N,

48 i
tg GN ie ne
He

ts
Ve eat

RER

Ban

7

"
ie

H

ae

nM

pene

‘tj

nat
AL

- Sal
er: nalen
i:

ie

THE EXTERNAL MALE GENITALIA
AND THE PHYLOGENY OF BLATTARIA
AND MANTODEA

by

KLAUS-DIETER KLASS

BONNER ZOOLOGISCHE MONOGRAPHIEN, Nr. 42
1997

Herausgeber:
ZOOLOGISCHES FORSCHUNGSINSTITUT
UND MUSEUM ALEXANDER KOENIG
BONN

BONNER ZOOLOGISCHE MONOGRAPHIEN

Die Serie wird vom Zoologischen Forschungsinstitut und Museum Alexander Koenig
herausgegeben und bringt Originalarbeiten, die für eine Unterbringung in den „Bonner
zoologischen Beiträgen” zu lang sind und eine Veröffentlichung als Monographie
rechtfertigen.

Anfragen bezüglich der Vorlage von Manuskripten sind an die Schriftleitung zu richten;
Bestellungen und Tauschangebote bitte an die Bibliothek des Instituts.

This series of monographs, published by the Zoological Research Institute and Museum
Alexander Koenig, has been established for original contributions too long for inclu-
sion in „Bonner zoologische Beiträge”.

Correspondence concerning manuscripts for publication should be addressed to the
editor. Purchase orders and requests for exchange please address to the library of the
institute.

UInstitut de Recherches Zoologiques et Museum Alexander Koenig a Etabli cette serie
de monographies pour pouvoir publier des travaux zoologiques trop longs pour étre
inclus dans les „Bonner zoologische Beiträge”.

Toute correspondance concernante des manuscrits pour cette série doit étre adressée a
l’éditeur. Commandes et demandes pour échanges adresser a la bibliotheque de l’insti-
CUES SY AV. Ds

BONNER ZOOLOGISCHE MONOGRAPHIEN, Nr. 42, 1997
Preis: 85,— DM
Schriftleitung/Editor: G. Rheinwald

Zoologisches Forschungsinstitut und Museum Alexander Koenig
Adenauerallee 150—164, D-53113 Bonn, Germany
Druck: JFe CARTHAUS, Bonn

ISBN 3-925382-45-3
ISSN 0302-671 X

THE EXTERNAL MALE GENITALIA
AND THE PHYLOGENY OF BLATTARIA
AND MANTODEA

by

KLAUS-DIETER KLASS

BONNER ZOOLOGISCHE MONOGRAPHIEN, Nr. 42
1997

Herausgeber:
ZOOLOGISCHES FORSCHUNGSINSTITUT
UND MUSEUM ALEXANDER KOENIG
BONN

Die Deutsche Bibliothek — CIP-Einheitsaufnahme

Klass, Klaus-Dieter:

The external male genitalia and the phylogeny of blattaria and mantodea /
by Klaus-Dieter Klass. Hrsg.: Zoologisches Forschungsinstitut und Museum

Alexander Koenig, Bonn. — Bonn: Zoologisches Forschungsinst. und Museum
Alexander Koenig, 1997

(Bonner zoologische Monographien ; Nr. 42)

ISBN 3-925382-45-3

Manuscript received August 1996

CONTENTS

PAINS tll Ge N ee ariel ee RE ea een ia
FAN SAIMENKASSUN EEE ee ee ln Ele BE Tee engere re ee te lea
I, Ininedieiion: esse SNS Hoe ee es WEN 28 EN EEE EIER EINER EAN RIESEN
NETO NEGATES TUES I ee nD te tse Mn Rr RR RA a
2. Material and mes ee ee ha, cece aca A manta Nal ee
3. General description and morphological discussion of the postabdomen and of the male
zenial FEIN OE IDWS OPO 5a a el EN ce
ER BWIhegcutieulargelementser se ran ir. N RR N EN we es
2.2, Te SANS. ee SS NER OLY cere nee
&, Terminologies at <A DEE V AAU ONSH ee ee Coduayey bows ohne tua wake
Zalesslihesierminolosyzior. the selerites of the phallomere complex >. .-2.25......::
Aye Nobteviations for other selerites/ofithe postabdomen . „2... .........2......
OM CmehMNNOLO Sy tonne Tormative Clements =... oes bee oe ee ee es
Rm CeLC HIN OLO SW torsthesmuseles Fr Seen nn... nem cl ecg
PUI CHCKMNOLOcymOnine artlCUlatlOMSy eats. to. ce te Bosal ene onen.
Zus unveyzolsthestenminologies used Yurnıza. <a bie Ce ete kee ee ee ae 8
ATs NODE VIITIOMS TY WAYS, ERS UN RAS PES ZILA 00.1. u. eee an ea ne ee a
DeDeschHpuonszoisphallemeres.and postabdomma = 2.526 440-.--.-- 0452505522.
lems prodromantisesps (Mantodea, Mantidae)) 25.5. ....-.5.-.2. 82205. 55s sss:
ye liciolyicus violaceus (Mantodea, Metallyticidae) 2. -2.-......:--.-4.....
DomGniacicessa caudata(Nantodea, Chaeteessidae) 3 02s. 0... soe. ee.
Dee antoiaa schradert (Mantodea, Mantoididae)i 2. 2.2. 2. ee ee ae
DS \rchiblattarhoeveni.(Blattania, Blattidaes Blattmac) {25-55 5-55.25+-.+-45..-
5.6. Eurycotis floridana (Blattaria, Blattidae, Polyzosteriinae) ...................
Dy PUB ONICuS pan us (Blattanta., Blattidaes Mnyonicimae) 9 2 2.2. f.4 4426425454050 0-
5.8. Polyphaga aegyptiaca (Blattaria, Polyphagidae, Polyphaginae) ...............
eon Cm piocercusipunciulatus (Blattana, Cryptocercidae) 22.2... -.2. 624-05 a:
5.10. Lamproblatta albipalpus (Blattaria, Blattidae, Lamproblattinae)..............
DalerAnaniccia sp. Blattaniar Blattellidaes Anapleetinae) 9.5.2... 022. 2. 2222.
SAN ahublattellassp. (Blattania, Blattellidae, Plectoptermae) ... 22.2... 2.2.22.
BBenunsoblaitanlatan Blattanar Blattellidas Blattellinae) 222. mn menu...
SAILwRlabenusseranüferBlattania, Blabemdae)) > 2. 5.55. fae a ee ok
De MUIR CIASECICS sey ee ot wa gtrh are sOt ates Week he a dee ae ha
Oma omolorvanclatonsnand character stateS= ye nn. ce ee ee et ee
6.1. Left complex I: Main sclerites LI and L6 and associated elements ............
02a eeiacomplexs1l-2Main’seleriteE2 andlassociatedelements 7. 2... 2.2.2.2...
6.3. Left complex III: Main sclerites L4 and L10 and associated elements ..........
Or eetmcomplexmly- Mamrsclente E3 andiassociated elements 95.022 .....4.../..
Gmambeiacomplexs VA kurtnher main sclerites and muscles). 94.440. 5... 4.
6.6. Left complex VI: The position of the phallomere-gland opening ..............
OS Aulihegelementszofäthernishtphallomere su. u. cats le ose se sk es
6.8. The muscles connecting the left complex and the right phallomere ............
On malihesphallomenro=stemalMUSCleS ops u 0 u. ote N Shai Wem olde ae
CmlOeMhessubsenitalaplateyand associated stmetures 2.02 545425555 000004.
Oallnespenipheralmusclesn rw laut etait ae ee
Gar Suihestermmaltpantsoisthe abdomen nn nun
OaleRaliherasymmenyzotsthesphallemerercomplex Sn nenn

WwW N

NYY DD WY WL
“SW

~

267

7. The ground-plan and the evolution of the phallomere complex and the phylogeny
of, Blattaria-and: Mäntodea.-.. ......er-:..22 = Ses eee ee ER
7.1. The common ground-plan of the phallomere complex of Blattaria and Mantodea
7.2. The evolution of the phallomere complex and the phylogeny in Mantodea
(= subgroup 1). - en es oe See er
7.3. The evolution of the phallomere complex and the phylogeny ın Blattaria
(= subgroup 2.) 2-22 28 Bar ee ne ee ee eer
742 Survey2oß phylogeny and aut/synapomorphiese Sr ei erence ee
7.5. Remarks onthe polarity and ecvolutiontof some characters 2.2 2 cree
7.62. Contlicts mytherdistibuttonnotecharactemstate sar: tne een a
7.7. Conelusions anitemms of phylogeny as ee eee
7.8. Conclusions in terms of the side-reversal of the phallomere complex ..........
7.9. Remarks on the procedure in the phylogenetic analysis and on character lists
and.charaeter state qmatrices' . . ... 2... ve 2 ea ee eee eee
8. Homology relations according to Mizukubo & Hirashima (1987) and general remarks
onstheranalysis-of homologyzrelationse. 2 2.2 Da 5 aoe ee
9. Homology relations according to Grandcolas (1994) and the phylogenetic position of
Cryplöcercus:.. 2 2022 20 a ee es Se eee ee,
9.1. Discussion of the homology relations assumed by Grandcolas ...............
9:2: The phylogenetic: position of Enypiocercus “215405 5 oe eee
Conclusions... 36 She We ee i ae ice N oes

ABSTRACT

The external male genitalia of Blattaria and Mantodea (phallomeres, phallomere complex)
are highly complicated structures, which are always extremely asymmetrical. They are
provided with many sclerites and muscles. Their cuticular surface is complexly folded,
and there are many distinct in- and evaginations (the formative elements), which may have
the shape of spines, lobes, bulges, pouches, apodemes, tendons, etc.. The knowledge of
phallomere morphology is extremely incomplete, and the potential for phylogenetic
research inherent in these structures has so far hardly been used.

In 4 species of Mantodea and 10 species of Blattaria the sclerites, muscles, and formative
elements of the phallomere complex and some other parts of the male postabdomen have
been investigated in detail. Most of the subgroups of Blattaria (subfamilies in the system
of McKittrick 1964) and four families of Mantodea (of the system of Beier 1968) are
represented in this sample. Certain parts of the phallomeres are described for some further
species of Blattaria.

A detailed homology hypothesis is presented for the sclerites, muscles, and formative
elements of the phallomeres, which includes the homologies between Blattaria and
Mantodea. The common ground-plan of Blattaria and Mantodea has been reconstructed.
Phallomere characters have been evaluated in terms of phylogeny.

The resulting phylogenetic hypothesis is roughly as follows: In Mantodea, the basal
dichotomy is between Mantoididae and the other families; the second one is between
Chaeteessidae and the remaining families. In Blattaria, the basal dichotomy is between
Blattinae + Polyzosteriinae and the remainder. These remaining Blattaria can be divided
into three groups: The first consists of Tryonicinae only. The second contains Cryptocerci-
dae as well as Lamproblattinae and Polyphaginae, the two latter taxa being especially
closely related. The third group comprises Blattellidae and Blaberidae. Blattellidae are
clearly paraphyletic, with Blaberidae as a rather subordinate subgroup. The first offshoot
within Blattellidae (+ Blaberidae) are the Anaplectinae. The subsequent offshoots are
various species of Plectopterinae, which is a paraphyletic taxon, too. Blaberidae,
Nyctiborinae, Blattellinae and Ectobiinae together form a holophyletic group. Nyctiborinae
and Blaberidae are possibly sister-groups.

Some other important results are:

(1) The asymmetry of the phallomere complex is homologous in Blattaria and Mantodea,
and the morphology of each side is quite similar in the two groups. In Mantodea the hook-
process hla (sclerite L3 of McKittrick 1964) is missing; this might be the consequence of
a derived copulation procedure.

(2) In the common ground-plan of Blattaria and Mantodea asymmetry is already as extreme
as in the extant species. The opinion of Mizukubo & Hirashima (1987) that the stem-
species of Blattaria still had symmetrical phallomeres is refuted.

(3) The ground-plan morphology is most extensively retained in the Mantodea Mantoididae
(left side) and Chaeteessidae (right side). In Blattaria, Blattinae have retained many
ground-plan features, but in some other phallomere characters they are rather derived. The
phallomeres of Cryptocercidae are not close to the Blattarian ground-plan as it is the
opinion of McKittrick (1964).

6

(4) The hypothesis of Bohn (1987) that the side-reversed similarities of the phallomeres
of Blaberidae on the one hand and of some subgroups of Blattellidae on the other are due
to homology is highly supported. They are not due to parallel evolution as it is the view
of Mizukubo & Hirashıma (1987).

(5) Homologies between the left and the right side of the phallomere complex can be
recognised in only very few respects. Probably, most of the complex morphology of the
phallomeres has evolved when asymmetry had already been established. The concept of
side-homologous subregions in Mizukubo & Hirashima (1987) and the assumptions of
side-homologies in Grandcolas (1994) are refuted.

ZUSAMMENFASSUNG

Die äußeren männlichen Genitalien der Blattaria und Mantodea (Phallomeren,
Phallomerenkomplex) sind äußerst komplizierte Strukturen. Sie sind mit vielen Skleriten
und Muskeln ausgestattet und sind immer vollig asymmetrisch. Ihre Cuticula ist stark in
sich verfaltet, und es finden sich viele markante Ein- und Ausstülpungen (formative
Elemente), die die Form von Stacheln, Lappen, Beulen, Taschen, Apodemen, Sehnen 0.4.
haben können. Die Phallomeren waren bislang noch kaum vergleichend-morphologisch
bearbeitet, und die Möglichkeiten, die sie der Phylogenie-Forschung bieten könnten,
wurden bislang noch kaum genutzt.

An 4 Arten der Mantodea und 10 Arten der Blattaria wurden die Sklerite, Muskeln und
formativen Elemente des Phallomerenkomplexes sowie einige weitere Teile des
Postabdomens eingehend untersucht. Die meisten höherrangigen Teilgruppen der Blattaria
(Unterfamilien im System von McKittrick 1964) und vier Familien der Mantodea (nach
dem System von Beier 1968) sind in dieser Auswahl vertreten. Teilbereiche der Phallo-
meren wurden an weiteren Arten der Blattaria untersucht.

Für die Sklerite, Muskeln und formativen Elemente der Phallomeren wird eine detaillierte
Homologiehypothese vorgestellt, die auch die Homologiebeziehungen zwischen Blattaria
und Mantodea einschließt. Der gemeinsame Grundbauplan der Blattaria und Mantodea
konnte weitestgehend rekonstruiert werden. Die Phallomeren-Merkmale wurden in
Hinblick auf die Phylogenie ausgewertet.

Aus den Merkmalsverteilungen ergibt sich die folgende Phylogenie-Hypothese: Innerhalb
der Mantodea besteht die basalste Dichotomie zwischen Mantoididae und den restlichen
Familien, die nächstfolgende zwischen Chaeteessidae und den verbleibenden Familien.
Innerhalb der Blattaria ist eine basale Schwestergruppenbeziehung zwischen Blattinae +
Polyzosteriinae und den ganzen restlichen Gruppen anzunehmen. Diese restlichen Blattarıa
lassen sich in drei Gruppen gliedern: Der ersten gehören nur die Tryonicinae an. Die zweite
Gruppe umfaßt die Cryptocercidae, Polyphaginae und Lamproblattinae, wobei die beiden
letzteren besonders enge Beziehungen zeigen. Der dritten Gruppe sind die Blattellidae und
die Blaberidae zuzurechnen. Die Blattellidae sind eindeutig paraphyletisch: die Blaberidae
sind eine untergeordnete Teilgruppe dieser Familie. Innerhalb der Blattellidae (+
Blaberidae) sind die Anaplectinae der basalste Seitenzweig. Die nachfolgenden
Abzweigungen werden von verschiedenen Vertretern der Plectopterinae repräsentiert,

womit auch dieses Taxon als paraphyletisch anzusehen ist. Blaberidae, Nyctiborinae,
Blattellinae und Ectobiinae bilden gemeinsam eine holophyletische Gruppierung.
Nyctiborinae und Blaberidae sind möglicherweise Schwestergruppen.

Weitere bedeutsame Ergebnisse sind:

(1) Die Asymmetrie des Phallomerenkomplexes ıst bei Blattarıa und Mantodea homolog,
und die Morphologie jeder Seite ist ziemlich ähnlich. Bei Mantodea fehlt der Hakenfortsatz
hla (L3-Sklerit in McKittrick 1964), was sich vielleicht als Folge eines apomorphen
Kopulationsverhaltens interpretieren läßt.

