Content uploaded by Jarmila Kukalová-Peck
Author content
All content in this area was uploaded by Jarmila Kukalová-Peck on Nov 28, 2014
Content may be subject to copyright.
Palaeodiversity 2: 169–198; Stuttgart, 30.12.2009. 169
Carboniferous protodonatoid dragonfly nymphs and the synapo-
morphies of Odonatoptera and Ephemeroptera (Insecta: Palaeoptera)
JARMILA KUKALOVÁ-PECK
Abstract
Three extremely rare fossil protodonatoid dragonfly nymphs are described from the middle Pennsylvanian
(Moscovian) of Mazon Creek, Illinois: Dragonympha srokai n. gen., n. sp. (Meganisoptera), a large, nearly com-
plete young nymph with an extended labial mask and uplifted wing pads; Alanympha richardsoni n. gen., n. sp.
(Meganisoptera), a nymphal forewing with two articular plates attached to it; and Carbonympha herdinai n. gen., n.
sp. (Eomeganisoptera), a detached nymphal forewing. Plesiomorphic states in Dragonympha n. gen. indicate ho-
mologies unresolved in modern Odonata. The segmented head bears 3
rd
tergum ventrally invaginated. The extended
labial mask still shows limb segments. The prothorax bears a pair of winglets. The short wing pads are fully articu-
lated, twisted, uplifted and streamlined with body. The mesothoracic anepisternum is placed between acrotergite
and prescutum. The abdominal leglets form long, segmented, serial gill filaments. In the ontogenesis of modern
dragonflies, the wing and articulation disc occurs just above subcoxal pleuron and far from tergum. Wing sclerites
are arranged in eight rows protecting eight blood pathways running towards eight wing veins. The sistergroup of
Odonatoptera has not yet been convincingly resolved with computer cladistic approaches. Reasons are examined
and discussed. More accurate, evolution-based character evaluations are shown with examples. The role of a correct
model of the pan-arthropod limb and the origin of insect wings is discussed. Groundplan characters in dragonflies
and mayflies are compared in their Paleozoic and modern states, their obscurity is clarified and complex synapo-
morphies are proposed. Palaeoptera is confirmed as a monophyletic group and the following sistergroup relation-
ships are suggested: Pterygota = Palaeoptera + Neoptera; Palaeoptera = Palaeodictyopteroida + Hydropalaeoptera;
Hydropalaeoptera = Odonatoptera + Ephemeroptera.
K e y w o r d s : Pterygota, Palaeoptera, Ephemeroptera, Odonatoptera, Meganisoptera, fossil dragonfly
nymphs, phylogeny, insect wing, Carboniferous, Mazon Creek.
Zusammenfassung
Drei extrem seltene fossile Larven protodonatoider Libellen werden aus dem mittleren Pennsylvanium (Mos-
covium) von Mazon Creek in Illinois beschrieben: Dragonympha srokai n. gen., n. sp. (Meganisoptera), eine große,
nahezu vollständige junge Larve mit ausgestreckter Fangmaske und hochgestellten Flügelscheiden; Alanympha
richardsoni n. gen., n. sp. (Meganisoptera), ein larvaler Vorderflügel mit zwei ansitzenden Flügelgelenkplatten; und
Carbonympha herdinai n. gen., n. sp. (Eomeganisoptera), eine isolierte Flügelscheide eines larvalen Vorderflügels.
Plesiomorphe Merkmalszustände von Dragonympha n. gen. weisen auf unerkannte Homologien zu modernen Odo-
naten hin. Der gegliederte Kopf weist ein ventral invaginiertes 3. Tergum auf. Die ausgestreckte Fangmaske zeigt
noch die Segmentierung von Beinen. Der Prothorax trägt ein Paar Flügelchen. Die kurzen Flügelscheiden sind
vollständig gelenkig, gedreht, aufgerichtet und stromlinienförmig mit dem Körper. Das mesothorakale Anepister-
num liegt zwischen dem Acrotergit und dem Prescutum. Die abdominalen Beinanhänge bilden lange, segmentierte,
serielle Kiemenfilamente. In der Ontogense moderner Libellen bildet sich die scheibenförmige Embryonalanlage
des Flügels und der Flügelgelenkung direkt oberhalb des subcoxalen Pleurons, weit entfernt vom Tergum. Die
Flügelgelenkstücke sind in acht Reihen angelegt, welche die acht Blutlakunen schützen, die in die acht Flügeladern
verlaufen. Die Schwestergruppe der Odonatoptera konnte bislang noch nicht überzeugend mittels computerkladis-
tischer Methoden ausfindig gemacht werden. Die Gründe hierfür werden untersucht und diskutiert. Genauere,
evolutionsbasierte Merkmalsuntersuchungen werden mit Beispielen vorgestellt. Die Bedeutung eins zutreffenden
Modells des ursprünglichen Beins der Pan-Arthropoden und des Ursprungs der Insektenflügel wird diskutiert.
Grundplanmerkmale von Libellen und Eintagsfliegen werden in ihren paläozoischen und modernen Ausprägungen
verglichen, ihre Unklahrheiten werden aufgeklärt und komplexe Synapomorphien werden vorgeschlagen. Die Pa-
laeoptera werden als monophyletische Gruppe bestätigt und folgende Schwestergruppenbeziehungen werden vorge-
schlagen: Pterygota = Palaeoptera + Neoptera; Palaeoptera = Palaeodictyopteroida + Hydropalaeoptera; Hydropa-
laeoptera = Odonatoptera + Ephemeroptera.
Contents
1. Introduction ............................................................................ 170
1.1. Morphological concepts and terms ...................................................... 172
1.2. Collecting, fossilization, matrix, and observation methods ................................... 174
2. Systematic paleontology .................................................................. 176
Infraclass Palaeoptera M
ARTYNOV, 1923 ...................................................... 176
Order Meganisoptera M
ARTYNOV, 1932 ................................................... 176
Family incertae sedis ............................................................. 176
170 PALAEODIVERSITY 2, 2009
1. Introduction
Three extremely rare protodonate fossil nymphs from
Pennsylvanian (Late Carboniferous) nodules at the locali-
ty Mazon Creek, Illinois, USA, are described. Protodonate
dragonflies from the Late Paleozoic orders †Geroptera,
†Eomeganisoptera and †Meganisoptera lived in the tropi-
cal belt of Europe, Asia and North America and in the
temperate zone of South America. †Meganisoptera had a
wingspan of up to 73 cm and specialized in hunting large
prey (probably †Palaeodictyoptera and †Megasecoptera).
Dragonflies at the Mazon Creek locality were the top
aerial predators in luscious moist forests of a large river
delta. Adult wing fragments are quite common in Late
Carboniferous and Permian freshwater deposits, while
bodies are mostly lost to predation or rapid decomposition.
Immatures – against all expectation – are almost entirely
absent. Besides the three remnants described here, only
one poorly preserved and uncertain nymphal wing is
known, Schlechtendaliola nympha H
ANDLIRSCH, 1919,
from the Stephanian C of Wettin, East Germany (see also
H
ANDLIRSCH 1906–1908). This was later misnamed
Schlechtendaliana by B
RIDGES (1994). BRAUCKMANN &
Z
ESSIN (1989) regarded its position uncertain, but REN et al.
(2008) considered it a protodonate. This scarcity of fossil
nymphs is inconsistent with their occurrence in aquatic
habitats. In contrast, the juveniles of Ephemeroptera and
Pleconeoptera are found in many Paleozoic localities and
Early Permian sediments at Elmo, Kansas are in places
littered with their small, isolated winglets, which became
detached from bodies at the articular membrane (winglets
never detach in modern nymphs as they are secondarily
fused to terga). While conspicuously missing from sedi-
ments in the Paleozoic, dragonfly nymphs are relatively
well represented in Mesozoic and Tertiary freshwater de-
posits. This discrepancy is discussed below.
