute,
Russian Academy of Sciences,
Profsoyuznaya 123, Moscow 117997, Russia
e-mail: aban@paleo.ru
123
Pala?ontol Z
DOI 10.1007/s12542-009-0031-7
sich auch bei diesem um einen Vertreter der Chaetodonti-
dae handelt. Der andere Fund kann aufgrund seiner
schlechten Erhaltung nur unter Vorbehalt hierzu gestellt
werden. Das Vorkommen von Tholichthys- bzw. Tholich-
thys-a?hnlichen Larven gibt Anlass die ehemaligen Lebens-
bzw. Umweltbedingungen im Bereich der oben genannten
Lokalita?ten zu u?berdenken.
Schlu?sselwo?rter Morphologie, Pala?oo?kologie 

Tholichthys, Chaetodontidae 
 Unter Oligoza?n 

Frauenweiler, Su?d-Deutschland 
 Przemys?l, Su?dost-Polen
Introduction
There are only a few fossil records of the Chaetodontidae.
Bannikov (2004) showed that taxa from the Eocene of
Monte Bolca that had been referred to the Chaetodontidae
by previous authors do not possess the diagnostic features
of that family and that no Eocene chaetodontids were
known. A recent review of the fossil record of the Chae-
todontidae by Carnevale (2006) revealed how very meager
it is, with the earliest known chaetodontid being a single
species (Chaetodon hoeferi) from the lower Oligocene
(Rupelian) of Slovenia (a second Oligocene skeletal taxon
is questionable, and Eocene otolith taxa are not valid).
Carnevale (2006) further stated that there is only a single
unquestioned chaetodontid species (Chaetodon ficheuri),
from the Miocene (Miocene otoliths either have been
referred to an extant taxon or are invalid), which he rede-
scribed, and that there is only a single Pleistocene record, a
specimen that had been referred to as Chaetodon hoefleri.
This investigation was prompted by the recovery of two
small but very well-preserved larval fishes from the lower
Oligocene of S Germany. It was immediately evident that
they exhibit many features of the armor-headed tholichthys
pre-settlement stages that characterize extant Chae-
todontidae. They also are well within the size range of the
larvae of this family, and thus can be regarded as one of the
earliest reliable records of this group. Two previously
found but less well-preserved specimens from the Polish
Carpathian are of similar size and share many morpho-
logical traits with the German specimens; we also consider
these to be tholichthys larvae.
General information
The German fossils were collected some years ago in the
Frauenweiler clay pit, which is officially designated as
??Grube Unterfeld??. It is located in the vicinity of the small
village of Frauenweiler, about 13 km S of Heidelberg
(Baden-Wu?rttemberg, S Germany). In addition to plant and
invertebrate fossils (Wagner-Klett 1919; Trunko` 1997),
finely preserved vertebrates have been found, especially
birds (Mayr 2000, 2004; Mayr et al. 2002; Mayr and
Manegold 2004). There is also a very rich fossil fish fauna
that is significant because of its broad systematic diversity
and excellent preservation (Micklich and Parin 1996; Parin
and Micklich 1996a, b; Micklich 1998; Pharisat and
Micklich 1998; Hovestadt and Hovestadt-Euler 1999,
2002; Sakamoto et al. 2003, 2004; Parin and Astakhov
2007). Many specimens are completely articulated because
they were deposited in a stratified water body with anoxic
conditions in the uppermost layers of the bottom mud,
which did not permit decay of organic matter.
Today the open mine covers a total area of scarcely
0.1 km2. It has already been extensively backfilled with
building rubble and ground excavation materials, and only
a very restricted area is accessible for excavations today.
The sediments consist mainly of dark grey or black,
sometimes marly, argillaceous clay stones, which are very
well-laminated in some sections. They represent the so-
called ??Fish Shales?? (Trunko` and Munk 1998; Grimm
et al. 2002; Micklich and Hildebrand 2005). Basically, the
exposed sequence terminates in a typical series of bitumi-
nous and well-laminated strata, with a total thickness of
about 20 cm. These clays overlie another characteristic
layer, which is about 8 cm thick and consists of concretions
with high carbonate content. The concretions probably
correspond with the ??Untere Geodenhorizont?? of Trunko`
and Munk (1998). Together with the overlying clays, these
layers have been well known for a long period of time to
amateur paleontologists, who labeled them as layers 23?25
in their own stratigraphic documentation. Both fossil tho-
lichthys described herein were recovered from layer num-
ber 23 of this peculiar sequence.
The presumed tholichthys specimens from Poland
originate from the Przysietnica and Knia _zyce fossil sites in
the Polish part of the Outer Carpathians (West Carpathian,
SE Poland). These are located about 45 km from each other
and are correlated with the stratotype profile of the Kre?pak
localities of the Skole Unit, which are natural outcrops that
expose the full profile of the Menilite?Krosno Series
(Jerzman?ska and Kotlarczyk 1976; Kotlarczyk and Jerz-
man?ska 1988). According to the nannoplankton, the sites
belong in the biozone NP23 of Martini (1971).
Przysietnica is a small village about 8 km from Brzozo?w
city, located SE of Rzeszo?w, the capital city of the Pod-
karpackie Voivodeship. The outcrops Przysietnica (PS1?
10) occur in the middle of the village and are located along
the nearby country road ??Droga pod Krzy _zami?? (??Road
under the Crosses??). They were explored during the years
1982?1998. The outcrop with tholichthys specimen
ZPALWr. A/4001 was about 15 m in maximum extension
and exposed a profile section of about 10 m, which was
N. R. Micklich et al.
123
mainly characterized by grey mudstones and brown marly
shales in the upper part and laminated limestones and
yellow marly shales in its lower part. The fossil was found
in the middle part of this section, which mainly consisted
of brown siliceous?argillaceous shales (thickness 35 cm).
The associated ichthyofauna represents the fish assemblage
IPM4A zone of the Menilite?Krosno Series and was
analogous to that of the main outcrop in Przysietnica vil-
lage (Kotlarczyk et al. 2006).
Knia _zyce is another small village about 10 km SW of
Przemys?l city, which also is located SE of Rzeszo?w. The
outcrops Knia _zyce (KN1?4) occur along a small stream,
which is the western tributary of the Wiar River. They
represent a separate sedimentary unit and were explored
during the years 1987?1988. The outcrop with the pre-
sumed tholichthys ZPALWr. A/4000 was about 10 m in
maximum extension and exposed brown siliceous?argilla-
ceous shales (thickness ca. 7 cm), grey mudstone (thick-
ness ca. 3 cm) and sandstones (thickness ca. 20 cm). The
fossil was found within a 7 cm thick layer of brown sili-
ceous?argillaceous shales in the middle of this sequence.
The ichthyofauna probably represents the IPM3 zone of the
Menilite?Krosno Series, and it is considered to be analo-
gous with the North Caucasus locality of Borislav at the
Belaya River.
Materials
Fossils
HLMD-T 410a (head to left) and HLMD-WT 410b, part
and counterpart, 17.0 mm SL. Frauenweiler clay pit
(properly ??Grube Unterfeld??), Rauenberg municipality,
Baden-Wu?rttemberg, S Germany (491601500N, 84002400E);
lower Oligocene (Rupelian, Fish Shales), 30.1 MYA.
Referred to in the descriptive text simply as 410a and 410b.
