A revision of the Mesozoic dragonfly family Tarsophlebiidae, with a discussion on the phylogenetic positions of the Tarsophlebiidae and Sieblosiidae (Odonatoptera: Panodonata).

Abstract

The Upper Jurassic-Lower Cretaceous dragonfly family Tarsophlebiidae is revised. The type species of the type genus Tarsophlebia Hagen, 1866, T. eximia (Hagen, 1862) from the Upper Jurassic Solnhofen Limestones, is redescribed, including important new information on its head, legs, wings, anal appendages and male secondary genital apparatus. The type specimen of Tarsophlebiopsis mayi Tillyard, 1923 is regarded as an aberrant or unusually preserved Tarsophlebia eximia. One new species of Tarsophlebia and three new species of Turanophlebia are described, i.e. Tarsophlebia minor n. sp., Turanophlebia anglicana n. sp., T. mongolica n. sp., and T. vitimensis n. sp. A new combination is proposed for Turanophlebia neckini (Martynov, 1927) n. comb. The phylogenetic relationships of the Mesozoic Tarsophlebiidae are discussed on the basis of new body and wing venation characters. The present analysis supports a rather derived position for the Tarsophlebiidae, as sister group of the the Epiproctophora rather than of (Zygoptera + Epiproc- tophora). Also, through the present discussion, the Oligo-Miocene family Sieblosiidae seems to be more closely related to the Epiproctophora than to the Zygoptera. But the present study and previous analyses suffer of the lack of informations concerning the more inclusive groups of Odonatoptera, viz. Protozygoptera, Triadophlebiomorpha, Protanisoptera, etc. The significance of the tarsophlebiid secondary male genital apparatus for the reconstruction of the evolution of odonate copulation is discussed.

Full text

33
GEODIVERSITAS • 2004 • 26 (1) © Publications Scientifiques du Muséum national d’Histoire naturelle, Paris. www.geodiversitas.com
A revision of the Upper Jurassic-Lower
Cretaceous dragonfly family Tarsophlebiidae,
with a discussion on the phylogenetic positions
of the Tarsophlebiidae and Sieblosiidae
(Insecta, Odonatoptera, Panodonata)
Günther FLECK
Département Histoire de la Terre, USM 0203 and CNRS-Muséum UMR 5143,
Muséum national d’Histoire naturelle, 45 rue Buffon, F-75231 Paris cedex 05 (France)
fleck@mnhn.fr
Günter BECHLY
Staatliches Museum für Naturkunde Stuttgart, Rosenstein 1,
D-70191 Stuttgart (Germany)
bechly.smns@naturkundemuseum-bw.de
Xavier MARTÍNEZ-DELCLÒS
Departament d’Estratigrafia, Paleontologia i Geociències Marines,
Facultat de Geologia, Universitat de Barcelona,
E-08071 Barcelona (Spain)
xdelclos@natura.geo.ub.es
Edmund A. JARZEMBOWSKI
Maidstone Museum and Bentlif Art Gallery, St Faith’s St,
Maidstone, Kent ME14 1LH (UK)
and Postgraduate Research Institute for Sedimentology & PRIS,
University of Reading, Reading RG6 2AB (UK)
ed@mbcmus1.demon.co.uk
André NEL
Département Histoire de la Terre, USM 0203 and CNRS-Muséum UMR 5143,
Muséum national d’Histoire naturelle, 45 rue Buffon, F-75231 Paris cedex 05 (France)
anel@mnhn.fr
Fleck G., Bechly G., Martínez-Delclòs X., Jarzembowski E. A. & Nel A. 2004. — A revision of
the Upper Jurassic-Lower Cretaceous dragonfly family Tarsophlebiidae, with a discussion
on the phylogenetic positions of the Tarsophlebiidae and Sieblosiidae (Insecta,
Odonatoptera, Panodonata). Geodiversitas 26 (1) : 33-60.

ABSTRACT
The Upper Jurassic-Lower Cretaceous dragonfly family Tarsophlebiidae is
revised. The type species of the type genus Tarsophlebia Hagen, 1866, T.
eximia (Hagen, 1862) from the Upper Jurassic Solnhofen Limestones, is
redescribed, including important new information on its head, legs, wings,
anal appendages and male secondary genital apparatus. The type specimen of
Tarsophlebiopsis mayi Tillyard, 1923 is regarded as an aberrant or unusually
preserved Tarsophlebia eximia. One new species of Tarsophlebia and three new
species of Turanophlebia are described, i.e. Tarsophlebia minor n. sp.,
Turanophlebia anglicana n. sp., T. mongolica n. sp., and T. vitimensis n. sp. A
new combination is proposed for Turanophlebia neckini (Martynov, 1927) n.
comb. The phylogenetic relationships of the Mesozoic Tarsophlebiidae are
discussed on the basis of new body and wing venation characters. The present
analysis supports a rather derived position for the Tarsophlebiidae, as sister
group of the the Epiproctophora rather than of (Zygoptera + Epiproc-
tophora). Also, through the present discussion, the Oligo-Miocene family
Sieblosiidae seems to be more closely related to the Epiproctophora than to
the Zygoptera. But the present study and previous analyses suffer of the lack
of informations concerning the more inclusive groups of Odonatoptera, viz.
Protozygoptera, Triadophlebiomorpha, Protanisoptera, etc. The significance
of the tarsophlebiid secondary male genital apparatus for the reconstruction
of the evolution of odonate copulation is discussed.
RÉSUMÉ
Révision des Tarsophlebiidae du Jurassique supérieur et Crétacé inférieur, discus-
sion sur les positions phylogénétiques des Tarsophlebiidae et des Sieblosiidae
(Insecta, Odonatoptera, Panodonata).
La famille de libellules mésozoïques des Tarsophlebiidae est révisée. L’espèce
type du genre type Tarsophlebia Hagen, 1866, T. eximia (Hagen, 1862) du cal-
caire lithographique du Jurassique supérieur de la Bavière, est redécrite, in-
cluant de nouvelles données sur les structures de la tête, des pattes, des ailes, des
appendices anaux et des pièces génitales secondaires du mâle. Le type de
Tarsophlebiopsis mayi Tillyard, 1923 est considéré comme un spécimen
aberrant ou curieusement préservé de Tarsophlebia eximia. Une nouvelle espèce
de Tarsophlebia et trois nouvelles espèces de Turanophlebia sont décrites :
Tarsophlebia minor n. sp., Turanophlebia anglicana n. sp., T. mongolica n. sp. et
T. vitimensisn. sp. Une nouvelle combinaison est proposée pour Turanophlebia
neckini (Martynov, 1927) n. comb. Les relations phylogénétiques des
Tarsophlebiidae sont discutées sur la base des nouveaux caractères du corps et
des ailes. Notre analyse conclut à une probable position des Tarsophlebiidae
comme groupe frère des Epiproctophora, plutôt que des (Zygoptera +
Epiproctophora). Au travers de la présente discussion, il ressort que la famille
cénozoïque des Sieblosiidae semble plus proche des Epiproctophora que des
Zygoptera. Il apparaîtaussi clairement que toutes ces analyses phylogénétiques
souffrent du manque de données concernant les groupes plus basaux, comme
les Protozygoptera, Triadophlebiomorpha ou Protanisoptera. L’intérêt des
pièces génitales secondaires des mâles des Tarsophlebiidae pour l’analyse de
l’évolution de la copulation chez les Odonatoptera est discuté.
Fleck G. et al.
34 GEODIVERSITAS • 2004 • 26 (1)
KEY WORDS
Insecta,
Odonatoptera,
Zygoptera,
Epiproctophora,
Tarsophlebiidae,
Sieblosiidae,
Upper Jurassic-Lower
Cretaceous,
Cenozoic,
fossil,
copulation,
phylogeny,
revision,
new species.
MOTS CLÉS
Insecta,
Odonatoptera,
Zygoptera,
Epiproctophora,
Tarsophlebiidae,
Sieblosiidae,
Jurassique supérieur-
Crétacé inférieur,
Cénozoïque,
fossile,
copulation,
phylogénie,
révision,
nouvelles espèces.

Revision of Tarsophlebiidae (Odonatoptera)
35
GEODIVERSITAS • 2004 • 26 (1)
INTRODUCTION
The Tarsophlebiidae Handlirsch, 1906 is one of
the most enigmatic Mesozoic family of Odona-
toptera. Its phylogenetic relationships with the Zy-
goptera and the Epiproctophora remain very un-
certain. Nel et al. (1993) considered them as the
most inclusive lineage within the “Anisozygoptera”
+ Anisoptera (= Epiproctophora). Bechly (1996)
and Rehn (2003) proposed to place them as sister
group of the (Zygoptera + Epiproctophora). After
our recent studies at PIN and MCZ, we could re-
vise and describe several species. This study also re-
vealed that the alleged calopterygoid-like anal ap-
pendages of Tarsophlebia, with apparently two
pairs of claspers, are clearly based on misinterpreta-
tions due to artifacts of preservation. Indeed Tarso-
phlebia definitely does not possess zygopteroid but
rather unique appendages. There are no visible
paraprocts, the epiproct must have been reduced
or inconspicuous since it is not visible in any speci-
men, and the cerci are very long, with a double-
barrelled basal petiole and a curious distal plate-
like expansion. The broken double-barrelled
petioles of the two cerci have been commonly mi-
sinterpreted as two pairs of claspers, while the dis-
tal plates have been overlooked or regarded as pre-
servational artifacts. This new result is of particular
phylogenetic relevance, since the previous inter-
pretation of the appendages would have represen-
ted a putative synapomorphy with Zygoptera. We
comment on the characters that are used for the
phylogenetic position of the Tarsophlebiidae.
ABBREVIATIONS
Institutions
BSPGM Bayerische Staatssamlung für Paläon-
tologie und Historische Geologie,
Munich;
JME Jura-Museum, Eichstätt;
MCZ Museum of Comparative Zoology,
Harvard University, Cambridge;
PIN Arthropod Laboratory, Palaeonto-
logical Institute, Academy of Science
of Russia, Moscow;
Names of body and wing structures
AA Analis Anterior;
Ax1 and Ax2 primary antenodal cross-veins;
Bq series of cross-veins between RP1+2,
IR2 and oblique vein;
C costal vein or costa;
Cr nodal cross-vein;
CuA Cubitus Anterior;
CuAb first posterior branch of Cubitus
Anterior;
CuP Cubitus Posterior;
IR intercalary vein of radial area;
MA Median Anterior;
MAb first posterior branch of Median
Anterior;
MP Median Posterior;
RA Radius Anterior;
RP Radius Posterior;
ScA Subcosta Anterior
ScP Subcosta Posterior;
MATERIAL AND METHODS
All drawings were made with a camera lucida and
a binocular microscope. The nomenclature of the
dragonfly wing venation is based on the interpre-
tations of Riek & Kukalová-Peck (1984), amen-
ded by Nel et al. (1993) and Bechly (1996). The
systematic analysis is based on the principles of
consequent phylogenetic systematics (sensu
Hennig 1966, 1969, 1981), especially character
polarisation based on outgroup comparison.
SYSTEMATICS
Order ODONATOPTERA Martynov, 1932
Suborder PANODONATA Bechly, 1996
Family TARSOPHLEBIIDAE Handlirsch, 1906
Tarsophlebiidae Handlirsch, 1906: 467, 468, 580,
581.
TYPE GENUS.— Tarsophlebia Hagen, 1866, original
designation of Handlirsch (1906) by monotypy.
INCLUDED GENERA.— Tarsophlebia Hagen, 1866 (=
Tarsophlebiopsis Tillyard, 1923) and Turanophlebia
Pritykina, 1968. Since Bechly (1996, 1997) demon-
strated that the genera Euthemis Pritykina, 1968
(Upper Jurassic) and Sphenophlebia Bode, 1953
(Upper Liassic) are isophlebioid “anisozygoptères” and
thus not related to Tarsophlebiidae, there is no more