The Pterygota clade originated with the origin of wings
and diversified into divisions, lineages and orders with the
adaptation of the wing pairs to different types of flight,
wing flexing, wing folding, and other functions. Thus, the
relationships between the major lineages with respect to
their monophyletic origin from Pterygota are most clearly
identified by their veinal systems. The synapomorphies
include similarities between the pairs of wings, veinal
braces and fusions near wing base, wing areas, fields,
flexion lines, and folds, and in fusions and reductions of
the wing articular sclerites. As a result, the wing organ-
system is a treasure chest of multiple character series,
which offer easily observed synapomorphies in the ptery-
gote higher categories: divisions, lineage and orders.
Based on the wing organ, for most of the last century sys-
tematists consistently classified Odonatoptera as the sister
lineage of Ephemeroptera, under the superlineage Hydro-
palaeoptera R
OHDENDORF, 1968 and the division Palae-
optera M
ARTYNOV, 1923 (MARTYNOV 1923, 1925; ROHDEN-
DORF 1962; SHAROV 1966; HENNIG 1969, 1981; CARPENTER
1992 and before).
The diversification of the limb-wing organ in Ptery-
gota into divisions, lineages and orders was thematically
studied from three decades by this author and co-authors,
while following closely the phylogenetic rules and criteria
theoretically outlined by H
ENNIG (1969, 1981): monophyly,
full homology, search for the groundplan level of character
states in all higher taxa, using this research as background
to identify synapomorphies, and turning these into a sys-
tematic hierarchy. Over the years, this study repeatedly
verified the reality of Palaeoptera and showed Odona-
toptera as the sistergroup of Ephemeroptera (K
UKALOVÁ-
Genus Dragonympha n. gen. ................................................... 176
Genus Alanympha n. gen. ...................................................... 183
Order Eomeganisoptera R
OHDENDORF, 1963 ................................................ 186
Family incertae sedis ............................................................. 186
Genus Carbonympha n. gen. ................................................... 186
3. Discussion .............................................................................. 187
3.1. Labial mask ........................................................................ 187
3.2. Gill filaments ....................................................................... 188
3.3. Thoracic pleuron, pleural inflexions, double wing pivot ..................................... 189
3.4. Arthropod wing homologue ........................................................... 189
3.5. Articular sclerites follow blood pathways to wing veins ..................................... 191
3.6. Difference in articulation: Palaeoptera versus Neoptera .....................................192
3.7. Comparative-morphological analysis of the pterygote wing articulation ........................192
3.8. Homologous muscular insertions .......................................................192
3.9. Anterior articular plate ...............................................................193
3.10. Posterior articular plate ...............................................................194
3.11. Paleoptery is a derived adaptation ....................................................... 194
3.12. Nymphal winglets in diverse resting positions ............................................. 195
3.13. Evolution of the veinal system in Odonatoptera and Ephemeroptera. . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
4. Conclusion .............................................................................195
5. References .............................................................................196
KUKALOVÁ-PECK, CARBONIFEROUS DRAGONFLY NYMPHS 171
PECK 1978, 1983, 1985, 1991, 1992, 1998, 2008; RIEK &
K
UKALOVÁ-PECK 1984; KUKALOVÁ-PECK & BRAUCKMANN
1990; S
HEAR & KUKALOVÁ-PECK 1990; HAAS & KUKALOVÁ-
P
ECK 2001; KUKALOVÁ-PECK & LAW R ENCE 2004, including
the list of modern genera used in systematic consider-
ations).
Evolutionary morphological facts (e. g., in the limb or-
gan system) clearly show Crustacea as the sistergroup of
Atelocerata, under Mandibulata (S
NODGRASS 1935; KUKA-
LOVÁ-PECK 1998). Entomologists sometimes do not fully
realize that, since Crustacea occur in the Cambrian so in
all probability did the ancestor of Myriapoda and Hexapo-
da. This pushes the divergence of Pterygota into divisions
(Palaeoptera and Neoptera), superlineages (such as †Pa-
laeodictyopteroida and Hydropalaeoptera) and even lin-
eages (Odonatoptera and Ephemeroptera) way back before
the Devonian. It is probably safe to suggest that terrestrial
arthropods evolved with first algal mats floating on water
and thriving on swampy shores. Ancestral Dicondylia
must have had functional outer rami ancestral to wings
ever since their divergence, when the vegetation was very
short, dense, and surrounded by water. Since freshwater
deposits in the Cambrian, Ordovician and Silurian were
almost completely destroyed by erosion, positive evidence
may never be found. The fact is that all Carboniferous
insects known to this author already belong to modern
divisions and superlineages, with a single exception: the
extinct superlineage †Palaeodictyopteroida (= †Diapha-
nopterodea, †Palaeodictyoptera, †Megasecoptera, †Permo-
themistida).
Within Pterygota, Odonatoptera are one of the oldest
lineages, and it also bears the most autapomorphic wing
organ. This combination of extreme attributes made the
full homologization of the wing characters exceptionally
confusing. The odonate wing articulation was fully ho-
mologized only in 1983 (by this author), and the venation
only in 1984 (by R
IEK & KUKALOVÁ-PECK). Here, shared
characters of articular plates, venational fusions and brac-
es are analyzed in more detail. New, even more complex,
synapomorphies are identified, first in the Paleozoic drag-
onflies and mayflies, and then again in their now obscured
state in modern higher taxa.
During the last 20 years, mainstream entomological
systematics (concerned mainly with species and genera)
has progressed remarkably in accuracy and objectivity
with the introduction of computerized systematics. In this,
similarities in as many as possible character series are re-
corded (usually) by classical comparative morphology
(S
NODGRASS 1935) and then processed by sophisticated
computer programs. The programs statistically separate
reversals and other secondary variants from genuine
(groundplan-level) synapomorphies, and deliver objective,
defendable phylogenies in species, genera and in some
families. However, when systematists tried to find the re-
lationships in the pterygote higher taxa (orders, lineages,
divisions) the resulting phylogenies were uncertain and
confusing. First, Odonatoptera were shown as the sister of
Neoptera (K
RISTENSEN 1975, 1981, 1991; many followers),
at first mainly because these taxa bear similar broad man-
dibles manipulated by shearing musculature and a posteri-
orly shifted hypopharynx, plus some isolated similarities.
Surprisingly, the wing organ including by far the largest
number of synapomorphies was never properly analyzed,
but sidestepped and the existing data supporting Palae-
optera + Neoptera were not refuted, but ignored. The out-
dated model for the limb organ was the same as used by
M
ANTON (1977, and before). Soon, a different set of simi-
larities and some molecular data showed Odonatoptera as
the sister of Ephemeroptera + Neoptera. Then, another
morphological and molecular data set confirmed their pre-
vious classification as the sistergroup of Ephemeroptera.
In the last decade, the merits of these three different data-
sets were intensely scrutinized, but final consensus was
never reached (see H
OVMÖLLER et al. 2002, OGDEN & WHIT-
ING 2003, and WILLKOMMEN 2008 for details and referenc-
es).
Some entomologists expressed hope that even larger
datasets and/or improved molecular data will solve the
Odonatoptera problem. Others came to believe that the
culprit is morphology, which in higher taxa is “incapable”
of delivering characters explicit enough to resolve its own
morphological phylogenies! If accepted as true, such an
embarrassing inability of evolution to leave its marks in
the largest animal group (winged insects) is the kind of
ammunition creationists are waiting for. The truth is quite
different: the problem is not new at all and it had been
theoretically analyzed and resolved decades ago by H
EN-
NIG (1969, 1981). For a detail review of different systematic
problems and solutions in the lower and the higher hexa-
pod taxa, see K
UKALOVÁ-PECK (2008).