(2) Im gemeinsamen Grundbauplan der Blattaria und Mantodea ist der
Phallomerenkomplex bereits im selben Ausmaß (und in derselben Weise) asymmetrisch
wie bei den rezenten Vertretern. Die Ansicht von Mizukubo & Hirashima (1987), die
Phallomeren seien bei der letzten gemeinsamen Stammart der Blattaria noch symmetrisch
gewesen, wird zurückgewiesen.

(3) Die Morphologie des Grundbauplans wird am umfangreichsten bei den Mantodea
Mantoididae (linker Teil) und Chaeteessidae (rechter Teil) beibehalten. Innerhalb der
Blattaria sind besonders bei den Blattinae viele Merkmale des Grundbauplans erhalten, in
manchen Merkmalen der Phallomeren ist diese Gruppe allerdings bereits stark abgeleitet.
Die Meinung von McKittrick (1964), daß die Phallomeren der Cryptocercidae dem
Grundbauplan der Blattaria besonders nahestünden, ist abzulehnen.

(4) Die Hypothese von Bohn (1987), daß die seitenverkehrten Ähnlichkeiten der
Phallomeren der Blaberidae und mancher Teilgruppen der Blattellidae als Homologien
anzusehen sind, wird umfassend bestätigt. Eine Entstehung dieser Ähnlichkeiten durch
parallele Evolution, wie sie Mizukubo & Hirashima (1987) annehmen, erscheint äußerst
unwahrscheinlich.

(5) Homologien zwischen linker und rechter Seite des Phallomerenkomplexes lassen sich
nur bezüglich sehr weniger Elemente begründen. Es ist zu vermuten, daß ein großer Teil
der komplexen Phallomeren-Morphologie erst ausgebildet wurde, als bereits eine deutliche
Asymmetrie etabliert war. Das Konzept seitenhomologer Subregionen von Mizukubo &
Hirashima (1987) und auch die Annahmen von Seitenhomologien in Grandcolas (1994)
lassen sich widerlegen.

1. INTRODUCTION

Blattaria, Mantodea, and Isoptera form a holophyletic group called Dictyoptera (s. lat.;
Kristensen 1995) or Blattopteroidea (Hennig 1969). The relationships between the three
groups are unresolved. Hennig (1969) regards Mantodea as the sister-group of Blattaria +
Isoptera, and he points out the possibility that Isoptera might be a subgroup of Blattarıa
— closely related to the Blattarian family Cryptocercidae. These assumptions probably
reflect the most parsimonious solution, but the arguments are scarce. Thorne & Carpenter
(1992) assume that Isoptera are the sister-group of Blattaria + Mantodea. However, their
results are not very convincing, since for many characters the assumed polarities are
questionable (Kristensen 1995; Klass 1995).

Concerning the internal phylogeny of Blattaria and of Mantodea, the current ideas are
based on the extensive investigations of McKittrick (1964) and McKittrick & Mackerras
(1965) (Blattaria) and on the survey in Beier (1968) (Mantodea). In terms of systematics,
these authors will be followed in this paper. In some aspects the ideas of these authors
are well-founded, but many points are still debatable.

Beier (1968) divides the Mantodea into 8 families: Chaeteessidae, Metallyticidae, Manto-
ididae, Amorphoscelididae, Eremiaphilidae, Hymenopodidae, Mantidae, and Empusidae.
These are not grouped into higher-ranked categories. Chaeteessidae are more primitive
than all other families in that their hind-wings have a complete second anal-vein and in
that their fore-legs are beset with stout setae rather than thorns. In Metallyticidae the
second anal vein is vestigial, and in the other families the vein is completely missing.
Hence, Chaeteessidae seem to be the first offshoot and Metallyticidae the second.

McKittrick (1964) divides the Blattaria into two sister-groups, Blattoidea and Blaberoidea.
The Blattoidea, comprising Cryptocercidae and Blattidae, do not reveal a single feature
that could be unambiguously regarded as a synapomorphy of the two families. The
phylogenetic position of Cryptocercidae has been reanalysed by Grandcolas (1994), who
assumes that they are a subgroup of Polyphaginae. However, this assignment is also not
very convincing, since many of the homology assumptions upon which this assignment is
based are debatable. The Blattidae of McKittrick, comprising Blattinae, Polyzosteriinae,
Tryonicinae, and Lamproblattinae, are based on features most of which can be suspected
to be plesiomorphic for Blattaria, and the family might be para- or even polyphyletic. The
Blaberoidea of McKittrick, including Polyphagidae, Blattellidae, and Blaberidae are
founded on the presence of a pair of special compound sclerites in the ovipositor, the
crosspieces. However, crosspieces are simple gonangula strictly homologous with those
of the other Blattaria (Klass 1995, in press), and the holophyly of Blaberoidea is thus
highly questionable. McKittrick’s assumption that Blattellidae and Blaberidae are closely
related is well-founded. However, the exact relations between the two families are
uncertain. Interpreting the morphological results of McKittrick (1964) concerning the male
and female genitalia from the viewpoint of phylogenetic systematics and parsimony, the
Blaberidae would have to be regarded as a rather subordinate subgroup of Blattellidae;
however, not all features are consistently supporting this view. The relations between the
various subgroups of Blattellidae, which are Anaplectinae, Plectopterinae, Blattellinae,
Ectobiinae, and Nyctiborinae, are also rather unclear.

The external male genitalia of Blattaria and Mantodea (the phallomere complex composed
of the phallomeres) have a highly complicated morphology. The knowledge of these
structures is extremely scarce. However, a large potential for phylogenetic research can
be supposed to be inherent in them, and this will be used in this paper to contribute to
the solution of the basic problems of Blattarian and Mantodean phylogeny.

The phallomere complex, or at least its anterior part, is concealed within a genital pouch.
Abdominal sternite 9 is a saucer-shaped subgenital plate, and the pouch is mainly formed
from the intersternal membrane between the sternites 9 and 10, which is deeply invaginated
anteriad. The ejaculatory duct opens into this pouch, and the phallomeres are evaginations
surrounding the genital opening. The phallomere complex is provided with many sclerites

9

and muscles, and with many distinct in- and evaginations of the cuticle (formative elements
such as processes, lobes, pouches, apodemes, or tendons). The structure as a whole is
always completely asymmetrical. Its morphology is quite variable within Mantodea
(LaGreca 1954) and extremely so within Blattaria (McKittrick 1964). In Isoptera the phal-
lomeres are said to be missing (Weidner 1970), a situation which is, according to Matsuda
(1976), an element of the neotenic traits generally observable within this taxon.

The phallomere morphology of Mantodea has been studied by Snodgrass (1937) in
Tenodera sinensis, by Levereault (1936, 1938) in Stagmomantis carolina, and by
LaGreca & Rainone (1949) in Mantis religiosa. In each of these studies the cuticular
elements and the musculature are described. The three species are closely related
(Mantidae, Mantinae, Mantini in Beier 1968), and their phallomeres are rather similar.
LaGreca (1954) compared the cuticular elements of the phallomeres of several Mantodea.
In this selection all families of Beier (1968) are represented, except for those regarded as
most primitive: Chaeteessidae, Metallyticidae, and Mantoididae.

Regarding the phallomeres of Blattaria, the paper of Snodgrass (1937) is the most
important of the earlier contributions. The sclerotisations and — in part — the musculature
of Periplaneta americana, Blatta orientalis (both: Blattidae, Blattinae), Blattella
germanica (Blattellidae, Blattellinae), and Ectobius lapponicus (Blattellidae, Ectobiinae)
are described. The phallomeres of Blattidae and Blattellidae are very different from each
other, and assumptions on homology relations are made only to a very small extent.

McKittrick (1964) investigated the phallomere sclerites in 24 genera of Blattaria and gives
a homology hypothesis. However, the descriptions are not very detailed, and the
musculature has not been studied. Thus, this homology hypothesis is not very convincing
in many points. McKittrick introduced a new terminology for the phallomere sclerites:
The terms are composed of several letters and numbers, each giving some information
about the position and the homology relations of the respective sclerite. McKittrick regards
the phallomeres of Cryptocercus (Cryptocercidae) as primitive within Blattaria.

Grandcolas (1994) studied the phallomere sclerites of Cryptocercus and some
Polyphaginae and Blattinae. He finds many synapomorphies for Cryptocercus and
(subgroups of) Polyphaginae and assigns Cryptocercus to Polyphaginae. However, the
homology relations assumed for the sclerites are disputable in many cases.

Mizukubo & Hirashima (1987) studied the phallomere sclerites and muscles of Periplaneta
(various species; Blattinae), Blattella (various species; Blattellinae), and Opisthoplatia
orientalis (Blaberidae). They use — with some modifications — the terminology of
McKittrick. The authors homologise the phallomere sclerites according to their relative
positions and their mutual relations. Furthermore, they introduce a new topic into the
discussion: Homologies between elements of the left and of the right half of the phallomere
complex are considered. In their analysis they dismiss the musculature as a valuable
reference system for homologising. The phallomeres of the stem-species of Blattaria
(excluding Mantodea) are supposed to be still symmetrical. Accepting this view, the
asymmetry of the phallomere complex would have to be regarded as non-homologous in
Blattaria and Mantodea. In the case of Blattellidae and Blaberidae, which families show
obvious but side-reversed similarities in their phallomere morphology, Mizukubo &

10

Hirashima assume a parallel evolution of these similarities, and they also assume a still
symmetrical morphology in the common stem-species of these two families. In McKittrick
(1964), the same view is indicated through the designation of the sclerites. In contrast,
Bohn (1987) supposes that the phallomeres of Blaberidae have undergone a change of
their left-right-asymmetry and that the similarities concerned are homologous. The point
of discussion is the same in the case of Plectopterinae, whose phallomeres also show side-
reversed similarities with the other subfamilies of Blattellidae.

The knowledge of the morphology of the phallomeres and the other parts of the male
postabdomen of Blattaria and Mantodea is extremely incomplete. The few hypotheses
concerning homology relations (between species and between the left and right halves of
the phallomere complex), the ground-plan, and the evolution of the phallomeres are not
very convincing. Furthermore, nothing is known about homology relations between the
phallomeres of Blattaria and Mantodea. Thus, there is compelling need for a large-scaled
morphological investigation of the phallomere complex.

To do this is the intention of this paper. Phallomere morphology will be analysed in
Blattarian and Mantodean species representing the various subgroups. This investigation
should be as detailed as possible in order to avoid misinterpretations due to superficial
observation and in order to get as many arguments as possible for assumptions and
conclusions. The homology relations between the various species will be worked out in
detail, and possible homologies between the left and the right side of the phallomere
complex will be considered. The ground-plan features of the phallomeres will be recon-
structed — focused on the common ground-plan of Blattaria and Mantodea, if there is one.
The special condition of the phallomere elements in the various species and their evolution
will be discussed in detail. The characters of the phallomeres will be evaluated in order
to establish a phylogenetic hypothesis for Blattaria and Mantodea. The terminology for
the phallomere elements will be based on the common ground-plan of Blattaria and
Mantodea — in accordance with the homology relations assumed. A standardised, well-
founded, and well-defined use of the terminology might also be valuable for taxonomic
research and the description of species.

ACKNOWLEDGEMENTS

I wish to thank Horst Bohn from the Zoologisches Institut der Ludwig-Maximilians-
Universität, Munich, for giving me the opportunity to do my investigations in his
laboratory, for supplying me with the animals necessary for my studies, and for
commenting on the manuscript. Greatly appreciated in terms of the English language was
also the help of Teresa Saks from the same institution. Ulrike Aspöck from the
Naturhistorisches Museum, Vienna, Christine A. Nalepa from the North Carolina State
University, Raleigh, and Louis R. Roth from the Museum of Comparative Zoology of
Harvard University are acknowledged for kindly providing me with specimens of various
Blattaria and Mantodea.

11

2. MATERIAL AND METHODS

Species investigated

Mantodea
Mantoididae Mantoida schraderi Rehn, 1951
Chaeteessidae Chaeteessa caudata Saussure, 1871
Metallyticidae Metallyticus violaceus Burmeister, 1838
Mantidae Sphodromantis Stal, 1871 (sp. indet.); Mantis religiosa (Linné, 1758)

Blattaria
Blattoidea

Cryptocercidae Cryptocercus punctulatus Scudder, 1862

Blattidae
Blattinae Archiblatta hoeveni Sn. v. Vollenhoven, 1862; Blatta orientalis Linné,
1758; Deropeltis Burmeister, 1838 (sp. indet.); Periplaneta americana (Linné,
1758)
Polyzosteriinae Eurycotis floridana (Walker, 1868)
Tryonicinae Tryonicus parvus (Tepper, 1895); Tryonicus angustus (Chopard,
1924)
Lamproblattinae Lamproblatta albipalpus Hebard, 1919

Blaberoidea

Polyphagidae
Polyphaginae Polyphaga aegyptiaca (Linné, 1758); Ergaula capensis (Saussure,
1893); Ergaula capucina (Brunner v. W., 1893)

Blattellidae
Anaplectinae Anaplecta Burmeister, 1838 (sp. indet.)
Plectopterinae Nahublattella Bruijning, 1959 (sp. indet.); Euphyllodromia
angustata (Latreille, 1811); Supella longipalpa (Fabricius, 1798)
Blattellinae Parcoblatta lata (Brunner v. W., 1865); Loboptera decipiens
(Germar, 1817)
Ectobiinae Ectobius sylvestris (Poda, 1761)
Nyctiborinae Nyctibora Burmeister, 1838 (sp. indet.)

Blaberidae Blaberus craniifer Burmeister, 1838; Byrsotria fumigata (Guérin-

Meneville, 1857); Blaptica Stal, 1874 (sp. indet.); Nauphoeta cinerea (Olivier,

1789)

The assignment of the respective genera to the various taxa is adopted from McKittrick
(1964) and Beier (1968). The only exceptions are: The assignment of Tryonicus is taken
from McKittrick & Mackerras (1965). Polyphaga and Ergaula are assigned to
Polyphaginae according to Grandcolas (1994). Euphyllodromia is assigned to
Plectopterinae by Roth (1967). Archiblatta is in its phallomere morphology rather close
to Deropeltis (McKittrick 1964, fig.108), which is assigned to Blattinae by McKittrick
(1964). Nahublattella is in its phallomere morphology very close to Lophoblatta
(McKittrick 1964, fig.113), which is assigned to Plectopterinae by McKittrick (1964).

In the subsequent text these species will be named by their generic name alone. (Tryonicus

12

is always T. parvus, Ergaula is always E. capensis. T. angustus and E. capucina will be
addressed by their complete names).

Sphodromantis, Mantis, Deropeltis, Periplaneta, Blatta, Eurycotis, Supella, Parcoblatta,
Loboptera, Blaberus, Nauphoeta, Blaptica, Byrsotria, Ergaula capucina, and Polyphaga
have been reared in the laboratory and were available as freshly killed specimens.
Lamproblatta, Anaplecta, Nahublattella, Euphyllodromia, Nyctibora, and Mantoida have
been stored in 4% formaldehyde, Cryptocercus and Ectobius in 70% isopropanol.
Chaeteessa, Metallyticus, Archiblatta, the two Tryonicus-species, and Ergaula capensis
were dried specimens.

Preparation

For the examination of the musculature the abdomina were cut off and stored in 75%
isopropanol for at least three days. For the study of cuticular elements the soft tissues
were removed by treating the abdomina with 10% KOH for 4-20 hrs. at 40°C. The
remaining cuticular structures were then washed in distilled water and stored in 75%
isopropanol. Descriptions of morphological structures are always based on preparations of
several specimens, with the exception of some species of which only one or two specimens
were available (Mantoida, Chaeteessa, Metallyticus, both species of Tryonicus, Ergaula
capensis, Archiblatta). Preparation was performed with sharp forceps and iris scissors. In
the observation of small and weakly sclerotised structures it was sometimes useful to
underlay the object with a piece of aluminium foil.

Remarks on the figures 1-319

— In all figures anterior is towards the top and posterior is towards the bottom of the
sheet.

— The cuticle generally has an internal surface, which is in contact with the epidermis,
and an external surface. In all figures the cuticle is partly seen from internally and partly
from externally.