evidence for a Lower Jurassic occurrence of
Tarsophlebiidae and also no need for a further redun-
dant taxon Tarsophlebioidea or Tarsophlebioptera.
STRATIGRAPHIC AND GEOGRAPHIC RANGE. — Only
known from the Upper Jurassic and Lower Cretaceous
of Eurasia.
EMENDED DIAGNOSIS. — Nel et al. (1993) proposed a
diagnosis of the Tarsophlebiidae. We here emend this
diagnosis with the following characters: 1) primary
antenodal braces Ax1 and Ax2 stronger than the sec-
ondary antenodal cross-veins; 2) in all fore and hind
wings, there are pairs of secondary longitudinal con-
cave veins “above” and “below” the convex veins CuA,
MA, and IR2, and closely parallel to them. We pro-
pose to name them respectively antero-CuA, postero-
CuA, antero-MA, postero-MA, antero-IR2 and
postero-IR2 intercalary veins. These veins are more or
less long and defined in the different taxa. The postero-
intercalaries are always longer than the associated
antero-intercalaries; and 3) in male, one pair of well
separated anal appendages visible, of very particular
shape, basally strongly sclerotized, with high humps
and a non-sclerotized paddle-like distal part. No me-
dian anal appendage visible. These body characters are
only known for Tarsophlebia eximia and Tura-
nophlebia vitimensis n. sp.
The monophyly of Tarsophlebiidae is supported by
several strong autapomorphies (Nel et al. 1993), such
as: hypertrophied hind wing subdiscoidal cell, devel-
oped as “pseudo discoidal cell”; in hind wing, fusion of
veins [MAb + MP + CuA] for a considerable distance
before separation of MP and CuA; vein AA strongly
bent at insertion of CuP-crossing; extremely acute dis-
tal angles of fore wing discoidal and subdiscoidal cell.
The characters “distinctly prolonged legs, with very
long tarsi” and “male cerci with strange distal expan-
sions” are present in both T. eximia and T. vitimensis
n. sp., and probably also in other Tarsophlebiidae.
The “extremely prolonged female ovipositor” could be
an autapomorphy of the family, too, but it is only
known from T. eximia. A further alleged autapomor-
phy mentioned by Nel et al. (1993) rather seems to be
a symplesiomorphy (Bechly 1996), viz. “veins
[RP + MA] - MAb - CuA aligned”. It is also present in
the Epiophlebiidae Muttkowski, 1910, Isophlebio-
ptera and Heterophlebioptera.
Genus Tarsophlebia Hagen, 1866
Tarsophlebia Hagen, 1866: 65.
TYPE SPECIES.— Heterophlebia eximia Hagen, 1862,
original designation of Hagen (1866) by monotypy.
OTHER SPECIES. — We transfer Tarsophlebia neckini
Martynov, 1927 into the genus Turanophlebia. Tarso-
phlebia westwoodii Hagen, 1850 from the Lower
Liassic of Gloucestershire in England was correctly
transferred to the heterophlebioid genus Liassophlebia
by Tillyard (1925).
EMENDED DIAGNOSIS. — Nel et al. (1993) proposed a
diagnosis, erroneous in some points. We emend it as
follows: primary antenodal braces stronger than sec-
ondaries, as in other Tarsophlebiidae (contra Nel et al.
1993); cubito-anal areas of fore and hind wings with
four rows of cells or less; 16 or less postnodal cross-
veins; less than ten secondary antenodal cross-veins in
hind wing; and IR1 relatively short.
REMARKS
Nel et al. (1993) noted that the genus Turano-
phlebia is very close to Tarsophlebia.
Tarsophlebia eximia (Hagen, 1862)
(Figs 1; 2)
Heterophlebia eximia Hagen, 1862: 102, 106.
Euphaea longiventris Hagen, 1862: 106, 121, pl. 13,
figs 7, 8.
Tarsophlebia eximia – Hagen 1866: 65, pl. 2, figs 1-6,
11.
Agrion latreillei [sensu Meunier] Meunier, 1896: pl. 1.
Agrion exhaustum [sensu Meunier] Meunier, 1896: pl. 2.
Tarsophlebia major Handlirsch, 1906: 580.
?Tarsophlebia longissima Handlirsch, 1906: 581.
Tarsophlebiopsis mayi Tillyard, 1923: 149 (n. syn.).
HOLOTYPE OF HETEROPHLEBIA EXIMIA. — Specimen No.
AS-VI-44a-b (BSPGM). The holotype of T. major is in
the collection of the National Museum in Prague (Malz
& Schröder 1979: 26). Tarsophlebiopsis mayi is known
by the holotype specimen, only represented by three
wing fragments, located in the collection of Sedgwick
Museum, Cambridge, UK.
GEOLOGICAL SETTING. — Upper Jurassic (“Oberer
Weißjura”, Malm ξ2b), lower Tithonian, Hybono-
tum-Zone, Solnhofen Formation (Solnhofen Litho-
graphic Limestone), Eichstätt area, near Solnhofen,
southern Frankonian Alb, Bavaria, SW Germany. The
type specimen of T. mayi was found at the Boulder
Clay (probably derived from the Ampthill Clay,
Corallian), Hertfordshire, UK (Tillyard 1923: 146).
DIAGNOSIS. — That of the genus. The fore wing length
varies from 35-41 mm, and the hind wing length from
30-39 mm. The wing lengths of Tarsophlebia major
Handlirsch, 1906 (39 mm) and Tarsophlebia longissima
Handlirsch, 1906 (42 mm) still fit in this continuous
Fleck G. et al.
36 GEODIVERSITAS • 2004 • 26 (1)

range of variability. Since there are no other diagnostic
characters known, we tentatively concur with the synon-
ymization of Nel et al. (1993), even though the unusu-
ally large range of size would well allow the recognition
of a distinct larger species of Tarsophlebia.
REDESCRIPTION
Head
A well preserved head is only known from speci-
men No. 6129 in coll. Carpenter of MCZ, which
is labelled “Tarsophlebia eximia Hag. - mas. -
Revision of Tarsophlebiidae (Odonatoptera)
37
GEODIVERSITAS • 2004 • 26 (1)
Counterpart - Solenhofen - Dr. Krantz” and
indicated as “Type”. Hagen (1866) figured this
specimen (reproduced by Nel et al. 1993:
fig. 53a-b). Our re-examination of the specimen
showed that the original figures were rather
imprecise. We give a new figure, after the direct
examination of the holotype (Fig. 1C).
Thorax
Nel et al. (1993: 85) remarked that all specimens
with the synthorax preserved are in dorsal aspect,
so that it is not possible to quantify exactly the
IV III
II
I
hamuli post. hamuli ant.
dorso-median keel
vesicula spermalis
ligula
A
B C
D
FIG.1.— Tarsophlebia eximia (Hagen, 1862); A, male specimen (SOS 1720, JME), secondary genital apparatus; B, male holotype
(BSPGM AS-VI-44b), counterpart, right hind leg; C, male (No. 6129, coll. Carpenter, MCZ), head; D, male (No. 6222, coll. Carpenter,
MCZ), male genital appendage. Scale bars: A, B, D, 5 mm; C, 10 mm.

degree of skewness (“prognathisme” sensu Nel et
al. 1993) of the thorax. However, they estimate
that the skewness was probably more important
than in recent Zygoptera, Lestidae Calvert, 1901.
The very elongated humeral region (mesepister-
num) of the synthorax with a very long dorsal
carina (visible in specimens No. SOS 1705/1720
in JME) strongly supports a hypertrophied skew-
ness.
Legs
The very long legs and tarsi stipulated the generic
name Tarsophlebia. Legs of similar relative length
are only known from the fossil protomyrmeleon-
tid Malmomyrmeleon viohli Martínez-Delclòs
& Nel, 1996 from the same locality (Martínez-
Delclòs & Nel 1996). Hagen (1866) described a
four-segmented tarsus, which was accepted by all
subsequent authors, except Nel et al. (1993) who
rejected this interpretation, since the type speci-
men is not sufficiently preserved. In fact, the tarsi
of most specimens are far too poorly preserved to
recognize the true number of tarsomeres. It is ex-
tremely hard to identify the different tarsomeres
of the legs of these fossils. As example, only three
tarsomeres could be identified on all tarsi of the
specimen No. 1951/73K (JME) (Nel et al. 1993:
86), with the following putative lengths: the fore-
leg length of first tarsomere 3.0 mm, length of se-
cond tarsomere 1.5 mm, length of third tarsomere
2.0 mm; median leg length of first tarsomere
2.5 mm, length of second tarsomere 1.2 mm,
length of third tarsomere 1.0 mm; hind leg length
of first tarsomere 3.0 mm, length of second tarso-
mere 2.0 mm, length of third tarsomere 1.0 mm.
Nevertheless, we found two well preserved speci-
mens which clearly confirm the original descrip-
tion of four-segmented tarsi because they show
four segment borders and the segments are angled
and not aligned with each other. The right hind
leg of the holotype specimen AS-VI-44b in
BSPGM indeed shows four tarsomeres, with the
following dimensions: length of first tarsomere
1.46 mm, length of second tarsomere 1.40 mm
(both could correspond to the putative first tarso-
mere of specimen No. 1951/73K), length of third
tarsomere 1.45 mm, length of fourth tarsomere
1.68 mm. The left foreleg of specimen No. 1960/
66K in JME also shows four tarsomeres of more or
less equal length.
The tarsi of all specimens show a pair of elongate
tarsal claws that do not seem to have a subapical
tooth (putative plesiomorphy, even though a
reduction or artifact of preservation cannot be
excluded).
Wings
They are extensively described in Nel et al. (1993),
except for the following emendations: primary an-
tenodal braces stronger than the secondaries; pre-
sence of a relatively long not zigzagged secondary
vein (“postero-CuA vein”) closely parallel to distal
part of CuA, in cubito-anal area and another one
in area between MP and CuA (“antero-CuA
vein”); presence of a relatively long not zigzagged
secondary vein (“postero-MA vein”) closely paral-
lel to distal part of MA, in postdiscoidal area and
another one in area between RP3/4 and MA (“an-
tero-MA vein”).
Abdomen
Several specimens (e.g., specimen No. SOS 1720
in JME) show the presence of a longitudinal medio-
dorsal carina on the abdominal terga, which is cov-
ered with a row of small spines. Specimen No. SOS
1720 also shows numerous tiny spines on the later-
al parts of the abdominal terga, very similar to extant
odonates (e.g., Calopteryx Leach, 1815).
Male secondary genital apparatus
There is a single male specimen (No. SOS 1720 in
JME) that shows the male secondary genital appa-
ratus, even though the specimen is preserved in
dorsal aspect, since the ventral side is pressed
through. Nel et al. (1993: fig. 54) figured this spe-
cimen, but the genitals were apparently difficult
to interpret. The alleged male “auricles” of Tarso-
phlebia eximia, described by Nel et al. (1993:
fig. 68) from the second abdominal segment of
specimen No. 1650/57a-b in JME, are obviously
based on a misinterpretation of the hamuli poste-
riors (Bechly 1996; Rehn 2003: 201). We propose
a new interpretation of the male secondary genital
apparatus (see Fig. 1A).
Fleck G. et al.
38 GEODIVERSITAS • 2004 • 26 (1)

Revision of Tarsophlebiidae (Odonatoptera)
39
GEODIVERSITAS • 2004 • 26 (1)
Female ovipositor
The female ovipositors are very long, extending
well beyond the abdomen. They were already
figured and described in Nel et al. (1993), on the
basis of specimen No. SOS 3609, in JME. Fleck
& Nel (2003) noted that, despite the fact they
both have long ovipositor, its structure is diffe-
rent in the extant family Cordulegastridae
Hagen, 1875 and the Mesozoic Aeschnidiidae
Needham, 1903. In the latter group, the valvula
1 is distinctly longer, narrower and weaker than
in Cordulegastridae. The exact structure of the
ovipositor cannot be established in Tarsophlebia
because of the poor state of preservation of the
available material.
Male anal appendages
The presence of a long ovipositor in female speci-
men supports the hypothesis that the specimens
without such a structure are male. The anal
appendages were always interpreted as caloptery-
goid-like, thus consisting of two pairs of long and
curved claspers (cerci and paraprocts). Also, Nel
et al. (1993: fig. 52) described such appendages
from the holotype specimen. The discovery of
very curious anal appendages of Turanophlebia
vitimensis n. sp. (see below) stipulated the thor-
ough re-examination of these structures in T.exi-
mia. Indeed we could identify appendages of
the same type in specimen No. 6222 in coll.
Carpenter of MCZ, which is also labelled “Tar-
sophlebia eximia Hag. - mas. - Counterpart -
Solenhofen - Dr. Krantz” and indicated as
“Type”. The previous misinterpretations are
based on artifacts of preservation.
THE STATUS OF THE ENIGMATIC
TARSOPHLEBIOPSIS MAYI TILLYARD, 1923
Tillyard (1923) described Tarsophlebiopsis mayi
based on two wing fragments that were found in
the body chamber of an ammonite from the
Upper Jurassic of England. Fraser (1955) revised
this specimen and provided a reconstruction of the
complete wing. Nel et al. (1993: 92) considered
that the holotype of this taxon could be just an
aberrant specimen in the plesiomorphic character
state “basal separation of the stems of M and Cu
with several cross-veins between their stems”. We
completely agree with this statement and now for-
mally synonymize Tarsophlebiopsis mayi with
Tarsophlebia eximia. The reasons are as follows:
Nel et al. (1993: 92) falsified two of the four di-
agnostic characters mentioned in the original de-
scription of Tillyard (1923). Nel et al. (1993: 92)
accepted the better-defined primary antenodal
cross-veins Ax1 and Ax2 as valid diagnostic charac-
ter. This state also occurs in Tarsophlebia eximia.
Thus, except for two character states (“M and Cu
basally separated”; “submedian space with two
rows of cells”), all preserved characters (including
size and fore wing venation) of Tarsophlebiopsis
mayi are absolutely identical with Tarsophlebia exi-
mia. Both taxa are from the Upper Jurassic of
Middle Europe. The first mentioned difference is
very strange, since this character state (“M and Cu
basally separated”) would represent a unique
plesiomorphy within Panodonata (= Tarso-
phlebiidae + Odonata), which is only known from
older and more basal stemgroup representatives of
Odonata (e.g., Meganisoptera, Protanisoptera,
Triadophlebiomorpha, and Protozygoptera). All
other tarsophlebiids agree with the derived charac-
ter state of extant odonates. The second character
state (“submedian space with two rows of cells”) is
in the same area of the wing and could be related
to the aberrant structure of veins M and Cu.
Therefore, both character states of the holotype of
Tarsophlebiopsis mayi could be related to an indivi-
dual atavistic aberration (also known from other
wing venational structures in extant Odonata), or
even as an artifact of preservation. The latter case
could occur when the two membranes of the wing
are separated and partly detached, which is for
example a common phenomenon in amber dam-
selflies.
FIG.2.— Tarsophlebia minor n. sp., holotype (No. 55, coll.
Carpenter, MCZ), left hind wing. Scale bar: 10 mm.