As generally acknowledged, the synapomorphies of
pterygote divisions, lineages and orders in living species
are obscured by post-groundplan adaptations. These cre-
ated convergencies and parallelisms, which cannot be
separated from genuine synapomorphies by classical com-
parative morphology alone. When computers process false
data, no matter which programs are used, the results are
falsified as well. Since homoplasies in the higher taxa are
significantly more frequent than readily recognizable syn-
apomorphies, more data make the situation even worse,
and never better. The consensus is that comparative mor-
phology alone is, simply, too limited for the old modern
higher taxa! For more information see K
UKALOVÁ-PECK
(2008).
Arthropods are all legs (S
MITH 1970; MARSHALL et al.
1994; R
AFF et al. 1991; RAFF 1996; WILLIAMS & CARROLL
1993; S
HUBIN et al. 1997; KUKALOVÁ-PECK 2008). As a cau-
tionary tale, remember that 40 years ago the all-arthropod
172 PALAEODIVERSITY 2, 2009
limb (now presented as polyramous with exites and en-
dites and 11-segmented: K
UKALOVÁ-PECK 2008, fig. 1) was
interpreted by classical comparative morphology as with-
out rami, and bearing only 6–7 podites (lacking 3–4
podites) (S
NODGRASS 1935). Since all arthropod appendag-
es could not be derived from such an impoverished ances-
tral model, phylum Arthropoda had to be (perfectly logi-
cally!) disassembled. The limb appendages most frequent-
ly cited were mandibles, erroneously interpreted as one
segment (M
ANTON 1977 and before, references). Instead,
all arthropod mandibles include three segments plus two
endites, which only may look like one segment! As a re-
sult, Hexapoda which actually do show three-segmented
mandibles as a plesiomorphy (in modern Archaeognatha)
were transferred to a polyphyletic phylum “Uniramia” and
it took 17 years to reunite them convincingly back with
Arthropoda (K
UKALOVÁ-PECK 1992, 1998, 2008; WÄGELE
1993, 1996). In the last two decades, the limited compara-
tive method and its disastrous ancestral “panarthropod”
limb model (based on highly derived thoracic legs of
grasshoppers!) again sneaked back into the higher taxa of
Hexapoda on the coat tails of a genuinely progressive
computerized systematic method, only to confuse the un-
derstanding of Parainsecta, Diplura (the sister of Insecta),
and the basal split of Pterygota (K
UKALOVÁ-PECK 1983,
1987, 1991, 1992, 1998, 2008; K
UKALOVÁ-PECK & RICHARD-
SON 1983). As a fact verified then and now again, defend-
able synapomorphies are found only when the systematic
rules outlined by H
ENNIG (1969, 1981) calling for full ho-
mologization at the groundplan level, are followed. The
remedy is to approach modern Odonata, Ephemerida and
Neoptera much more broadly, with evolutionary evidence
and with data known in genetics, developmental genetics,
ontogeny, physiology, arthropod anatomy, etc., to reach
this required level of character attributes.
The purpose of this account is to show, by concrete
examples, that this thematic goal is doable, and that it of-
fers objective and repeatable systematic results. Once the
all-pterygote protowing ancestor was established (based
on a long-term research), the groundplan-level synapo-
morphies in the limb/wing organ system of Odonatoptera
and Ephemeroptera are quite clearly recognizable in the
Paleozoic ancestors. It is their expression in the modern
higher taxa, Odonata and Ephemerida, which makes them
opaque. Most of them “blend” together as phylogenetically
uninformative states and are easily overlooked. However,
by applying Hennigian principles, they can be clearly
identified with help from fossils.
As a very important step in the phylogenetic research
of the higher taxa, all data and identifications in morpho-
logical organ systems must be thoroughly crosschecked,
supplemented, and scrutinized for support by the evolu-
tionary processes indicated in other biological fields (such
as, by dragonfly ontogeny: see evidence below). The natu-
ral urge to trust explicitly the “direct” evidence available
in modern fauna to indicate truthfully the grand evolu-
tionary-morphological events is rarely the best choice. As
evolutionists use to caution: the adage “seeing is believ-
ing” packs lots of visual risks of unbelievable errors.
1.1. Morphological concepts and terms
To avoid possible confusion, the concepts and terms in
the limb-wing organ system used in the following text, are
shortly reviewed:
N o m e n c l a t u r e . – The higher classification of
dragonflies has become quite complex, and is not always
reflecting evolution. The internal phylogeny of Odona-
toptera was significantly improved by B
ECHLY (1996, 1999)
and by B
ECHLY et al. (2001) and is followed here. Odona-
toptera and Ephemeroptera is a name used for the lineage.
Protodonate dragonflies are considered a grade. Odonate
dragonflies are a clade and a synonym for Odonatoclada
B
ECHLY, 1999. Odonatoclada includes the modern order
Odonata. Paleozoic Ephemeroptera belong to the order
†Syntonopterida and †Protereismatida. Order Protereis-
matida is the sistergroup of the modern order Ephemeri-
da.
M o n o p h y l y . – The Hennigian evolutionary phylo-
genetic approach and the groundplan systematic method
(K
UKALOVÁ-PECK 2008) are followed. The wing organ sys-
tem in a monophyletic Pterygota must be flawlessly deriv-
able from a single protowing model and by using exactly
the same criteria for separating plesiomorphies from apo-
morphies. The limb/wing character states used here were
homologized and evaluated with respect to a single all-
arthropod limb model and the all-pterygote protowing
(K
UKALOVÁ-PECK 2008, figs. 1, 17). Reliable synapomor-
phies shared by divisions, lineages and orders are found
only in their groundplan character states (H
ENNIG 1969,
1981). For a review of evolutionary models and the appli-
cation of the groundplan method, see K
UKALOVÁ-PECK
(2008, figs. 1–21).
P r o t o w i n g . – The first evolutionary model of the
pre-flight pterygote protowing was reconstructed in 1983,
by this author. This was based on the step-wise transfor-
mation series of veinal systems recorded in all pterygote
lineages, from the Paleozoic to modern times. The least
modified (fused, reduced) states in eight principal veins
were selected from a significant sample in each higher
taxon, compared, and the least modified reconstructed
into a single protowing shared by all Pterygota. Against
expectation, this protowing model was later found in fossil
prothoracic wings, which were never used in powered
flapping flight! These shared similar plesiomorphic veinal
system in Carboniferous Odonatoptera and Palaeodicty-
optera, which bear very different meso- and metathoracic
KUKALOVÁ-PECK, CARBONIFEROUS DRAGONFLY NYMPHS 173
flying wings. Thus, the protowing model used here is
based both on the extremely broad data from modern in-
sect orders, and on the evidence from fossil record (K
UKA-
LOVÁ-PECK 1983, figs. 13–15; 2008, fig. 17). The plesiomor-
phic wing articulation arranged in separate rows following
blood pathways was found in Palaeodictyopteroida: †Di-
aphanopterodea. This also was verified against significant
samples from all modern lineages (K
UKALOVÁ-PECK 1998;
2008, fig. 18). The pteryogote articulation models used
here include articular sclerites in their least fused and re-
duced state.