— Dark areas are sclerotised; white areas are membranous.

— Muscles are hatched longitudinally, in correspondence with the course of their fibers.

— In each figure only those structures are shown which can be directly seen by the
observer and which are not covered by other structures. Thus, e.g. sclerites covered by
membrane are not shown, even if they can be easily seen through the membrane in an
original preparation.

— Mostly the cuticle is very thin, and in drawing it is regarded as a convoluted plane
without thickness. Only in some cases when the cuticle is strongly thickened its
thickness is considered in drawing.

— In drawing, the geometry of the cuticular foldings and of the other elements is strictly
held to. Continuous black lines represent edges. Edges are understood throughout as
lines along which the cuticle or the surface of a muscle curves beneath itself and thus
vanishes from the observer’s view. What appears as an edge is dependent on the angle
of view. Edges of the cuticle can be external or internal: External edges are directed to
the exterior, and along them the external surface of the cuticle is visible; internal edges
are directed to the interior of the body, and along them the internal surface of the cuticle
is visible. Edges beneath the visible surface are sometimes drawn as broken lines.

13

— The boundaries of the insertion areas of muscles are also shown by continuous black
lines if they are directly visible. Parts of these boundary lines which are covered by
the muscle itself or by other structures are drawn as broken lines. In some cases the
insertion areas alone are drawn without the respective muscles (mainly in the figures
showing the subgenital plates, e.g. fig.5); the boundary of the insertion area is again
shown by a continuous black line.

— Undulate lines are cutting lines (through cuticle of normal thickness) or bound cut
surfaces (of muscles or strongly thickened cuticle cut through).

— The series of figures pertaining to a certain species has as a whole been worked out
with the intention of showing all relative positions of all elements of the phallomeres,
including all the membranous foldings.

— The series of figures for the various species are designed for the best possible
comparability. For example, in the overall views of the postabdomina (compare fig.1,
2, 3 and fig.58, 59, 60) the cutting lines have the same course in each species (their
courses are, so to speak, homologous), and, if present, the same muscles are shown.
Or, the right phallomeres of Blattaria are always shown in the same four aspects.

Remarks on homology

As a principle, elements regarded as homologous will be given the same name, and
elements given the same name are regarded as homologous. Minor exceptions to this rule,
mainly due to a not very high probability of homology, will be explicitly mentioned in
the text.

Mainly the criteria of the relative position and of the special structure (1. and 2. major

criteria of Remane 1952) will be used in this paper. These will be applied to the following

kinds of structures, whose relative position and special structure will be comprehensively
discussed in the homology analysis:

— Sclerotised areas of the cuticle (sclerites).

— Articulations or other special relations between sclerites.

— Formative elements: more or less distinct evaginations or invaginations of the cuticular
surface of the phallomere complex (processes, ridges, pouches, tendons, apodemes,
lobes, etc.).

— Insertion areas of muscles.

— The opening of the ejaculatory duct or genital opening.

— The opening of the phallomere-gland (a gland within the left part of the phallomere
complex).

The sclerotisations of the phallomeres will be divided into areas which are strictly

homologous in the different species. These areas are the main sclerites (or sclerite groups,

if these main sclerites have split into several isolated sclerites) and — as their subunits —
the sclerite regions. Some difficulties arise in this demarcation of homologous areas within
the cuticular surface of different species and in the standardisation of this procedure. The

following example shall illustrate these problems: Provided: In two compared species A

and B homology is certain (as much as it can be) for a sclerite as well as for a muscle.

In species A the muscle inserts on the sclerite, in species B the muscle inserts on the

14

membrane next to the sclerite. The situation in A is primitive, the situation in B is derived.
The derivation which B shows as compared with A can be interpreted in two different
ways: (1) In B the insertion of the muscle has shifted from the sclerite to the membrane.
(2) In B the sclerite has diminished and has “lost” the insertion of the muscle. According
to (1) the sclerites of the two species are homologous in a strict sense. According to (2)
they are homologous only in part, since in species B part of the sclerotisation has been
lost. In one peculiar case the special circumstances can suggest an interpretation according
to either (1) or (2). In many cases, however, an objective decision in favor of one of the
two alternatives is hardly possible, and it is debatable whether a discussion of such a case
is of importance at all. The interpretation will then be done in that way which seems to
be more suitable for the explanation of homology relations.

3. GENERAL DESCRIPTION AND MORPHOLOGICAL DISCUSSION OF THE
POSTABDOMEN AND OF THE MALE GENITAL REGION OF DICTYOPTERA

The postabdomen of male Dictyoptera comprises the abdominal segments 9-11 and the
telson, which contains the anus (Snodgrass 1937). Matsuda (1976) postulates a twelfth
segment for the ground-plan of insects, and this would also affect the interpretation of the
Dictyopteran postabdomen. According to Matsuda himself, p.52, however, this “segment”
contains neither mesoderm rudiments, nor ganglion rudiments, nor appendage rudiments.
Thus, it does not fulfil either criterion to be regarded as a segment. This “twelfth segment”
could be regarded as a (highly reduced) segment, if it is demonstrated to be homologous
with a true segment (containing mesoderm) of another group of Arthropoda, having lost
its segmental organs secondarily. This, however, has not been shown. Therefore, the
twelve-segment-theory of Matsuda is not followed here.

Subsequently the general morphology of the postabdomen and the phallomeres of Blattaria
and Mantodea will be described. This will essentially be a description of the common
ground-plan of Blattaria and Mantodea, whose reconstruction will be substantiated step
by step later on in this paper.

3.1. The cuticular elements

Abdominal segment 9

The sternite of segment 9 (subgenital plate, S9 in fig.320, 321) always forms a large lobe-
like extension to the posterior, which reaches or even exceeds the morphological posterior
end of the body (with the anus Af in fig.320, 321c). Most species have a large membranous
or only weakly sclerotised area in the anterior half of the subgenital plate (M in fig.320,
321b,d,k). The (heavier) sclerotisation is continuous in antero-posterior direction only in
the lateralmost parts, to the left and to the right of area M (fig.321k). The lateral parts of
the subgenital plate (S91 in fig.321k) curve upwards. The posterior edge of the subgenital
plate bears styli (S9s in fig.320, 321b,d,k).

Along the anterior margin of the subgenital plate the intersternal membrane connecting
sternites 8 and 9 adjoins and bends back sharply to the posterior margin of sternite 8

15

(compare fig.320). A more or less extensive anterior part of the subgenital plate is thus
concealed by the posterior part of sternite 8 from ventrally and can be regarded as a broad
apodeme or internal apophysis. The paired parts which project especially deeply anteriad
(S9a in fig.321b,d,k) are the lateral sternal apodemes or apophyses; these paired parts will
be designated as the apophyses of the subgenital plate subsequently. The summits of what
I call apophyses can reasonably be regarded as homologous in the various species though
they can take various positions from far lateral (like in fig.265) to far medial (like in
fig.22). However, strict homology is certainly not true for the whole apophyses (= paired
parts): The apophyses can be separated from each other to the far posterior (like in fig.265),
the paired parts being very long, or the whole median part of the subgenital plate is
produced far anteriad and only the anteriormost parts show the paired condition (like in
fig.296). The apophyses present in the latter situation seem to correspond only to the
anteriormost parts of the apophyses present in the former situation, and probably some
median fusion has taken place in the posterior part. Hence, the term “apophysis” as used
here is not intended to claim strict homology.

According to Walker (1922) and other authors, the subgenital plate is not the sternite of
segment 9 only but is probably composed of: (1) the true sternite 9 (the part anterior to
M); (2) the paired but medially fused coxites of segment 9 — probably serially homologous
with the thoracic coxae or with more extensive basal parts of the thoracic appendages. If
this composition is true, the subgenital plate is a coxosternite. The styli S9s sitting upon
the coxites are probably serially homologous with distal parts of the thoracic legs.

The tergite of segment 9 (T9 in fig.320, 321a) resembles the more anterior tergites, but
like tergite 8 it is rather short. Its lateral parts (= paratergites, T9p in fig.321b) incline
ventrad from the dorsal main part — often along a distinct edge. The ventral margins of
the paratergites overlap in most cases, and to a varied extent, the lateral parts of S9.

The position of the phallomere complex

The intersternal membrane between sternites 9 and 10 is deeply invaginated anteriad to
form the walls of the funnel-shaped genital pouch: The ventral wall of this pouch extends
anteriad from the lateral and posterior edges of the subgenital plate (Vw in fıg.321k, left
half, fig.320). The dorsal wall extends anteriad from the anterior margins of the paraprocts
Pp and Pv-sclerites (Dw in fig.320, 321b). The lateral walls extend anteriad from the
posterior edge of the pleural membrane between tergite 9 and subgenital plate (Sw in
fig.321b,d).

Deep in the genital pouch the cuticle turns posteriad again and forms the walls of the
phallomere complex. The edge or line of turning, along which the walls of the genital
pouch meet the walls of the phallomere complex, will be called the basal line (BI in

fig.320, 321b,d). Hence, the phallomere complex seems to be exclusively an elaboration
of the intersternal area between sternites 9 and 10. The ejaculatory duct (D in
fig.321b,d,e,g) opens on the phallomere complex.

In many Blattaria and Mantodea, the ventral wall of the genital pouch (Vw in fig.320,
321k), which covers the posterior part of the subgenital plate from dorsally, contains a
sclerotisation (S9d in fig.320, 321b,d,k). S9d is regarded as a dorsal sclerotisation of the

16

subgenital plate and is possibly the sclerotisation of the dorsal walls of the fused coxites
and hence a part of the appendages of segment 9. This S9d can be either separated from
or connected with the ventral main sclerotisation of the subgenital plate (around the lateral
and posterior edges of the plate, as in fig.321k), and it may either occupy an extensive
part of the ventral wall of the genital pouch or is restricted to the marginal areas close to
the edges of the subgenital plate. The sclerotisations comprised in S9d are certainly not
homologous in a strict sense throughout the species.

The phallomere complex

The phallomere complex will be divided into two main parts belonging to the left and to
the right half of the body: left complex and right phallomere. This major division is shown
in fig.32le and f, where the two parts are separated (compare fig.321d). Both are
complicated structures with intensively folded cuticle and with sclerotised and
membranous areas. Left complex and right phallomere are extremely asymmetrical in all
Blattaria and Mantodea. The phallomere-gland (penis-gland, conglobate gland; P in
fig.32le) opens on the left complex; at least its outlet channel is cuticulised. Since the
morphology of the phallomere complex is highly variable within Blattaria and Mantodea,
a description valid for all subgroups is impossible. The following description corresponds
to the common ground-plan of Blattaria and Mantodea. In addition, some important
derived states wıll be mentioned.

Left complex

Several left-lateral and ventral sclerites are designated L4 (fig.321e,g,i): A large crescent-
shaped L4-sclerite occupies the left edge of the left complex, including the adjacent
margins of the dorsal and ventral walls, and the anteriormost ventral wall. Along most of
this sclerite there runs an apodeme (swe in fig.32le,g), which is groove-like posteriorly
but solid and beam-like anteriorly (the groove is filled in by the cuticle becoming
thickened). The posteriormost part of the sclerite occupies a short process (pda in
fig.32le,g). As a derived condition, the dorsal part of the sclerite can be strongly expanded
to the right, and the dorsal and the ventral parts of the sclerite can be separated. A second,
plate-like L4-sclerite lies in the right ventral wall. Another L4-sclerite in the anterior left
ventral wall bears a node-like process (nla in fig.321i; present in Blattaria only). These
three L4-sclerites can be separated from or connected with each other in the anterior
ventral wall.

In the central and right parts of the left complex there are two pouches invaginated
anteriad, which lie one above the other. The walls of the dorsal pouch (pne in fig.321e,n)
are largely occupied by the hood-shaped sclerite L1. The phallomere-gland (P in fig.321e)
opens into this pne-pouch. The ventral pouch (Ive in fig.321e,g) contains the L2-sclerite,
which is often restricted to the dorsal wall of the pouch and extends like an arch along
its anterior and lateral margins. The left posterior part of sclerite L2 leaves the lve-pouch
and extends onto a process (paa in fig.321e,g) immediately to the right of the pda-process.
The sclerotisations of L4 and L2 are connected in between the processes pda and paa.
The right end of sclerite L2 articulates with sclerite L1 (articulation A2 in fig.321e,n).

17

The ventral wall of the Ive-pouch is almost completely membranous, and it is at the same
time the left-anterior part of the dorsal wall of a large ventral lobe (vla in fig.32le,g,i).
The ejaculatory duct (D in fig.320, 321g) opens into this wall. In some species this wall
contains a small sclerite L5 (fig.32le,g). The ventral wall of the vla-lobe is part of the
ventral wall of the whole left complex and is largely occupied by the right-ventral L4-
sclerite (fig.321g,1).

All Blattaria, but not Mantodea, have a large hook-like evagination from the left ventral
wall of the left complex (hla in fig.3211). The hla-hook is largely occupied by sclerite
L3, but a more or less extensive basal part is membranous (30 in fig.321i). This membrane
can be introverted, which procedure results in a more or less deep retraction of the hook
into the left complex.

Right phallomere

The anteriormost ventral wall is occupied by the plate-like sclerite R3 (fig.321f,h). Along
the lateral and anterior margins of R3 (parts of the basal line Bl) the sclerotisation of R3
extends somewhat into the wall of the genital pouch, and these margins of sclerite R3
form a groove- (as seen from externally) or fold-like (as seen from internally) apodeme
age (fig.321f,h). Like in the swe-apodeme, parts of this age-apodeme can be filled in by
the cuticle becoming thickened, and the respective parts of age are beam-like.

Behind the central part of R3 the ventral wall is extensively invaginated dorsad and
anteriad (cbe in fig.321f,h), and this invagination is partly sclerotised in its dorsal wail
(anterior part of sclerite R1 in fig.321f). Blattaria, but not Mantodea, have a sclerite R2
left-ventral to the cbe-invagination (fig.321f,h), which articulates with the left posterior
end of R3 (articulation A7 in fig.321f) and with the left end of RI (articulation A6 in
fig.321f). Sclerite R2 and the posterior margin of the anterior part of R1 often form tooth-
or ridge-like cuticular evaginations (on Ri: pva in fig.321h).

The part of the right phallomere posterodorsal to the cbe-invagination is composed of a
large dorsal lobe (fda in fig.321f) and a ventral tooth (pia in fig.321h, which is in most
species much smaller than the fda-lobe). The fda-lobe and the pia-tooth are confluent
along the right edge of the right phallomere, and they diverge to the left. The dorsal wall
of the fda-lobe — and often parts of its ventral wall, too — as well as the dorsal and ventral
walls of the pia-tooth are occupied by the posterior part of sclerite R1. The sclerites R1
and R3 articulate with each other at the anterior right edge of the right phallomere
(articulation A3 in fig.321f,h). In the anteriormost dorsal wall of the fda-lobe, part of the
cuticle is invaginated to form a hollow, long and narrow, membranous tendon (tre in e.g.
fig.74; not shown in fig.321), which is present in some Blattaria only.

Discussion of the basic division of the phallomere complex

I propose this division of the phallomere complex into a left complex and a right
phallomere. However, earlier suggestions for a basic division differ from this hypothesis:
Snodgrass (1936, 1937) divides the phallomere complex of Blattinae into a ventral, a right,
and a left phallomere. Beier (1970) follows Snodgrass regarding Blattaria as well as
Mantodea, and he terms these main divisions hypophallus, right epiphallus, and left

18

epiphallus. The ventral phallomere (= vla-lobe in the previous description) lies ventral to
the genital opening (in a strict morphological sense anterior to it); 1.e. the genital opening
is in its anteriormost dorsal wall. The right phallomere (= right phallomere in the previous
description) and the left phallomere (= left complex minus the vla-lobe) have their bases
in the areas dorsal (in a strict sense: posterior) and lateral to the genital opening. Snodgrass
(1937) deduces this basic division from his investigations of the ontogenetic stages of the
phallomere complex in Periplaneta americana and Blatta orientalis (both: Blattinae): In
medium-sized nymphs the phallomere complex consists of three distinct lobes, which
hardly reveal any further elaborations. One lobe is situated medioventral to the prospective
genital opening (prospective ventral phallomere), the other two take positions dorsolateral
to the genital opening (prospective right and left phallomeres). Thus, a composition of a
medioventral, a right-dorsal, and a left-dorsal basic element seems plausible, and according
to Snodgrass (1937) the ventral phallomere is an unpaired medioventral element.