Fleck G. et al.
40 GEODIVERSITAS • 2004 • 26 (1)
Tarsophlebia minor n. sp.
(Fig. 2)
HOLOTYPE. — Specimen No. 55 in coll. Carpenter of
MCZ, which is labelled “Agrion spec. - Solenhofen -
Dr. Krantz”.
ETYMOLOGY. — Named after the relatively small size
(hind wing length only about 26 mm), compared to
the other species of the same genus.
GEOLOGICAL SETTING. — Hybonotum-Zone,
Solnhofen Formation (Solnhofen Lithographic
Limestone), lower Tithonian, Upper Jurassic (“Oberer
Weißjura”, Malm ξ2b), Eichstätt area, near Solnhofen,
southern Frankonian Alb, Bavaria, SW Germany.
DIAGNOSIS. — This new species is distinguished by its
distinctly smaller size than T. eximia (fore wing about
26 mm long instead of 35-41 mm long in T. eximia).
Also, the angle between MAb and MP + CuA is dis-
tinctly more opened than in the fore wing of T. eximia.
DESCRIPTION
The holotype is an isolated left fore wing with
missing apex. Length of preserved part, 25.0 mm,
probable length of wing, about 26.0 mm;
although the venation is rather poorly preserved,
some characters can be determined with accu-
racy. Wing probably hyaline, pterostigma dark
brown; distance from base to arculus, 3.2 mm;
from arculus to nodus, 7.8 mm; from nodus to
pterostigma, 10.3 mm, pterostigma elongated
and narrow, about 3.5 mm long, 0.5 mm wide,
not basally recessed; pterostigmal brace oblique
and strong, opposite pterostigma base; median
and submedian areas free of cross-veins; CuP
strongly curved, basal of Ax2, basally closing sub-
discoidal space; at least primary antenodal brace
and Ax2 stronger than secondary antenodal
cross-veins; arculus slightly opposite Ax2; only
few secondary antenodal cross-veins preserved
distal of Ax2, probably less than 10 in the living
animal; all secondary antenodal cross-veins not
aligned with the cross-veins of second rank be-
tween ScP and RA; MP + CuA strongly curved
just before its fusion with MAb; a sharp angle
between MP + CuA and MAb, but more opened
than in T.eximia; no fusion between MAb and
MP + CuA before CuA separates from MP. Thus
it is a fore wing, as the tarsophlebiid hind wings
have such a long fusion; RP + MA, MA and
MAb, MP + CuA + MAb, and basal free part of
CuA well aligned in arculus, as in other
Tarsophlebiidae; discoidal space basally opened;
subdiscoidal area divided into two cells by a
cross-vein; AA without any strong posterior
branches; anal area with two or three rows of
cells; posterior wing margin rounded; petiole
short, about 0.8 mm long; AA reaching free part
of CuA at sharp angle; no CuAb (sensu Fleck
et al. 2003); CuA without strong posterior
branches; less than five rows of small cells be-
tween CuA and posterior wing margin; “postero-
CuA vein” and “antero-CuA vein” not preserved;
CuA reaching posterior wing margin just basal to
nodus level; area between MP and CuA with one
row of cells in its basal part but greatly widened
in its distal half; postdiscoidal area slightly wi-
dened distally; bases of RP3/4 and IR2 between
arculus and nodus, distinctly nearer to nodus,
base of RP3/4 3.5 mm from nodus; base of IR2
apparently on RP3/4; nodal Cr and subnodus
strongly oblique; base of RP2 aligned with sub-
nodus; oblique vein “O” three small cells distal of
base of RP2; numerous Bq cross-veins; less than
11 postnodal cross-veins between C and RA, not
aligned with the postsubnodal cross-veins; IR1
well defined, not zigzagged and only slightly
curved; one row of cells between RP1 and IR1;
area between RP2 and IR2 distinctly widened
distally, “antero-IR2” and “postero-IR2” veins
not preserved; area between IR2 and RP3/4 dis-
tally widened; area between RP3/4 and MA dis-
tally widened; “antero-MA” and “postero-MA”
veins not preserved.
Genus Turanophlebia Pritykina, 1968
Turanophlebia Pritykina, 1968: 42.
TYPE SPECIES.— Turanophlebia martynovi Pritykina,
1968 by monotypy.
OTHER SPECIES INCLUDED.— Turanophlebia sibirica
Pritykina, 1977, Turanophlebia anglicana n. sp.,
Turanophlebia mongolica n. sp., Turanophlebia viti-
mensis n. sp., Turanophlebia neckini (Martynov, 1927)
n. comb.
EMENDED DIAGNOSIS. — Differs from the Upper
Jurassic Tarsophlebia eximia and Tarsophlebia minor

n. sp. in its denser wing reticulation, mainly visible
through: 1) the presence of more than 25 postnodal
cross-veins (against 16 in T. eximia and around 11 in T.
minor n. sp.); 2) six (or more) rows of cells between CuA
and posterior hind wing margin (against less than five
rows in T. eximia and T. minor n. sp.); 3) more than 10
secondary antenodal cross-veins in hind wing (against
less than 10 in T. eximia); 4) IR1 longer than in T.
eximia; and 5) presence of long secondary longitudinal
not zigzagged veins in area between IR2 and RP2.
REMARKS
Jarzembowski (1990) considered that Tarso-
phlebia and Turanophlebia differ in: 1) general
density of venation; 2) width of cubito-anal area;
and 3) form of “discal cell” in hind wing
(“pseudo-discoidal cell” sensu Nel et al. 1993).
The shape of the “pseudo-discoidal” cell is
variable in both Tarsophlebia eximia and the dif-
ferent species of Turanophlebia, thus the charac-
ter 3) is not constant.
Turanophlebia martynovi Pritykina, 1968
(Fig. 3)
Turanophlebia martynovi Pritykina, 1968: 43, 44, text-
fig. 14, pl. 3, fig. 4. — Carpenter 1992: 73,
fig. 46.4. — Nel et al. 1993: 89, 90, fig. 70.
HOLOTYPE. — Specimen No. 2554/21 (PIN).
GEOLOGICAL SETTING. — Upper Jurassic, Callovian-
Kimmeridgian or Oxfordian-Kimmeridgian
(Zherikhin & Gratshev 1993; Mostovski & Martínez-
Delclòs 2000), Karatau, Chimkent region, Southern
Kazakhstan, C.I.S.
DIAGNOSIS.— T. martynovi differs from T. sibirica
and T. anglicana n. sp. in the following character
states: pterostigma covering six cells; arculus opposite
Ax2; two rows of cells between C and RA distal of
pterostigma. T. martynovi has two rows of cells in hind
wing anal area, unlike T. neckini n. comb. T. marty-
novi has a strong angle in its hind wing arculus; its
subdiscoidal space is divided into three small cells,
unlike other Turanophlebia species. It differs from T.
vitimensis n. sp. in the presence of only six rows of cells
in cubito-anal area between CuA and posterior wing
margin. It differs from T. mongolica n. sp. in the pres-
ence of its vein “O” oblique.
?Turanophlebia sibirica Pritykina, 1977
Turanophlebia sibirica Pritykina, 1977: 84, 85, text-
fig. 2, pl. 1, fig. 2. — Nel et al. 1993: 91.
HOLOTYPE. — Specimen No. 1289/1258 (PIN).
GEOLOGICAL SETTING. — Lower Cretaceous, Zaza
Formation, Neocomian or Barremian-Aptian (Zheri-
khin et al. 1999; Mostovski & Martínez-Delclòs
2000). Baissa, Vitim river, Transbaikalia, C.I.S.
DIAGNOSIS.— ?T. sibirica differs from T. martynovi in
the presence of only one row of cells in the area
between C and RA distal of pterostigma. It differs
from T. anglicana n. sp. in its longer wings. It differs
from T. neckini n. comb. in the presence of two or
three rows of cells in the hind wing anal area.
REMARKS
The type specimen of ?T. sibirica is very poorly
preserved and incomplete. It is nearly impossible
to be accurate of its attribution to the genus Tura-
nophlebia: its cubito-anal area is too fragmentary to
determine the exact number of cells rows between
CuA and posterior wing margin (Pritykina’s wing
reconstruction is optimistic); its vein IR1 is badly
preserved. The only argument is the large number
of cells in the apical part of its wing.
Turanophlebia neckini (Martynov, 1927) n. comb.
(Fig. 4)
Tarsophlebia neckini Martynov, 1927: 757, 758, figs 1,
2. — Nel et al. 1993: 87.
HOLOTYPE. — Specimen No. 2452/3 (PIN).
GEOLOGICAL SETTING. — Upper Jurassic, Callovian-
Kimmeridgian or Oxfordian-Kimmeridgian
(Zherikhin & Gratshev 1993; Mostovski & Martínez-
Delclòs 2000), Karatau, Chimkent region, Southern
Kazakhstan, C.I.S.
DIAGNOSIS.— T. neckini n. comb. differs from the
other Turanophlebia species in the presence of only
one or two rows of cells in the hind wing anal area.
Revision of Tarsophlebiidae (Odonatoptera)
41
GEODIVERSITAS • 2004 • 26 (1)
FIG.3.— Turanophlebia martynovi Pritykina, 1968, holotype (PIN
2554/21). Scale bar: 10 mm.

REDESCRIPTION
There is some imprecision in the original descrip-
tion of Martynov (1927), thus a redescription is
necessary. Imprint of a hind wing, with the distal
half partly destroyed. Wing probably hyaline,
pterostigma missing; wing about 39.0 mm long;
7.3 mm wide; distance from base to arculus,
6.0 mm; from arculus to nodus, 10.2 mm;
median and submedian areas free of cross-veins;
CuP strongly curved, nearly opposite Ax2, basally
closing subdiscoidal space; primary antenodal
braces Ax1 and Ax2 stronger than secondary
antenodal cross-veins, without cross-veins be-
tween them, 2.0 mm apart; Ax1 is 3.8 mm from
wing base; arculus opposite Ax2; 14 secondary
antenodal cross-veins distal of Ax2, not aligned
with the cross-veins of second rank between ScP
and RA; numerous cross-veins in the area be-
tween RA and RP, between arculus and nodus; a
long “gap” without cross-veins between arculus
and RP3/4 in the area between RP and MA;
MP + CuA strongly curved just before its fusion
with MAb; a sharp angle between MP + CuA and
MAb; presence of a long fusion between MAb
and MP + CuA before CuA separates from MP,
1.1 mm long, characteristic of the Tarso-
phlebiidae (Nel et al. 1993); RP + MA, MA and
MAb, MP + CuA + MAb, and basal free part of
CuA well aligned in arculus, as in other Tarso-
phlebiidae (Nel et al. 1993); discoidal space
basally opened; presence of a two-celled “tarso-
phlebiid pseudo-discoidal space” just distal of
MAb in postdiscoidal area; subdiscoidal area
divided into two cells by a cross-vein; AA without
any strong posterior branches; anal area with one
or two rows of cells; posterior wing margin round-
ed; a short petiole, 1.4 mm long; AA reaching
free part of CuA at sharp angle; no CuAb (sensu
Fleck et al. 2003); CuA without strong posterior
branches; six rows of small cells in cubito-anal
area; CuA reaching posterior wing margin just
distal of nodus level; area between MP and CuA
with one row of cells in its basal part but greatly
widened in its distal half; postdiscoidal area with
two rows of cells in its basal part, narrowed in its
mid part and slightly widened distally; bases of
RP3/4 and IR2 between arculus and nodus, mid-
way between arculus and nodus, 5.2 mm from
nodus; apparent base of IR2 on RP3/4; nodal Cr
and subnodus strongly oblique; base of RP2 ali-
gned with subnodus; oblique vein “O” three small
cells distal of base of RP2; numerous Bq cross-
veins, but apparently no cross-vein in basal part of
area between RA, RP, RP3/4 and IR2; numerous
postnodal cross-veins between C and RA, not
aligned with the postsubnodal cross-veins; base of
IR1 about 10 cells distal of that of RP2; IR1 well
defined, long, not zigzagged and only slightly
curved; one row of cells between RP1 and IR1;
area between RP3/4 and MA distally widened.
DISCUSSION
This fossil has the diagnostic characters of the
genus Turanophlebia, i.e. cubito-anal area broad,
with six rows of cells between CuA and posterior
wing margin, more than 25 postnodal cross-
veins, more than 10 secondary antenodal cross-
veins in hind wing, IR1 longer than in
Tarsophlebia eximia. Thus, we propose to transfer
it into the genus Turanophlebia.
Turanophlebia anglicana n. sp.
(Fig. 5)
HOLOTYPE. — Specimen No. 018531, part and coun-
terpart, Booth Museum of Natural History, Brighton,
UK.
ETYMOLOGY. — After the latinised name of England.
GEOLOGICAL SETTING. — Lower Cretaceous, Lower
Weald Clay, Upper Hauterivian, Clockhouse
(Butterley) Brickworks, UK, National Grid Reference
TQ 175385.
Fleck G. et al.
42 GEODIVERSITAS • 2004 • 26 (1)
FIG. 4. — Turanophlebia neckini (Martynov, 1927) n. comb.,
holotype (PIN 2452/3). Scale bar: 10 mm.

DIAGNOSIS. — Nel & Jarzembowski (1996: 91,
fig. 4) attributed erroneously this specimen to a
Campterophlebiidae genus and species incertae sedis.
T. anglicana n. sp. differs from T. martynovi in the
following character states: pterostigma covering only
three cells instead of six; arculus slightly distal of
Ax2; Ax1 shifted more distally; only one row of cells
between C and RA distal of pterostigma. It differs
from ?T. sibirica in its smaller size (wing about
40 mm long against 45 mm for ?T. sibirica). It differs
from T. neckini n. comb. in the presence of two to
three rows of cells in the anal area instead of one to
two rows in T. neckini n. comb. T. anglicana n. sp.
differs from T. mongolica n. sp. in its vein “O”
oblique and of better defined secondary longitudinal
veins parallel to CuA. It differs from T. vitimensis
n. sp. in the presence of only six rows of cells between
CuA and posterior wing margin, instead of nine or
10, and of one row of cells in the most narrow part of
postdiscoidal area instead of two.
DESCRIPTION
Imprint and counterimprint of a nearly complete
hind wing. The counterpart was previously de-
scribed alone by Nel & Jarzembowski (1996) and
erroneously attributed to the Campterophlebiidae
Handlirsch, 1920. The error was due to its poor
state of preservation. Fortunately, the imprint is
very well preserved. The present redescription is
based on its study. Wing probably hyaline, ptero-
stigma dark brown; wing 39.7 mm long; 8.3 mm
wide; distance from base to arculus, 6.4 mm; from
arculus to nodus, 11.4 mm; from nodus to ptero-
stigma, 15.5 mm, to apex, 22.1 mm; pterostigma
elongated and narrow, 3.7 mm long, 0.5 mm wide,
covering three cells, not basally recessed; ptero-
stigmal brace oblique and strong, opposite
Revision of Tarsophlebiidae (Odonatoptera)
43
GEODIVERSITAS • 2004 • 26 (1)
antero-IR2
postero-IR2
antero-MA postero-MA
antero-CuP postero-CuP
pseudo-discoidal cell
Ax2 Ax1
CuP
A
B
C
FIG. 5. — Turanophlebia anglicana n. sp., holotype (No. 018531, Booth Museum of Natural History, Brighton, UK); A, complete wing,
other wing is that of a Libelluloidae; B, reconstruction of the hind wing; C, detail of hind wing nodus. Abbreviations: Ax1, Ax2,
primary antenodal cross-veins; CuP, Cubitus Posterior; IR2, intercalary vein of radial area; MA, Median Anterior. Scale bars: A,
10 mm; B, 5 mm; C, 1 mm.