A r t i c u l a r s c l e r i t e s . – These are fragments of
the muscled epicoxal pleuron (= first arthropod limb
podite flattened and imbedded into the pleural membrane)
(K
UKALOVÁ-PECK 2008, fig. 17). This is articulated dorsally
to the tergum and ventrally to the subcoxa (shaped either
as a cylindrical podite (in mouthparts and male genitalia)
or as a pleuron (on thorax and on pregenital abdominal
segments) (K
UKALOVÁ-PECK 1991, 2008). Horizontally, the
sclerites are arranged in rows protecting the limb blood
pathways, delivering blood to principal veins of the wing
ramus (Figs. 6A, 11–15). Each of the eight rows, named
after their veins include three sclerites: proxalare (PR),
axalare (AX), and fulcalare (F). The fulcalare always ar-
ticulates with the same wing basivenale (B). Vertically,
wing sclerites form columns (proxalar, axalar, fulcalar,
and basivenal). Proxalaria and axalaria originated from
subdivided epicoxal pleuron, and fulcalaria, probably
from the muscled base of the wing ramus. On the wing-
side from the wing flexion line, which runs distad from
fulcalaria, basivenalia (B) are the sclerotized blood sinus-
es of the eight principal veins. Wing basivenalia have no
muscular insertions. Wings are moved by muscles at-
tached to the thoracic epicoxal sclerites and indirectly, by
the movement of thoracic segments.
These evolutionary conditions indicate that wing scler-
ites, which are articulated in rows, inserted with muscles
and arranged in regular columns are ancestral to all Ptery-
gota and thus plesiomorphic. Fusions between sclerites,
either in rows, in columns, between columns, to the ter-
gum, to the subcoxal pleuron, or to basivenalia are derived
and thus offer higher-level synapomorphies. Paleopterous
articular plates and composite sclerites, and neopterous
mobile axillary sclerites, wing processes, and medial and
humeral plates are identified below as non-homologous
clusters composed of different row-sclerites. Therefore,
they are derivable only from the shared common ancestral
model (the muscles are homologous, but the sclerite clus-
tering is different).
The comparative morphological interpretation based
directly on the modern fauna is quite different. Wing
sclerites are seen as irregularly distributed cuticular frag-
ments, which are mostly considered directly homologiz-
able. Much effort over many years has been spent to find a
convincingly shared pattern (W
ILLKOMMEN & HÖRN-
SCHEMEYER 2007; WILLKOMMEN 2008, and for references),
but no concensus was reached.
Articulation symmetry and blood
f l o w . – The relationship between the limb blood flow
into the wing appendage and the alignment of articular
sclerites with wing basivenalia and veins was researched
by the insect physiologist A
RNOLD (1964) (KUKALOVÁ-PECK
1978, 1983). The original row-sclerites are still clearly vis-
ible in fossil dragonflies (Figs. 5, 6B, 13), fossil and mod-
ern mayflies (Figs. 11, 12), and in most parts of the
neopteran articulation (Fig. 14) (H
AAS & KUKALOVÁ-PECK
2001). In all lineages of modern Palaeoptera and Neoptera,
the row arrangement is retained and can be recognized
(K
UKALOVÁ-PECK & LAW R ENCE 2004, figs. 29–33; KUKA-
LOVÁ-PECK 2008, figs. 18, 19). This shows that all modern
lineages, no matter how modified their wing articulation
may be, retained the underlaying row arrangement of
sclerites.
Ve n a t i o n s y s t e m . – At the wing base, each
blood channel forms a sclerotized blood sinus called a ba-
sivenale (B), which gives rise to a pair of separate veinal
sectors, each dichotomously branched two to three times
(each vein is a fully dichotomous branching system adapt-
ed to fill out the space). The protowing shows a slightly
reinforced anterior wing margin (indicating motion), but
completely lacks the veinal fusions and braces (which are
necessary for powered flight; these were added later and
are not shared by all Pterygota). But, even the precosta
bears two veinal sectors (in some modern Ephemerida:
personal communication by T. S
OLDA´ N). This ancestral
model rules that the veinal stems, crossvein braces and
veinal fusion braces, which characterize modern higher
taxa, are all derived and potentially synapomorphic. Note
that braces close to the wing base mark the diversification
of early flight, and deliver clear and testable division- and
lineage-level synapomorphies (see below). Example: The
division Palaeoptera shares two veinal stems (the stem of
M and of Cu), while the division Neoptera does not share
any veinal stems, which can be found in all lineages! For
the history of fully homologized venation in Odonatoptera
wings, see R
IEK & KUKALOVÁ-PECK (1984), KUKALOVÁ-
P
ECK (1983, 1991, 1992, 1998, 2008), HAAS & KUKALOVÁ-
P
ECK (2001), KUKALOVÁ-PECK & LAW R ENCE (2004), and the
text below.
Groundplan character states in high-
e r t a x a . – In modern higher taxa, groundplan character
states are those which are the least derived (with respect to
the protowing) and present in a significant sample of all
representatives (living and fossil). This requires some ex-
planation. A taxon without a groundplan cannot exist
(H
ENNIG 1981). Groundplan characters are always present,
either visibly or obscured (most often by reductions, e. g.,
in size) and fusions. In the higher taxa the groundplan-
174 PALAEODIVERSITY 2, 2009
level fusions never become “unfused” and reductions “un-
reduced” (reversals are completely absent; any changes in
the groundplan would change the taxon itself). However,
post-groundplan adaptations added at the family, genus
and species levels are often impossible to eliminate with-
out evolutionary clues (see examples below). Therefore,
since only the groundplan characters contain reliable syn-
apomorphies, their convincing separation and identifica-
tion is crucial for unveiling the relationship between the
higher taxa.
Predictability of the wing character
t r a n s f o r m a t i o n s e r i e s . – With reversals absent
in the higher taxa (H
ENNIG 1969, 1981; KUKALOVÁ-PECK
1983 to 2008), and the ancestral protowing lacking fu-
sions, braces, and reduction between wing veins and wing
sclerites, all these later adaptation can only be added, but
never removed. This predictability is the Rosetta stone of
higher-level phylogenetic analysis, because it provides the
systematic tool for an objective recognition of the ple-
siomorphic, apomorphic and synapomorphic states. They
stand out when the specimens are compared with the pro-
towing. This so-called groundplan method (K
UKALOVÁ-
P
ECK 2008), applied in a significant representative sample,
is first used to identify the groundplan character states in
all higher taxa (of Pterygota), and second to select the reli-
able synapomorphies which they share (see examples be-
low). If it appears to be a long-term, labor-intensive pro-
cess, this impression is correct. But, the foundation has
been finally established and now it is time for fine-tuning.
Comparative morphology approach in the limb/wing or-
gan system based on modern and close to modern fauna is
a very considerable shortcut, but it does not resolve phy-
logeny of the modern higher taxa with Paleozoic origin.
1.2. Collecting, fossilization, matrix, and observation
methods
M a z o n C r e e k n o d u l e s . – The middle Penn-
sylvanian (Moscovian) fauna of Mazon Creek offers three-
dimensional remnants of terrestrial and aquatic arthro-
pods with well-preserved delicate morphological details.
The decomposing organisms attracted iron ions percolat-
ing in mud, and these formed around them hard ironstone
nodules composed of pyrite and limonite. An insect rem-
nant is found in about every 115 nodules. Freshly exca-
vated nodules cannot be split open without severely dam-
aging the fossil. They should be kept submerged in water
in buckets for two or more winters, subjected to repeated
freezing, and only then are they ready to split into two
perfect halves along the plane of weakness occupied by a
fossil. This lengthy process makes casual collecting un-
productive, and the abandoned mining area is rapidly dis-
appearing under vegetation and new housing development.