Quadri (1940) studied the ontogeny of the phallomere complex of Blatta orientalis in more
detail. In first instar nymphs he finds one pair of lobes with an invagination between them
(rudiment of ejaculatory duct). In the third instar each lobe is divided into a dorsal and a
ventral secondary lobe, and thus four lobes surround the prospective genital opening. Later,
the two left lobes form the left phallomere (more or less by fusion, without any clear
border remaining). The ventral right lobe shifts to the left, into a position beneath the
genital opening, and becomes the ventral phallomere. The dorsal right lobe maintains its
position and becomes the right phallomere. Thus, according to Quadri, the ventral
phallomere is a ventral part of the right half of the phallomere complex. Later Snodgrass
(1957) took over the opinion of Quadri but still used the tripartite division in his
terminology.

Concerning the assignment of the ventral phallomere, or ventral lobe vla, my own
observations as well as the mode of innervation (Pipa 1988) are in conflict with the views
of both Snodgrass (1937) and Quadri (1940):

In some aberrant specimens of Blattaria the phallomere complex is completely
symmetrical. I could find two such specimens: (1) Polyphaga aegyptiaca (Polyphaginae)
with two “right” phallomeres being mirror-images of each other; there was no trace of a
ventral phallomere, which is present in normal specimens. Unfortunately, the specimen
had been dried and macerated, and the relations to the internal genitalia and the presence
of an ejaculatory duct could not be investigated. (2) Deropeltis sp. (Blattinae) with two
“left” phallomeres and two complete ventral phallomeres, both pairs being mirror-images
of each other. The phallomere-gland is paired. The ejaculatory duct is, as usual, unpaired.
It opens in the median plane — in that area where the dorsal walls of the left and the right
ventral phallomeres are confluent with each other. Thus, the location of the genital opening
— in relation to the ventral phallomeres — is the same as in normal specimens, and the two
ventral phallomeres are arranged in a way that this relative position is true of both of
them.

Pipa (1988) studied the innervation of the male postabdomen in Periplaneta americana:
From the posterior part of the last abdominal ganglion — a compound ganglion formed
from the ganglion rudiments of abdominal segments 7 to 11 — one pair of nerves runs to

19

the phallomeres (phallic nerve = nerve 5a in Pipa). Their basal branchings are symmetrical.
After entering the phallomeres, where the branches innervate the phallomere muscles, the
branching pattern becomes completely asymmetrical. The ventral phallomere gets its
innervations exclusively from branches coming from the left nerve.

The morphology of the two symmetrical specimens and the innervation pattern suggest
that the ventral phallomere is neither an unpaired ventromedian element of the phallomere
complex nor a part of its right half but a lobe-like part of its left half. I term it the vla-
lobe, and the left and ventral phallomeres together I term the left complex. There is another,
more practical (though not decisive) reason for this concept: The morphological relations
between the left phallomere and the ventral phallomere are often very close, and the border
between them is in many cases not very distinct. And this is with high probability the
ground-plan situation (like in fig.321e,1). However, the question of the correct assignment
of the ventral phallomere or vla-lobe is certainly not finally settled.

The homology relations between the phallomere elements of Blattaria and Mantodea on
the one hand and the elements of the external genitalia of other Ectognathan taxa on the
other are completely unclear. Only the earliest rudiments or primary phallic lobes can be
reliably regarded as homologous.

The abdominal segments 10 and 11 and the telson

This morphologically posteriormost part of the body lies dorsal to the phallomere complex
and covers it completely (Blattaria) or partly (Mantodea) (fig.320, 321a,b). For many
sclerotisations of this region it is unclear whether they belong to abdominal segment 10
or ll or to the telson, or to the segment 12 proposed by Matsuda (1976).

Description of morphology

The principal morphology of this area is in Blattaria and Mantodea always quite similar:
Tergite 10 (T10 in fig.320, 321a,b) is somewhat triangular by a more or less pronounced
expansion of its median part to the posterior. Along the posterior edge of tergite 10 (X in
fig.320, 321a,b) the cuticle bends sharply ventrad and anteriad, and the sclerotisation of
T10 often — and to a varied extent — follows this bend and forms the ventral sclerotisation
of tergite 10 (T10v in fig.320, 321b). In some species tergite 10 is longitudinally divided
along its midline by a stripe of membrane (a derived condition).

The cerci (E11 in fig.321a,b) are the appendages of abdominal segment 11. They have
their bases laterally beneath the posterior edge of tergite 10. The basal article of each
cercus has at its dorsal basal margin a distinct articulation with a node-like thickening on
the posterior margin of tergite 10 (articulations A98 in fig.321b and e.g. fig.58). Median
to the cercal bases there may be some further sclerotisations (three pairs at most; not
shown in fig.321): The crescent-shaped Ca-sclerites (e.g. in fig.263) are close to the cercal
bases and often lie upon distinct bulge-like evaginations. The Cb- and Cc-sclerites take
more median positions (e.g. in fig.169, 170).

The anterolateral parts of tergite 10 curve ventrad and then mesad; these parts are the
paratergites (T10p in fig.321b,c), which take a position posterolateral to the phallomere

20

complex. The ventromedian ends of the paratergites T10p are in most species distinctly
articulated with the lateral ends of the paraprocts Pp (articulations A99 in fig.321b,c).

Median to the paratergites T10p and beneath the T10v-sclerotisation there is on each side
a transverse (or oblique) bulge, the subanal lobe (sbl in fig.321b,c). The paraprocts Pp
are always present as one pair of sclerites (fig.321b,c). From their lateral ends at the
articulations A99 they extend mesad and sclerotise more or less extensive parts of the
subanal lobes sbl. Laterally the paraprocts are restricted to the ventral sbl-walls, medially
they curve more and more into their dorsal walls (fig.321b,c). Consequently, the posterior
parts of the paraprocts are curved upwards and back anteriad, and the paraprocts have a
ventral anterior margin and a dorsal anterior margin. (The latter will subsequently be
designated as the posterior margin, which is true in a strict morphological sense). The
median tips of the subanal lobes (Y in fig.321b,c) lie on both sides of the anus (Af in
fig.321c) and are either membranous or also sclerotised by the paraprocts. The median
walls of the subanal lobes continue anteriad into the lateral walls of the rectum (Re in
12321bXo):

In front of the (ventral) anterior margins of the paraprocts there is often another pair of
ribbon-shaped sclerites (Pv in fig.320, 321b,c). These Pv-sclerites are either completely
free (like in fig.321) or connected with the paraprocts laterally. In some species separate
Pv-sclerites are missing, and in these cases they seem to have fused to the anterior margins
of the paraprocts.

The membrane anterior to the ventral sclerotisation of tergite 10 (T10v in fig.321b) is
evaginated to form an unpaired supraanal lobe (spl in fig.320, 321b,c), whose ventral wall
continues anteriad into the dorsal wall of the rectum (Re in fig.320, 321b,c). In Mantodea
the supraanal lobe bears a sclerotisation in its dorsal wall, the epiproct (Ep in fig.320,
321b,c). In many Blattaria the supraanal lobe is still distinct but never has a sclerotisation.
In other Blattaria the supraanal lobe is no longer distinct from other small membranous
foldings in the anal region, and its presence is questionable.

Discussion of morphology

Concerning all these elements, only the assignment of the anterior part of tergite 10 (T10
including T10p) to abdominal segment 10 and the consideration of the cerci as the
appendages of segment 11 is generally accepted. Regarding the other elements there are
various opinions. Snodgrass (1933, 1936, 1937) regards the cerci and the subanal and
supraanal lobes as elements of segment 11, the paraprocts being the medially divided
sternite and the epiproct being the tergite of segment 11. Sternite 10 is assumed to be
missing, tergite 10 is the true tergite 10, with no other elements incorporated, and the
telson is only a small membranous ring surrounding the anus. Chopard (1917), Walker
(1922), and Ford (1923) differ from Snodgrass only in assuming a participation of sternite
10 within the anterior margins of the paraprocts. This is said to be indicated by the
articulations between the paraprocts and the paratergites of segment 10 (A99 in fig.321b,c)
and by some muscle insertions. In complete contrast to these authors, Heymons (1895)
and Matsuda (1976) consider segment 11 as strongly reduced — the cerci being its only
persisting products — and regard the subanal lobes and the paraprocts as well as the

21

supraanal lobe and the epiproct as elements of the telson (Heymons) or of a twelfth
segment (Matsuda; the only difference to Heymons is that Matsuda regards this
posteriormost part of the body as a segment). Matsuda regards tergite 10 of Mantodea as
a proper one, but tergite 10 of Blattaria is supposed to contain the epiproct.

These differences in the interpretation of the terminal elements are accompanied by a
confused situation in the terminology for these structures. This concerns the usage of e.g.
the terms subanal lobe, supraanal lobe, epiproct, paraproct, tergite 11, sternite 11, and
telson. The comparison of the results of the various authors is thus rather difficult. For a
correct interpretation of the elements concerned some clarifying investigations of ontogeny
and morphology would be necessary. To do this is not the purpose of this paper, and the
terminology for the respective elements subsequently practised is a descriptive one, not
the morphologically correct one.

These controversial opinions, however, have to be discussed as far as homology relations
within Dictyoptera are involved. This concerns the elements called tergite 10 T10 and
supraanal lobe spl in the above description (fig.320, 321), which are, according to Matsuda
(1976), both not homologous in Mantodea on the one hand and in Blattaria and Isoptera
on the other. (This difference is the same for females). Matsuda’s opinion is as follows:
Blattaria and Isoptera show in their ontogeny a very early differentiation of a supraanal
lobe (meaning of supraanal lobe here: the dorsal part of the embryonic telson — Heymons
— or segment 12 — Matsuda; not the structure called spl-lobe above!). By the extensive
reduction of segment 11 during embryonic development this supraanal lobe comes into a
position immediately behind abdominal tergum 10. The dorsal segmental border between
supraanal lobe and tergum 10 then vanishes and these regions become fused. Thus, in the
imago the sclerite called “tergite 10” T10 above is regarded as a compound sclerite
containing the true tergite 10 and the epiproct (the latter considered as the tergite of the
telson or of the twelfth segment, respectively). In Mantodea, however, the differentiation
of this supraanal lobe is delayed until postembryonic development. No fusion of supraanal
lobe and tergum 10 takes place. Thus, in the Mantodean imago “tergite 10” T10 is the
true tergite 10, and the epiproct Ep is still situated behind it as a separate sclerite on an
independent supraanal lobe.

If this is true, the element I call supraanal lobe spl (fig.321b) would be: (1) the supraanal
lobe sensu Heymons and Matsuda in Mantodea (dorsum of telson or segment 12); (2) a
posterior part of the supraanal lobe sensu Heymons and Matsuda in Blattaria / Isoptera (a
lobe-like posterior part of the dorsum of the telson or segment 12). The condition in
Mantodea, if not regarded as a neotenic trait, would be more primitive than the situation
in Blattaria and Isoptera.

Matsuda (1976) refers to the results of earlier workers: Heymons (1895), Wheeler (1889),
and Cholodkowsky (1891) for Blattaria; Graber (1890), Hagan (1917), Görg (1959), and
Walker (1919, 1922) for Mantodea. From the data contained in these papers the following
view results:

— Looking at the descriptions in Heymons, the fusion between tergum 10 and the dorsal
part of the telson (or segment 12) really seems to take place in Blattaria.

22

— In Graber, Hagan, and Görg, however, | could not find any observation contradicting
the same fusion in Mantodea: None of these authors treats the development of the region
concerned in sufficient detail.

— Matsuda agrees with Walker and also Snodgrass (1933), p.73, and (1936), p.42, about
Mantodea: Supraanal lobe and tergite 10 are not fused, and tergite 10 of the adults is
a proper one. The two latter authors (the only ones from whom Matsuda could have
adopted his assumption for Mantodea), however, make the same assumption for
Blattaria, too. They regard — as I did in the above description — the membranous lobe
(spl in fig.320, 321b) of Blattaria as homologous with the spl-lobe of Mantodea. Thus,
the opinions of Walker and Snodgrass for Mantodea cannot — in a comparison with the
results of Heymons for Blattaria — serve to state a difference between Blattaria / Isoptera
and Mantodea.

— Accepting Heymons’ results, in Blattaria the supraanal lobe sensu Walker and Snodgrass
(= spl-lobe in my terminology) is posterior to or a posterior part of the supraanal lobe
sensu Heymons. According to Matsuda, in Mantodea the former (spl-)lobe is
differentiated in a postembryonic stage. Such a late elaboration of the spl-lobe is
possibly also true of Blattaria (and Isoptera?); at least, to my knowledge, an embryonic
rudiment of this lobe is not mentioned in the literature.

— Thus, no argument comes from the data used by Matsuda to contradict the homology
of the spl-lobes of Blattaria and Mantodea. The assumption of a difference between
Blattaria and Mantodea is based upon a comparison of non-comparable data.

Hence, the elements I call supraanal lobe spl and tergite of segment 10 T10 might both
be regarded as homologous in Blattaria and Mantodea — whatever structures these may be
in a strict morphological sense. Moreover, there are some arguments supporting these
homologies: (1) The supraanal lobe of Mantodea and the membranous lobe found in many
Blattaria (spl) show exactly the same relations to surrounding elements — namely those
shown in fig.320, 321b,c. (2) My own investigations of the musculature of the respective
region in Sphodromantis (Mantodea), Lamproblatta, Eurycotis, and Cryptocercus
(Blattaria) had the result that muscle insertions are present neither on the lobe of
Sphodromantis nor on that of the Blattarian species, and the relations of these lobes to the
surrounding muscles are the same in both groups. (3) Investigations in the same species
show that the tergites 10 are provided with the same sets of muscle insertions. An unpaired
muscle running from the posterior part of tergite 10 to the rectum (present in all these
species) could be the musculus epiprocto-analis (Weidner 1982). The position of its dorsal
insertion might support the view that the true epiproct has been incorporated into tergite
10 in both Blattaria and Mantodea.

3.2. The musculature

Most muscles are compact, and the insertion areas are well-defined. Some others, however,
are rather diffuse, and it is not possible to exactly demarcate their insertion areas. (Such
a diffuse condition will be mentioned in the muscle lists in chapter 5.). The data given in
the figures must be considered with these reservations.

23

According to their morphological arrangement, the following groups of muscles can be
distinguished:

Phallomere muscles: Intrinsic muscles of the phallomere complex. All muscles of this
group will be studied. Three subgroups will be distinguished:

a) Intrinsic muscles of the left complex.

b) Intrinsic muscles of the right phallomere.

c) Muscles connecting the left complex and the right phallomere.

Phallomero-sternal muscles: Muscles connecting parts of the phallomere complex or the
ventral and lateral walls of the genital pouch with the subgenital plate. All muscles of this
group will be studied.

Ventral muscles: Muscles connecting successive sternites (mainly the respective anterior
margins). These muscles are, compared with the more anterior segments, quite reduced in
the postabdomen. Only the muscles belonging to abdominal segment 9 (running from
sternite 9 to the — possibly obsolete — sternite 10) will be studied.

Dorsal muscles: Muscles connecting successive tergites. In some species lateral parts of
the dorsal muscles of segment 9 have shifted in a way that they can hardly be recognised
as dorsal muscles but seem to be muscles from the tergite to the phallomere complex.
Only these parts of the dorsal muscles will be studied.

Dorsoventral muscles: Muscles connecting tergite and sternite of the same segment. If
there really are vestiges of appendages contained in the subgenital plate, some of the
muscles included here might be muscles from the tergite to the appendages. These muscles
will be considered only in part.

Rectal muscles: Muscles from the ectodermal rectum to e.g. the anterior margin of the
subgenital plate, the tergite 10, or the paraprocts, which function as dilators or suspensors
of the rectum. These have in most cases clearly demarcated insertion areas on the
respective parts of the exoskeleton, but the fibers diverge like a fan on their way to the
rectum, and the rectal insertions are composed of many small insertion areas, which are
often widely separated from each other. Only those muscles inserting on the subgenital
plate will be considered.