pterostigma base; median and submedian areas free
of cross-veins; CuP strongly curved, nearly oppo-
site to Ax2, basally closing subdiscoidal space; pri-
mary antenodal braces Ax1 and Ax2 stronger than
secondary antenodal cross-veins, with no visible
cross-veins between them, 1.5 mm apart; Ax1 is
4.4 mm from wing base; arculus slightly distal of
Ax2 (0.4 mm); 13 secondary antenodal cross-veins
distal of Ax2, not aligned with the cross-veins of
second rank between ScP and RA; 12 cross-veins
in the area between RA and RP, between arculus
and nodus; a long “gap” without cross-veins be-
tween arculus and RP3/4 in the area between RP
and MA; MP + CuA strongly curved just before its
fusion with MAb; a sharp angle between MP +
CuA and MAb; presence of a long fusion between
MAb and MP + CuA before CuA separates from
MP, 1.4 mm long, characteristic of the
Tarsophlebiidae; RP + MA, MA and MAb, MP +
CuA + MAb, and basal free part of CuA well a-
ligned in arculus, as in other Tarsophlebiidae; dis-
coidal space basally opened; presence of a two-
celled “tarsophlebiid pseudo-discoidal space” just
distal of MAb in the postdiscoidal area; subdis-
coidal area divided into two cells by a cross-vein;
AA without any strong posterior branches; anal
area with two or three rows of cells; posterior wing
margin rounded; petiole short, 1.5 mm long; AA
reaching free part of CuA at sharp angle; no CuAb
(sensu Fleck et al. 2003); CuA without strong pos-
terior branches; six or seven rows of small cells be-
tween CuA and posterior wing margin; a relatively
long not zigzagged secondary vein (“postero-CuA
vein”) closely parallel to distal part of CuA, in
cubito-anal area and another one in area between
MP and CuA (“antero-CuA vein”); CuA reaching
posterior wing margin just distal to nodus level;
area between MP and CuA with one row of cells
in its basal part but greatly widened in its distal
half, with about thirteen rows of cells along poste-
rior wing margin; postdiscoidal area with two rows
of cells in its basal part, narrowed in its mid part
and slightly widened distally, with five or six rows
of cells between MA and MP near posterior wing
margin; bases of RP3/4 and IR2 between arculus
and nodus, distinctly nearer to arculus, base of
RP3/4 5.0 mm from nodus; base of IR2 apparently
on RP3/4; nodal Cr and subnodus strongly
oblique; base of RP2 aligned with subnodus;
oblique vein “O” three small cells distal of base of
RP2; numerous Bq cross-veins, but apparently no
cross-vein in the basal part of areas between RA
and RP, and between RP3/4 and IR2; 27 post-
nodal cross-veins between C and RA, not aligned
with the 20 postsubnodal cross-veins; base of IR1
11 cells distal of that of RP2; IR1 well defined, not
zigzagged and only slightly curved; one row of
small cells in the area between C and RA distal of
pterostigma; one row of cells between RP1 and
IR1; five rows of cells in the area between IR1 and
RP2, in its widest part; area between RP2 and IR2
distinctly widened distally, “antero-IR2” and “pos-
tero-IR2” veins long; a secondary longitudinal vein
closely parallel to RP2; area between IR2 and
RP3/4 distally widened; area between RP3/4 and
MA distally widened; “antero-MA” and “postero-
MA” veins long.
DISCUSSION
Because of the structures of the anal area, the
basally opened discoidal space, the strongly
curved MP + Cu, the long MAb + MP + CuA,
the sharp angle between MP + Cu and MAb and
the presence of a “tarsophlebiid pseudo-discoidal
space”, this fossil is clearly a tarsophlebiid hind
wing (Nel et al. 1993). It can be attributed to
the genus Turanophlebia rather than to Tarso-
phlebia, because of its broad cubito-anal area,
long IR1, numerous postnodal and antenodal
cross-veins.
NOTES
1) As this fossil is not a Campterophlebiidae but
a Tarsophlebiidae, the Isophlebioidea Hand-
lirsch, 1906 remains unknown in the British
Weald Clay.
2) Nel & Jarzembowski (1996) described
Proeuthemis pritykinae from the Lower Weald
Clay (Upper? Hauterivian), UK. Bechly (1997)
attributed it to the Sphenophlebiidae Bechly,
1997 (in Isophlebioptera, Parazygoptera). Bechly
(1997) also transferred the Euthemistidae
Pritykina, 1968 from the Tarsophlebioidea to the
Isophlebioptera Bechly, 1996.
Fleck G. et al.
44 GEODIVERSITAS • 2004 • 26 (1)

3) Jarzembowski (1990) described two fossil fore
wings from the British Lower Weald Clay
(Durlston Bay), he attributed to “Tarsophlebia?”
rather than to Turanophlebia, on the basis of the
presence of only one row of cells in the area be-
tween C and RA, unlike Turanophlebia martyno-
vi. As we can see above, this character alone is not
sufficient to separate the two genera. Thus, the
generic placement of these two fore wings within
the Tarsophlebiidae is uncertain.
Turanophlebia mongolica n. sp.
(Fig. 6)
HOLOTYPE. — Specimen No. 3559/69 (PIN).
ETYMOLOGY. — Named after the country of Mongolia.
GEOLOGICAL SETTING. — Lower Cretaceous, Barre-
mian-Aptian (Mostovski & Martínez-Delclòs 2000),
Bon-Tsagaan series, Bon-Tsagaan, Bayanhongor
Aimak, Central Mongolia.
DIAGNOSIS.— T. mongolica n. sp. differs from all
other Turanophlebia species as follows: 1) presence of a
complete secondary antenodal cross-vein between Ax1
and Ax2; 2) vein “O” not oblique, visible because of an
angle of RP2; and 3) postdiscoidal area not distinctly
narrowed in its mid part. It differs from T. martynovi
and T. anglicana n. sp. in its base of vein IR1 only
three cells distal of base of RP2, instead of more than
10 cells and the presence of only 15 rows of cells
between MP and CuA along posterior wing margin,
instead of more than 25 in T. martynovi and T. angli-
cana n. sp. It differs from ?T. sibirica in its cells of the
cubito-anal area distinctly transverse. It differs from
T. neckini n. comb. in the presence of two to three
rows of cells in the anal area instead of one to two rows
in T. neckini n. comb. The veins “antero-CuA” and
“postero-CuA” are less well defined than in other
Turanophlebia species.
DESCRIPTION
Imprint of a hind wing with the apical third
missing and the region of discoidal cell partly
destroyed. Wing probably hyaline; preserved part
of wing 29.5 mm long; wing probably about
41.0 mm long; 9.7 mm wide; distance from base
to arculus, 5.7 mm; from arculus to nodus,
12.0 mm; median and submedian areas free of
cross-veins; CuP strongly curved, just distal of
Ax2, basally closing subdiscoidal space; primary
antenodal braces Ax1 and Ax2 2.9 mm apart,
stronger than secondary antenodal cross-veins,
with a complete secondary antenodal cross-vein
between them; Ax1 is 2.0 mm from wing base;
arculus distinctly distal of Ax2 (0.5 mm); 11 sec-
ondary antenodal cross-veins distal from Ax2, not
aligned with the cross-veins of second rank
between ScP and RA; 12 cross-veins in the area
between RA and RP, between arculus and nodus;
a long “gap” without cross-veins between arculus
and RP3/4 in the area between RP and MA; MP
+ CuA strongly curved just before its fusion with
MAb; a sharp angle between MP + CuA and
MAb; presence of a long fusion between MAb
and MP + CuA before CuA separates from MP,
about 1.5 mm long, characteristic of the
Tarsophlebiidae; MP + CuA + MAb and basal
free part of CuA well aligned in arculus, as in
other Tarsophlebiidae; discoidal space basally
opened; presence of the two-celled “tarsophlebiid
pseudo-discoidal space” just distal of MAb in the
postdiscoidal area; subdiscoidal area divided into
two cells by a cross-vein; AA without any strong
posterior branches; anal area with three rows of
elongate transverse cells; posterior wing margin
rounded; a short petiole, about 1.5 mm long; AA
reaching free part of CuA at sharp angle; no
Revision of Tarsophlebiidae (Odonatoptera)
45
GEODIVERSITAS • 2004 • 26 (1)
A
B
FIG. 6. — Turanophlebia mongolica n. sp.; A, holotype (PIN
3559/69); B, holotype, reconstruction of hind wing. Scale bar:
A, 10 mm; B, 5 mm.

CuAb (sensu Fleck et al. 2003); CuA with no
strong posterior branches; seven rows of small
cells between CuA and posterior wing margin;
“antero-CuA” and “postero-CuA” veins present,
relatively long but weakly zigzagged; CuA reach-
ing posterior wing margin just distal of nodus
level; area between MP and CuA with one row of
cells in its basal part but rapidly greatly widened
in its distal part, with about 15 rows of cells along
posterior wing margin; postdiscoidal area with
two rows of cells in its basal half, not distinctly
narrowed in its mid part, and slightly widened
distally, with a distal secondary longitudinal
straight vein closely parallel to MA and six rows
of cells between MA and MP near posterior wing
margin; bases of RP3/4 and IR2 midway between
arculus and nodus, 6.0 mm from nodus; base of
IR2 apparently on RP3/4; nodal Cr and sub-
nodus strongly oblique; base of RP2 aligned with
subnodus; vein “O” not oblique, visible because
of an angle of RP2, two cells distal of base of
RP2; numerous Bq cross-veins, but no cross-vein
in basal part of area between RA, RP, RP3/4 and
IR2; numerous postnodal cross-veins between C
and RA (12 of them being preserved), not aligned
with the numerous postsubnodal cross-veins;
base of IR1 only three cells distal of that of RP2;
IR1 well defined, basally zigzagged but straighter
distally; area between RP2 and IR2 widened dis-
tally; area between IR2 and RP3/4 distally
widened; “antero-IR2 vein” not preserved; “pos-
tero-IR2 vein” elongate; area between RP3/4 and
MA distally widened; “antero-MA” and “postero-
MA” veins elongate.
DISCUSSION
This fossil is clearly a tarsophlebiid hind wing,
because of the structures of the anal area, the
basally opened discoidal space, the strongly
curved MP + Cu, the long common stem MAb +
MP + CuA, the sharp angle between MP + Cu
and MAb and the presence of a “tarsophlebiid
pseudo-discoidal space”. It can be attributed to
the genus Turanophlebia rather than to Tarso-
phlebia, because of its broad cubito-anal area,
long IR1, and numerous postnodal and anteno-
dal cross-veins.
Turanophlebia vitimensis n. sp.
(Fig. 7)
HOLOTYPE. — Specimen No. 2361/1, part and coun-
terpart (PIN).
ETYMOLOGY. — Named after Vitim River.
GEOLOGICAL SETTING. — Lower Cretaceous, Zaza
Formation, Neocomian or Barremian-Aptian,
Romanovka, right side of Vitim River downstream
Romanovka village, Eravna district, Buryat Republic,
Russia (Zherikhin pers. comm.).
DIAGNOSIS.— T. vitimensis n. sp. differs from all
other Turanophlebia species in: 1) presence of nine to
10 rows of cells in the cubito-anal area between CuA
and posterior wing margin; 2) CuA reaching posterior
wing margin distinctly distal of nodus level, as in T.
neckini n. comb.; 3) oblique vein “O” four cells distal
of base of RP2. It shares with T. mongolica n. sp. the
presence of two or more rows of cells in postdiscoidal
area, instead of one in other Turanophlebia species;
and 4) base of RP3/4 closer to nodus than to arculus.
DESCRIPTION
Part and counterpart of a body with the bases of
the two fore wings and a nearly complete hind
wing. Wings hyaline.
Preserved part of hind wing 20.0 mm long; hind
wing probably about 47.0 mm long; 10.1 mm
wide; distance from base to arculus, about
7.0 mm; from arculus to nodus, 15.0 mm;
median and submedian areas free of cross-veins;
CuP not preserved; only primary antenodal brace
Ax2 preserved, stronger than secondary anteno-
dal cross-veins; Ax2 about 6.0 mm from wing
base; arculus distinctly distal of Ax2 (0.6 mm);
14 secondary antenodal cross-veins distal of Ax2,
not aligned with the cross-veins of second rank
between ScP and RA; more than 10 cross-veins in
the area between RA and RP, between arculus
and nodus; a long “gap” without cross-veins be-
tween arculus and RP3/4 in the area between RP
and MA; MP + CuA not preserved basal of its
fusion with MAb; long fusion between MAb and
MP + CuA before the separation between CuA
and MP, 2.5 mm long, characteristic of the
Tarsophlebiidae; MA, MAb, MP + CuA + MAb
and basal free part of CuA well aligned in arculus,
as in other Tarsophlebiidae; discoidal space pro-
Fleck G. et al.
46 GEODIVERSITAS • 2004 • 26 (1)

bably basally opened; presence of the two-celled
“tarsophlebiid pseudo-discoidal space” just distal
of MAb in the postdiscoidal area; subdiscoidal
area divided into two cells by a cross-vein; AA
without any strong posterior branches; anal area
with three rows of cells; posterior wing margin
rounded; AA reaching free part of CuA at sharp
angle; no CuAb; CuA with no strong posterior
branches; nine or 10 rows of small cells between
CuA and posterior wing margin; “antero-CuA”
and “postero-CuA” veins elongate; CuA reaching
posterior wing margin distinctly distal of nodus
level; area between MP and CuA with one row of
cells in its basal part but rapidly greatly widened
in its distal part; postdiscoidal area with three
rows of cells in its basal part, narrowed with two
rows in its mid part, and slightly widened distal-
ly, with a distal secondary longitudinal straight
Revision of Tarsophlebiidae (Odonatoptera)
47
GEODIVERSITAS • 2004 • 26 (1)
A
BC
D
E
F
FIG. 7. — Turanophlebia vitimensis n. sp., holotype (PIN 2361/1); A, imprint; B, right fore wing; C, left fore wing; D, left hind wing;
E,F, apex of the abdomen; E, counterimprint; F, imprint. Scale bars: A, 10 mm; B, D-F, 5 mm; C, 3 mm.