SMNS 87457/283a (head to left) and SMNS 87457/283b,
part and counterpart, 15.5 mm SL. Same locality data as
above. Referred to in the descriptive text simply as 283a and
283b. ZPALWr. A/4000a (head to left) and ZPALWr. A/
4000b (older numbers in the collection are N/6420a, b), part
and counterpart, 17.5 mm SL. Knia _zyce, southwest of
Przemys?l, Voivodeship Podkarpackie (Subcarpathian),
Boryslav-Pokuttya Nappe, Outer Carpathians, SE Poland
(49400N, 22500E); Lower Oligocene (Rupelian), probably
IPM3 zone of Menilite Formation, 30?29 MYA. Referred
to in the descriptive text simply as A/4000. ZPALWr.
A/4001 (older number in collection is N/6421), single plate
(head to left, and fragments of counterpart), ca. 12.5 mm
SL. Przysietnica southwest of Przemys?l, Voivodeship
Podkarpackie (Subcarpathian), Subsilesian Nappe, Outer
Carpathians, southeast Poland (49440N, 22030E); Lower
Oligocene (Rupelian), IPM4 zone of Menilite Formation,
29?28 MYA. Referred to in the descriptive text simply as
A/4001.
Extant tholichthys
We have examined tholichthys larvae of great morpho-
logical diversity and representing many genera of Chae-
todontidae, but we list here only those of Chaetodon sp.
because they are the most similar to the fossil tholichthys.
USNM 391660, 9.6 mm SL; USNM 152964, 18.6 mm
SL; USNM 266895, 19.3 mm SL; Dana Collection station
38871, 5.8 and 6.2 mm SL; NSMT-PL-389, 7.5 mm SL;
NSMT-PL-390, 8.3 mm SL.
Extant adult Chaetodontidae
One-hundred and seventy radiographs or cleared and
stained specimens representing 34 species of nine genera of
Chaetodontidae and ten specimens of three genera of out-
group taxa from the collections of the USNM, as listed in
Fig. 9 and Table 1.
Methods
Preparation
The tholichthys specimens from Frauenweiler were pre-
pared by the transfer method. This was developed espe-
cially for fossils that are preserved in a fragile or otherwise
unstable original matrix, for example clays, marls or oil
shales, which are frequently at great risk of drying out and
disintegrating because of their high water content. The
process involves transferring the bony substance of the
fossil on to an artificial matrix, which in this case consists
of epoxy resin. It also has the advantage that the fossil is no
longer exposed in interior view, but, rather, in external
view and often shows very delicate morphological details.
For the specimens presented herein, this method was
modified so that the fossils were transferred to pre-for-
matted slabs of fiber cement, using an epoxy resin paste
(for methodological details see Kaiser and Micklich 1995;
Roth and Micklich 2006; Micklich and Drobek 2007 pp 19
and 20).
Institutional abbreviations
HLMD, Hessisches Landesmuseum Darmstadt; NSMT,
National Science Museum, Tokyo; SMNS, Staatliches
Museum fu?r Naturkunde Stuttgart; USNM, collections of
the former US National Museum; collections now part of
the National Museum of Natural History, Smithsonian
First fossil records of the tholichthys larval stage of butterfly fishes
123
N. R. Micklich et al.
123
Institution; ZPALWr., Department of Palaeozoology,
Institute of Zoology, University of Wroclaw.
Anatomical abbreviations and numbering
Fr, frontal; O, opercle; Po, preopercle; Ps, pelvic spine; Pt,
posttemporal; PU, preural centrum; Sc, supracleithrum; SL,
standard length; So, supraoccipital; VIS, vacant interneural
space. The numbering of vacant interneural spaces (spaces
between neural spines below the dorsal-fin in which there
are no ventral shafts of dorsal pterygiophores) follows
Birdsong et al. (1988), Baldwin and Johnson (1993),
Bannikov and Tyler (1995), Tyler et al. (2003), Baciu et al.
(2005), etc., in which the space takes the number of the
vertebra whose neural spine borders the space anteriorly
(i.e., the first interneural space is between the neural spines
of the first and second vertebrae; the preneural space is that
just anterior to the first neural spine).
Miscellaneous abbreviations
IPM1-6 zones??ichthyofauna, Paleogene, Menilite?Krosno
Series; for the six fish-bearing zones in the Oligocene of
the Outer Carpathian Mountains of southeastern Poland
(see Kotlarczyk and Jerzman?ska 1976, 1988; Jerzman?ska
and Kotlarczyk 1976).
Morphological descriptions of fossils
HLMD-WT 410a, b (Figs. 1, 2) and SMNS 87457/
283a, b (Figs. 3, 4)
The teeth are clearly preserved in 283a; they are long,
slender, and typically setiform. In 283b there are impres-
sions of the teeth, but teeth are not preserved in 410a, b.
The premaxilla has a long ascending process that extends
posteriorly to the level of the anteroventral edge of the
frontal plate in both specimens, and in both the articular is
larger and longer than the dentary. The expanded upper end
of the maxilla and its articulation with the palatine is rel-
atively clearly seen in 283b. The lower rounded end of the
maxilla meets (283) or somewhat overlaps (410) the upper
edge of the lower jaw. The lachrymal is large (best seen in
410a), and a large infraorbital shelf is present; we cannot
determine which infraorbitals are involved in the compo-
sition of the shelf. Situated between the regions of the
articular head of the maxilla and the upper anterior end of
the lachrymal is the shaft-like anterior process of the pal-
atine that articulates with the maxilla, but most of the
palatine is obscured from view by the lachrymal. In 283a,
the laterally expanded anterior end of the vomer is evident
just below the anterior end of the palatine. The vertically
oriented lateral surface of the lateral ethmoid is relatively
clearly defined (especially in 283a and 410a), and the less
well-defined bony mass just to the front and rear of this
surface is the large ethmoid block. The parasphenoid in
both specimens is a strong horizontal shaft without a
ventral flange. The nasal bone is preserved in 283a just
below the posterior end of the premaxilla and the antero-
ventral edge of the frontal plate, and in 283b grooves and
surface sculpturing are evident on the surface of the nasal.
In both fossils there is a huge bony plate, with a rugose
surface and various grooves and other surface sculpturing
that extends from the front of the head in the region above
the upper jaw to the rear of the skull below the supraoc-
cipital region. There is so much surface sculpturing on this
plate (with far more and smaller rounded humps than can
be shown in the ink illustrations of the fossil tholichthys)
that we cannot distinguish the individual component bones,
but the supraoccipital crest is distinct even if the lower
regions of the supraoccipital are not distinct from the sur-
rounding bones, and the anterior two-thirds of this plate
obviously corresponds to the frontal. We presume that this
plate in the fossils appears especially large and broad
mostly because the heads of the fossils are flattened into
single planes and the frontals are being seen at a right angle
to the eye rather than obliquely, as in the lateral view of the
cleared and stained extant specimen. Nevertheless, it could
be that some of the great breadth of the frontal region of
this plate in the fossils is a specialized feature of this
particular taxon. In the lower posteroventral region of the
plate in 410a, at the level of the rear of the top of the orbit,
there are several limited areas of thickened and somewhat
differently surfaced bone that probably indicate the sphe-
notic, pterotic, and epiotic regions of the plate, at least in
part. Just behind the rear of the orbit in 283b there is
evidence of one or more of the extrascapular bones, with
deep grooves and indentations for canal openings, but the
number of elements involved cannot be determined.