Muscles of the ejaculatory duct: The ejaculatory duct is covered by a mat of fibers
showing a ring-like, spiral, or longitudinal arrangement. This musculature will not be
investigated in detail.

4. TERMINOLOGIES AND ABBREVIATIONS
4.1. The terminology for the sclerites of the phallomere complex

Mantodea

The most elaborate terminology is that of LaGreca (1954). It is the only one that is based
on quite detailed investigations of phallomere morphology and that has already been
applied to a broader sample of species. However, some disadvantage lies in the fact that
LaGreca uses some names for sclerites as well as for the formative elements of the cuticle

24

(e.g. processes or pouches) on or in which these sclerites are situated. For example, lamina
ventrale (= lv) designates the sclerite I call L2 and, at the same time, the pouch Ive, which
contains sclerite L2 (fig.321g). This ambiguity makes LaGreca’s terminology rather
impractical. The other terminologies put forward (e.g. Beier 1968) are not very handy
because of their long terms, and they are by far not detailed enough for my purposes.

Blattaria

McKittrick (1964) has developed a new, very simple and handy terminology, which has
been adopted by most of the subsequent workers. Mizukubo & Hirashima (1987) also
employ it but propose some changes. In both terminologies, the names for sclerites are
short sequences of letters and numbers, each position containing certain information. Some
of these terms have already been applied in the description of the phallomeres in 3.1., e.g.
L2, R3.

McKittrick regards the phallomere complex of Cryptocercus punctulatus as the most
primitive and takes it as the reference type for her terminology. She adopts the tripartition
of Snodgrass (1936) into left, right and ventral phallomeres, and according to this major
division McKittrick basically distinguishes left, right and ventral sclerites, which get L,
R, or V, respectively, in the first position of their names. Then, on the left and on the
right phallomere, the sclerites are numbered separately. The ventral phallomere has only
one sclerite. In this way seven main sclerites are distinguished (L1, L2, L3, V, R1, R2,
R3). No assumptions concerning side-homologies are intended in this terminology. What
McKittrick — starting fom this situation in Cryptocercus — regards as a product of a
secondary “subdivision” or as a special region or “elaboration” of a main sclerite is
expressed by the addition of one or two small letters (d = dorsal, v = ventral, I = lateral,
m = median, vm = ventromedian). Sclerites of certain species regarded as completely new
elements not present in Cryptocercus are given the next free number of the respective
phallomere. This terminology is very handy and clear and contains a lot of information.

Mizukubo & Hirashima state side-homologies for the elements of the left and of the
right half of the phallomere complex and integrate these assumptions in their terminology.
For that purpose, they modify the terminology of McKittrick in two ways:

(1) According to the assumption of a plane of symmetry, they basically distinguish right
and left elements (R or L in first position); then both R and L are grouped into dorsal
and ventral elements (D or V in second position). The left-dorsal elements LD and the
left-ventral elements LV compose the left phallomere. The right-ventral elements RV
correspond to the ventral phallomere (vla-lobe in my terminology). The right-dorsal
elements RD correspond to the right phallomere. Thus, the basic division into LD, LV,
RV, and RD essentially conforms with the division of the phallomere-complex proposed
by Quadri (1940).

(2) As regards the numbers and small letters, Mizukubo & Hirashima adopt the
terminology of McKittrick, but changes are made in order to get side-homologous elements
provided with the same names — except for R or L in the first position. These changes
are, compared with McKittrick, not very extensive.

25

The terminology for Blattaria and Mantodea used in this paper

The terminology of Mizukubo & Hirashima will not be employed since I do not agree
with the assumptions of these authors (discussion in chapter 8.). I will use a modified
version of McKittrick’s terminology and apply it to both Blattaria and Mantodea. There
are three reasons why the terminology of McKittrick is not adopted unchanged, and I will
procede in the following way:

1. Reason: The tripartition in McKittrick’s terminology (L, V, R) reflects the earlier view
of Snodgrass (1936) that the ventral phallomere is a medioventral basic element of the
phallomere complex. In my view the ventral phallomere = ventral lobe vla is a ventral
part of the left half of the phallomere complex (= left complex).

1. Procedure: All sclerites of the left complex will get L, all sclerites of the right
phallomere will get R in the first position of their names.

2. Reason: Like Mizukubo & Hirashima (1987), I cannot accept the view of McKittrick
that the phallomeres of Cryptocercus are closest to the primitive Blattarian type and should
be used as a reference type. I have taken the common ground-plan of the phallomeres of
Blattaria and Mantodea (compare in 3.1.) as the basis of my terminology. The ground-
plan pattern I assume for the phallomere sclerites is rather different from that proposed
by McKittrick (compare fig.32le-i and McKittrick 1964, fig.106).

2. Procedure: Each sclerotisation that is assumed to be present as one isolated and
undivided sclerite in its most primitive condition within the taxon comprising all Blattaria
and Mantodea and their last common stem-species is designated as a main sclerite. Hence,
these main sclerites can be (1) sclerites of the common ground-plan of Blattaria and
Mantodea or (2) sclerites formed de novo (not by the division of sclerites already present
before) in any subgroup of Blattaria or Mantodea. Each main sclerite will get its own
number in the second position of its name. In the description in 3.1. these are the sclerites
L1, L2, L3, L4, L5, R1, R2, and R3. Numbering is arbitrary. If any of these main sclerites
becomes divided, the whole of its descendants is called sclerite group L1, L2, etc..

Unfortunately, for many sclerotisations the most primitive condition and the evolution are
not completely clear, and there is in many cases, and to various extents, some uncertainty
about whether a certain sclerotisation fulfils the definition of a main sclerite. As regards
the sclerotisations shown in fig.321, L1, R1, and R3 are assumed to be isolated and
undivided sclerites in the common ground-plan of Blattaria and Mantodea (fig.321e-i).
L3, L5, and R2 are also isolated and undivided in their most primitive states, but they
are possibly not yet present in this ground-plan; however, if a later origin is really true
for them, they can be at least assumed to be new sclerotisations, not split off descendants
of ground-plan sclerites. To regard L2 and L4 as two main sclerites is somewhat

subjective: (1) L2 and L4 are primitively connected in between the processes paa and
pda (fig.321e,g); the interspace between paa and pda is here defined as the border between
L2 and L4. (2) For L4 it is not clear whether it has been present as one, two, or three
sclerites in the ground-plan (the latter alternative is shown in fig.32le,g,i). Apart from
these early evolved sclerites, there are several main sclerites which are undoubtedly
apomorphic for certain subgroups (L6...., R4....).

26

3. Reason: McKittrick names certain parts of her main sclerites with small letters in the
third position, no matter whether these parts are (1) products of a complete division or
(2) only certain regions of a main sclerite. However, these two situations represent two
different aspects of a sclerotisation and its evolution: (1) On the one hand, a main sclerite
is composed of one or more separate individual sclerites. Since divisions or fusions of
sclerites can take place, the state of subdivision of a main sclerite is subjected to
evolutionary changes. (2) On the other hand, a main sclerite, irrespective of its state of
subdivision, consists of several regions each of which is characterised by taking a certain
position, by having a certain shape, by occupying certain in- or evaginations of the cuticle,
or by bearing certain muscle insertions or articulations. The properties of these regions
undergo evolutionary changes, too. The special state of subdivision (1) and the special
properties of the regions (2) of a main sclerite are largely independent of each other, and
evolution works on both these aspects and has to be considered from both viewpoints.

3. Procedure: I strictly separate these two aspects (1) and (2) in my terminology, and for
the designation of different parts of main sclerites two terminologies completely
independent of each other will be used:

The first terminology serves to designate individual sclerites having originated by a
division of a main sclerite: If any main sclerite is divided into two or more sclerites
completely separated from each other by membrane, each of the sclerites will get one
capital letter in the third position of its name (e.g. L2D, RIC). (The main sclerite has
become a sclerite group. An individual sclerite as defined here may, however, be connected
with parts of another main sclerite). If any of these individual sclerites undergoes a further
division or a fusion, all resulting sclerites involved in this process will get a new capital
letter. Equal designation of sclerites of different species means the assumption of
homology. Different designation, however, does not always mean complete non-homology:
Some small sclerites of one species can as a whole be homologous with one large sclerite
of another species, and none of the sclerites will have the same letter. In the third position
of these terms, I, O, and Q will not be used (danger of confusion with “1” and “0” in the
figures); R will not be used on the right side, L on the left side. Among other things, this
terminology serves for a clear reference between the text and the figures.

The meaning of a term designating a certain individual sclerite, e.g. L2D, is hence as
follows:

L 2 D
a sclerotisation belonging to being the indi-
of the Left side, main sclerite 2, vidual sclerite D

The second terminology serves to designate certain regions of the main sclerites or
sclerite groups: This regioning is essentially independent of any natural sub-division of
the respective sclerotisations. To name these regions one small letter will be added in the
third position (e.g. L2d, Ric). Equal designation of sclerotisations of different species
means the assumption of homology. Different designation means complete non-homology.
These names mainly serve for the demarcation of homologous areas on the main sclerites
/ sclerite groups of different species and for a description of their evolution as regards the
properties listed above in 3. problem (2). Therefore, this dividing into regions, or

2]

regioning, will be performed with practicability as the main point of interest (the best
possible way to explain homology relations). It is in general arbitrary. (The definition of
the regions, however, will in some cases be done in correspondence with a concrete
division of the respective main sclerite into smaller sclerites in a certain, arbitrarily selected
species). Only the more complicated main sclerites will be divided into regions. The small
letters are in most cases abbreviations of typical attitudes (e.g. position) of the regions.
These abbreviations and the definition of the various regions will be given in the homology
discussion of the respective main sclerites (chapter 6.).

The meaning of a term designating a certain sclerite region, e.g. Rle, is hence as follows:

R 1 c
a sclerotisation of belonging to being the region c, which
the Right side, main sclerite 1, takes a rather central position

In Blattellidae Plectopterinae, some other Blattellidae, and Blaberidae the phallomere
complex is side-reversed (Bohn 1987), being essentially a mirror-image of the phallomere
complex of the other Blattarian subgroups. In the terminology for the sclerites, this fact
will be taken into account by adding ’ at the end of the term. This will be done in the
terms for individual sclerites as well as in the terms for sclerite regions. (For example,
L4U’ is a LAU on a side-reversed phallomere complex, and it is on the right side). This
procedure is different from Bohn (1987), who adds ’ after the first letter (L4U’ would be
L’4U), but the meaning is the same. (Of course, since homology should be the basis of
the terminology, the left complex will still have this name, if it is, after a reversal of the
phallomere complex, on the right side of the body; the same practice will be applied to
the right phallomere.)

4.2. Abbreviations for other sclerites of the postabdomen

The terminology for the sclerites of the postabdomen is largely pre-set by the earlier
literature. Abbreviating is done according to the same principle as in the phallomere
sclerites. A capital letter in the first position designates the category (S = sternite; T =
tergite; E = sclerotisation of an appendage = extremity). A number in the second position
designates the abdominal segment the sclerite belongs to. A small letter in the third position
(not obligatory) serves to designate a special region. As discussed in 3.1., there are many
problems concerning the correct morphological assignment and designation of postab-
dominal sclerotisations. In such problematical cases neutral abbreviations will be used (e.g.
Pp = paraproct; Ep = epiproct; all elements of the subgenital plate are called S9 despite
the possibility of true appendages being involved).

4.3. The terminology for the formative elements

The phallomere complex contains many areas where the cuticular body wall forms
pouches, apodemes, tendons, hook-, spine-, tooth-, or lobe-like processes, or comparable

28

structures. These more or less distinct invaginations (or infoldings) and evaginations (or
outfoldings) of the cuticle are responsible for the shape of the phallomere complex, and
I will designate them as the formative elements. These will receive special names serving
for a clear and practicable reference and for a clear presentation of homology relations.
The terms are composed of three small letters. The first two letters are the individual name
of the respective formative element, the third letter stands for one of the two possible
categories (a = evagination, German “Ausstülpung”; e = invagination, German “Einstül-
pung’’). A certain area of the cuticle can belong to two or more neighboring formative
elements, and, in such a case, the respective terms “overlap” each other. For example, a
certain area of the cuticle can be e.g. the dorsal wall of a specific evagination and the
ventral wall of an immediately adjacent invagination at the same time, or a large
evagination can fork into smaller ones in its distal part.

The first two letters are essentially selected arbitrarily but are mostly abbreviations for
characteristic attitudes (e.g. vla = ventral lobe, evagination). In other cases the
abbreviations of the terms of LaGreca (1954) have been used (e.g. Ive = lamina ventrale,
invagination; LaGreca’s name for sclerite L2 and for the pouch containing L2; compare
fig.321g). The formative elements of the phallomere complex termed in this paper are
listed in 4.7.

Similar terms are used for formative elements outside the phallomere complex (e.g. sbl =
subanal lobe), but the letter in third position is part of the abbreviation and does not stand
for a category.

4.4. The terminology for the muscles

Muscles are named by a small letter determining the category, followed by a number for
the individual muscle. Numbering is arbitrary. The categories are as follows:

1 Intrinsic muscles of the left complex (left)

r Intrinsic muscles of the right phallomere (right)

b Muscles connecting the left complex and the right phallomere (between)
s Phallomero-sternal muscles (sternal)
p Muscles not inserted on the phallomere complex but taken

into account (some ventral, dorsal, dorsoventral, and rectal muscles) (peripheral)

In some cases it can be reliably assumed that a muscle has divided. In other cases
homology can be assumed for groups of muscles of different species, without the
possibility to ascertain the homology relations for the single muscles of the group. In both
these situations all muscles of the group get the same number, and the single muscles are
distinguished by adding a small letter in the third position. In these cases equal letters do
not mean homology of the respective muscles by principle; which homologies are assumed
will be explicitly mentioned in the text.

For the muscles of the categories I and r it is not clear which of them have originally
been pairs or if there are pairs at all. The muscles of category b are probably unpaired
transverse muscles. The phallomero-sternal muscles s will be separately numbered on the

29

left and on the right side, though pairs can be identified in many cases; those inserting on
the left half of the subgenital plate are given odd numbers, those inserting on the right
half are given even numbers. (In species having side-reversed phallomeres the reverse is
true). Only the “peripheral” muscles p will be designated pairwise.

4.5. The terminology for the articulations

Articulations between sclerites are named by A (“Articulation”) and an added number (e.g.
A5, A98). The numbers 1-79 are reserved for the articulations within the phallomere
complex, the numbers 80-99 for the articulations outside the phallomere complex.
Numbering is arbitrary. The term articulation is taken in its widest sense: each case of
close contact between two sclerites. Even if the contact has become looser by evolutionary
changes, the remaining vicinity of the sclerites will in many cases still be designated as
an articulation (in order to designate the assumed homology of the respective vicinity).
Places where articulations have been lost by a fusion of the sclerites concerned are often
given the name of the respective articulation, but * is added (e.g. A99*).

4.6. Survey of the terminologies used

Main sclerites / sclerite groups of the phallomere complex:

IL, 2)
Capital letter Number
(GIGS IL, 18) (main sclerite)

Regions of the main sclerites / sclerite groups of the phallomere complex:

IL, 2 d
Capital letter Number Small letter
(side: L, R) (main sclerite) (specific region)

Individual sclerites of the main sclerites / sclerite groups of the phallomere complex:

L 2 D
Capital letter Number Capital letter
(side: L, R) (main sclerite) (individual sclerite)

Postabdominal sclerites outside the phallomere complex:

S 9 d
Capital letter Number Small letter
(category: S, T, E) (abdominal segment) (specific region)

Formative elements of the phallomere complex:

p n e
Small letter Small letter Small letter
(part of individual name) (part of individual name) (category: a, €)
Muscles:

ie 4 b

Small letter Number Small letter

(category: |, r, b, s, p) (individual muscle) (separate part of the muscle)

30

Articulations:
A
Capital letter A

(articulation)

6
Number

(individual articulation)

4.7. Abbreviations in the figures 1-321

CL = Capital letter; SL = Small letter; NR = Number; facultative parts of the terms are put in brackets;
the abbreviations for the formative elements are listed separately.