vein closely parallel to MA and five or six rows of
cells between MA and MP near posterior wing
margin; bases of RP3/4 and IR2 between arculus
and nodus, nearer to nodus, 5.5 mm from nodus;
nodal Cr and subnodus oblique; base of RP2 a-
ligned with subnodus; vein “O” oblique, four
cells distal of base of RP2; numerous Bq cross-
veins; numerous postnodal cross-veins between C
and RA (15 of them being preserved), not aligned
with the numerous postsubnodal cross-veins;
base of IR1 only six to eight cells distal of that of
RP2; IR1 well defined, basally zigzagged but
straighter distally; area between RP2 and IR2
widened distally; area between IR2 and RP3/4
distally widened; “postero-IR2” vein elongate;
area between RP3/4 and MA distally widened;
“antero-MA” and “postero-MA” veins elongate.
Preserved part of fore wing 17.0 mm long;
6.7 mm wide (in its preserved part); distance from
base to arculus, 7.0 mm; median and submedian
areas free of cross-veins; CuP distinctly curved;
Ax1 and Ax2 preserved, only slightly stronger
than the secondary antenodal cross-veins; two
secondary antenodals basal of Ax1 and one be-
tween Ax1 and Ax2; arculus between Ax1 and
Ax2; numerous secondary antenodal cross-veins
distal from Ax2, not aligned with the cross-veins
of second rank between ScP and RA; numerous
cross-veins in the area between RA and RP, be-
tween arculus and nodus; a long “gap” without
cross-veins between arculus and RP3/4 in the
area between RP and MA; MP + CuA strongly
curved just before its fusion with MAb; presence
of a very short fusion between MAb and MP +
CuA before CuA separates from MP, better pre-
served on the right fore wing; MA, MAb, MP +
CuA + MAb and basal free part of CuA well a-
ligned in arculus, as in other Tarsophlebiidae;
discoidal space basally opened; subdiscoidal area
divided into two cells by a cross-vein; AA without
any strong posterior branches; anal area with two
rows of cells; AA reaching free part of CuA at
very sharp angle; no CuAb; CuA with no strong
posterior branches; more than seven rows of
small cells in cubito-anal area; area between MP
and CuA with one row of cells in its basal part
but rapidly widened in its distal part; postdiscoid-
al area with three rows of cells in its basal half,
narrowed with two rows in its mid part.
Thorax thin and elongate. Metathoracic leg well
preserved, very long, as long as that of Tarso-
phlebia eximia, femora distinctly shorter than
tibia, tibia 12.5 mm long, tarsus 8.5 mm long;
tarsus very long and slender, with three brakes,
probably corresponding to three visible tarso-
meres.
Abdomen long and slender, about 45.0 mm long
and 3.7 mm wide, the widest part being at the
level of segment eight; presence of two lateral
carinae on the posterior part of the abdominal
segments. One pair of well separated anal appen-
dages visible, of very particular shape, basally
strongly sclerotized, with high humps and a non-
sclerotized paddle-like distal part. No median
anal appendage visible. The presence of only a
pair of anal appendages shows that this specimen
is a male, as the female Tarsophlebiidae have a
very long ovipositor (Nel et al. 1993).
DISCUSSION
Because of the structures of the anal area, the
basally opened discoidal space, the strongly curved
MP + Cu, the long common stem MAb + MP +
CuA, the sharp angle between MP + Cu and MAb
and the presence of a “tarsophlebiid pseudo-
discoidal space”, this fossil is clearly a tarsophle-
biid hind wing. It can be attributed to the genus
Turanophlebia rather than to Tarsophlebia, because
of its broad cubito-anal area, long IR1, and numer-
ous postnodal and antenodal cross-veins.
PHYLOGENETIC POSITION
OF THE TARSOPHLEBIIDAE
Nel et al. (1993) proposed that the Tarsophlebii-
dae are the most inclusive clade of the “Anisozy-
goptera” + Anisoptera. The alleged putative syn-
apomorphies mentioned by Nel et al. (1993), e.g.,
the less separated and relatively large eyes, the pre-
sence of two cephalic sutures, and the small leg
spines, are at least characters of uncertain primary
polarity because they are unknown in the Proto-
zygoptera, sister group of the Panodonata.
Fleck G. et al.
48 GEODIVERSITAS • 2004 • 26 (1)

Bechly (1996) excluded the Tarsophlebiidae
from the Epiproctophora (= “Anisozygoptera” +
Anisoptera) and considered them as the sister
group of the Odonata (= Zygoptera +
Epiproctophora) within the Panodonata (=
Tarsophlebiidae + Odonata), the sister group of
the Panodonata being the Protozygoptera, and
the sister group of the (Panodonata +
Protozygoptera) being the Triadophlebiomorpha
(see Fig. 8). More recently, Rehn (2003) prop-
osed a new analysis that supports Bechly’s hypo-
thesis on the position of the Tarsophlebiidae.
DISCUSSION ON THE CHARACTERS
We propose to re-examine some of the characters
that would be available to clarify the problem of
the phylogenetic position of the Tarsophlebiidae.
(A) Number of adult tarsomeres
Bechly (1999) indicated that the Protozygoptera
(sister group of the Panodonata) have four tarso-
meres. In all the described Protozygoptera, the
body is not preserved or too poorly preserved to
observe the number of tarsomeres. Thus, we
consider that it is still unknown in this group.
After our observations, Tarsophlebia eximia seems
to have four tarsomeres but Turanophlebia viti-
mensis n. sp. seems to have only three tarsomeres
(see Figs 1B and 7A). Thus, some uncertainty
remains concerning the exact number of tarso-
meres in Tarsophlebiidae. The body structures of
fossil insects from Solnhofen Lithographic
Limestone are frequently poorly preserved, and
difficult to interpret.
Under the hypothesis that the Tarsophlebiidae
had four tarsomeres, this would correspond to a
symplesiomorphy of the family with the more
inclusive groups of Odonatoptera even though
this character is unknown in the sister group of
Panodonata and all more inclusive groups be-
tween Protozygoptera Tillyard, 1925 and Mega-
neuridae Handlirsch, 1906. It is under a different
state in the Meganisoptera, Meganeuridae, which
have five tarsomeres (clearly visible in specimen
MNHN-LP-R.52938 of Meganeura monyi
Brongniart, 1884; Nel pers. obs.). Thus, three
evolutionary scenarios are possible: 1) a progressi-
ve reduction of the number of tarsomeres from
five in Meganeuridae to four in Protozygoptera
and Tarsophlebiidae and then to three in
Odonata. In this case, this character state would
be a striking symplesiomorphy of the Tarso-
phlebiidae; 2) a reduction from five tarsomeres in
Protozygoptera and more inclusive lineages to
four in the stemgroup of Panodonata, main-
tained in the Tarsophlebiidae lineage, and a
parallel reduction from four to three in
Zygoptera and Epiproctophora. This hypothesis
is supported by the present phylogenetic analysis
(see below); and 3) this theoretical possibility
would be a reversal from three tarsomeres to four
in Tarsophlebiidae, but it is not only less parsi-
monious, but also seems to be an evolutionary
impossibility because no such example is known
from any other insect. The problem will be defi-
nitely solved when the tarsal structures of the
Protozygoptera will be known.
Under the hypothesis that the Tarsophlebiidae
had three tarsomeres, Bechly (1996) suggested
Revision of Tarsophlebiidae (Odonatoptera)
49
GEODIVERSITAS • 2004 • 26 (1)
Geroptera
Eomeganisoptera
Meganisoptera
Protanisoptera
Triadophlebiomorpha
Protozygoptera
Tarsophlebiidae
Zygoptera
Epiproctophora
FIG. 8. — P hylogeny of the major groups of Odonatoptera,
hypothesis of Bechly (1996).

that the presence of a very long basal tarsomere in
all Tarsophlebiidae could correspond to the
fusion of two basal protozygopterid tarsomeres.
Thus, it could correspond to an intermediate
situation towards the three-segmented legs with a
short basal tarsomere of the Odonata. There is no
evidence of such a fusion in available fossils. In
Meganeura monyi and in all Panodonata, except
Tarsophlebiidae, the first tarsomere is the short-
est. There is still no way to determine if some of
the meganeurid’s five tarsomeres have fused or
were atrophied during the evolution.
Note. Under both the hypotheses of four- or
three-segmented tarsi, all tarsomeres are very long
in Tarsophlebiidae. This is a unique character in
known Odonatoptera. It could represent a syn-
apomorphy of this family.
(B) Tarsal claws
Bechly (1996) indicated that the Odonata have
“tarsal claws with a ventro-apical claw-hook”
(autapomorphy) and that this structure is “appa-
rently absent” in Tarsophlebiidae. The observa-
tions we have made support this hypothesis.
Nevertheless, such structures could be hard to
observe in some fossils. As example, the species of
the modern libellulid genus Oligoclada have very
reduced ventro-apical claw-hook that could be
hardly visible in fossils. The basal odonatopterid
family Meganeuridae have simple tarsal claws,
thus this structure could correspond to a plesio-
morphy. But it is still unknown in Protozy-
goptera. Thus, it is not possible to conclude with
certainty in which clade it appeared. Never-
theless, it suggests a position inclusive of the
Odonata for the Tarsophlebiidae.
(C) Male secondary genital apparatus of second and
third abdominal segment
Recent and fossil Panodonata have a complex
secondary genital apparatus of second abdominal
segment that has no homology in other insects.
Such a structure can be observed in some
Protozygoptera with the body preserved (pers.
obs.), but it is not possible to interpret and com-
pare correctly the secondary genital apparatus of
the Protozygoptera to those of the recent
Zygoptera, Epiophlebiidae and Anisoptera. The
secondary genital apparati of these three lineages
strongly differ but these differences concern
structures that are nearly impossible to interpret
in fossil specimens, even if very well preserved
and closely related to well known recent taxa.
Bechly (1996) indicated that the Odonata have
their “male vesicula spermalis more advanced and
anteriorly elongated” (synapomorphy). After the
present observations, this structure is indeed dis-
tinctly shorter and not anteriorly elongated in
Tarsophlebiidae (Fig. 1A). Nevertheless, this
structure remains unknown in Protozygoptera
and all the more inclusive lineages within
Odonatoptera. Thus, it is not possible to primari-
ly polarize this character by outgroup compari-
son. However, from the viewpoint of
evolutionary biology the vesicula originated as
sternal outgrowth and therefore a shorter vesicula
would have to be considered as plesiomorphic.
The secondary male genital apparatus of
Tarsophlebiidae turns out to be of high signifi-
cance for the understanding of the evolution of
odonate copulation. In the groundplan of
Odonatoptera (e.g., Namurotypus sippeli
Brauckmann & Zessin, 1989) there was no
secondary genital apparatus but primitive prima-
ry genitalia that were most similar to those of
wingless insects who still deposit external sperma-
tophores (Bechly et al. 2001). Such a deposition
of spermatophores, first on the substrate and later
on the basal male abdominal venter, is also the
only possibility for the evolution of the complex
secondary apparatus and mating wheel in
Odonata (Bechly et al. 2001). In extant Odonata
the secondary genital apparatus is without doubt
homologous and is always composed by the
following substructures: a pair of hamuli ante-
riores, a pair of hamuli posteriores, an unpaired
median sternal process of segment II (ligula), and
a pouchlike and unpaired median sternal out-
growth of segment III (vesicula spermalis). All
these structures are present in all three major sub-
groups of Recent Odonata, viz. Zygoptera,
Epiophlebiidae and Anisoptera. It is a most
curious phenomenon that in each of these three
groups a different structure of this apparatus is
Fleck G. et al.
50 GEODIVERSITAS • 2004 • 26 (1)

developed as copulatory organ for sperm transfer
and removal of foreign sperm (ligula in
Zygoptera, hamuli posteriores in Epiophlebiidae,
and vesicula spermalis in Anisoptera). Based
on the well established phylogenetic hypothesis
that Zygoptera form the sister group of Epi-
proctophora, which comprises the Epio-
phlebiidae and Anisoptera, a hypothetical
groundplan of the secondary genital apparatus
can be reconstructed:
1) Hamuli anteriores: they are platelike in
Zygoptera and Anisoptera, but hooklike in
Epiophlebiidae. The most parsimonious interpre-
tation is that the state in Epiophlebiidae is derived,
so that platelike hamuli anteriores belong to the
common groundplan.
2) Hamuli posteriores: they are platelike and
small in Zygoptera and Anisoptera, and greatly
enlarged and modified as copulation organs in
Epiophlebiidae. The most parsimonious interpre-
tation again is that the state in Epiophlebiidae is
derived, so that small platelike hamuli posteriores
belong to the common groundplan.
3) Ligula: the ligula is simple and small in
Epiophlebiidae and Anisoptera, but greatly en-
larged, subsegmented, and modified as copulation
organ in Zygoptera. It can be regarded as very pro-
bable that the ligula is derived from the processus
caudalis of this segment, a posteromedian sternal
process that is still present on other abdominal seg-
ments of Epiophlebiidae and Anisoptera (probably
reduced in Zygoptera). This process is simple,
small and unmodified, so that the state in
Zygoptera would have to be considered as derived,
by using these serial homologous structures as kind
of “outgroup”. Consequently, a small and simple
ligula seems to belong to the common ground-
plan. Nevertheless, this structure remains un-
known in more inclusive groups of Odonatoptera,
viz. Protozygoptera, Triadophlebiomorpha,
Protanisoptera, etc. Thus, it is not possible to pro-
pose a primary polarisation of this character on the
basis of the outgroup comparison.
4) Vesicula spermalis: it is relatively short and
unsegmented in Zygoptera, somewhat elongated
but unsegmented in Epiophlebiidae, and greatly
elongated, subsegmented and modified as copu-
lation organ in Anisoptera. The most parsimo-
nious interpretation is that the state in Zygoptera
is plesiomorphic. Therefore, a relatively short,
unsegmented and unmodified vesicula would
belong to the common groundplan.
To sum up, the secondary genital apparatus in
the hypothetical reconstruction of the odonate
groundplan includes small and platelike hamuli
anteriores and posteriores, a small and simple
ligula, and a short and unmodified vesicula sper-
malis. However, this groundplan reconstruction
implies a surprising problem: there is no structure
left that would be suited to function as a copula-
tion organ for sperm transfer and sperm removal.
If this reconstruction would be correct, one
would be forced to make two assumptions: 1) the
stemspecies of Odonata did not transfer liquid
sperm and did not remove foreign sperm, but did
only partly conceal a spermatophore in the vesi-
cula that was fetched by the female from this
place; and 2) the transfer of liquid sperm and the
mechanism of sperm removal developed three
times as parallelism in Zygoptera, Epio-
phlebiidae, and Anisoptera. The latter hypothesis
would explain why very different and non-homo-
logous structures were developed as functional
penis in each of these groups. The assumption
that any of these highly specialised organs repre-
sents the primitive condition would hardly be
explainable from the viewpoint of evolutionary
biology, because it would imply the reduction of
a perfect organ for sperm transfer and removal,
only to be subsequently replaced by another
organ that fullfills the same function. This clearly
would not make sense and can be considered as
unlikely.
What does secondary genital apparatus of
Tarsophlebiidae contribute to this discussion: the
secondary genital apparatus of Tarsophlebiidae
exactly agrees with the hypothetical groundplan
reconstruction, and therefore strongly confirms
the latter! Vice versa, the primitive secondary
genital apparatus of Tarsophlebiidae could be
congruent with the assumed inclusive position of
this fossil group relative to all extant Odonata,
but would also be compatible with a position at
the very base of Epiproctophora. However, the
Revision of Tarsophlebiidae (Odonatoptera)
51
GEODIVERSITAS • 2004 • 26 (1)