Projecting back from the rear of the head of the fossils,
in the region above and behind the arch of the pectoral
girdle, is a large knobby plate that reaches posteriorly to
the level of about the second to third from the last
abdominal vertebral centra. Again, because of the exten-
sive surface sculpturing, we cannot detect a division of this
plate into what we presume to be a more dorsal portion
formed by the posttemporal and a ventral portion formed
by the supracleithrum. That such a division of the large
plate is indeed present, however, is indicated in 410b, in
which only a single much smaller plate is preserved in this
Fig. 1 HLMD-WT 410a, b; 17.0 mm SL. Pre-settlement pelagic
tholichthys larva of a chaetodontid, general view. Lower Oligocene
(Rupelian, about 30.1 MYA), Frauenweiler clay pit (Grube Unter-
feld), Baden-Wu?rttemberg, S Germany. a Main part. b Counterpart.
Each scale bar 5 mm
b
First fossil records of the tholichthys larval stage of butterfly fishes
123
region (displaced slightly ventrally). This small plate
obviously corresponds to the (posttemporal) upper part of
the compound plate of the 410a counterpart and to that of
283; the rest of the plate in 410b represents the (supra-
cleithrum) lower part of the plate, even though we cannot
distinguish a suture between the two pieces.
The large preopercular spine in the fossils tapers to a
point well behind the estimated rear of the skull; the spine
reaches posteriorly to about the level of the pectoral-fin
base in 410, whereas it reaches to beyond the pectoral-fin
base in 283b and to between the two halves of the pelvis in
283a.
There are two supraneural bones in the fossil tholich-
thys. Both supraneurals in 410 have expanded distal heads
that project both anteriorly and posteriorly to the relatively
Fig. 2 HLMD-WT 410a, b.
Reconstructions, with labeling
of most of the characteristic
features; same specimen as in
Fig. 1
Fig. 3 SMNS 87457/283a, b; 15.5 mm SL. Pre-settlement larval
stage of a chaetodontid, general view. Lower Oligocene (Rupelian,
about 30.1 MYA), Frauenweiler clay pit (Grube Unterfeld), Baden-
Wu?rttemberg, S Germany. a Main part. b Counterpart. Each scale bar
5 mm
c
N. R. Micklich et al.
123
First fossil records of the tholichthys larval stage of butterfly fishes
123
vertical shafts of the bones, and the ends of these heads are
only slightly out of contact with one another and with the
supraoccipital and first dorsal pterygiophore. In 283,
however, only the anterior projections of the heads of the
supraneurals are evident, and no posterior projections are
indicated; it is not clear if the posterior projections are
present and simply not exposed or if they are absent. In
either case, the apparent absence of a posterior projection
of the heads of these two supraneurals in 283 leaves a
significant gap between the two heads and between the
supraneural heads and the supraoccipital (the shaft of the
second supraneural is close to the front edge of the first
dorsal pterygiophore so very little space separates them).
The pelvis and pelvic fins of 410 are turned 90 and are
exposed in dorsoventral view, with the pelvic spine from
each side evident, and with an indeterminate number of
articulated branched rays between the spines. A dark,
thickened area (not indicated in the reconstruction)
between the bases of the pelvic spines probably represents
a short posterior process of the pelvis (basipterygium). The
pelvis and pelvic fins of 283 are preserved in lateral view.
Both pelvic spines and an indeterminate number of
Fig. 4 SMNS 87457/283a, b.
Reconstructions, with labeling
of most of the characteristic
features, same specimen as in
Fig. 3
N. R. Micklich et al.
123
branched rays are visible in 283b, whereas only one spine
is exposed in 283a. The two halves of the pelvis in 283a are
separated for most of their lengths, and the posterior tip of
the preopercular process is situated between their anterior
regions; a short posterior process of the pelvis seems evi-
dent in 283a.
In 410, the postcleithra (presumably ventral) from both
sides are positioned vertically below the pectoral-fin, but it
cannot be determined if dorsal postcleithra are present. In
283, a single long postcleithrum is seen in both parts, but
no distinction between ventral and dorsal components is
evident; in 283a there is a wide flange with vertical stria-
tions along the posterior edge of the shaft of the postclei-
thrum. The pectoral arch is best seen in 410a, but the
individual components are not distinguishable; about 13
pectoral-fin rays are evident.
The ribs, as exposed in 410, are sturdy and very long,
reaching almost to the ventral midline. Although these ribs
are relatively wide, we cannot see sufficient detail to
determine if they have developed the characteristic chae-
todontid anteromedial flange from the main shaft. An
anteromedial flange is more clearly indicated in the ribs of
283, which, as exposed, are even wider than those in 410.
However, in 283 the ribs do not extend as far ventrally as in
410, reaching to no more than about three-fourths down the
abdominal cavity; we suppose that the ventral ends of the
ribs in 283 are not fully exposed and that they are probably
as long as those in 410.
There are 13 dorsal-fin spines and three anal-fin spines
in both specimens. The total number of fin rays cannot be
determined in either the dorsal or anal fins of the fossils.
The first two dorsal and anal spines are in supernumerary
association with the first dorsal and anal pterygiophores,
both of which have lateral expansions buttressing the basal
articulations of these first two spines. The dorsal region of
the first anal pterygiophore is situated along the anterior
edge of the haemal spine of the first caudal vertebra, and
the dorsal end of the second anal pterygiophore is situated
in the space between the ventral ends of the first and second
haemal spines of the caudal vertebrae (410a) or at least on
that trajectory.
Most of the abdominal vertebrae are obscured from view
by the posttemporal and supracleithrum head plates, with
only the last two or three centra fully exposed, so we can
only presume that there are ten abdominal vertebrae,
whereas we can determine with certainty that there are 14
caudal vertebrae in both specimens. The neural spines of
the last three abdominal vertebrae are exposed behind the
head plates for at least most of their lengths. In both
specimens there is a faint indication of the dorsal end of the
neural spine of the fourth from last abdominal vertebra in
the region between the ventral ends of the fifth and sixth
dorsal pterygiophores, but none of the length of the neural
spines more anteriorly is exposed because of the presence
of the head plates.
Whereas the head plates obscure the neural spines of the
more anterior abdominal vertebrae, the neural spines pos-
terior to the plate in the middle region of the body are well-
preserved and do not appear displaced or misaligned. In
both the part and the counterpart of 283 there is no ventral
shaft of a dorsal-fin pterygiophore situated between the
neural spines of the last (presumed tenth) abdominal ver-
tebra and the first caudal vertebra; therefore, the tenth in-
terneural space is vacant in 283. In 410a we are confident
that there is no ventral shaft of a dorsal-fin pterygiophore
situated between the neural spines of the presumed ninth
and tenth abdominal vertebrae; therefore, it is the ninth
interneural space that is vacant. In 410b the relationships of
the ventral ends of the pterygiophores situated near the
dorsal ends of the neural spines of the ninth and tenth
abdominal vertebrae are not as clearly indicated as in 410a,
but, after examination under many lighting and moisture
conditions, we believe that the ninth interneural space is
also vacant in 410b. The neural spines of the ninth and
tenth abdominal vertebrae in 410, bordering the vacant
ninth interneural space, are positioned much closer to one
another than are the other neural and haemal spines, and
the same is true of the neural spines of the tenth abdominal
and first caudal vertebrae, which border the vacant tenth
interneural space in 283. Thus, the two Frauenweiler
specimens differ from one another with regard to which
interneural space is vacant at the posterior end of the
abdominal series of vertebrae. As discussed in the com-
parative section (below), we have reason to believe there is
another vacant interneural space situated more anteriorly in
the abdominal vertebral series of the Frauenweiler
specimens.
The caudal skeleton is not well preserved in its entirety
in either of the specimens, and the last two centra are
displaced in 410. What can be determined is best indicated
in 283: PU2 has a low neural spine, and the haemal spines
of PU2, PU3, and the parhypural are autogenous; there are
probably five hypurals and 17 principal caudal-fin rays,
with about four procurrent rays both above and below;
epurals are not clearly indicated. There is evidence of a
large uroneural in the somewhat disarticulated caudal
skeleton of 410a.