Dr. E. S. R
ICHARDSON, the legendary former custodian of
the Chicago Field Museum, met the challenge by recruit-
ing help from more than 100 volunteers living near the
coal mining area. For three decades, these amateur pale-
ontologists spent their free time collecting and processing
Mazon Creek nodules. The fossils were then brought to
the Field Museum, where they were expertly preparated,
identified and catalogued. Dr. R
ICHARDSON provided in-
structions and encouragement, organized meetings and
exhibits, and arranged lectures by visiting scientists. Pri-
vate collectors competed with each other in making im-
portant discoveries. When Dr. R
ICHARDSON was active, al-
most all private collections were donated or willed to the
Field Museum. Later, many specimens were sold abroad
or ended up in other museums willing to interact with col-
lectors. During my frequent visits, I was able to study
fossil insects in private homes. Large homoiopterid Pa-
laeodictyoptera (K
UKALOVÁ-PECK & RICHARDSON 1983,
figs. 3–6) opened the path towards the full homologiza-
tion of the wing articulation in Odonatoptera, Ephe mero-
ptera and Neoptera (K
UKALOVÁ-PECK 1983, 1985, 1998).
The owner refused to donate or sell this historical holo-
type, which was later knocked down by a cat and broken.
Good casts made by Dr. R
ICHARDSON as insurance, are
available in the Field Museum as well as in my collection.
P r e s e r v a t i o n o f f o s s i l i n s e c t s . – The rare
attribute of Mazon Creek fossils is that they are preserved
more or less three-dimensionally, with appendages in a
nearly natural position (N
ITECKI 1979). Mouthparts, anten-
nae, legs, leglets, and genitalia are usually buried in the
matrix and must be dug out. Heads and bodies were often
desclerotized and deformed by pressure before petrifica-
tion, and then thinly coated with strongly adhering white
kaolinite. The splitting plane often runs partly outside and
partly through the middle of insect bodies, so that the im-
print and counter-imprint do not exactly match (labial
mask of the dragonfly nymph described here is preserved
only in one half of the nodule). Genitalia are often seen in
dorsal view, from inside the abdomen. Exites, endites and
vesicles are delicate features smaller than the grains of
matrix, yet they are sometimes distinctly preserved. De-
sclerotization makes some structures semi-transparent
and visible through the skeleton. There is often some free
space around fossils, with protruding patches of kaolinite
and lumps of pyrite. These occur on the natural inner sur-
face and not as the sign of in-depth preparation, as some
paleontologists hypothesize without attempting hands-on
preparation of Mazon Creek nodules.
Preparation and coating. – Mazon Creek
fossils have yielded fine details of limb morphology, wing
articulation, genitalia, cercopods, head structures and
mouthparts, which may not have been otherwise homolo-
gized correctly in modern insects (K
UKALOVÁ-PECK 2008).
Minute limb podites, exites, endites, sclerites and sutures
KUKALOVÁ-PECK, CARBONIFEROUS DRAGONFLY NYMPHS 175
are usually coated with kaolinite glued to cuticle. When
chipped away, it often scalps carbonized chitin so that fea-
tures look “artificially sculptured”. The hard matrix is best
removed by a heavy-duty vibro-engraver with a chisel-
shaped sharp iridium point. Preparation of a good speci-
men is slow and usually takes more than one day. Small
debris constantly conceal newly uncovered sutures and
morphological details and must be brushed away with a
fine brush and cleaned. Brush strokes almost instantly
produce a reflective sheen. This does not harm the fossil
but is unsightly and lowers the visibility of the details un-
der a microscope. Application of a light (grey) coating of
ammonium chloride completely solves the problem and
enhances fine details. The coating is hygroscopic and must
be washed off by water after each session to prevent forma-
tion of hydrochloric acid. Preparation of hard matrix re-
quires considerable practice, a morphological background
and lots of patience since fine details are easily “prepared
off” rather than out.
I l l u m i n a t i o n a n d o b s e r v a t i o n . – Fine de-
tails in fossil insects are best seen when illuminated at a
very low angle from 10 o’clock and/or from 2 o’clock, and
by two gooseneck illuminators, which are used one at a
time. For best results, the fossil is held in the hand, turned
slowly around 360 degrees, and rocked back and forth to
catch the best light. Sunlight should be always used as well,
as it sometimes brings out colour shades and a significant
visual enhancement. An exact angle of light is often need-
ed to make visible the weakly preserved exites, mouth-
parts, segmentation in the head and leglets, etc., which
tend to be overwhelmed by the grainy, uneven and reflec-
tive matrix. Observers not familiar with Mazon Creek fos-
sils often need help with illumination. Individuals with a
naturally low ability to distinguish details on the uneven
surface, or lacking stereoscopic vision, may miss fine de-
tails altogether. Unshakeable belief in faulty panarthropod
limb model (with 6–7 segments and no rami) appears to
lessen observation abilities as well. A correct limb model is
absolutely crucial for full homology and evaluation of the
limb-derived appendages, which deliver by far the largest,
most informative dataset for arthropod-insect phylogenies.
It is unfortunate that insect specialists are often not famil-
iar with the complexity and phylogenetic importance of the
old arguments about how many muscled segments, exites
and endites are included in the ancestral pan-arthropod
limb. Hence, they often overlook the epicoxal pleuron, ex-
ites, prefemur, patella and basitarsus rather often visible in
Paleozoic insect palps, legs, leglets and gonostyli. The sole
reason is that they are not expected to be there! But, sur-
prisingly, they are still retained in some modern insects
and are unrecognized (see a review by this author, 2008).
Unfortunately, a denial of their presence is fatal for realis-
tic arthropod and insect phylogenies.
Museum specimens and posterity. – Af-
ter the surface of Dragonympha srokai n. gen., n. sp. be-
came reflective, the Field Museum authorities decided to
treat it with ultra-sound to remove the sheen so that the
holotype would be in good shape when returned to the Il-
linois State Museum in Springfield. In a split second, the
ultrasound dulled the irreplaceable, finely preserved mor-
phological details and almost completely erased the weak-
ly imprinted labial mask. During a later visit to Spring-
field, I obtained permission and removed the thin, dulled
surface layer with some success, but the sheen returned.
The red flag is out that the ultra-sound technique is ex-
tremely harmful to the Mazon Creek fossil, and should be
completely avoided.
The purposes of keeping insect collections in the mu-
seums are multiple and complex. The dilemma is to which
extent should specimens be preserved for posterity in a
pristine original state? Should they be, ever, exposed to a
potentially harming preparation and dissections by com-
petent specialists? In 2001, F. H
AAS (the world foremost
functional morphologist in earwig wing folding) and I
(with two decades of evolutionary study of the pterygote
wing articulation and wing base braces on my credit) co-
authored a detailed phylogenetic analysis of the earwig
wing base. Yet, we were not allowed to open the wings of
earwig specimens from two rare families for fear that they
may be damaged! The opportunity that two narrow spe-
cialists in two extremely broad fields meet again to com-
pletely analyze the utmost complex insect wing folding
from an all-pterygote evolutionary view, may not repeat
itself any time soon. Since posterity now has four vials of
completely undamaged earwigs, the pertinent question is
– for which better purpose? In the specimen of Drago-
nympha srokai n. gen., n. sp., the preparation uncovered
unique scientific evidence. This was later diminished for a
parallel reason – so that posterity had a better-looking
holotype.
Acknowledgements
My sincere thanks go to my past and present co-authors and
friends, G. B
ECHLY (Staatliches Museum für Naturkunde, Stutt-
gart), J. G. P
ETERS (Florida A & M University, Tallahassee), E. L.