fact that the vesicula spermalis is even shorter
than in Zygoptera rather supports the former
alternative. It is obvious that Tarsophlebiidae
were unable to transfer liquid sperm or to remove
foreign sperm during an internal copulation.
Therefore, we here propose that Tarsophlebiidae
did still transfer spermatophores that were at-
tached within the vesicula spermalis and fetched
by the female in wheel position. The other struc-
tures of the secondary genital apparatus most
probably served as holding and guiding devices
for the female ovipositor.
(D) Male auricles
Nel et al. (1993) misinterpreted the hamuli pos-
teriors as possible epiproctophorid male auricles,
as noted by Bechly (1996).
(E) Male genital appendages
In males of both Tarsophlebia eximia and
Turanophlebia vitimensis n. sp., no median anal
structure is visible thus the epiproctal process is
certainly absent in Tarsophlebiidae, which could
exclude them from Epiproctophora sensu Bechly
(1996) if this structure would be primitively
absent and not reduced. Rehn (2003: 193) consi-
dered that the “complete or nearly complete
absence of the epiproct” is a “unique zygopteran
apomorphy, that seemingly provides strong sup-
port for the monophyly of the suborder” but as
this character is unknown in more inclusive
groups (Protozygoptera, Triadophlebiomorpha,
Protanisoptera, etc.), it is not possible to propose
a correct primary polarization of this character.
After the present phylogenetic analysis (see
below), the absence of the epiproct would appear
plesiomorphic, contra Rehn (2003).
In both Zygoptera and Epiproctophora, the cerci
are always well developed. In Tarsophlebia eximia,
the structures that were interpreted as two pairs
of appendages by Nel et al. (1993) or Bechly
(1996) correspond to the strongly sclerotized
parts of the two appendages of Turanophlebia
vitimensis n. sp. These two appendages strongly
resemble the cerci of a Campterophlebiidae
(Epiproctophora, Isophlebioidea) from the
Mesozoic of China (Fleck & Nel 2002). Some
Anisoptera (Petaluridae) also have basally sclero-
tized and spatulate cerci but no Odonata have
spatulate paraprocts. Thus, these tarsophlebiid
appendages are probably the cerci. Three inter-
pretations of tarsophlebiid male genital appen-
dages remain possible: 1) the Tarsophlebiidae
have no developed paraprocts and no epiproct.
Thus they would have “lost” the paraprocts
(which are a groundplan feature of insects) but
not yet developed the epiproctal process of
Epiproctophora. The very strongly sclerotized
cerci would fullfill the function of both para-
procts and epiproct, for grasping the female
during mating; 2) the paraprocts are not visible
on these fossils, hidden below the cerci.
Nevertheless, nothing in these fossils supports
this hypothesis; and 3) the tarsophlebiid appen-
dages are uniquely shaped paraprocts and the
cerci are reduced. The loss of the cerci would
then be a unique autapomorphy of the
Tarsophlebiidae.
All three hypotheses concerning the paraprocts
would be compatible with any of the discussed
phylogenetic positions for Tarsophlebiidae.
(F) Hind wing discoidal cell
Bechly (1996) proposed to consider the basal clo-
sure of the hind wing discoidal cell as a strict
synapomorphy of the Odonata, not shared by the
Tarsophlebiidae. Rehn (2003) also considered
that the absence of the posterior part of arculus in
Tarsophlebiidae is plesiomorphic. This character
would support a very inclusive position for this
family, as sister group of Odonata. After the pre-
sent phylogenetic analysis (see below), the sister
group relationship with the Epiproctophora
implies the convergent closure of the discoidal
cell in Zygoptera and Epiproctophora, or the loss
of the closure of the discoidal cell in the
Tarsophlebiidae, which are of course less parsi-
monious solutions than the former hypothesis.
Nevertheless, the Tarsophlebiidae have developed
a closed “pseudo-discoidal cell” that has the func-
tion of the close discoidal cell of the Odonata.
The basal opening of the discoidal cell could be
“correlated” with the development of this highly
specialized structure.
Fleck G. et al.
52 GEODIVERSITAS • 2004 • 26 (1)

(G) Vein CuAb
Fleck et al. (2003) reinterpreted the structure of
the fusion of vein AA with vein CuA. In
Protozygoptera, Zygoptera and Tarsophlebiidae,
there is no postero-proximal branch CuAb of
CuA (plesiomorphy). Such a branch is present in
all Epiproctophora (the Steleopteridae being
excluded, Fleck et al. 2001). This vein CuAb is
fused with the distal part of AA in Epio-
phlebiidae, Heterophlebiomorpha and An-
isoptera. These veins are more or less separated in
Isophlebioptera. Thus, this character is congruent
with a very inclusive position for the Tarso-
phlebiidae, either as sister group of Odonata or as
sister group of Epiproctophora.
(H) Nodal structures
Bechly (1996) also characterized the Zygoptera
by the following three potential synapomorphies
in the wing venation, i.e. characters (1) “reduc-
tion of the terminal kink of the CP”, (2) “reduc-
tion of the nodal furrow”, and (3) “obliteration
of the tubular sclerotized canal of ScP along the
venter of the postnodal costal margin”. The
Tarsophlebiidae and Epiproctophora have the
contrary character states (see Fig. 5C and Nel
et al. 1993). Bechly (1996) polarized these cha-
racters after the assumption that Tarsophlebiidae
are the sister group of Odonata. The more inclu-
sive clades Protanisoptera, Triadophlebiomorpha
and Protozygoptera have the same states for cha-
racters (1) and (2) as Zygoptera (Nel et al. 2001;
Huguet et al. 2002). The principle of parsimony
suggests that the polarities of these characters are
contrary to that proposed by Bechly (1996) and
that the character states are plesiomorphic in
Zygoptera. As he apparently neglected the works
of Nel et al. (1993) or Huguet et al. (2002), Rehn
(2003) ignored the state of these important cha-
racters (coded “?”) for the most inclusive clades of
Odonatoptera (Geroptera, Meganisoptera, Protan-
isoptera, Protozygoptera, even Tarsophlebiidae)
in his analysis. After the present analysis, these
characters support the clade (Tarsophlebiidae +
(Sieblosiidae + Epiproctophora)) (see below).
Concerning the third character, the situation is
different because the distal wing margin is a
fusion of the three longitudinal veins CA, ScA
and ScP (Bechly 1996). Consequently, the pres-
ence of the tubular sclerotized canal of ScP is pro-
bably a plesiomorphy. The Tarsophlebiidae and
the Epiproctophora would share only one putati-
ve synapomorphy, i.e. “presence of a strong kink
of CP aligned with nodal Cr”, “correlated” with
“a strong nodal furrow”. While the state “ScP
ventrally visible along costal margin” must repre-
sent a symplesiomorphy. The Sieblosiidae
Handlirsch, 1906 also have the same structures
proper to Epiproctophora (contra Nel &
Paicheler 1994).
Note: in Tarsophlebiidae, the nodus is distally
shifted (between 44 and 47% of wing length), as
in Epiproctophora. In Sieblosiidae, it is in basal
position as in Protozygoptera and Zygoptera
(probable plesiomorphic state).
(I) Vein CuP
In the most inclusive lineages of the Epiprocto-
phora, i.e. Epiophlebiidae, Isophlebioptera, He-
terophlebioptera, the Anisoptera, Liassogomphi-
dae Tillyard, 1935, some Gomphidae Rambur,
1842 and Aeshnidae Leach, 1815 (but less pro-
nounced), the vein CuP is strongly curved and
seems to begin on AA rather than on MP + Cu.
The same character is present in Tarsophlebiidae.
In Zygoptera (except the Cenozoic family Sieblo-
siidae Handlirsch, 1906), Protozygoptera and
more inclusive groups of Odonatoptera, the vein
CuP is straight, which is thus probably the plesio-
morphic character state. Their curved CuP sug-
gests close relationships between the Tarsophle-
biidae, the Sieblosiidae, and the Epiproctophora.
Within the Sieblosiidae, the CuP of the Steno-
lestes species, Parastenolestes Nel & Paicheler,
1994 and Paraoligolestes Nel & Escuillié, 1993
are of epiproctophorid type. The exact structure
of CuP is still unkown in Oligolestes Schmidt,
1958 because of some preservation problems
(Schmidt 1958; Nel & Escuillié 1993; Nel &
Paicheler 1994). Bechly (1996) included the Sie-
blosiidae within the Zygoptera in a very inclusive
position, close to the Eucaloptera Bechly, 1996.
As we have seen above, the nodal structures of the
Sieblosiidae are of epiproctophorid-type and not
Revision of Tarsophlebiidae (Odonatoptera)
53
GEODIVERSITAS • 2004 • 26 (1)

of zygopterid-type. Bechly (1996) proposed the
following venational synapomorphies of the Zy-
goptera, which could be potentially visible on the
fossil Sieblosiidae:
1) “both wing pair distinctly stalked with a petio-
lus that is at least somewhat longer than broad”.
In the Sieblosiidae, the petioles are short, shorter
than those of some Epiproctophora, Steno-
phlebiidae (Fleck et al. 2003). Furthermore, in
the Calopterygidae, the petiole can be extremely
short, but it is long in the “basal” calopterygid
Caliphaea Hagen, 1859 and the calopterygid sis-
ter group Dicteriadidae. Also, the protozygopte-
rid petioles are very long. This wing petiolation
has been convergently acquired at least six times
by very different groups of Odonatoptera. Thus,
it is highly homoplastic;
2) “both wing pair of identical shape and vena-
tion”. The Triadophlebiomorpha and Proto-
zygoptera (sister group of Panodonata) have this
character state as convergence. Thus, this charac-
ter is homoplastic as well;
3) “on the ventral wing surface the posterior part
of the basal brace (Ax0) is obliterated or covered
by a rather extensive sclerotization of the wing
base”. This character is unknown in Triado-
phlebiomorpha and Protozygoptera, thus its pri-
mary polarization is uncertain. Furthermore, this
character occurs also in Anisoptera, Aeshnidae
(Aeshna cyanea (Müller, 1764), Cordulegaster bol-
toni (Donovan, 1807), etc.), in which this struc-
ture is identical to what occurs in the Zygoptera
Mecistogaster linearis Olivier, 1792 and Pseudo-
stigma Selys, 1860. Lastly, in Sieblosiidae, this
sclerotization is absent and Ax0 is in a rather dis-
tal position. Thus it is probably homoplastic;
4) “reduction of the long spines of the dorsal sur-
face of the RP and MP”. This character cannot
be observed on the available material of Sieblosii-
dae. Furthermore, in the Epiproctophora, Hete-
rophlebiidae Needham, 1903, these spines are re-
duced: on very well preserved specimens from the
Liassic of Bascharage (Grand-Duché-du-Luxem-
bourg), the spines of RA and of the net of trans-
verse veins are clearly visible but none is visible
on RP and MP. In the Anisoptera, Liassogom-
phidae from the same outcrop, the spines of RP
and MP are clearly visible. Thus, these spines are
absent on the RP and MP of the Heterophle-
biidae. It is possibly “correlated” with the smaller
diameters of these veins in these taxa. Also, in the
recent Epiproctophora Epiophlebia superstes (Selys,
1889), there are very few and very small spines on
RP. Thus, this character cannot be used to attri-
bute the Sieblosiidae to the Zygoptera, it is ho-
moplastic and of uncertain primary polarity, as it
is unknown in Protozygoptera and more inclu-
sive Odonatoptera;
5) “significant increase of spine-density at the
apical costal margin”. This character is unknown
in the Triadophlebiomorpha and Protozygoptera,
thus, its exact polarity remains unknown. Fur-
thermore, the Sieblosiidae have no such increase
(clearly visible on the very well preserved speci-
men MNHN-LP-R.10375 of Stenolestes coulleti
Nel & Papazian, 1986) (Nel & Escuillié 1992).
The Tarsophlebiidae have no such increase either
(visible on type specimen of Turanophlebia angli-
cana n. sp.), which could confirm the polarity as-
sumed by Bechly (1996) if Tarsophlebiidae in-
deed represent the sister group of all crowngroup
Odonata. Other autapomorphies of the Zygo-
ptera proposed by Bechly (1996) are not visible
in the known specimens of Sieblosiidae.
In conclusion, the Sieblosiidae do not share any
known unambiguous apomorphy with the
Zygoptera. Furthermore, the sieblosiid nodal and
CuP structures suggest closer affinities with the
Epiproctophora than with the Zygoptera.
In addition to these potential synapomorphies,
the Sieblosiidae have the arculus shifted basally in
a position between Ax1 and Ax2. Nevertheless, it
is very close to Ax2. In Zygoptera, the arculus is
aligned with Ax2 or very slightly basally recessed.
In Protozygoptera, the arculus is in a rather
variable position but generally aligned with Ax2
or in a distal position. Furthermore, this poten-
tial synapomorphy of the Sieblosiidae with the
Epiproctophora is in fact homoplastic, contra
Rehn (2003: 201): 1) Fleck et al. (2003) de-
scribed a Epiproctophora, Stenophlebiidae having
the arculus aligned with Ax2, as in Zygoptera;
2) also, the Calopterygoidea, Dicteriadidae
(Heliocharis amazona Selys, 1853) has its arculus
Fleck G. et al.
54 GEODIVERSITAS • 2004 • 26 (1)