Minute cycloid scales with maximum diameters
between 3.0 and 3.5 mm are clearly visible above and
below the vertebral column in the abdominal region of both
Frauenweiler specimens (Fig. 5).
ZPALWr. A/4000 (Fig. 6) and ZPALWr. A/4001
These two small specimens from the Polish Carpathians
were identified many years ago by the late Anna
First fossil records of the tholichthys larval stage of butterfly fishes
123
Jerzman?ska as being young chaetodontids similar to the
genus Forcipiger, even though there is no clearly dis-
cernable long snout to be seen in either specimen. They are
listed in Kotlarczyk et al. (2006, p 34), as Forcipiger sp.
under Chaetodontidae and, unnecessarily, just below this
section of the list, as two specimens of undetermined
Chaetodontidae.
A/4000 is the better preserved of the two specimens.
Many of the bony plates on the top of the head are difficult
to distinguish individually, but two are distinct, as is the
preopercular spine. The two easily distinguished plates are
somewhat displaced behind the head in the region above
the preopercular spine. The uppermost and somewhat
smaller of the two clearly corresponds to the posttemporal
plate of a tholichthys, and the lower corresponds to the
supracleithrum plate. The posttemporal and supracleithrum
plates have rugose and knobby surfaces with deep broad
grooves along their lengths and with circular radiating
ridges, which is similar to those of the Frauenweiler
specimens. The preopercular spine is very long (somewhat
more so than in the Frauenweiler specimens) and tapers to
a point. The dorsal-fin spines cannot be fully counted
(posterior to the sixth spine they become increasingly dif-
ficult to distinguish individually), but there are three anal
spines and 14 caudal vertebrae (the haemal spine of the first
caudal vertebra, situated behind the first anal pterygio-
phore, is fully exposed, whereas only the distal end of its
neural spine is exposed, situated above the posterodorsal
end of the supracleithrum plate); the pelvic spines from
both sides are evident, but the rays cannot be counted. In
the abdominal region of the less complete counterpart
(head to left), there is a part of a long shaft of bone that is
relatively wide and may represent a chaetodontid-like rib.
A/4001 is too poorly preserved to distinguish many of the
seemingly enlarged bones in the head region, but present in
the area below the rear of the eye is a large, long preop-
ercular spine that tapers to a point at the rear of the
abdomen. There is also a rather similar long tapering bone
that reaches just as far posteriorly from the top of the head
as does the preopercular spine. This second bone could be
the preopercular spine that has been displaced from the
other side, or it could be a long tapered extension of a
posttemporal plate. If the latter, it is rather more elongate
and tapered than the posttemporal extension in both the
other Polish specimen and in the Frauenweiler specimens.
The ribs in this specimen are sufficiently preserved to be
sure that they are very long, reaching close to the mid-
ventral line, and there is some evidence that the ribs are
relatively wide and stout, which would represent the
chaetodontid condition.
Extant tholichthys morphology (Figs. 7, 8)
It has long been recognized that the tholichthys is a spe-
cialized pelagic larval stage of chaetodontids (Day 1870;
Fig. 5 Detailed view of cycloid scales in the ventral part of the abdominal region, directly in front of the haemaxial complex. a HLMD-WT
410a. b SMNS 87457/283a. Each scale bar 0.5 mm
N. R. Micklich et al.
123
Gunther 1871) that can reach 60 mm SL before settlement
(Burgess 1978; Leis 1989). Leis (1989) discussed the
confusion that existed for more than 100 years wherein
the tholichthys stage was erroneously attributed to the
Pomacanthidae and to the Chaetodontidae, until this was
rectified by Burgess (1978). A superficially similar larva
with hypertrophied head bones, one of which projects
behind the head, characterizes the Scatophagidae, and these
larvae have often been referred to as ??tholichthys.??
Johnson (1984) briefly discussed the most essential
differences between these two ??so-called?? tholichthys
stages, emphasizing that in chaetodontids the posttemporal
and supracleithrum are rugose and expanded posteriorly as
large laminar plates, and that the preopercle is expanded
and extended posteriorly into a broad spine, whereas in
scatophagids the supracleithrum is unmodified, the post-
temporal extends posteriorly as a very thick blunt spine, the
preopercle is expanded although not extended posteriorly,
and the pterotic has a rugose protuberant covering. Johnson
(1984) noted that the similarities and differences between
the chaetodontid and scatophagid larval stages needed
further investigation to assess the possible phylogenetic
implications for these two families.
The most extensive description of the morphological
features of the chaetodontid tholichthys was given con-
cisely in an unpublished dissertation by Blum (1988, pp
89?91). He restricted the term tholichthys to chaetodontids,
and he used this stage as one of the seven synapomorphies
documenting the monophyly of the family. According to
this dissertation, typical chaetodontid tholichthys larvae are
characterized by the combination of the following char-
acteristics: a dramatically hypertrophied preopercle, which
is expanded anterodorsally almost to the margin of the
orbit, and ventromedially under the head, so that the left
and right antimeres are juxtaposed midventrally, and which
bears a large posterior projection that is attenuated into
either a sharp or blunt spine; posttemporals and supra-
cleithra with large, plate-like expansions that extend pos-
teriorly over the trunk; frontals that are slightly expanded
such that they cover the parietals posteriorly and more of
the orbit laterally than they do in post-metamorphic indi-
viduals; nasals, lacrymals, and dermosphenotics (last cir-
cumorbital bone) that are hypertrophied such that they fill
the gaps between the larger bones already mentioned and
form an almost complete casing of dermal armor around
the head. Blum (1988, pp 89?91) described numerous other
differences, including those previously given by Johnson
(1984), between the tholichthys of chaetodontids and the
comparable pre-settlement larval stage of scatophagids. We
agree with most, but not all, of these differences. His
characterization of the chaetodontid tholichthys stage is
adequate for our purposes here, and a subsequent paper
(Johnson et al., in preparation) will present a more com-
prehensive comparison of diversity within the family.
Blum (1988) concluded that scatophagid larvae show only
superficial similarity to chaetodontid larvae and lack
almost all of the specific characteristics mentioned above.
Fig. 6 ZPALWr. A/4000a, b; 17.5 mm SL. Pre-settlement larval
stage of a chaetodontid, general view. Lower Oligocene, probably
IPM3 zone of Menilite Formation (Rupelian, 30?29 MYA), Outer
Carpathians, Knia _zyce, southwest of Przemys?l, SE Poland. a Main
part. b Counterpart. c Reconstruction of counterpart, with labeling of
most of the characteristic features. Each scale bar 2 mm
First fossil records of the tholichthys larval stage of butterfly fishes
123
Fig. 7 a USNM 159264, 18.6 mm SL. Cleared and stained extant tholichthys stage specimen of an unidentified species of the genus Chaetodon,
with labeling of most of the characteristic features. b NSMT-PL-390, 8.3 mm SL. Same, alcohol preserved, specimen. Each scale bar 2 mm