S
MITH (The California Academy of Sciences, San Francisco),
and J. C
OOK (Carleton University, Ottawa), for their invaluable
help. T. F. A
LLEN (Westport, Ontario) devoted much time to for-
matting the illustrations. My work in Canada was supported by
the discovery grant from NSERC, Canada and by S
TEWART B.
P
ECK (Carleton University, Ottawa, Canada). S. B. PECK, O. KU-
KAL and two unknown reviewers suggested improvements in the
manuscript. G. B
ECHLY helped to verify some facts and citations,
generously formatted the manuscript for the journal, and pro-
vided encouragement which helped to bring this paper to com-
pletion. All these colleagues and family members have my sin-
cere thanks and gratitude.
176 PALAEODIVERSITY 2, 2009
2. Systematic paleontology
Infraclass Palaeoptera M
ARTYNOV, 1923
Hydropalaeoptera R
OHDENDORF, 1968
Superorder Odonatoptera M
ARTYNOV, 1932
Order Meganisoptera M
ARTYNOV, 1932
Family incertae sedis
Genus Dragonympha n. gen.
Typus gener is: Dragonympha srokai n. sp. This ge-
nus is monotypic and is known only from the holotype (obverse
and reverse).
D e r i v a t i o n o m i n i s : From dragon (English) as in
dragonfly, and nympha (Latin); feminine.
D i a g n o s i s . – Young nymph of a large adult. Head:
Segmentation distinct. Antennae moderately long and
heavy, with relatively long joints. Labial mask with very
narrow basal portion, centrally incorporating a long and
narrow sternum; prementum includes serrated, very broad
paraglossae (almost as long as palps), and slender and
pointed labial palps (glossae not preserved). Thorax: Pro-
tergum large and complete, with acrotergite, prescutum,
and reduced prothoracic wings fused to it laterally. Pro-
thorax equally long as mesothorax but shorter than
metathorax. In the mesothorax, the acrotergite and pre-
scutum distinctly preserved. In the mesothorax and
metathorax, the pleural sulcus associated with a pleural
inflexion running parallel to it anteriorly, and dissecting
katepisternum. Another parallel dorso-ventral inflexion
dissects epimeron. Ventral wing process (VWP) distinctly
developed, two-pronged. Wings: Nymphal wings articu-
lated, at rest held in uplifted position. The size is relatively
much larger than in any modern nymph of a comparable
developmental stage. Wing veins distinct, veinal system
protodonate and bearing short intercalated branches as in
Meganisoptera. Basalar and subalar sclerites are sclero-
tized. Thoracic legs long, robust, with patella (PAT) sepa-
rated from tibia (TI) by a deeply incised suture. Abdomen
is similar as in adults, slender and much longer than in
modern nymphs. Abdominal terga flanked by lateral ridg-
es (epicoxal pleura fused to terga?) followed by long and
narrow plates representing subcoxal-to-trochanteral pleu-
ra (exact homology uncertain). Last pleural plate bears
serial posteriorly articulated leglets, elongated and adapt-
ed as respiratory filaments. Abdominal leglets (incom-
plete) composed of tubular, movably articulated podites.
R e l a t i o n s h i p . – Dragonympha n. gen. is a young
nymph of a large adult, which bears articulated and fully
movable wings with protodonate veinal system, a long and
slanted CuP crossing, and veinal branches with short vein-
al supplements. These character states occur in the gigan-
tic †Meganisoptera, which is the most probable order of
provenance. The wings and the abdomen of Dragonympha
n. gen. are morphologically much closer to those of adults
than are those in the nymphs of modern Odonata. The la-
bial mask is strongly autapomorphic. These two character
states exclude the nymph from Odonatoclada (which in-
clude modern Odonata).
R e m a r k s . – In Paleozoic †Palaeodictyopteroida and
Ephemeroptera, the abdominal epipleura form ridged side-
lobes separated from the terga by deeply incised sutures.
In †Diaphanopterodea, which are closest to the all-ptery-
gote groundplan, ridged abdominal epicoxal pleura are
followed by three flattened coxopodal pleura: the subcoxal
(SCX), coxal (CX) and trochanteral (TR) pleuron. Ab-
dominal leglets (telopodites) are articulated distally to the
trochanteral pleuron (K
UKALOVÁ-PECK 1983; 1991; 1992,
figs. 27, 37; 2008, figs. 15, 20). But, in Dragonympha n.
gen., there is no suture separating the epicoxal pleuron
from the tergum and there are probably only two pleural
plates, to the second of which the gill filament is articu-
lated; therefore, an accurate homologization of the ab-
dominal pleuron probably includes pleural fusions and is
uncertain.
Dragonympha srokai n. sp.
Figs. 1–3
H o l o t y p e : Specimen No. ISM 004 ab, Illinois State Mu-
seum, Springfield, Illinois, USA.
D e r i v a t i o n o m i n i s : In recognition of paleontologist
Dr. S
TEVEN D. SROKA, who found the unique nymph and gener-
ously donated it to the collection of the Illinois State Museum in
Springfield, Illinois.
S t r a t u m t y p i c u m : Pennsylvanian (Moscovian).
L o c u s t y p i c u s : Mazon Creek area, Will-Kankakee
County, Francis Creek shale, Pit 11, Peabody Coal Company, Il-
linois, USA.
D i a g n o s i s . – Same as genus (monotypic).
P r e s e r v a t i o n . – The holotype of Dragonympha
srokai n. gen., n. sp. is preserved in an ironstone nodule
with some free space left inside, containing crystals of
pyrite, limonite, and kaolinite. The nodule split in two
halves along the plane of weakness containing the fossil.
The holotype is slightly 3-dimensional, with the exo-skel-
eton softened and deformed before petrification. The im-
print and counter-imprint do not match. Head segmenta-
tion, antennae, tergal sclerites, abdominal pleura, and ab-
dominal gill filaments are preserved in both halves,
thoracic pleura are much better preserved in the reverse,
and the labial mask occurs only on the obverse half of the
nodule. Note that because of these differences, some pho-
tographs do not fully match with the figures. To provide
full interpretation, a figure was made on transparent pa-
per, flipped back and forth, and the details were drawn in
two colors. This method allows one to decide which detail
in which half is better preserved, and to combine them
into one drawing as accurately as possible. This is a cus-
tomary, practical procedure for Mazon Creek 3-dimen-
KUKALOVÁ-PECK, CARBONIFEROUS DRAGONFLY NYMPHS 177
sional fossils. Consequently, the photographs of both
halves do not completely match with the combined figure!
In the past, some of my combined figures were unjustly
criticized as “inexact” and “controversial” (repeatedly by
A. G. R
ASNITSYN), even when the figured structures still
exist in living insects. Combined figures are much more
accurate and user-friendly than two non-matching mirror
images accompanied by a verbal interpretation. Figs. 1A
and 2 are composite presentations of all morphological
features in obverse and reverse halves, combined.
D e s c r i p t i o n . – Young nymph in full lateral view
preserved on two halves of an ironstone nodule shows a
segmented head, antenna, extended labial mask, thorax
with incompletely preserved legs, and seven abdominal
segments with posteriorly articulated filamentous gills.
The shape of the head, extended labial mask and the pres-
ence of respiratory abdominal filaments prove that Drago-
nympha srokai n. gen., n. sp. is a true protodonate nymph
and not a newly emerged adult in which the wings have not
yet expanded.
D i m e n s i o n s : Length without antennae 38.1 mm;
head, length 3.3 mm, height 3.8 mm; prothorax, length 2
mm; pterothorax, length 4.8 mm, height 4 mm; abdomen,
length of the fragment 28 mm; 2
nd
abdominal segment,
length 4.7 mm, height 3.4 mm.
For better clarity, the description of features below is
immediately followed by remarks, which offer broader
interpretations and additional evidence.