A +NR

Af

b + NR (+ SL)
Bl

D

Dw

E + NR (4+ SL)
Ell

Ep

L+NR
L+NR+CL
L+NR+SL
I+NR (+ SL)
M

P

p + NR (4+ SL)
Pp

Pv

R+NR
R+NR+ CL
R + NR + SL
r+ NR ( SL)
Re

S + NR (+ SL)
S9a

S9d

S9l

S9s

s + NR (4+ SL)
sbl

spl

Sw

T + NR ( SL)
T9p

Articulation

Anus

Muscle between left complex and right phallomere (or part of it)
Basal line

Ejaculatory duct

Dorsal wall of genital pouch

Sclerotisation of an appendage (or part of it)

Sclerotisation of cerci

Epiproct

Main sclerite / sclerite group of left complex

Separate sclerite of main sclerite / sclerite group of left complex
Region of main sclerite / sclerite group of left complex
Muscle of left complex (or part of it)

Membranous or weaker sclerotised field of subgenital plate
Phallomere-gland

Muscles not inserted on phallomere complex

Paraproct

Ribbon-like sclerotisation anterior to paraproct

Main sclerite / sclerite group of right phallomere

Separate sclerite of main sclerite / sclerite group of right phallomere
Region of main sclerite / sclerite group of right phallomere
Muscle of right phallomere (or part of it)

Rectum

(Coxo-) Sternite (or part of it)

Apophysis of subgenital plate

Dorsal sclerotisation of subgenital plate

Anterolateral part of subgenital plate

Stylus on subgenital plate

Phallomero-sternal muscle (or part of it)

Subanal lobe

Supraanal lobe

Lateral wall of genital pouch

Tergite (or part of it)

Paratergite of abdominal segment 9

Paratergite of abdominal segment 10

Ventral sclerotisation of abdominal tergite 10

3]

Pleura between abdominal tergite 9 and sternite 9

Ventral wall of genital pouch

Posterior edge of abdominal tergite 10

Median part of subanal lobe

Formative elements of the phallomere complex

Column 2: LG = term derived from LaGreca (1954). Column 3: left = left complex; right = right
phallomere. Column 4: If the formative element is sclerotised entirely or partly, the respective
sclerotisation is given; if it is membranous, memb is listed; if both situations occur in different species,
the more primitive situation is listed first.

Term

afa
age

are
ate

boe
cbe

cla
cwe

dca

dla
dte
fda

fpe
goa
gta
hge
hla
ipe

Iba
loa

lve
mila
nla
paa

Meaning of term

apofisi falloide = af (LG)

anterior groove

apodeme, right
anterior tendon
bulb, opposite
central bulge

central lobe
cleft “Wulst”

dorsal cushion

dorsal lobe

dorsal trough
fallomero dorsale
di destra = fd (LG)
fence, posterior
genital opening
hook groove

hook, left

inter, phallomere-
glands

lobe, between

lobo membranoso =
lo (LG)

lamina ventrale = lv (LG)

median lobe
node, left

Position

left
right

right
left
left
right

left
right

left

right
left
right

left
left
left
left
left
left

left
left

left
left
left

processo apicale = pa (LG) left

Sclerotisation Shape

memb, Lim
R3

R3

memb, L4n?
L2d

Rit

memb
Rit

memb, L1

memb, R4
L2d + L4l
Rl

memb
memb
L2p
L3

L3
memb

17
Lim

12
memb
L4n
L2d

lobe, process
fold/beam-like
apodeme

fold-like apodeme
tendon

hood-shaped apodeme
large, shallow
invagination

lobe

curved cuticular
thickening

l-more cushions,
processes

lobe

shallow invagination
lobe

fold-like invagination
1-2 cushions, lobes
process

groove

hook

fold-like invagination

lobe
slender process

flat pouch
lobe

node, bulge
process

pbe pouch, between left memb flat pouch
pda processo distale = pd (LG) left L4l process
pia piastra ventrale = pi (LG) right Rlv tooth, ridge
pne processo anteriore = left Lla pouch
pn (LG)
pra prong, right right Rid process
psa posterior sting left L2m bifid process
pva processo ventrale right Rit tooth, ridge
sclerificato = pv (LG)
rge right groove right Rlc groove
sbe sting bulb left L4l bulb + channel
sla sting, left left L4d sting-like process
sra Sting, right right Rld sting-like process
swe Sclerotised “Wulst” left L4l fold/beam-like
apodeme
tre tendon, right right memb long, narrow tendon
tve tendon, virga left L4l tendon, apodeme
vfa ventral fold left memb broad lobe
vge virga groove left L4l groove
via virga left L2d + L4l process
vila ventral lobe left L4v lobe
vpe ventral pouch left memb broad, flat pouch
vsa ventral sting left L4 sting-like process
vte ventral tendon left memb broad, flat tendon
xla —- left L2m short process

5. DESCRIPTIONS OF PHALLOMERES AND POSTABDOMINA

The complicated morphology of the phallomere complex and the courses of all investigated
muscles will be described in detail. The morphology of the subgenital plate and of the
posteriormost part of the abdomen can be largely seen from the figures and from the
general descriptions in 3.1.. Only the condition of the following elements will be
mentioned: Pv-sclerites; sclerites Ca, Cb, and Cc median to the cercal base; ventral
sclerotisation of tergite 10 T10v; articulations A98 (cercal base — posterior margin of
tergite 10) and A99 (paratergite 10 — paraproct); some peculiarities will also be pointed
out.

Figs.1,2: Sphodromantis sp. (Mantodea, Mantidae) — 1: Male postabdomen in dorsal view; with
phallomere complex, subgenital plate, marginal parts of abdominal tergites 9 and 10, supraanal lobe,
epiproct, subanal lobes, paraprocts, distal part of rectum, basal parts of cerci, and part of musculature.
— 2: Same as in fig.1, after removal of further parts of abdominal tergites 9 and 10, parts of right
paraproct, and supraanal lobe with epiproct. Distal part of rectum and basal parts of cerci cut open.
Another part of musculature shown. Posterior to transverse line: like in fig.1. — Scale: 2mm.

33

Sphodromantis

S

34

5.1. Sphodromantis sp. (Mantodea, Mantidae)

Left complex

The dorsal wall is largely occupied by sclerite L4B (fig.9), the ventral wall by the large
plate-like sclerite L4A (fig.6, 9-11). L4A extends posteriorly onto the process pda (fig.6,
12). On the left edge of the left complex L4A and L4B articulate (A1 in fig.6, 10). L4B
has a solid keel-like apodeme (swe in fig.10; cross-section in fig.11) at its anterior left
margin.

From beneath the posterior and right margins of L4B, two deep pouches invaginate to the
left and anteriad and lie one above the other (pne and Ive in fig.10, 11). The dorsal pouch
pne (fig.10) is largely occupied by sclerite L1A (fig.10), which lies mainly in the ventral
wall but bends anteriorly into the dorsal wall around the left edge of the pouch. The
phallomere-gland (P in fig.10) opens into the membranous anteriormost ventral wall of
the pne-pouch. The ventral pouch Ive (fig.10-12) deepens extensively to the left. (The
edge along the bottom of the Ive-pouch is labelled 7 in fig.11). Its dorsal wall is occupied
by sclerite L2. The ejaculatory duct (D in fig.11, 12) opens into its membranous ventral
wall. The edge between the pouches pne and Ive (1 in fig.10, 11) bears three processes:
most anteriorly a hammer-shaped one (afa in fig.9-11); far behind this a long and narrow
lobe (loa in fig.10); immediately behind the latter a stout hook (paa in fig.10). The anterior
part of this edge 1 is occupied by sclerite L1B (fig.10, 11), which is separated from both
L1A and L2 by narrow stripes of membrane (2 in fig.10, A2 in fig.11). L1B also sclerotises
the afa-process and has a strip-like posterior extension into the dorsal wall of the loa-
process. The posterior part of sclerite L2 extends onto the paa-process (fig.11) and
occupies most of its surface.

The membranous ventral wall of the Ive-pouch with the genital opening is at the same
time the dorsal wall of the large ventral lobe vla (= ventral phallomere; fig.10-12). Close
to the genital opening there are two small membranous lobes (goa in fig.12). The ventral
wall of the vla-lobe is part of the ventral wall of the left complex and is largely sclerotised
by L4A (fig.6).

Right phallomere

The anterior ventral wall is occupied by sclerite R3 (fig.6, 13): Its left part is expanded,
its narrow right part curves posteriad and articulates (A3 in fig.6, 8, 14) with the
posteroventral sclerite R1B. The ventral wall behind the left and central parts of R3 is
broadly invaginated dorsad and anteriad (cbe in fig.13). The ventral wall behind the right
part of R3 bears two teeth, which are both sclerotised by R1B: the anterior pva and the
posterior pia (fig.6, 14). pia and pva are both on the crest of a leftward projecting
evagination of the ventral wall (fig.8, 14). The posterior part of the right phallomere is a
large lobe (fda in fig.6, 13), with its dorsal wall completely occupied by sclerite R1A.
Around the anterior right edge of the right phallomere R1A curves into the ventral wall,
where it is connected with R1B by a stripe of very weak sclerotisation (4 in fig.6, 14).

35

Sphodromantis
sp.

Figs.3-5: Sphodromantis sp. (Mantodea, Mantidae) — 3: Male postabdomen in dorsal view; with
phallomere complex, subgenital plate, and lateral parts of abdominal tergite 9. — 4: Left margin of
subgenital plate (compare fig.3); with insertion of muscle p6. — 5: Subgenital plate in dorsal view;
with insertion areas of muscles (except p6). — Scale: 2mm.

36

The whole anterior and right margins of R3 form a groove-like apodeme age (fig.6; cross-
section in fig.8, 14). The cuticle along the bottom of the age-groove is thickened. In the
left part of R3 the age-groove abruptly deepens and, in addition, curves posteriad and
finally back to the right (fig.6, 8, 13, 14). By this deepening and curving the left part of
the age-apodeme is a crescent-shaped plate, which is sclerotised dorsally and ventrally
(fig.6-8, 13, 14). The ventral wall of the genital pouch becomes evaginated by the posterior
part of the crescent and forms a membranous pouch (5 in fig.6, 7, 8, 13, 14). In the right
posterior part of R3 the age-apodeme bears an additional apodeme fold are (fig.6; cross-
section in fig.8), which bears another small keel-like apodeme (3 in fig.6, 13). The groove-
like shape of the sclerotisation extends posteriad beyond articulation A3 (where R3 ends
and the groove is no longer called age) onto sclerite R1B (fig.6, 8, 13).

Subgenital plate and posterior abdominal segments

Fig.1,2 (posterior segments); fig.5 (subgenital plate S9). The paraprocts Pp and the
epiproct Ep are just weakly sclerotised and do not have distinct boundaries to the
surrounding membranes; it is therefore not possible to determine the exact outlines of
these sclerotisations, the presence of articulations A99 (A99* in fig.2), and the presence
of separate Pv-sclerites (compare fig.321b,c). The data given in fig.1, 2 must be considered
with these reservations. A ventral sclerotisation of tergite 10 T10v was not found. The
sclerites Ca, Cb, and Ce are missing (or strongly obsolete?). The articulations A98 are
well-developed.

Musculature
Muscles Positions of insertions in fig.
1 L1A (anteriorly on pne-pouch) — right anterior dorsal wall of left
complex 17
12 L1A (dorsally on pne-pouch) — L4B (right part) 15
13 L1A (posteroventrally on pne- pouch) — L2 (left part) 16
14 L2 (left-posterior part) — L4B (central part) 15
15 L4A (left-anterior part) — L2 (left-anterior part) 15

eee Cy)

Figs.6-8: Sphodromantis sp. (Mantodea, Mantidae) — 6: Phallomere complex in ventral view. — 7:
Anterior part of phallomere complex in ventral view; with some muscles; ventral wall of genital
pouch more complete than in fig.6; left anterior part of sclerite R3 removed. — 8: Right phallomere
with transition to left complex in ventral view; some ventral parts removed. — Scale: 2mm.

— p.38

Figs.9,10: Sphodromantis sp. (Mantodea, Mantidae) — 9: Left complex in dorsal view. — 10: Left
complex in dorsal view; some dorsal parts removed. — Scale: 2mm.

——> p.39

Figs.11-14: Sphodromantis sp. (Mantodea, Mantidae) — 11,12: Left complex in dorsal view; with
further removal of its parts (mainly of dorsal ones). — 13: Right phallomere in dorsal view. — 14:
Right phallomere in dorsal view; some parts removed (mainly dorsal ones). — Scale: 2mm.

87]

I

Sphodromant

SAAN

38

Sphodromanti

39

SS SEG

fe

hodromanti

a
>>
mn
loa

SKI :

SCG QS

< S

40

SAN
Ys

AVAL eRe
AAA
AAA
AAA

NE

Figs.15,16: Sphodromantis sp. (Mantodea, Mantidae) — Left complex and right phallomere in dorsal
view; each figure with some muscles; parts of phallomere complex removed to various extents. —
Scale: 2mm.

Sphodromantis
sp.

\
x
.

x
“3

Figs.17-19: Sphodromantis sp. (Mantodea, Mantidae) — 17: Detail of left complex in dorsal view
(compare fig.10); with some muscles. — 18: Left complex in dorsal view; with muscle 16b; dorsal

parts largely removed. — 19: Right phallomere in dorsal view; with some muscles; dorsal parts largely
removed. — Scale: 2mm.

41

16a L4A (anterior part) — L2 (right-anterior part) 16
16b L4A (central part) — dorsal wall of vla-lobe behind genital opening 16, 18
17 L4A (left-posterior part) — L4B (left-posterior part) 15
18 Membrane behind left posterior end of L4A — L2 (posterior part);
only very few fibers; completely missing in some specimens 16
rl R3 (right margin) — RIA (dorsal wall of fda-lobe) It, IS)
r2 R3 — RIB on pva-tooth and membrane of cbe-invagination 16, 19
r3 RIA (right part) —- R1B on pia-tooth and right ventral wall of fda-lobe 16, 19
r4 RIA (left-anterior part) and membrane behind it — left ventral wall of
fda-lobe 15
bl L4A (anterior margin) — R3 (left anterior part) IS, 10, 1D
b2 L4A (anterior margin) and vla-lobe (anteriormost part) — R3
(left-anterior part) IS), ID
b3 Ejaculatory duct D - left ventral wall of right phallomere 15
b4 (19?) Longitudinally within right dorsal wall of left complex 107
sl S9 left side (medially and most anteriorly) — L4A (anterior margin) 25 Suu les
s2 S9 right side (medially and most anteriorly) — R3 (lateral anterior
margin) 2 Seca
s3 S9 left side (medially and anteriorly) — anteriormost left ventral wall
of genital pouch Ss)
sda S9 right side (medially and anteriorly) — R3 (left ventral wall of
age-apodeme) Dy
s4b S9 right side (medially and anteriorly) — anteriormost median ventral
wall of genital pouch S301
s5 S9 left side (laterally and anteriorly) — anterior left wall of genital pouch 2, 5, 7, 15
s6 S9 right side (laterally and anteriorly) — anterior right wall of genital
pouch 255 Te ales)
pl (pair) S9 — paraproct Pp (anterior margin, corresponding to Pv-sclerotisation) 1,5
p3 (pair) S9 — rectum (ventral wall); in most specimens p3 penetrates pl. 11,3
p4 (pair) T9 (lateralmost anterior margin, possibly corresponding to paratergite
T9p) — anterior margin of paratergite T10p 1
p5 (pair) T10 (lateralmost anterior margin) — paraproct Pp (lateral anterior
margin, corresponding to Pv-sclerotisation) i
p6 (pair) T9 (lateralmost part) — S9 (lateral margin) 1,4
p7 (pair) Membrane anterior to paraproct Pp — anterior margin of paraproct Pp
and paratergite T10p (border between Pp and T10p not clear) 2

5.2. Metallyticus violaceus (Mantodea, Metallyticidae)

Left complex

The left complex resembles that of Sphodromantis. However, sclerite L4B (fig.23) is
ribbon-shaped and restricted to the anterior dorsal wall. The ventral sclerite L4A extends

Figs.20-23: Metallyticus violaceus (Mantodea, Metallyticidae) — 20: Phallomere complex in ventral
view. — 21: Right phallomere with transition to left complex in ventral view; some ventral parts
removed. — 22: Subgenital plate in dorsal view; left stylus missing. — 23: Left complex and right
phallomere in dorsal view. — Scale: 1mm.