Fig. 8 Details of the squamation in the anterior part of the body
flank in extant pre-settlement larvae (heads are to the left): a USNM
391660, 9.6 mm SL; tholichthys larva of Chaeotodontidae. b USNM
391661, 9.6 mm SL; corresponding larval stage of Scatophagus
argus. Each scale bar 0.2 mm
N. R. Micklich et al.
123
Interneural spaces in extant adult Chaetodontidae
(Fig. 9, Table 1)
To assess the intraspecific and intrageneric variability of
the distribution of the dorsal-fin pterygiophores in the in-
terneural spaces, we have examined a comprehensive
sample of cleared and stained or radiographed specimens
representing 20 species among nine genera of extant
chaetodontids. In all extant species the neural spine of the
first vertebra is situated anterior to the shaft of the first
dorsal pterygiophore, with the tip of the neural spine
between the bases of the two supraneurals, and the neural
spine of the second vertebra is situated between the shafts
of the first and second dorsal pterygiophores; subsequently,
a single neural spine is situated between adjacent ventral
shafts of the dorsal pterygiophores until the position of the
vacant interneural space is reached, whereupon two neural
spines are accommodated between adjacent pterygiophore
shafts. A single interneural space is vacant in the great
majority of these taxa; however, two taxa (Chelmon mar-
ginalis and Parachaetodon ocellatus) have no vacant
spaces. In Chelmon rostratus there is significant variability
in the condition of the vacant interneural space. It has
either the sixth, the seventh, or the eighth space vacant, or
there is no vacant space at all. Most extant taxa typically
have the sixth space vacant (Fig. 9a), but a minority of
specimens of those taxa may have either the seventh space
vacant or, even less frequently, the fifth rather than the
sixth space vacant (the sixth space is vacant in the only
well-preserved skeletal fossil chaetodontid, from the Mio-
cene; Carnevale 2006). Some taxa typically have the sev-
enth space vacant (Fig. 9b), but a minority of specimens of
these taxa may have the sixth or the eighth rather than the
seventh space vacant. The eighth space is never vacant as
the typical condition of any extant taxon. Two species
(Parachaetodon ocellatus and Chelmon marginalis) have
no vacant interneural space (Fig. 9c), and one (Chelmon
rostratus) is exceptionally variable in its interneural spaces,
with most specimens having the seventh vacant but sig-
nificant minorities having either the sixth, eighth, or none
vacant. In none of the extant materials examined is there a
vacant interneural space posterior to the eighth space.
Comparison of fossil tholichthys with extant taxa
The morphology of the fossil tholichthys, as best seen in
the relatively complete and well-exposed German speci-
mens, is basically very similar to that of the tholichthys of
most genera of extant Chaetodontidae, i.e., exclusive of the
tholichthys of the long-snouted Forcipiger and of Cora-
dion, which have some apparently secondarily reduced
features (Johnson et al., in preparation). As an exemplar of
a typical extant chaetodontid tholichthys, we use a cleared
and stained specimen of an unidentified species of the
genus Chaetodon (USNM 152964, 18.7 mm SL; Fig. 7).
The fossil and extant tholichthys possess the following
comparable features: setiform teeth (although in the fossils
we can see only the outer row of teeth and not the pre-
sumed full array of separate bands); a large lachrymal that
obscures most of the body of the palatine; a large ethmoid
block that extends both anterior and posterior to the lateral
ethmoid (although in the fossils we cannot determine if the
ethmoid has lacunae); a parasphenoid without a ventral
flange; the pelvis with only a short posterior process; a
laterally expanded buttress on the first dorsal pterygiophore
for support of the first two (supernumerary) dorsal spines;
13 dorsal and three anal-fin spines in the fossils, which is
the norm in many species of many genera of chaetodontids;
14 caudal vertebrae; and a short neural spine on PU2. The
preopercular spine in the fossils is slightly longer and more
acutely tapered than in our Chaetodon tholichthys. The
heads of the supraneurals in the fossils are not as closely
spaced to one another and to the crest of the supraoccipital
and the first dorsal pterygiophore as is the case of the close
sequential articulation of these elements in extant tholich-
thys. However, in specimen 410 at least, most of this dif-
ference in spacing could be attributed to lack of full
exposure of the supraneural heads and to some slight dis-
placement of the supraneurals (we also suspect that the
heads of the supraneurals are not fully exposed in specimen
283). Our examination of cleared and stained specimens
and radiographs of chaetodontids indicates that whereas
most specimens of most species have full sequential
articulation of the elements between the supraoccipital
crest and the first dorsal pterygiophore, a minority of
specimens have a less than fully integrated series (espe-
cially in many specimens of Parachaetodon), to which
condition the spacing in specimen 410 is similar. The
length of the ribs in 410 extends downward close to the
ventral midline, or to at least 80% the depth of the abdo-
men. Our examination of extant chaetodontids indicates
that in most specimens the ribs extend ventrally to about
90% the depth of the abdomen (about as shown for the
chaetodontid skeleton in Tyler et al. 1989, Fig. 31), but that
in a significant minority of specimens the ribs are shorter
and an extension to no more than 80% the depth of the
abdomen is not unusual (especially in Coradion). In 283
the ribs extend ventrally less than 80% of the abdomen, but
we believe that this is because the ventral ends of the ribs
are not fully exposed. The anteroventral flange of the ribs,
which is characteristic of chaetodontids, is evident in 283.
The size and surface sculpturing of the supracleithrum and
posttemporal plates in the fossils are comparable with those
of extant tholichthys, whereas the apparent larger size of
the frontal plate in the fossils is probably because the plates
First fossil records of the tholichthys larval stage of butterfly fishes
123
Fig. 9 Radiographs showing
the vacant interneural spaces in
extant Chaetodontidae. a
Chaetodon trichrous, USNM
392506; sixth interneural space
vacant. b Chaetodon reticularis,
USNM 392440; seventh
interneural space vacant. c
Parachaetodon ocellatus,
USNM 266892, no interneural
space vacant. Each scale bar
1.0 mm
N. R. Micklich et al.
123
are being viewed at a right angle rather than obliquely, as
in the illustration of the extant tholichthys.
The most significant difference between the morphology
of the fossils and that of all of the extant taxa is in the
position and number of the vacant interneural spaces. In
none of the extant materials examined is there a vacant
interneural space posterior to the eighth space, as is the
case in both German specimens. If we assume that the
arrangement of dorsal pterygiophores and interneural
spaces that we described above as a constant feature of
extant chaetodontids is present in the fossils also, and that
the fossils share ten abdominal vertebrae with extant
chaetodontids, then there must be an additional vacant
interneural space somewhere between the third to seventh
neural spines. The third to sixth neural spines are entirely
obscured by the posttemporal and supracleithral plates, but
the seventh neural spine is faintly indicated just at the rear
edge of the head plates; the more posterior neural spines
are clearly exposed. In 410 we suggest, on the basis of the
space available between the ventral shafts of the dorsal
pterygiophores, that it is probably the fourth and fifth
neural spines that are accommodated there (i.e., the fourth
interneural space is vacant in addition to the ninth space).