H e a d , c r a n i u m : For an updated evolutionary
reference model of insect head see K
UKALOVÁ-PECK (2008,
fig. 6). The epipleuron (= the first flattened segment of the
limb-derived mouthparts in Arthropoda) is fused to the
ventral end of all head terga. The labrum, mandible, max-
illa and labium in Hexapoda are formed by the entire
coxopodite; this is articulated to epipleuron by its subcoxa
and bears two endites (coxal and trochanteral) (S
HEAR et
al. 1998; K
UKALOVÁ-PECK 1998, 2008, figs. 6, 12). Acron (a
flap bearing sensory organs, not a segment) shifted far
dorsally and into tergum I and II, which became reduced
to a thin rope loop running around the eyes (as a ring) and
Fig. 1. Dragonympha srokai n. gen., n. sp., holotype (†Meganisoptera). Young nymph of a large adult. Labial mask extended, wings
uplifted (flexible, not fused to body) and streamlined with body, abdominal gills derived from segmented leglets, thorax with mul-
tiple reinforcements, anepisternum not expanded. Pennsylvanian (Moscovian), Mazon Creek, Illinois, USA. – A. Composite figure,
obverse and reverse combined. B. Obverse of the holotype, uncoated, the end of abdomen missing. – Length of the remnant without
antennae = 38.1 mm. Original.
178 PALAEODIVERSITY 2, 2009
the dorsal acron, above which they leave no noticeable
trace. Acron small and clearly delimited, bearing three
ocelli close to each other and eyes. Eyes are middle sized,
laterally placed and broadly framed by tergum III. Ventral
ends of tergum I meet ventrally from acron, abut laterally,
and fuse one to another (this fusion also brings together
epicoxal pleura and coxopodites I, fused into labrum).
Tergum I forms a large clypeus, dorsally extended and
ventrally fused to labral epipleura. Coxopodites I fuse
laterally one to another and into the labrum. Antennal ap-
pendage shifted upward into the ventral portion of tergum
II. It bears a very short scape (SCX), slightly longer pedi-
cel (CX), and narrower flagellum (TR through PT is flag-
ellated). Antennae are probably medium long, much more
robust than in modern fauna. Eye pushed deep into the
tergum III, which forms around it a distinctive strip,
broader anteriorly and above, narrower posteriorly, and
invaginated ventrally. Dorsal portion of tergum III is not
delimited and is possibly fused to tergum IV without a
suture. Ventrally, epicoxal pleuron fused to tergum III in-
vaginates inside the head and its limb appendages fuse
together to form hypopharynx. Tergum II and tergum IV
Fig. 2. Dragonympha srokai n. gen., n. sp., holotype (†Meganisoptera), morphological interpretation. Head tergum I, II, III forming
a strip surrounding the eye, tergum III not delimited above the eye, epipleuron III invaginated to bear hypopharynx, tergum+epipleuron
II and IV meet at the invagination scar. Labial mask: subdivision of CX into CX1 and CX2 serving as a hinge; ST VI flanked by SCX
fused to CX1; CX2 narrow, flanked by a protruding TR; telopodite-based palpus narrow, pointed and barbed; very broad paraglossae
with small even teeth, glossae unknown. Mesothorax: anepisternum (AS) placed between acrotergite and prescutum; pleural sulcus
subdivides katepisternum (KS); parallel pleural inflexion subdivides epimeron (EM); wings articulated and uplifted; thoracic legs
with distinct TR, PFE slanted and fused to FE, PAT slanted and fused to TI. Abdomen: boundaries between abdominal pleural plates
(EP to TR) uncertain; telopodites (PFE to PT) with movable podites serving as respiratory filaments. – Symbols: AS = anepisternum;
BA = basalare; CX, CX1, CX2 = coxa, subsegment 1 and 2; EP = epipleuron; ES = episternum; FE = femur; KM = katepimeron; KS
= katepisternum; PAT = patella; PFE = prefemur; PT = pretarsus; SCX = subcoxa; SP= spiracle; ST = sternum; TI = tibia; VWP =
ventral wing process. Original.
Fig. 3. Dragonympha srokai n. gen., n. sp., holotype (†Meganisoptera) coated with ammonium chloride. Symbols as in Fig. 2. – A.
Obverse showing labial mask, but head originally covered by matrix. B. Reverse lacking head and not matching obverse. C. A close-
up of 3A with light at a different angle. D. Labial mask: narrow palpus is steeply slanted towards horizontal paraglossa. E. Head: eye
surrounded by tergum I, II, III; tergum III narrowing into a suture ventrally from the eye; prothoracic pleuron with three invaginated
sulci. F. Labial mask in a different light angle than in 3D (protruding parts falsely appear as sunken). G. Obverse of meso- and
metathorax: acrotergite, bulging prescutum, and three pleural inflexions distinctly preserved. H. Reverse of meso- and metathorax:
Sulci shown in different light angle. Katepistermum (KS) dissected by an inflexion parallel to pleural sulcus. – Abbreviations as in
Fig. 2. Original.
KUKALOVÁ-PECK, CARBONIFEROUS DRAGONFLY NYMPHS 179
180 PALAEODIVERSITY 2, 2009
fuse together ventrally with a deep scar under tergum III,
marking invagination of epipleuron III with hypopharynx.
Tergum IV is narrow dorsally, encircling posterior part of
the eye, broadened ventrally and extended proximally to
meet epicoxal pleuron II at an invagination scar of epi-
coxal pleuron III. Tergum V + VI fully fused without a
suture (a synapomorphy of Dicondylia), sinuous, very nar-
row dorsally but broadening latero-ventrally and ventral-
ly.
R e m a r k s . – The strip-like portion of tergum III
anteriorly from the eye was first noticed and figured (with-
out interpretation) in the Permian Ephemeroptera (K
UKA-
LOVÁ-PECK 1985, fig. 34B) and †Diaphanopterodea (KUKA-
LOVÁ-PECK 1985, fig. 33B). Later, the strip was recognized
as part of tergum III (by K
UKALOVÁ-PECK 2008, figs. 6, 7).
Dragonympha srokai n. gen., n. sp. is the first fossil,
which shows that the tergum III actually makes a full loop
under the eye. A conspicuous, broad, crescent-shaped ter-
gum III also occurs ventrally and anteriorly from the eye
in †Monura (Insecta: †Monocondylia, the sister-order of
Archaeognatha) (S
HAROV 1966; KUKALOVÁ-PECK 1998,
2008, fig. 7). In all Hexapoda, the epicoxal pleuron III
(fused to head tergum III, with limb appendages III articu-
lated to it) invaginated into the head capsule under the eye
and fused together into the hypopharynx. Only interpret-
ing correctly the coxopodites forming labrum, mandible,
maxilla and labium (rather than calling them – incorrectly
– “coxae”) makes their characters useful in phylogeny. In
predatory Odonatoptera, mandibles must generate strong
bite. For this adaptation, they are broad and hypopharynx
is shifted posteriorly to make place; a strong, permanent,
socketed anterior articulation is formed, for shearing and
opening sideways like a door on two pivoting condyles;
and, some muscles are strengthened and other reduced.
This adaptation yielded in Odonatoptera five convergen-
cies shared with Neoptera, which group evolved similar
shearing mandibles to chew hard vegetation (see “Conclu-
sion” below). In contrast, in the maxillary coxopodite sep-
arate coxal and trochanteral endites (inner rami: lacinia
and galea) are shared by all Arthropoda. In all Palaeoptera
= (†Palaeodictyopteroida) + (Odonatoptera +
Ephemeroptera) they are uniquely fused together into lac-
inio-galea. In Neoptera and other Hexapoda, these endites
are separate (a symplesiomorphy: K
UKALOVÁ-PECK 2008,
figs. 6, 10, 11).