43

Metallyticus violaceu

44

posteriorly onto two processes (fig.20, 27): the left pda-process and the right vla-lobe.
L4A and L4B articulate on the left edge of the left complex (Al in fig.20, 24). L4A bears,
anterior to Al, a small keel-like apodeme (swe in fig.23, 24; cross-section in fig.25).

Like Sphodromantis, Metallyticus has a dorsal pouch pne and a ventral pouch Ive, which
contain the sclerites L1 and L2 (fig.23-25). Within the pne-pouch (fig.24, 25) L1 occupies
the ventral wall and parts of the dorsal wall. In the ventral wall L1 is partly divided by
a stripe of membrane (2 in fig.25). A phallomere-gland, whose opening were to be expected
within the pne-pouch, was not found. The ventral pouch lve (fig.26) deepens strongly
anteriad. Its dorsal wall is occupied by the ribbon-shaped L2-sclerite. The edge between
the pouches pne and Ive (1 in fig.25, 26) has its anterior starting point in the ventral wall
of the pne-pouch. The anteriormost part of edge 1 bulges to the right (afa in fig.25, 26).
Most posteriorly edge 1 bears a somewhat hook-like process paa, whose ventral wall is
sclerotised by a posterior extension of L2 (fig.20, 25, 26). Another sclerotised process loa
(fig.23-25) arises from the posterior ventral wall of the pne-pouch (fig.25). Anterior to
the loa-process, in the area of the bulge afa, L1 bends around the edge 1 into the dorsal
wall of the Ive-pouch, where it has a hinge-like articulation with the right margin of L2
(A2 in fig.26).

The membranous ventral wall of the Ive-pouch is at the same time the left part of the
dorsal wall of the ventral lobe vla (= ventral phallomere; fig.27). The ejaculatory duct (D
in fig.26, 27) opens into the anterior ventral wall of the vla-lobe. Dorsal to the genital
opening there are two membranous lobes (goa in fig.25, 26). Posterior to the genital
opening the dorsal vla-wall contains a small sclerite L5, with some small but distinct folds
to the left of it. The ventral vla-wall is completely sclerotised by a part of L4A (fig.20).

Right phallomere

Sclerite R3 occupies the anterior ventral wall (fig.20, 24). The left end of R3 has a short
extension to the posterior (fig.20, 26). The anterior and right margins of R3 form a groove-
like apodeme age, which deepens abruptly in its left part but does not curve like in
Sphodromantis (fig.20, 24; cross-section in fig.21). Next to its right-posterior end the age-
apodeme bears a small keel-like apodeme (3 in fig.20, 24). The right posterior end of R3
articulates (A3 in fig.21, 24, 26) with the posteroventral sclerite RIC. Sclerite R1D is
situated to the left of RIC and posterior to R3; it is in close contact with R1C and bears
a dental ridge (pva in fig.20, 25, 26). The ventral wall of the right phallomere posterior
to R3 and RID is extensively invaginated dorsad and anteriad (cbe in fig.21, 24). Sclerite
RIC extends from articulation A3 posteriad onto a large ventral tooth (pia in fig.20, 26),
occupying mainly its dorsal wall. Like in Sphodromantis, the area of articulation A3 and
the adjacent part of RIC are somewhat groove-like — a posterior extension of the age-
apodeme on R3. The posterior part of the right phallomere is composed of a large dorsal
lobe (fda in fig.23, 24) and of the ventral pia-tooth. The dorsal wall of the fda-lobe is

Figs.24-27: Metallyticus violaceus (Mantodea, Metallyticidae) — Left complex and right phallomere
in dorsal view; with successive removal of their parts (mainly of dorsal ones); fig.27: all parts of
right phallomere removed. — Scale: 1mm.

45

S

icu
laceus

Metallyt
vio

GGG

46

completely occupied by sclerite RIA. Around the right edge of the right phallomere RIA
curves ventrad and sclerotises the ventral wall of the pia-tooth (fig.20, 26). RIA is
completely separated from RIC.

Subgenital plate

Fig.22. In the specimen examined only the right stylus S9s was preserved.

5.3. Chaeteessa caudata (Mantodea, Chaeteessidae)

Left complex

Sclerite L4 occupies the whole ventral and dorsal walls (fig.28, 31). The dorsal and ventral
parts of L4 are firmly connected along the left edge of the left complex. Within the ventral
part of L4, the left, right, and anterior margins as well as an anterior transverse bridge
are distinctly heavier sclerotised (fig.28, 32).

A dorsal pne-pouch and a ventral Ive-pouch invaginate from beneath the right margin of
the dorsal L4 and take a position in the center of the left complex. These pouches contain
the sclerites L1 and L2 (fig.32, 34). The pne-pouch has its anterior part deeply invaginated
anteriad, its posterior part has the shape of a groove open to the right. L1 occupies the
whole ventral wall of the pne-pouch; only within the anterior part of pne L1 curves around
the left edge of the pouch into the dorsal wall; here it has a short extension directed right-
posteriad (fig.32). The phallomere-gland (P in fig.32, 34) opens into the membranous right
wall of the pne-pouch. The ventral pouch lve deepens strongly anteriad. The ribbon-shaped
sclerite L2 occupies most of its dorsal wall (fig.34). The edge 1 (fig.34) between the
pouches pne and Ive bears in its posterior part a long process paa, whose ventral wall is
sclerotised by a posterior extension of L2 (fig.28, 34). Immediately anterior to the paa-
base, L1 curves from the ventral wall of the pne-pouch into the dorsal wall of the Ive-
pouch (around edge 1 in fig.34, 35), where it articulates with L2 (A2 in fig.34, 35).
Anterior to this Ll-curvature edge 1 bears a membranous process afa. Edge 1 forks
immediately anterior to afa. Between the two branches the small membranous pouch pbe
deepens to the left. pbe is situated between the pouches pne and Ive.

The membranous ventral wall of the Ive-pouch is at the same time the left dorsal wall of
a large ventral lobe vla (= ventral phallomere; fig.28, 32). The ejaculatory duct (D in
fig.32, 35) opens into the anterior dorsal wall of the vla-lobe — far on the right side and
outside the Ive-pouch. The ventral wall of the vla-lobe is sclerotised by a part of the
ventral L4 (fig.28).

Figs.28-31: Chaeteessa caudata (Mantodea, Chaeteessidae) — 28: Phallomere complex in ventral view.
— 29a: Right phallomere with transition to left complex in ventral view; some ventral parts removed.
— 29b: Detail of right phallomere in ventral view (compare fig.29a); with sclerite R1B and tooth pva.
— 30: Subgenital plate in dorsal view; styli missing. — 31: Left complex and right phallomere in dorsal
view. — Scale: 0.5mm.

47

Kan. = 6 pe ne nen

a rn
SL Ss II x
un), a

2 IN

Chaeteessa
caudata

48

Right phallomere

Sclerite R3 in the anterior ventral wall is hatchet-shaped (fig.28, 31). Its anterior and right
margins form an apodeme age (fig.23, 32). The left part of age is groove-like, the right
part is beam-like (the groove is filled in by the cuticle being thickened; cross-section
through age in fig.29a, 33). The leftmost part of R3 bends back to the right (fig.29a, 32,
33)

Posterior to the left part of R3 the ventral wall of the right phallomere is invaginated
dorsad and anteriad (cbe in fig.29a, 32). The right posterior end of R3 has a broad
articulation A3 with the posteroventral sclerite RIB. From A3 sclerite R1B extends to the
left onto the anterior tooth pva, on which way it bends dorsad along the edge 16 (fig.28,
29a), as well as onto the posteroventral tooth pia. The dorsal and ventral walls of pva
and pia are completely sclerotised by RIB (fig.28, 29a, 33).

The posterior part of the right phallomere is composed of the large dorsal lobe fda (fig.31,
32) and the ventral tooth pia. The dorsal wall of fda is occupied by sclerite RIA. Around
the right edge of the right phallomere R1A curves into the ventral wall, where it is
restricted to the right margin and completely separated from R1B by membrane (4 and
17 in fig.28, 32).

Subgenital plate

Fig.30. The styli S9s have been lost in the examined specimen; only their points of
insertion are shown (S9s*).

5.4. Mantoida schraderi (Mantodea, Mantoididae)

Left complex

Sclerite L4 extends along the left, anterior, and right margins of the ventral wall (fig.41,
44). Along the whole left edge of the left complex L4 also curves into the dorsal wall
(fig.44, 45), where it is restricted to the left margin. Only a very distinct part of L4 in the
anterior dorsal wall extends farther to the right (L4d in fig.44). An apodeme swe runs
along the whole left arm of L4 (fig.44, 45). The anterior part of swe is beam-like by the
cuticle being thickened (cross-section of swe in fig.46). To the posterior this thickening
decreases, and swe is groove-shaped (cross-section in fig.45).

To the right of the dorsal part of L4, a dorsal pouch pne (fig.44, 45) and a ventral pouch
Ive (fig.45, 46) are invaginated anteriad, which contain the sclerites L1 and L2. Within

Figs.32-35: Chaeteessa caudata (Mantodea, Chaeteessidae) — 32: Left complex and right phallomere
(separated from each other) in dorsal view; some parts removed (mainly dorsal ones; compare fig.31).
— 33: Right phallomere in dorsal view; further parts removed (mainly dorsal ones; compare fig.32).
— 34: Left posterior part of left complex in dorsal view; further parts removed (mainly dorsal ones;
compare fig.32). — 35: Posterior part of left complex in dorsal view; right posterior part of left complex
with genital opening retained; further parts removed in the left half (mainly dorsal ones; compare
fig.34). — Scale: 0.5mm.

49

Chaeteessa

caudata

50

the pne-pouch, the hood-shaped L1 occupies most of the ventral wall and the left margin
of the dorsal wall (fig.45). The left-dorsal posterior margin of L1 articulates with a small
sclerite on the process loa (fig.45). The right posterior margin of L1 has a ribbon-like
extension. The phallomere-gland (P in fig.44, 45) opens into the anteriormost membranous
part of the pne-wall. The ventral pouch Ive (fig.46) is more tranversely extended (and less
antero-posteriorly as in Metallyticus and Chaeteessa). The edge along the bottom of the
Ive-pouch is labelled 7 in fig.46, 47. The ejaculatory duct (D in fig.46) opens into the
right part of Ive. Sclerite L2 extends like an arch along the margins of the dorsal wall of
Ive. In the anteriormost left edge of the pouch, however, it bends into the ventral wall of
Ive (fig.47), and from here it extends posteriad to join the left posterior end of L4. The
area where these posterior ends of L2 and L4 are interconnected is as a whole upcurved
(fig.45) and bears two short processes: the right, somewhat pointed paa, whose
sclerotisation belongs to L2, and the left, bulge-like pda, whose sclerotisation is part of
L4. The edge 1 (fig.45, 46) between the pouches pne and Ive is transversely orientated,
not longitudinally as in the previous species. Far to the right of paa the right posterior
ends of L1 and L2 contact each other (articulation A2 in fig.45, 46) — exactly in the edge
1. Immediately to the right of A2 the invagination of the Ilve-pouch — and thus also the
edge 1 — ends (fig.46), and immediately to the right of this point the membranous lobe
afa has its base.

The membranous ventral wall of the Ive-pouch is at the same time the anterior dorsal wall
of the ventral lobe vla (= ventral phallomere; fig.46, 47). The ejaculatory duct (D in fig.46,
47) opens most anteriorly, and quite far to the right, into this membrane. The ventral wall
of the vla-lobe is partly sclerotised by the right posterior part of L4 (fig.41, 47).

Right phallomere

Sclerite R3 in the anterior ventral wall is hatched-shaped (fig.41, 44). Its anterior and right
margins form an age-apodeme (fig.41, 44, 45), which is distinctly groove-like in its left
part but more beam-like in its right part (cross-section through age in fig.43, 45). In its
left part age is deeper. On the utmost right posterior part of R3 the age-apodeme bears a
small keel-like apodeme (3 in fig.43, 44).

The horseshoe-shaped sclerite R1D lies in the ventral wall posterior to the central part of
R3 and sclerotises the tooth pva (fig.41, 43, 45). Posterior to the left and central parts of
R3 and posterior to RID the ventral wall of the right phallomere is invaginated dorsad
and anteriad (cbe in fig.43-45). The right posterior end of R3 articulates (A3 in fig.41,
44) with the ventral part of sclerite RIE. The groove called age on R3 extends beyond

Figs.36,37: Mantoida schraderi (Mantodea, Mantoididae) — 36: Male postabdomen in dorsal view;
with phallomere complex, subgenital plate, marginal parts of abdominal tergites 9 and 10, supraanal
lobe, epiproct, subanal lobes, paraprocts, distal part of rectum, basal parts of cerci, and part of
musculature. — 37: Same as in fig.36, after removal of further parts of abdominal tergites 9 and 10,
parts of right paraproct, and supraanal lobe with epiproct. Distal part of rectum and basal parts of
cerci cut open. Another part of musculature shown. Posterior to transverse line: like in fig.36. — Scale:
Imm.

51

36

Bm m nenmreneeneaennsnnn] | om

Mantoida

schraderi

San Cay

52

S9a

39

a> s9d

4 Mantoida
schraderi

Figs.38-40: Mantoida schraderi (Mantodea, Mantoididae) — 38: Male postabdomen in dorsal view;
with phallomere complex, subgenital plate, and lateral parts of abdominal tergite 9. — 39: Left margin
of subgenital plate (compare fig.38); with insertion of muscle p6. — 40: Subgenital plate in dorsal
view; with insertion areas of muscles (except p6). — Scale: Imm.

53)

| Mantoida
| schraderi

Figs.41-43: Mantoida schraderi (Mantodea, Mantoididae) — 41: Phallomere complex in ventral view.
— 42: Anterior part of phallomere complex in ventral view; with some muscles; ventral wall of genital
pouch more complete than in fig.41. — 43: Right phallomere with transition to left complex in ventral
view; with some muscles; some ventral parts removed. — Scale: Imm.

54

Mantoida

schraderi

55

Mantoida
schraderi

47

Figs.46,47: Mantoida schra-
deri (Mantodea, Mantoididae)
— Left complex in dorsal view;
with successive removal of its
parts (mainly of dorsal ones).
— Scale: Imm.

Figs.44,45: Mantoida schraderi (Mantodea, Mantoididae) — 44: Left complex and right phallomere
in dorsal view. — 45: Left complex and right phallomere in dorsal view; some parts removed (mainly
dorsal ones). — Scale: Imm.

56

articulation A3 (where it ıs no longer called age) onto RIE (fig.41). The anterior margin
of RIE is in close contact with RID (fig.41).

The posterior part of the right phallomere is composed of the large dorsal lobe fda (fig.44,
45), with RIE in its dorsal wall, and the ventral tooth pia. Sclerite RIE extends from
articulation A3 left-posteriad onto the pia-tooth and occupies most of its dorsal and ventral
walls (fig.41, 43, 45). Another part of RIE extends narrowly to the right edge of the
phallomere, curves dorsad, and, becoming broader again, occupies most of the dorsal wall
of the fda-lobe (fig.41, 44).

Subgenital plate and posterior abdominal segments

Fig.36, 37 (posterior segments); fig.40 (subgenital plate S9). The ventral part of tergite 10
T10v is very narrow but distinct. Separate Pv-sclerites are missing; they are assumed to
have been incorporated into the anterior margins of the paraprocts Pp. The sclerites Ca,
Cb, and Ce are missing. The articulations A99 are missing (A99*: paratergites T10p and
paraprocts Pp have fused). The articulations A98 are well-developed.