In 283 we suggest that it could be either the third and
fourth neural spines together in the space between the
second and third dorsal pterygiophores, or the fifth and
Table 1 Synopsis of vacant
interneural space patterns in
extant Chaetodontidae
Genus Species Number of specimens Vacant Interneural Space
5th 6th 7th 8th None
Amphichaetodon melbae 2 2
Chaetodon adiergastros 1 1
Chaetodon auriga 2 2
Chaetodon baronessa 1 1
Chaetodon capistratus 8 7 1
Chaetodon citrinellus 2 2
Chaetodon flavirostris 1 1
Chaetodon guttatissimus 3 2 1
Chaetodon lunula 1 1
Chaetodon lunulatus 1 1
Chaetodon melanotus 1 1
Chaetodon mertensii 2 2
Chaetodon ocellatus 12 10 2
Chaetodon octofasciatus 7 1 6
Chaetodon plebeius 2 2
Chaetodon reticulatus 4 4
Chaetodon sedentarius 20 1 17 2
Chaetodon striatus 11 11
Chaetodon trichrous 2 2
Chaetodon trifasciatus 2 2
Chaetodon unimaculatus 1 1
Chelmon marginalis 8 None in 8
Chelmon rostratus 18 2 8 4 None in 4
Coradion chrysozonus 11 10 1
Forcipiger cyrano 1 1
Forcipiger flavissimus 4 1 3
Forcipiger inornatus 1 1
Forcipiger longirostris 4 2 2
Hemitaurichthys polylepis 2 1 1
Heniochus acuminatus 7 4 3
Heniochus chrysostomus 5 5
Parachaetodon ocellatus 5 None in 5
Prognathodes aculeatus 4 3 1
Prognathodes aya 14 1 9 4
First fossil records of the tholichthys larval stage of butterfly fishes
123
sixth neural spines together in the space between the fourth
and fifth; we think the latter is more likely because this
more posterior position would be in keeping with the more
posterior position of the other vacant space (tenth inter-
neural space vacant in 283 vs. ninth space vacant in 410).
Consequently, the fossil tholichthys from Germany as
preserved are unique among chaetodontids in having either
the ninth or tenth interneural space vacant, and in probably
having another space vacant more anteriorly, which we
speculate to be the fourth or fifth space. By contrast, in
extant chaetodontids there is never more than one space
vacant, and it is never more posterior than the eighth space.
Therefore, the German fossils could represent a separate,
extinct lineage of the Chaetodontidae. Nevertheless, at the
present state of investigation, we remain open to the pos-
sibility that the positions of the vacant interneural spaces in
these fossils could be taphonomic artifacts. In specimens
with a very wide and heavily armored skull such as the
tholichthys larvae, substantial mechanical stress can be
expected on the few most posterior abdominal and the few
most anterior caudal vertebrae during embedding. This
could result in dislocations, especially for specimens such
as 410a and b that are in a somewhat progressive state of
decay. However, the position of the vacant interneural
spaces is a relatively constant and an almost diagnostic
feature of many extant species of chaetodontids, and, in the
case of the German tholichthys, we believe that the ninth
and tenth interneural vacancies are natural and not artifacts
of preservation.
The squamation of the two Frauenweiler specimens is
very similar to that of a 9.6 mm SL cleared and stained
Chaetodon tholichthys (USNM 391660, Fig. 9a) in their
circularity; however, the latter has very weakly developed
ctenii along the posterior edge. Such ctenii cannot be
observed in the fossils, but this is probably because the
scales are too minute and weak. There is substantial varia-
tion in specimen size at which ctenii develop among extant
chaetodontid tholichthys. The 15.5 and 17.0 mm SL fossil
tholichthys from Frauenweiler are well within the range of
sizes at which ctenii first develop in extant tholichthys. We
note that the scales of the somewhat similar larvae of Sca-
tophagidae are very different, with upraised processes that
are very easy to see even in this early stage (Fig. 9b).
The general appearance of Polish specimen A/4000, and
of its three best preserved features of armor (the enlarged
posttemporal and supracleithrum plates and the preoper-
cular spine), are so similar to those of the Frauenweiler
specimens that we are confident that this specimen is also a
chaetodontid tholichthys of undetermined genus and spe-
cies, but it is not necessarily of the same taxon as the
Frauenweiler specimens. A/4001 is so poorly preserved
that it can only very tentatively be referred to the
Chaetodontidae.
Conclusions
Our comparisons of the fossils with larval stages of extant
chaetodontids clearly show that the two specimens
approximately 30.1 MYA old from the Lower Oligocene
of Frauenweiler clay pit are the first fossil records of the
chaetodontid tholichthys stage, and one of the earliest fossil
records of the Chaetodontidae. The German tholichthys
specimens are unique among the members of this family in
having either the ninth or tenth interneural space vacant,
and in almost certainly having another space vacant more
anteriorly (probably the fourth or fifth; either that or there
are only nine abdominal vertebrae), whereas in extant
chaetodontids there is never more than one space vacant,
and it is never more posterior than the eighth space.
Therefore, the Frauenweiler specimens cannot be accom-
modated in any extant genus of this family. Until we can
obtain additional materials of this taxon from this locality
to more fully document its morphological peculiarities, and
especially of its interneural space conditions, we place it as
an incertae sedis Chaetodontidae.
The Polish specimen A/4000 is probably a chaetodontid
tholichthys of an undetermined genus and species. The
specimen A/4001 is so poorly preserved that we only very
tentatively accept it as a possible tholichthys stage.
Paleoenvironmental implications
German specimens
Different theories have been presented regarding the pa-
leoenvironmental conditions of the Frauenweiler fossil site
(Weber 1951; Weiler 1966; Trunko` 1997; Micklich 1998;
Grimm et al. 2002). Nevertheless, it is obvious that the
fossil fish record of the site is dominated by juvenile
individuals, although rather early postlarval stages are not
rare. Unfortunately, it is extremely difficult to give precise
estimates concerning the relative percentages of larval or
pelagic stages. Such stages probably are under-represented
in the collections for various reasons: they may have been
overlooked during the excavations because of their small
size, which also makes their identification more difficult.
Additionally, there is a general preference to collect larger
and more attractive specimens, as indicated by the fact that
only about 20% of all Frauenweiler taxa are represented by
both juveniles and larger, presumably adult or subadult,
individuals. The preference for collecting larger individuals
has the consequence that very small individuals are not as
frequently represented in museum collections, even though
in reality they undoubtedly dominate the fossil fish record.
Layer number 23, where the tholichthys were found, is
one of the richest fish-bearing horizons in the Frauenweiler
N. R. Micklich et al.
123
clay pit. The fishes of layer 23 are characterized by com-
paratively numerous records of small-sized specimens,
including those of clupeids, syngnathids, centriscids,
??Serranus?? budensis (which, in fact, is not a serranid), and
carangids, and some rarer scorpaenids, scombroids, and
caproids. Many of these are less than 5 cm in total length,
and some are distinctively smaller (about 2 cm total
length).
The life style and distribution patterns of the pre-set-
tlement larvae of extant Chaetodontidae were extensively
described and discussed by Leis (1989), who documented
the rarity of these larvae in tropical plankton and midwater
trawling samples. Such larvae may prefer habitats that
differ from those that are most often sampled by conven-
tional methods. Although adult chaetodontids are charac-
teristic reef inhabitants, their larvae do not stay close to
reefs but, rather, seem to be neritic or oceanic. Size at
settlement varies widely, from 9 to 60 mm SL. Many
species settle at less than 20 mm SL, so the Frauenweiler
tholichthys may have been close to their time of settlement.
Taking into account the general aspects of the fish fauna,
a reconstruction of the Frauenweiler paleoenvironment
may be as follows: It is likely that this area was some kind
of shallow coastal bay, which partly and/or occasionally
was separated from the open sea by some sort of ??shel-
tering?? ridge or reef, and with water conditions that varied
substantially, depending upon external situations; certain
taxa probably preferred brackish waters, river mouths and/
or estuarine water bodies (Micklich 2005; Micklich and
Hildebrand 2005). Within this basin there may have been
seagrass meadows, which (together with estuaries and
mangrove swamps) are well known as nursery grounds for
a wide variety of marine fishes. This environment would
have been ideal for the young stages of many species of
fishes, including those that as adults prefer greater depths
(e.g., Laegdesgaard and Johnson 2001; Cocheret de la
Morinie`re et al. 2002). The Frauenweiler fishes may have
been killed by occasional phytoplankton blooms or by
rising temperatures and salinities, which also caused
stratification of the water column and loss of benthic life,
characteristics that are prerequisites for fine fossil preser-
vation. Occasional in-flows of fresh water may have peri-
odically rejuvenated the water quality.