closer to Ax1 than to Ax2; and 3) inversely, the
arculus of the Anisoptera Synthemis montaguei
Campion, 1921 is shifted well distally from Ax2.
Rehn (2003) proposed the following two main
synapomorphies for the Zygoptera (but he ig-
nored the group Sieblosiidae): “head transversely
elongate”; “eyes separated by more than their
own width” Although deformed, the head of the
Sieblosiidae are not very transversely elongate and
their eyes are not so separated. Other characters
he proposed are subject to numerous homoplasies
or are unknown in the Panodonata sister group.
The Sieblosiidae do not have the male anal angle
of the Epiproctophora. They do not have any
visible epiproctophorid vein CuAb either but
there is no visible furrow at the point of fusion
between AA and CuA, thus it is not possible to
determine if they have a CuAb completely fused
with AA as in modern Anisoptera or fossil
Stenophlebiidae, or a CuAb absent, as in
Tarsophlebiidae (Fleck et al. 2003). Thus, the
Sieblosiidae are probably in a very inclusive posi-
tion within the Epiproctophora.
(J) Other characters
1) In some Tarsophlebiidae, the arculus is between
Ax1 and Ax2, in others, it is opposite to Ax2 or
distal of Ax2. Thus, this character is nearly useless.
2) The eyes of Tarsophlebia eximia are more simi-
lar to those of the modern Anisoptera and
Epiophlebiidae and Permo-Triassic Proto-
zygoptera than to those of the Zygoptera, not
strongly separated of more than their diameter.
However, this could well be a symplesiomorphy,
because the widely separated eyes of Zygoptera
could be “correlated” with the apomorphic ham-
mer-shaped head of Zygoptera and therefore pro-
bably are derived rather than primitive.
3) The ocelli of Tarsophlebia eximia are more of
“zygopterid” type rather than of “anisopterid”
type, in equilateral triangle, with the lateral ocelli
posteriorly rejected. Unfortunately, the polarities
of these characters remain uncertain because the
head structures are still unknown in Protaniso-
ptera, Triadophlebiomorpha, and Protozygoptera.
4) Thoracic skewness. The skewness was proba-
bly more important in Tarsophlebiidae than in
recent Zygoptera, Lestidae. The presence of a
strong thoracic skewness in some Permian
Protozygoptera (among others, in the holotype of
Permolestes gracilis Martynov, 1932; Nel pers.
obs.) suggests that this similarity with Zygoptera
is not a synapomorphy but a symplesiomorphy,
or at least be due to convergence. Note that Rehn
(2003) erroneously indicated that the important
thoracic skewness is a unique synapomorphy of
the Zygoptera, but he ignored the structure of
the thorax of the Protozygoptera and the work of
Nel et al. (1993) who first established the impor-
tant skewness of the Tarsophlebiidae.
5) “Subdiscoidal cross-vein” sensu Rehn (2003).
Rehn (2003) amended the wing venation
nomenclature of Riek & Kukalová-Peck (1984)
as follows: he supposed that the transverse vein
named “CuP” by Riek & Kukalová-Peck (1984)
is the vein “CuA + CuP” and he named “subdis-
coidal cross-vein” the transverse vein called
“CuA” by Riek & Kukalová-Peck (1984). Rehn
Revision of Tarsophlebiidae (Odonatoptera)
55
GEODIVERSITAS • 2004 • 26 (1)
1234567891011121314151617
Tarsophlebia 01100000000001000
Epiophlebia 11101111110001011
Calopteryx 11110000001110100
Permolestes ????????0011?00?0
Bellabrunetia 11100011110001011
Lestes 11110000001110100
Stenolestes 00????0?00????0??
TABLE 1. — Matrix of taxa/characters used in the phylogenetic analyses (see Appendix 1).

(2003) apparently ignored the work of Nel et al.
(1993) who confirmed the hypothesis of Riek &
Kukalová-Peck (1984), after the study of the
situation in Isophlebia aspasia Hagen, 1866
(Epiproctophora), in which there is no distal
fusion between the vein AA + CuP and CuA
(both sensu Riek & Kukalová-Peck 1984).
Furthermore, the presence of the alleged “subdis-
coidal cross-vein” is supposed to be one the two
synapomorphies supporting the clade “Zygoptera
+ Epiprocta” sensu Rehn (2003) (= Zygoptera +
Epiproctophora), but as this vein is in fact CuA,
it is present in the more inclusive groups
Protozygoptera and Triadophlebiomorpha. The
character “CuA separating from MP at the point
of fusion of MAb (“discoidal vein” sensu Rehn
2003) and aligned with it” is shared by the Tarso-
phlebiidae and Zygoptera + Epiproctophora, as
already indicated in Nel et al. (1999). Thus, it
cannot constitute a synapomorphy of the
(Zygoptera + Epiproctophora).
6) Oblique vein “O” present. Rehn (2003: 201)
considered that the presence of the oblique vein
is a synapomorphy of the Epiproctophora (= for-
mer “Anisozygoptera” + Anisoptera, without the
Tarsophlebiidae), with a convergency with the
Zygoptera, Lestinoidea. Nel et al. (1993) and
Fleck et al. (2003) demonstrated that this vein is
absent in some Epiproctophora, Stenophlebiidae.
Thus it is subject to more homoplasies than cur-
rently suspected. It is also present in the
Tarsophlebiidae, as indicated in Nel et al. (1993)
and the present paper, contra Rehn (2003).
PHYLOGENETIC ANALYSIS
We made a computer phylogenetic analysis of the
relationships of the Tarsophlebiidae, based on the
set of characters discussed above. It is based on
17 characters, equally weighted and unordered (see
Table 1 and Appendix 1). The taxa we considered
are as follows: Tarsophlebia (Tarsophlebiidae),
Permolestes Martynov, 1932 (Protozygoptera),
Epiophlebia Calvert, 1903 (Epiproctophora,
Epiophlebiidae), Bellabrunetia Fleck & Nel, 2002
(Epiproctophora, Isophlebiomorpha, i.e. anisop-
terid lineage), Stenolestes Scudder, 1895 (Sieblo-
siidae), Lestes Leach, 1815 (Zygoptera, Lestidae),
and Calopteryx Leach, 1815 (Zygoptera,
Calopterygidae). The analysis was made using the
computer software PAUP4.0b10 (Swofford
1998), Branch and Bound option. Permolestes was
chosen as outgroup. The analysis gave one more
parsimonious tree T1: (Permolestes & (Lestes +
Calopteryx) & [Tarsophlebia + (Stenolestes +
(Epiophlebia + Bellabrunetia))]), with a length of
the minimal trees of 16 steps, Consistency Index
CI: 0.9375, CI excluding uninformative charac-
ters: 0.9231, Retention Index RI: 0.9375, and RC
(RC = CI ×RI): 0.8789 (see Fig. 9). Both clades
Zygoptera and [Tarsophlebia + (Stenolestes +
(Epiophlebia + Bellabrunetia))] are present. The
clade [Tarsophlebia + (Stenolestes + (Epiophlebia +
Bellabrunetia))], corresponding to the (Tarso-
phlebiidae + Epiproctophora), is supported by the
apomorphic character states “11 (0) terminal kind
of CP in nodus not reduced”, “12 (0) nodal fur-
Fleck G. et al.
56 GEODIVERSITAS • 2004 • 26 (1)
Permolestes
Calopteryx
Lestes
Stenolestes
Epiophlebia
Bellabrunetia
Tarsophlebia
FIG . 9. — Most parsimonious tree T1 (obtained with
PAUP4.0b10, Branch and Bound option), Consistency Index CI:
0.9375, CI excluding uninformative characters: 0.9231,
Retention Index RI: 0.9375, and RC (RC = CI ×RI): 0.8789.

row not reduced”, and by the two unpolarized cha-
racters “4 (0) male ligula simple and small”, “13
(0) tubular sclerotized canal of ScP not oblite-
rated”. The clade (Stenolestes + (Epiophlebia +
Bellabrunetia)) is supported by the character state
“17 (1) thoracic skewness not very important”. The
clade (Epiophlebia + Bellabrunetia), representing
the true Epiproctophora, is supported by the five
apomorphies “7 (1)”, “8 (1)”, “9 (1)”, “10 (1)”,
and “16 (1)”. The clade (Lestes + Calopteryx) is sup-
ported by the character state “15 (1) eyes of adults
strongly separated”, unknown in Triadophlebia.
The analysis does not support the clade
[Tarsophlebiidae + (Zygoptera + Epiprocto-
phora)] proposed by Bechly (1996) and Rehn
(2003).
After the present observations, the Tarso-
phlebiidae seem to have four-segmented tarsi,
which is probably a plesiomorphic condition
because the Meganeuridae have five-segmented
tarsi. Thus, their position in the same clade with
the Epiproctophora would imply that the pres-
ence of three-segmented tarsi in both Zygoptera
and Epiproctophora is either due to a convergen-
cy or that the “four-segmented tarsi” of the
Tarsophlebiidae corresponds to an autapomor-
phy. The fact that they have extremely particular,
very long tarsi would support this last hypothesis.
Nevertheless, we tested the hypothesis that the
Protozygoptera also had four-segmented tarsi
(state “0” for character “1”). We obtained the
same parsimonious tree T1 as above. We also
tested the hypothesis that the reduction of the
nodal furrow is “correlated” with the reduction of
the terminal kink of the CP in nodus, by exclu-
ding the character “12” from the analysis. We
obtained the same results as above.
The phylogenetic implications of the characters
examined above are summarized in the Table 2.
They are rather favouring a derived position of
Tarsophlebiidae close to the Epiproctophora
than a position basal to the (Zygoptera +
Epiproctophora) (Appendix 2). Nevertheless, the
phylogenetic position still remains a matter of
discussion, partly because of the uncertain prima-
ry polarization of some characters currently used
to define the Zygoptera as a monophyletic group.
The problem of the position of Tarsophlebiidae
is related to a better relative characterization of
Zygoptera and Epiproctophora. Also, this analy-
sis clearly demonstrates the weakness of our
knowledge on several characters that cannot be
correctly polarized because of the lack of infor-
mation in the Permo-Triassic Protozygoptera and
Triadophlebiomorpha. The same remark can be
made for the phylogeny proposed by Rehn
(2003) in which no less than 34 characters per
112 (about 30%) are unknown in all the most
inclusive (out)groups. Part of solution may come
through the discovery of better preserved material
Revision of Tarsophlebiidae (Odonatoptera)
57
GEODIVERSITAS • 2004 • 26 (1)
Character Phylogenetic implication Remark
A, number of adult tarsomeres Basal to Odonata Uncertain observations
B, tarsal claws Basal to Odonata Uncertain observations
C, male vesicular spermalis Basal to Odonata
D, male auricle Erroneous
E, male genital appendages Basal to Odonata
or to Epiproctophora
F, hind wing discoidal cell Basal to Odonata
G, vein CuAb Basal to Odonata
or to Epiproctophora
H, nodal structure Basal to Epiproctophora
I, vein CuP Basal to Epiproctophora
TABLE 2. — Implications of the studied structures on the phylogenetic position of the Tarsophlebiidae. Abbreviations: CuAb, first
posterior branch of Cubitus Anterior; CuP, Cubitus Posterior.