Like all other Palaeoptera, Dragonympha srokai n.
gen., n. sp. has prominent eyes. This attribute doubtlessly
coevolved with the prominently paleopteran trend of the
time toward improving flight. The head is attached to pro-
thorax, which is relatively short to very short, never elon-
gate. These are the Palaeoptera groundplan characters.
Therefore, a very small head with miniscule eyes on a long
prothorax, that W
ILLMANN (1999) interpreted as belonging
to a very large Carboniferous Ephemeroptera: †Syntonop-
terida: Lithoneura lameerei, is a neopterous remnant (pos-
sibly a young gerarid; see rebuttal continued below).
M o u t h p a r t s a n d a n t e n n a e : The clypeo-
labrum formed by limb appendages of tergum I, consist-
ing of two epicoxae I and two coxopodites I fused along
their lateral margins. Clypeo-labrum is transversely sub-
divided by secondary sulci and turned postero-ventrally,
as in modern odonate nymphs. Limb appendage II (an-
tenna) is heavier and much longer than in modern dragon-
flies (total length unknown). Antennae bear a small scape
(SCX), long pedicel (CX), and slightly shorter flagellated
antennal articles (seven articles preserved). Appendage III
(hypopharynx) fully invaginated with epicoxae III under
the eye, invisible. Appendage IV (mandible) includes co-
xopodite IV articulated by SCX (in Hexapoda, the man-
dibular telopodite is reduced); mandible is broad and ex-
tended obliquely antero-dorsally towards the secondary
permanent clypeo-tentorial condyle (an autapomorphy of
Odonatoptera). Appendage V (maxilla) is not preserved.
Appendages VI (labium) are fused with labial sternum
into a labial mask. This is divided into basal portion and
prementum by a transverse hinge. The hinge runs through
the labial coxae, subdividing them into two subsegments,
CX1 and CX2: CX1 is part of the basal portion, CX2 of the
premetum. Basal portion of the labial mask includes a long
labial sternum in the middle. This is flanked on each side
by a narrow subcoxa (SCX) fused without a suture with
long CX1. Basal portion is very narrow proximally and
broadening abruptly distally. Prementum bears a very
short and thickened base composed of fused coxal subdi-
visions CX2, flanked by two short trochanters (TR). La-
bial palp (= telopodite VI) articulated to trochanter (TR)
includes a long prefemur (PFE), weakly indicated femur
(FE), a patella (PAT) fully fused to tibia (TI) through pre-
tarsus (PT), ending in a pointed spike. Palp is equipped
with dense lateral spines. Glossae (coxendites VI) not pre-
served. Paraglossae (trochendites VI) almost as long as
palps, very broad, seemingly flat and with a sinuous inner
margin equipped with about ten teeth. Paraglossae and
palps appear to be partially fused.
R e m a r k s . – The labial mask is extended, detached
from the head and partially decomposed along the coxal
subdivision into CX1 and CX2, which functioned as a
hinge. Horizontal subdivisions in limb segments subcoxa
and coxa (marked by sutures) may be plesiomorphic (Note
that in pterygote thorax, the epicoxal pleuron is horizon-
tally subdivided into the proxalar and axalar column of
pteralia, and the subcoxal pleuron may be also subdivid-
ed.). Mandibles of Archaeognatha bear a coxa distinctly
subdivided by a suture (K
UKALOVÁ-PECK 1998, fig. 19.1;
2008, figs. 11, 12). All Arthropoda share the serially ho-
mologous (= homonomous) polyramous limb-derived ap-
pendages articulated to the epicoxal pleuron: the mouth-
parts (including labrum), antennae, legs, leglets, gonostyli,
KUKALOVÁ-PECK, CARBONIFEROUS DRAGONFLY NYMPHS 181
cerci, wings, winglets, vesicles, and genitalia. By mono-
phyly, these are all flawlessly derivable from a single an-
cestral polyramous limb model shared by all Arthropoda
(K
UKALOVÁ-PECK 2008, fig. 1). Evolutionary steps by which
they are derived (fusions, reductions, modifications) pro-
vide numerous reliable higher-level arthropod and hexa-
pod synapomorphies in the morphological system (K
UKA-
LOVÁ-PECK 1998).
P r o t h o r a c i c s p i r a c l e : The placement of this
non functional spiracle in the intersegmental membrane
(rather than on a plate within the mesothorax) is consid-
ered here a plesiomorphy.
Prothorax and prothoracic winglets:
Lateral cervical sclerite large. Prothoracic acrotergite and
prescutum crescent shaped, also relatively large. Prono-
tum is large and well sclerotized, with a protruding me-
dial ridge. Narrow prothoracic winglets fully fused to
pronotum, separated from it by a deeply incised suture.
Prothoracic pleuron very small, bearing a narrow epister-
num and an epimeron about twice as broad as episternum.
Pleural sulcus simple, not accompanied anteriorly by
pleural inflexion. Katepisternum and katepimeron are
missing.
R e m a r k s . – Prothoraces in Paleozoic Palaeoptera
nymphs often bear small crescent-shaped prothoracic win-
glets, which in adults become larger and fully veined.
These may occasionally retain all principal veinal sectors,
branched and lacking any fusions and braces, as in the
ancestral protowing venation shared by all Pterygota (K
U-
KALOVÁ-PECK & LAW R ENCE 2004, fig. 1). At the groundplan
level, sister lineages Odonatoptera and Ephemeroptera
share a short, winged prothorax, in nymphs as well as in
adults. The large Carboniferous mayfly Lithoneura
lameerei (†Syntonopterida; interpreted by K
UKALOVÁ-PECK
1985, figs. 11–13) bears a short prothorax with relatively
large wings preserved only in outline. W
ILLMANN (1999)
erroneously replaced it with a narrow neopteran protho-
rax, which accidentally settled nearby, and provided an
inadequate figure of this well preserved fossil. However,
the veinal fusions and braces, exactly as figured by me in
1985, are recognizable in good photographs that W
ILL-
MANN (1999) published with his emendation, but obviously
did not recognize himself. These are all shared with an-
other syntonopterid mayfly published in the same 1985
paper, Bojophlebia prokopi. Bojophlebia was accepted by
W
ILLMANN (2007) as the oldest known mayfly without any
doubt. Basal Carboniferous mayflies are very important
for an understanding of Odonatoptera since they help to
pinpoint and verify the groundplan characters (see the
analysis of sistergroup below). Large prothoracic winglets,
such as those present in Lithoneura, were until 1985 erro-
neously believed to be present only in †Palaeodictyoptera.
The 1985 paper introduced them for the first time as pres-
ent in several nymphs and adults. Perhaps, this novelty
alone is the reason that a well-preserved and phylogeneti-
cally important mayfly showing clearly all groundplan
character states of Ephemeroptera and Syntonopterida, is
being introduced in some publications as an “unresolved”
specimen. In any case, it is always the groundplan that
matters (H
ENNIG 1969, 1981). In quite the same line of
logic, yet another syntonopterid, the gigantic nymph very
probably of Bojophlebia prokopi (also with prothoracic
winglets and also perfectly fitting the Ephemeroptera
groundplan) was redescribed by K
LUGE (1994) erroneously
as a “silverfish” named after this author. The Bojophlebia
nymph died belly up, showing ventral abdomen ending in
three characteristically thin “tails”, bearing serial leglets
and very large and rounded (after preparation) abdominal
winglets with strong anterior margins, stacked like card
one upon another so that they are mostly shown only as
crescent-shaped slivers (unless prepa