Unfortunately, there are some contradictions to the
above hypothesis. Some species of extant fishes that are
especially typical of seagrass meadows, for example the
porgies (Sparidae), are rare in that fossil site and are rep-
resented only by disarticulated remains. One would also
expect to find fossilized seagrass remains, but such remains
are absent. Also absent are the remains of other typical reef
organisms. The presence of pre-settlement larvae also is
unlikely for a coastal, inshore habitat, with a reef-like
shelter from the open sea.
Polish specimens
According to Kotlarczyk et al. (2006), the Knia _zyce fauna
belongs to the fish assemblage IPM3 zone of the Menilite?
Krosno Series. This is characterized by the predominance
of deep-water fishes (54.4 vs. 18% of shallow-water
forms). The paleoenvironment was considered to be a
warm, deep-water basin. By contrast, the Przysietnica
fauna belongs to the IPM4A zone. This is characterized by
the presence of numerous shallow-water forms, which
clearly dominate those that prefer deep waters (27.3 vs.
5.9%). Both are associated with taxa that typically inhabit
reef or lagoonal environments. Highly characteristic are the
Centriscidae, which comprise up to 45% of the assemblage
but are absent in the other outcrops from Przysietnica and
also absent in Knia _zyce or very rare in the ichthyofaunal
zones in the other Polish outcrops from the Outer Carpa-
thians (Kotlarczyk et al. 2006). The paleoenvironment was
assumed to be a warm-water basin, distinctly shallower
than that of the IPM3 zone.
As indicated in the preceding section, extant tholichthys
larvae do not stay close to reefs but, rather, seem to be
neritic or oceanic. Fossil representatives are more likely to
be expected in the deep-water deposits of the IPM3 zone
than in the shallow-water localities of the IPM4A zone. As
in the Frauenweiler locality, it is, therefore, somewhat
surprising to find an at least presumable record of these
generally very rare fossils at IPM4A.
Acknowledgments We are grateful to Frank Sanzenbacher, of
Bondorf, Germany, a skillful amateur paleontologist who first found
one of the Frauenweiler tholichthys specimens, prepared it, and per-
mitted it to be purchased by the Hessisches Landesmuseum Darmstadt
under reasonable conditions. We also thank Annette and Harald
Oechsler of Wagha?usel, Germany, amateur paleontologists whose
ardent and arduous work at the Frauenweiler site during the past few
years of field work have obtained many valuable specimens for the
scientific community. The other Frauenweiler tholichthys specimen
was obtained by the Staatliches Museum fu?r Naturkunde Stuttgart and
was generously loaned to us for study in Darmstadt and in Wash-
ington, DC, through the cooperation of Ronald Bo?ttcher, Curator of
Paleontology. We appreciate the cooperation of Jeffery M. Leis,
Australian Museum, Sydney, and Keiichi Matsuura, National Science
Museum, Tokyo, for the loan of tholichthys specimens. We thank
James F. DiLoreto, Smithsonian Photographic Services, for scanning
our illustrations and transmitting them between the authors. We are
much indebted to Sandra J. Raredon, National Museum of Natural
History, for the many hours she devoted to taking digital radiographs
of extant chaetodontid fishes to provide us with an extensive survey of
vacant interneural space position in that family. Ai Nonaka, National
Museum of Natural History, provided technical assistance in sorting,
cataloging, photographing and taking data on numerous chaetodontid
larvae. Our identifications of chaetodontid specimens at the Smith-
sonian Institution were greatly aided by Gerry R. Allen of the Western
Australian Museum during his visit to the Smithsonian in 2008. The
photographs of the Polish specimen ZPALWr. A/4000 were made by
Pawe? Socha of the Paleozoology Department of Wroclaw University.
The research of Alexandre F. Bannikov was supported by the
Russian Foundation of Basic Research (grants nos 08-05-00654,
First fossil records of the tholichthys larval stage of butterfly fishes
123
09-05-00170). The manuscript was improved by the editing of
Diane M. Tyler. Useful suggestions were offered during the peer
review process by Giorgio Carnevale (Universita` di Pisa) and an
anonymous reviewer.
References
Baciu, D.-S., A.F. Bannikov, and J.C. Tyler. 2005. Revision of the
fossil fishes of the family Zeidae (Zeiformes). Bollettino del
Museo Civico di Storia Naturale di Verona, Geologia Paleont-
ologia Preistoria 29: 95?128.
Baldwin, C.C., and G.D. Johnson. 1993. Phylogeny of the Epinep-
helinae (Teleostei: Serranidae). Bulletin of Marine Science
52(1): 240?283.
Bannikov, A.F. 2004. Fishes from the Eocene of Bolca, northern Italy,
previously classified with the Chaetodontidae (Perciformes).
Studi e Ricerche sui Giacimenti Terziari di Bolca, Museo Civico
di Storia Naturale di Verona 10: 13?92.
Bannikov, A.F., and J.C. Tyler. 1995. Phylogenetic revision of the
fish families Luvaridae and 
Kushlukiidae, with a new genus and
two new species of Eocene luvarids. Smithsonian Contributions
to Paleobiology 81: 1?45.
Birdsong, R.S., E.O. Murdy, and F.L. Pezold. 1988. A study of the
vertebral column and median fin osteology of gobioid fishes with
comments on gobioid relationships. Bulletin of Marine Science
42: 174?214.
Blum, S.D. 1988. The osteology and phylogeny of the Chaetodontidae
(Teleostei: Perciformes). Unpublished doctoral diss., Honolulu:
University of Hawaii. http://research.calacademy.org/research/
informatics/sblum/ Accessed December 2008.
Burgess, W.E. 1978. Butterflyfishes of the world. Neptune City:
T.F.H. Publications.
Carnevale, G. 2006. Morphology and biology of the Miocene
butterflyfish Chaetodon ficheuri (Teleostei: Chaetodontidae).
Zoological Journal of the Linnean Society 146: 251?267.
Cocheret de la Morinie`re, E., B.J.A. Pollux, L. Negelkerken, and G.
Van der Velde. 2002. Post-settlement life cycle migration
patterns and habitat preferences of coral reef fish that use
seagrass and mangrove habitats as nurseries. Estuarine, Coastal
and Shelf Science 55: 309?321.
Day, F. 1870. Notes on some fishes form the western coast of India.
Proceedings of the Zoological Society of London 1870: 369?374.
Grimm, K., M. Grimm, A. Ko?the, and T. Schindler. 2002. Der
??Rupelton?? (Rupelium, Oligoza?n) der Tongrube Bott-Eder
bei Rauenberg (Oberrheingraben, Deutschland). Courier For-
schungsinstitut Senckenberg 237: 229?253.
Gunther, A. 1871. On the young state of fishes belonging to the family
Squamipinnes. Annals and Magazine of Natural History, 4th ser
8: 318?320.
Hovestadt, D.C., and M. Hovestadt-Euler. 1999. Weissobatis mickli-
chi n.gen., n.sp., an eagle ray (Myliobatiformes, Myliobatidae)
from the Oligocene of Frauenweiler (Baden-Wu?rttemberg,
Germany). Pala?ontologische Zeitschrift 73(3/4): 337?349.
Hovestadt, D.C., and M. Hovestadt-Euler. 2002. The remains of a
carcharhinid shark with a new triakid species in its digestive tract
from the Oligocene of Germany. Tertiary Research 21(1?4):
171?182.