of Tarsophlebiidae and above all, a better know-
ledge of the body structures of the Protozygopera,
Triadophlebiomorpha, and Protanisoptera.
Acknowledgements
We sincerely thank Dr Carsten Brauckmann
(Institut für Geologie und Paläontologie,
Technische Universität Clausthal, Germany),
and Dr Jörg Ansorge (Institut für geologische
Wissenschaften Ernst-Moritz-Arndt, Universität
Greifswald, Germany) for their kind and instruc-
tive referee comments.
REFERENCES
BECHLY G. 1995. — Morphologische Untersuchungen
am Flügelgeäder der rezenten Libellen und deren
Stammgruppenvertreter (Insecta; Pterygota;
Odonata), unter besonderer Berücksichtigung der
Phylogenetischen Systematik und des Grundplanes
der *Odonata. Petalura Böblingen, Special Volume 1,
341 p.
BECHLY G. 1996. — Morphologische Untersuchun-
gen am Flügelgeäder der rezenten Libellen und
deren Stammgruppenvertreter (Insecta; Pterygota;
Odonata), unter besonderer Berücksichtigung der
Phylogenetischen Systematik und des Grundplanes
der *Odonata. Petalura Böblingen, Special Volume
2, 402 p.
BECHLY G. 1997. — New fossil Odonata from the
Upper Triassic of Italy, with a redescription of
Italophlebia gervasutti, and a reclassification of Tri-
assic dragonflies. Rivista del Museo Civico di Scienze
Naturale E. Caffi Bergamo 19: 31-70.
BECHLY G. 1999. — Phylogeny and Systematics of Fossil
Dragonflies (Insecta: Odonatoptera) with Special Ref-
erence to some Mesozoic Outcrops. Ph.D. thesis, Eber-
hard-Karls-Universität Tübingen, Germany, x +
755 p.
BECHLY G., BRAUCKMANN C., ZESSIN W. & GRÖNING
E. 2001. — New results concerning the morpholo-
gy of the most ancient dragonflies (Insecta:
Odonatoptera) from the Namurian of Hagen-
Vorhalle (Germany). Journal of Zoological Systemat-
ics and Evolutionary Research Berlin 39: 209-226.
CARLE F. L. 1982. — The wing vein homologies and
phylogeny of the Odonata: a continuing debate.
Societas Internationalis Odonatologica, Rapid Com-
munications Utrecht 4: 1-66.
CARPENTER F. M. 1992. — Superclass Hexapoda, in
MOORE R. C. & KAESLER R. L. (eds), Treatise on
Invertebrate Paleontology. The Geological Society of
America and the University of Kansas, Boulder,
Colorado, (R), Arthropoda 4, 3/4, xxii + 655 p.
FLECK G. & NEL A. 2002. — The first isophlebioid
dragonfly (Odonata: Isophlebioptera: Camptero-
phlebiidae) from the Mesozioc of China.
Palaeontology 45: 1123-1136.
FLECK G. & NEL A. 2003. — Revision of the
Mesozoic family Aeschnidiidae (Odonata:
Anisoptera). Zoologica 153: 1-180.
FLECK G., NEL A., BECHLY G. & MARTÍNEZ-DELCLÒS
X. 2001. — Revision and phylogenetic affinities of
the Jurassic Steleopteridae Handlirsch, 1906
(Insecta: Odonata: Zygoptera). Insect Systematic
Evolution 32: 285-305.
FLECK G., BECHLY G., MARTÍNEZ-DELCLÒS X.,
JARZEMBOWSKI E. A., CORAM R. & NEL A. 2003. —
Phylogeny and classification of the Stenophlebioptera
(Odonata, Epiproctophora). Annales de la Société ento-
mologique de France (N.S.) 39 (1): 55-93.
FRASER F. C. 1955. — Note on Tarsophlebiopsis mayi
Tillyard (Odonata: Tarsophlebiidae). Psyche 62:
134, 135.
FRASER F. C. 1957. — A reclassification of the order
Odonata, based on some new interpretations of the
venation of the dragonfly wing. Handbook of the Royal
Zoological Society of New South Wales 12: 1-133.
HAGEN H. A. 1862. — Über die Neuropteren aus
dem lithographischen Schiefer in Bayern.
Palaeontographica 10: 96-145.
HAGEN H. A. 1866. — Die Neuroptera des litho-
graphischen Schiefers in Bayern. Palaeontographica
15: 57-96.
HANDLIRSCH A. 1906-1908. — Die fossilen Insekten
und die Phylogenie der rezenten Formen. Ein
Handbuch für Paläontologen und Zoologen.
Engelman V.W., Leipzig, 1430 p.
HENNIG W. 1966. — Phylogenetic Systematics.
University of Illinois Press, Urbana, 263 p.
HENNIG W. 1969. — Die Stammesgeschichte der
Insekten. Kramer, Frankfurt am Main, 436 p.
HENNIG W. 1981. — Insect Phylogeny. Wiley & Sons,
New York, xxii + 514 p.
HUGUET A., NEL A., MARTÍNEZ-DELCLÒS X., BECHLY
G. & MARTINS-NETO R. 2002. — Preliminary
phylogenetic analysis of the Protanisoptera (Insecta:
Odonatoptera). Geobios 35: 537-560.
JARZEMBOWSKI E. A. 1990. — Early Cretaceous
zygopteroids of Southern-England with the descrip-
tion of Cretacoenagrion alleni, gen. nov., spec. nov.
(Zygoptera, Coenagrionidae, “Anisozygoptera”:
Tarsophlebiidae, Euthemistidae). Odonatologica 19:
27-37.
MALZ H. & SCHRÖDER H. 1979. — Fossile Libellen-
biologisch betrachtet. Kleine Senckenberg-Reihe
Frankfurt am Main 9: 1-46.
MARTÍNEZ-DELCLÒS X. & NEL A. 1996. — Discovery
of a new Protomyrmeleontidae in the Upper
Jurassic from Germany (Odonatoptera, Odonata,
Archizygoptera). Archaeopteryx 14: 67-73.
Fleck G. et al.
58 GEODIVERSITAS • 2004 • 26 (1)

MARTYNOV A. V. 1927. — Jurassic fossil insects from
Turkestan. 7. Some Odonata, Neuroptera,
Thysanoptera. Bulletin de l’Académie des Sciences de
l’URSS, Classe des Sciences mathématiques et
naturelles 20 (13-14): 757-768.
MEUNIER F. 1896. — Les Agrionides fossiles des
musées de Munich et de Harlem. Annales de la
Société entomologique de France 65: 30-35.
MOSTOVSKI M. B. & MARTÍNEZ-DELCLÒS X. 2000.
— New Nemestrinoidea (Diptera: Brachycera)
from the Upper Jurassic-Lower Cretaceous of
Eurasia, taxonomy and palaeobiology.
Entomological Problems 31: 137-148.
NEL A. & ESCUILLIÉ F. 1992. — Présence du genre
Stenolestes Scudder, 1895 dans les laminites
oligocènes du Revest-des-Brousses (Lubéron,
France) (Insecta, Odonata, Zygoptera, Lestoidea,
Sieblosiidae). L’Entomologiste 48: 1-13.
NEL A. & ESCUILLIÉ F. 1993. — Découverte d’un
nouveau genre de Sieblosiidae dans le Miocène
supérieur de l’Ardèche (France) (Odonata,
Zygoptera, Lestoidea, Sieblosiidae). Nouvelle Revue
d’Entomologie (N.S.) 10: 233-242.
NEL A. & PAICHELER J.-C. 1994. — Les Lestoidea
(Odonata, Zygoptera) fossiles : un inventaire cri-
tique. Annales de Paléontologie 80: 1-59.
NEL A. & JARZEMBOWSKI A. E. 1996. — Description
and revision of some dragonflies (“Anisozygoptera”)
from the Lower Cretaceous of England (Odonata:
Stenophlebiidae, Campterophlebiidae?, Epiophle-
biidae, Euthemistidae). Cretaceous Research 17: 87-96.
NEL A., GAND G., FLECK G., BETHOUX O. &
LAPEYRIE J. 1999. — Saxonagrion minutus nov. gen.
et sp., the oldest damselfly from the Upper Permian
of France (Odonatoptera, Panodonata,
Saxonagrionidae nov. fam.). Geobios 32 (6): 883-
888.
NEL A., MARTÍNEZ-DELCLÒS X., PAICHELER J.-C. &
HENROTAY M. 1993. — Les “ Anisozygoptera ” fos-
siles. Phylogénie et classification (Odonata).
Martinia Bois d’Arcy, numéro hors série 3: 1-311.
NEL A., BETHOUX O., BECHLY G., MARTÍNEZ-
DELCLÒS X. & PAPIER F. 2001. — The Permo-
Triassic Odonatoptera of the “protodonate” grade
(Insecta: Odonatoptera). Annales de la Société ento-
mologique de France 37: 501-525.
PRITYKINA L. N. 1968. — Strekozy Karatau (Odona-
ta) [= Dragonflies from Karatau (Odonata)], in
PANFILOV D. V. (ed.), Yurskie nasekomye Karatau.
[= Fossil Insects from Karatau]. Academy of Sciences
of the SSSR, Nauka, Moscow: 26-55 (in Russian).
PRITYKINA L. N. 1977. — [New dragonflies from
Lower Cretaceous deposits of Transbaïkalia and
Mongolia], in Fauna, flora i biostratigrafiya mezo-
zoya i kaynozoya Mongolii [= Fauna, flora and bios-
tratigraphy of the Mesozoic and Cenozoic of
Mongolia]. Trudy Sovmestnaya Sovetsko-Mon-
gol’skaya Paleontologicheskaya Ekspeditsiya Moscow
4: 81-96 (in Russian).
REHN A. C. 2003. — Phylogenetic analysis of higher-
level relationships of Odonata. Systematic
Entomology 28 (2): 181-239.
RIEK F. & KUKALOVÁ-PECK J. 1984. — A new inter-
pretation of dragonfly wing venation based upon
early Upper Carboniferous fossils from Argentina
(Insecta: Odonatoidea) and basic character states in
pterygote wings. Canadian Journal of Zoology 62:
1150-1166.
SCHMIDT E. 1958. — Bemerkungen über Lestiden III.
(Odonata). Decheniana 111: 1-7.
SWOFFORD D. 1998. — PAUP*. Phylogenetic Analysis
Using Parsimony (* and Other Methods), Version 4.
Sinauer, Sunderland, Massachusetts, software.
TILLYARD R. J. 1923. — Tarsophlebiopsis mayi n. gen.
et n. sp., a dragon-fly, found in the body-chamber
of a Corallian ammonite. Geological Magazine 60:
146-152.
TILLYARD R. J. 1925. — The British Liassic dragon-
flies. British Museum (Natural History), Fossil Insects
London 1: 1-39.
TILLYARD R. J. & FRASER F. C. 1940. — A reclassifica-
tion of the order Odonata based on some new inter-
pretations of the venation of the dragonfly wing.
The Australian Zoologist 9: 124-396.
ZHERIKHIN V. V. & GRATSHEV V. G. 1993. —
Obrieniidae, fam. nov., the oldest Mesozoic weevils
(Coleoptera, Curculionoidea). Paleontological
Journal 27 (1A): 50-69.
ZHERIKHIN V. V., MOSTOVSKI M. B., VRANSKY P.,
BLAGODEROV V. A. & LUKASHEVICH E. D. 1999.
— The unique Lower Cretaceous locality Baissa
and other contemporaneous fossil insects sites in
North and West Transbaikalia. AMBA Projects
Publications No.AMBA/AM/PFCIM98/1.99:
Proceedings of the Ist Palaeoentomological Conference,
Moscow, 1998: 185-191.
Submitted on 7 October 2002;
accepted on 12 May 2003.
Revision of Tarsophlebiidae (Odonatoptera)
59
GEODIVERSITAS • 2004 • 26 (1)

Fleck G. et al.
60 GEODIVERSITAS • 2004 • 26 (1)
APPENDIX 1
List of characters used in the phylogenetic analysis (see Table 1). Abbreviations: AA, Analis Anterior; CP, Costa Posterior; CuA,
Cubitus Anterior; CuAb, first posterior branch of Cubitus Anterior; CuP, Cubitus Posterior; MP + Cu, Median Posterior + Cubitus;
ScP, Subcosta Posterior.
1. Number of adult tarsomeres: 4 or more, 0; 3, 1.
2. Tarsal claws: without ventral-apical claw-hook, 0; with a ventro-apical claw-hook, 1.
3. Male vesicula spermalis: short and not anteriorly elongated, 0; long and anteriorly elongated, 1.
4. Male ligula: simple and small, 0; greatly enlarged, subsegmented, and modified as copulation organ, 1.
5. Male hamuli posteriors: platelike and small, 0; greatly enlarged and modified as copulation organs, 1.
6. Male hamuli anteriores: platelike, 0; hooklike, 1.
7. Male auricles: absent, 0; present, 1.
8. Male epiproctal process: absent, 0; present, 1.
9. Hind wing discoidal cell: basally opened, 0; basally closed, 1.
10. Postero-proximal branch CuAb of CuA: absent, 0; present, even if more or less fused with AA, 1.
11. Terminal kink of the CP in nodus: not reduced, 0; reduced, 1.
12. Nodal furrow: not reduced, 0; reduced, 1.
13. Tubular sclerotized canal of ScP along the venter of the postnodal costal margin: not obliterated, 0; obliter-
ated, 1.
14. Vein CuP: straight, 0; strongly curved and seems to begin on AA rather than on MP + Cu, 1.
15. Eyes of adults: not strongly separated of more than their diameter, 0; strongly separated, 1.
Remark: we prefer to consider on the character that concerns the distance between the eyes because the char-
acter “head transversely elongate” is “geometrically” correlated to it.
16. Ocelli: “zygopterid” type, in equilateral triangle, with the lateral ocelli posteriorly rejected, 0; “anisopterid”
type, not in equilateral triangle, 1.
17. Thoracic skewness: very important, 0; not very important, 1.
APPENDIX 2
List of synapomorphies of Tarsophlebiidae and Odonata (= synapomorphies of Panodonata Bechly, 1996). Abbreviations: Ax0,1,2,
primary antenodal cross-veins; CP, Costa Posterior; CuA, Cubitus Anterior; IR2, intercalary vein of radial area; MAb, first posterior
branch of Median Anterior; MP, Median Posterior; RP, Radius Posterior; ScP, Subcosta Posterior.
Presence of a costal triangle as broad and strong sclerotisation of the basal costal margin (even the most basal part
of the CP is completely fused to the costal margin); the distal discoidal vein MAb (= distal side of discoidal cell)
and the subdiscoidal vein (origin of the CuA on MP) are aligned and dorsally enforced by a strong sclerotisation,
so that this structure appears to cross the vein MP and the concave fold along this vein (formation of a “discal
brace” sensu Carle [1982], which is not an autapomorphy of Zygoptera, contra Bechly [1995]; this discal brace is
aligned with the arculus in the groundplan, but this character state was only retained in some taxa with an open
discoidal cell and in the fore wings of Epiophlebia); the midfork (first fork of RP and base of IR2) is shifted basal-
ly, with the RP3/4 generally arising basal of the subnodus (reversed in some Platycnemidoidea) and RP2 arising
close to the subnodus (in the groundplan); more derived type of nodus, with a kink in ScP; the oblique basal
brace (still present in Protanisoptera and Protozygoptera) is transformed into a transverse “basal bracket” Ax0
which looks like a primary antenodal cross-vein. Note that the presence of two strong primary antenodal cross-
veins Ax1 and Ax2 is a synapomorphy also present in some Protozygoptera since they are aligned and enforced
by a chitinous bracket in some well preserved specimens (Upper Permian of Russia, specimen No. 1/276, PIN),
although difficult to observe in some fossils (Nel pers. obs.); pterostigma distinctly braced by an oblique post-
subnodal cross-vein beneath the basal margin of the pterostigma; presence of a tracheated “lestine oblique vein”
between RP2 and IR2 (secondarily absent in Caloptera, Coenagrionomorpha and Oreopterida); in the median
space (= basal space) the convex vestige of the Media-stem (“vestigial CuA” sensu Fraser 1957) is suppressed since
it is fused with the cubital stem to a common medio-cubital-stem (the alleged presence of this vestige in
Tarsophlebiopsis is either an individual aberration, teratology, or preservational artifact), convergent to some
Protanisoptera, Triadophlebiomorpha and Protozygoptera.