Jerzman?ska, A., and J. Kotlarczyk. 1976. The beginnings of the
Sargasso Assemblage in the Tethys? Palaeogeography Palaeo-
climatology Palaeoecology 20: 297?306.
Johnson, G.D. 1984. Percoidei: development and relationships. In
Ontogeny and systematics of fishes, Special Publication Number
1, American Society of Ichthyologists and Herpetologists, ed.
H.G. Moser, et al., 464?498. Lawrence: Allen Press.
Kaiser, T., and N. Micklich. 1995. Schnellpra?paration von Messel-
Fossilien zum Erstellen einer osteologischen Referenzsammlung.
Der Pra?parator 41(2): 69?72.
Kotlarczyk, J., and A. Jerzman?ska. 1976. Biostratigraphy of the
Menilite Beds of Skole Unit from the Polish Flysh Carpathians.
Bulletin de l?Acade?mie Polonaise des Sciences, Serie des
Sciences de la Terre 24(1): 55?61.
Kotlarczyk, J., and A. Jerzman?ska. 1988. Ichtiofauna w stratygrafii
Karpat. Przegla?d Geologiczny 6: 346?352.
Kotlarczyk, J., A. Jerzman?ska, E. S?widnicka, and T. Wiszniowska.
2006. A framework of ichthyofaunal ecostratigraphy of the
Oligocene-early Miocene strata of the Polish outer Carpathian
basin. Annales Societatis Geologorum Poloniae 76: 1?111.
Laegdesgaard, P., and C. Johnson. 2001. Why do juvenile fish utilise
mangrove habitats? Journal of Experimental Marine Biology and
Ecology 257: 229?253.
Leis, J.M. 1989. Larval biology of butterflyfishes (Pisces, Chae-
todontidae): what do we really know? Environmental Biology of
Fishes 25: 87?100.
Martini, E. 1971. Standard tertiary and quaternary calcareous
nannoplankton zonation. In Proceedings of the II Planktonic
conference, Roma 1970, ed. A. Farinacci, 738?785. Roma:
Edizioni Teenoscienza.
Mayr, G. 2000. A new mousebird (Coliiformes: Coliidae) from the
Oligocene of Germany. Journal of Ornithology 141: 85?92.
Mayr, G. 2004. Old World fossil record of modern-type humming-
birds. Science 304: 861?864.
Mayr, G., and A. Manegold. 2004. The oldest European fossil
songbird from the early Oligocene of Germany. Naturwissen-
schaften 91: 173?177.
Mayr, G., S.D. Peters, and S. Rietschel. 2002. Petrel-like birds with a
peculiar foot morphology from the Oligocene of Germany and
Belgium. Journal of Vertebrate Palaeontology 22: 667?676.
Micklich, N. 1998. New information on the fish fauna of the
Frauenweiler fossil site. Italian Journal of Zoology 65(supple-
ment): 169?184.
Micklich, N. 2005. The fish fauna of Frauenweiler clay pit and its
bearing on the paleoenvironmental reconstructions. Kaupia,
Darmsta?dter Beitra?ge zur Naturkunde 14: 89.
Micklich, N., and M. Drobek. 2007. Mining, excavations, and
preparation. In Messel??Treasures of the Eocene. Book to the
exhibition ??Messel on Tour??, ed. G. Gruber, and N. Micklich,
15?21. Darmstadt: Hessisches Landesmuseum.
Micklich, N., and L. Hildebrand. 2005. The Frauenweiler clay pit
(??Grube Unterfeld??). Kaupia, Darmsta?dter Beitra?ge zur
Naturkunde 14: 113?118.
Micklich, N., and N.N. Parin. 1996. The fishfauna of Frauenweiler
(Lower Oligocene, Rupelian; Germany): preliminary results of a
revision. Publicaciones Especiales Instituto Espanol de Ocean-
ographica 21: 129?148.
Parin, N.N., and N. Micklich. 1996a. Two aulostomid fishes
(Aulostomidae, Syngnathiformes) from the Lower Oligocene
of Germany. Palaeontologysyesky Zhurnal (Moscow) 1: 61?67.
[in Russian with English abstract].
Parin, N.N., and N. Micklich. 1996b. Fossil Gasterosteiformes from
the Lower Oligocene of Frauenweiler. I. New information on the
morphology and systematics of the genus Aeoliscus. Pala?onto-
logische Zeitschrift 70(3/4): 521?545.
Parin, N.V., and D.A. Astakhov. 2007. New fossil fish Musculope-
dunculus micklichi (Trichiuroidei) from the Lower Oligocene of
Germany. Journal of Ichthyology 47(8): 581?588.
Pharisat, A., and N. Micklich. 1998. Oligocene fishes in the western
Paratethys of the Rhine Valley Rift System. Italian Journal of
Zoology 65(supplement): 163?168.
Roth, R., and N. Micklich. 2006. Eine neue Methode zur Pra?paration
kleiner Messel-Fossilien. Der Pra?parator 52: 64?67.
N. R. Micklich et al.
123
Sakamoto, K., T. Uyeno, and N. Micklich. 2003. Oligoscophthalmus
weissi gen. et sp. nov., an Oligocene scophthalmid flatfish from
Frauenweiler, S-Germany. Bulletin of the National Science
Museum, Series C 29: 25?32.
Sakamoto, K., T. Uyeno, and N. Micklich. 2004. Oligopleuronectes
germanicus gen. et sp. nov., an Oligocene pleuronectid from
Frauenweiler, S-Germany. Bulletin of the National Science
Museum, Series C 30: 89?94.
Trunko`, L. 1997. Bericht zum Forschungsprogramm ??Sedimentolo-
gische Untersuchungen an Tertia?rsedimenten der Randfazies im
Mittleren Oberrheingraben?? gefo?rdert von der DFG in den
Jahren 1998?1991. Karlsruhe: Staatliches Museum fu?r Naturk-
unde. [unvero?ffentlichter Forschungsbericht].
Trunko`, L., and W. Munk. 1998. Geologische Beobachtungen in drei
tertia?ren Aufschlu?komplexen im Randbereich des Mittleren
Rheingrabens. Carolinea 56: 9?28.
Tyler, J.C., G.D. Johnson, I. Nakamura, and B.B. Collette. 1989.
Morphology of Luvarus imperialis (Luvaridae), with a
phylogenetic analysis of the Acanthuroidei (Pisces). Smithsonian
Contributions to Zoology 485: 1?78.
Tyler, J.C., B. O?Toole, and R. Winterbottom. 2003. Phylogeny of the
genera and families of zeiform fishes, with comments on their
relationships with tetraodontiforms and caproids. Smithsonian
Contributions to Zoology 618: 1?110.
Wagner-Klett, W. 1919. Das Tertia?r von Wiesloch in Baden, ein
Beitrag zu seiner tektonischen, stratigraphischen und pala?onto-
logischen Kenntnis. Jahresberichte und Mitteilungen des Ober-
rheinischen Geologischen Vereins, N.F. 8: 73?116.
Weber, H. 1951. Mikropala?ontologische Untersuchungen im Tertia?r
des Rheintalrandes bei Wiesloch in Baden. Erdo?l und Kohle 4:
543?549.
Weiler, W. 1966. Die Bedeutung der Fischfunde im Rupelton der
Tongrube Frauenweiler bei Wiesloch su?dlich von Heidelberg.
Zeitschrift der Rheinischen Naturforschenden Gesellschaft
Mainz 4: 1?37.
First fossil records of the tholichthys larval stage of butterfly fishes
123