Musculature
Muscles Positions of insertions in fig.
1 L1 (anteriorly on pne-pouch) — central dorsal wall of left complex

near right end of L4d 48
12 L1 (left-ventrally on pne-pouch) — L4 (dorsally on left arm, on

swe-apodeme) 49
13 L1 (posteroventrally on pne-pouch) — L2 (left-anterior part) 50
14 L2 (left-anterior part) — L4 (dorsally on left arm, on swe-apodeme) SI, Dil
15 Left anterior membranous ventral wall of left complex — L2

(left-anterior part) SD
16 Right anterior membranous ventral wall of left complex — L2

(right part) and dorsal wall of vla-lobe around genital opening 30); SD?
17 Left posterior membranous ventral wall of left complex — L4

(posterior ventral part of left arm) 32
19 Transversely within right dorsal wall of left complex 49
rl R3 (right margin) — RIE (dorsal wall of fda-lobe) 48
r2 R3 — RID on pva-tooth and membrane of cbe-invagination 49, 50
r3 RIE (right part) — RIE on pia-tooth and right ventral wall of fda-lobe 49, 50
r4 RIE (left-anterior part) — left ventral wall of fda-lobe 49
bl Membrane behind right anterior margin of L4 — R3 (left anterior margin) 43
b2 Dorsal wall of vla-lobe (right-anterior part) — R3 (left anterior margin) 49
b4a Dorsal wall of vla-lobe (right-anterior part) — RIE (left anterior margin) 36, 48
b4b Central dorsal wall of left complex — membrane posterior to anterior

margin of RIE (in fig.36 the insertion is beneath muscle b4a) 36, 48

Figs.48,49: Mantoida schraderi (Mantodea, Mantoididae) — Left complex and right phallomere in
dorsal view; each figure with some muscles; parts of phallomere complex removed to various extents.
— Scale: 1 mm.

57

ne nn een eee |

Mantoida

schraderi

en le

58

ee

a
ool
m

anne ei
Sf fy
Lf ff,
Lif)

$4

Mantoida
schraderi

a

Figs.50-52: Mantoida schraderi (Mantodea, Mantoididae) — 50: Left complex and right phallomere
in dorsal view; each figure with some muscles; parts of phallomere complex removed. — 51: Detail
of left complex in dorsal view (compare fig.46); with some muscles. — 52: Left complex in dorsal
view; with some muscles; parts of left complex removed (mainly dorsal ones). — Scale: 1mm.

59

sl S9 left side (medially and most anteriorly) — L4 (anterior margin) 37, 40, 42, 48
s2 S9 right side (medially and most anteriorly) — R3 (lateral anterior margin) 37, 40, 42, 48
s3 S9 left side (medially and anteriorly) — anteriormost left ventral wall of

genital pouch 40, 42
s4 S9 right side (medially and anteriorly) — R3 (left ventral wall of

age-apodeme) 40, 42, 43
s5 S9 left side (laterally and anteriorly) — anterior left wall of genital pouch 37, 40, 42, 48
s6 S9 right side (laterally and anteriorly) — anterior right wall of genital

pouch 37, 40, 42, 48
pl (pair) S9 — paraproct Pp (anterior margin, corresponding to Pv-sclerotisation) 36, 40
p3 (pair) S9 — rectum (ventral wall) 36, 40

p4 (pair) T9 (lateralmost anterior margin, also extending onto paratergite T9p) —

anterior margin of paratergite T10p (far medially); on both sides

completely divided into a dorsal (fig.36: on T9) and a ventral

(fig.37: on T9p) bundle BOS
p5 (pair) Not investigated; presence highly probable aa
p6 (pair) T9 (lateralmost part) - membrane immediately above lateral margin

of S9 30439
p7 (pair) Membrane anterior to paraproct Pp — anterior margin of paraproct Pp
and paratergite T10p (border between Pp and T10p not clear) 57

5.5. Archiblatta hoeveni (Blattaria, Blattidae, Blattinae)

Left complex

L4 is a group of five sclerites: The crescent-shaped L4C occupies the whole left edge and
the anteriormost left ventral wall of the left complex (fig.53). That part along the left edge
has a distinct dorsal extension to the right (L4d in fig.53), the end of which forms a spine
sla. Along the whole of L4C there runs an apodeme swe (fig.53-56), whose anterior part
is beam-like by the cuticle being thickened (anterior cross-section of swe in fig.54); to
the posterior this thickening decreases, and swe becomes more and more groove-shaped
(posterior cross-section of swe in fig.54). The sclerites L4D, L4E, L4F, and L4G lie in
the ventral wall of the left complex. L4D bears a node-like process nla (fig.56, 57). L4E
and L4F are simple plates. L4G occupies the ventral wall of a large ventral lobe vla (=
ventral phallomere; fig.53, 54, 57), and its posterior part is upcurved.

In the center of the left complex there are two pouches pne and lve (fig.53-56), which
contain the sclerites L1 and L2. The anterior part of L1 lies in the ventral wall of the
small and flat pne-pouch (fig.53, 54). The posterior part of L1 extends onto a broad lobe,
which is divided into three processes posteriorly (fig.54, 55): the two short, membranous
dca (left side) and the long, sclerotised loa (right side). Along the lateral and posterior
edges of this lobe the cuticle bends ventrad and anteriad (fig.55) as far as to the edge 6,
where it turns posteriad again. The phallomere-gland (P in fig.55, 56) opens immediately
posterior to edge 6.

The large sclerite L2 adjoins to the right of L1 (fig.53, 54), and the two sclerites articulate
(A2 in fig.53, 54). From A2 sclerite L2 extends to the right, curves ventrad and — becoming
narrower — back to the left; then it bends posteriad and runs to the posterior edge of the

60

via

at
goer

S&S

QQ,

S
a
Qa

Archiblatta
hoeveni

61

left complex; here it ends at the base of a small process paa (fig.56). This arch-like course
of L2 extends along the margins of a pouch lve, which is, like L2 itself, curved dorso-
ventrally (in contrast to Mantodea, whose Ive-pouches extend within one plane). Thus, the
left parts of the Ive-pouch and of L2 lie beneath the pne-pouch as in Mantodea, but their
right parts curve upwards into the same plane which also contains sclerite L1. The edge
along the bottom of this Ive-pouch is labelled 7 in fig.54, 55. The invagination of the Ive-
pouch starts immediately anterior to the paa-process (posterior end of 7 in fig.55), and it
ends in the anterior dorsal wall of the left complex (dorsal part of 7 in fig.54), where L2
leaves the pouch and approaches L1. According to the curvature of the Ive-pouch, it is
preferable to name that wall of Ive containing sclerite L2 the inner one (instead of dorsal)
and the opposite wall the outer one (instead of ventral). L2 is restricted to the inner Ive-
wall; only most posteriorly it bends into the outer (or ventral) Ive-wall. Then it leaves the
pouch and ends on the paa-process (fig.55, 56). paa is, except for another small distal
sclerite (probably a split off part of L2), membranous (fig.55-57).

The membranous outer (or ventral) wall of the Ive-pouch is at the same time the dorsal
wall of the vla-lobe (fig.53-56; in the figures vla is pulled to the right). The ejaculatory
duct (D in fig.53) opens into this wall. Dorsal to the genital opening there is a small
membranous lobe goa. The ventral wall of the vla-lobe is sclerotised by L4G. The
posterior edge of the vla-lobe continues leftward into the posterior edge of the remaining
left complex, where the paa-process follows (fig.53, 57).

The large hook hla is evaginated from the left anterior ventral wall of the left complex.
hla is, except for its basalmost walls (30 in fig.54-56), completely sclerotised by L3
(fig.53-55). hla is retractable for a very short distance, since the basal membranous walls
30 can be introverted (this state is shown in the figures).

Right phallomere

The right phallomere is only schematically shown in fig.330f. Differences to Eurycotis
(following species) will be explained in 6.7.4., 6.7.5., and 6.7.6.

5.6. Eurycotis floridana (Blattaria, Blattidae, Polyzosteriinae)

Left complex

The left complex resembles that of Archiblatta. L4 is a group of three sclerites: The
crescent-shaped L4H occupies the left edge and the left anterior ventral wall of the left
complex (fig.65, 66). That part of L4H in the left edge broadens in its posterior half. A
beam-like apodeme swe (fig.65, 66; anterior and posterior cross-section in fig.66) runs
along L4H. At the anterior end of the swe-apodeme L4H bends abruptly posteriad and
broadens into a plate bearing the node-like process nla (fig.68, 69) in its left half. The

Figs.53-57: Archiblatta hoeveni (Blattaria, Blattidae, Blattinae) — 53: Left complex in dorsal view. —
54-57: Left complex in dorsal view; with successive removal of its parts (mainly of dorsal ones). —
Scale: 1mm.

62

plate L4F lies in the posterior ventral wall. Along its anterior margin the membrane is
evaginated posteroventrad to form the broad lobe mla (fig.63, 69). L4G occupies most of
the ventral and right walls of a large ventral lobe vla (= ventral phallomere; fig.63, 65,

66).

lig

PO

S9d

| Eurycotis
58 floridana

Fig.58: Eurycotis floridana (Blattaria, Blattidae, Polyzosteriinae) — Male postabdomen in dorsal view;
with phallomere complex, subgenital plate, marginal parts of abdominal tergites 9 and 10, supraanal
lobe, subanal lobes (covered), paraprocts (mostly covered), Pv-sclerites, distal part of rectum, basal
parts of cerci, parts of vasa deferentia, and part of musculature. Supraanal lobe shown through a
window cut into the membrane anterior to ventral sclerotisation of abdominal tergite 10 T10v. — Scale:
2mm.

63

Sclerite L1 lies in the central dorsal wall of the left complex (fig.65-67). Its anterior part
occupies the ventral wall of a pouch-like invagination (pne in fig.65-67). Its posterior part
has a longitudinal furrow (8 in fig.67) and extends onto two processes (dca in fig.67, 68).
The phallomere-gland (P in fig.68) opens into the ventral wall of the right dcea-process.
Two small sclerites L6A and L6B lie in the left-dorsal wall of the pne-pouch; each bears

TS

Po}

| Eurycotis :
59 J ses floridana Ss

Fig.59: Eurycotis floridana (Blattaria, Blattidae, Polyzosteriinae) — Male postabdomen as in fig.58,
after removal of further parts of abdominal tergites 9 and 10 (especially T10v) and supraanal lobe.
Distal part of rectum, basal parts of cerci, and dorsal wall of right subanal lobe cut open. Another

part of musculature shown. — Scale: 2mm.

64

a spine (fig.65, 66). Sclerite L2 adjoins to the right of L1 (fig.67), and the two sclerites
articulate (A2 in fig.67). Like in Archiblatta, L2 is dorsoventrally curved (fig.67, 68), and
the anterior part of L2 lies in the inner wall of a large pouch (Ive in fig.65-68; the edge
along the bottom of the Ive-pouch is labelled 7 in fig.66-68). The anteroventral parts of
L2 and lve deepen strongly to the left to form a tongue-like apodeme (lve in fig.67). At
the posterior margin of L2 there are three processes, the largest of which is the completely
sclerotised paa (fig.67).

The outer wall of the Ive-pouch is completely membranous and is at the same time the
dorsal wall of the vla-lobe (fig.65-67; in the figures vla is pulled to the right). The
ejaculatory duct (D in fig.65, 66) opens into the anterior part of this wall. Dorsal to the
genital opening there is a membranous lobe goa (fig.66). In contrast to Archiblatta, the
main part of the vla-lobe is separated by a deep notch (9 in fig.63) from the remaining
ventral parts of the left complex. The hla-hook (fig.65-67) is like in Archiblatta.

Right phallomere

Sclerite R3 in the anterior (right-)ventral wall is a curved plate (fig.74-77). The right and
the right anterior margins of R3 form a groove-like apodeme age (fig.74, 77; cross-section
through age in fig.78). The right part of age bears a keel-like apodeme (3 in fig.74, 77).

Posterior to the left part of R3 sclerite R2 adjoins, and the two sclerites articulate (A7 in
fig.75-77). R2 forms a dental ridge (fig.74-77). Posterior to the central part of R3 the
ventral wall of the right phallomere is extensively invaginated dorsad and anteriad (cbe
in fig.74-76; compare fig.77 and 78). This cbe-invagination takes a position in the center
of the right phallomere.

Posterior to the right end of R3 sclerite RIF adjoins, and the two sclerites are articulated
(A3 in fig.74, 75, 77). RIF extends from its central part behind A3 in three directions:
The first arm bends left-dorsad (along edge 16, fig.77, 78) and occupies the whole right-
dorsal wall of the cbe-invagination (fig.74, 75, 78). This arm forms a dental ridge (pva
in fig.75, 78, 80) at its posterior margin and articulates with the left-dorsal end of R2 (A6
in fig.75, 76) at its median end. The second arm of RIF extends posteriad and sclerotises
the anterior part of a two-pointed ventral tooth (pia in fig.77). The third arm of RIF
extends posterodorsad (fig.74); its dorsal margin folds back to the right to form a sclero-
tised groove (rge in fig.74); the rge-groove is a posterior continuation of the age-groove
on R3.

The posterior part of the right phallomere is composed of the dorsal lobe fda (fig.74) and
the ventral tooth pia (fig.77, 78). fda and pia are confluent along the right edge of the
right phallomere and diverge towards the left. The dorsal wall of the fda-lobe and parts
of its ventral wall are occupied by sclerite RIH (fig.74, 76). The left part of RIH scle-

Figs.60-62: Eurycotis floridana (Blattaria, Blattidae, Polyzosteriinae) — 60: Male postabdomen in
dorsal view; with phallomere complex, subgenital plate, and lateral parts of abdominal tergite 9. —
61: Left margin of subgenital plate (compare fig.60); with insertion of muscle p6. — 62: Subgenital
plate in dorsal view; with insertion areas of muscles (except p6). — Scale: 2mm.

65

Eurycotis

floridana

66

Eurycotis
floridana

Figs.63-65: Eurycotis floridana (Blattaria, Blattidae, Polyzosteriinae) — 63: Phallomere complex in
ventral view. — 64: Phallomere complex in ventral view; with some muscles; ventral wall of genital
pouch more complete than in fig.63. — 65: Left complex in dorsal view. — Scale: 2mm.

67

Euryeotis

$
Hm = N a
SST SS Ss =
OS
S =
= 'S
os} % ;
N Yu.
\ ot >
“m 8 IH
= 3 27
U
o ;
= z

ST /
PLOY YEG?

’

1EW

dorsal vi

ın

) — Left complex

rıınae

Eurycotis floridana (Blattaria, Blattidae, Polyzoste

.

Figs.66-69

2mm.

with successive removal of its parts (mainly of dorsal ones). — Scale

68

S & | Ee © ee ©
na Se sa Se
Sis} SS) i! GS)
2 3S @ x iS) ns
£ Sao 5 &
Sa 8S & no oO
Mas asreoeupeer gs
= Bes aaSE
Su,
een
SQ x) -
Aes ae 2a 2 82%
Sm Oo 6 & 2 aS
_ m oO & ©) Gy ©)
ee SS] oy et, S|) we) GS SOE OD

-
nr
ee ge
I u
Be RE

ao
= ____.8
< : wd

si

Eurycotis
floridana

69

mean

x

ear

Den ae mann net ner Meran Eh Ma a Ne eet

A

Eurycotis

floridana

VPEr ete sie

Figs.72,73: Eurycotis floridana (Blattaria, Blattidae, Polyzosteriinae) — Left complex in dorsal view;

each figure with some muscles; parts of left complex removed to various extents. — Scale: 2mm.

70

rotises the spine sra; the left end of RIH is somewhat spoon-shaped. The right end of
RIH extends posteriorly onto the process pra (fig.74) and articulates anteriorly with the
dorsal arm of RIF (A8 in fig.74). The tendon tre (fig.74), a very long and narrow
invagination of the cuticle, has its origin in the anteriormost dorsal wall of the fda-lobe.
The ventral tooth pia is anteriorly sclerotised by the second arm of RIF (fig.77, 78);
posteriorly pia is sclerotised by R1G, which articulates with RIF in the dorsal wall of
pia (A9 in fig.78). RIG has a narrow extension to the right, which reaches the pra-process
and maintains a narrow connection with the dorsal RIH (fig.74, 77, 78). Anterior to this
sclerite bridge there is a large membranous area (17 in fig.74, 77).

Subgenital plate and posterior abdominal segments

Fig.58, 59 (posterior segments); fig.62 (subgenital plate S9). The ventral part of tergite 10
T10v is very extensive. Strip-like Pv-sclerites are present; they are completely free from
the paraprocts Pp. The Ca-sclerites lie on curved bulges immediately median to the cercal
bases. The Cc-sclerites are dorsoventrally curved plates on paired shallow outfoldings
beneath the anterior margin of T10v and above the subanal lobes sbl. Cb-sclerites are
missing. The articulations A98 and A99 are well-developed.

Musculature
Muscles Positions of insertions in fig.
12 Membranous left wall of pne-pouch — L4H (left part, on swe-apodeme) 70
l3a,b,c L1 and membrane to the left of L1 (posteroventrally on pne-pouch) — L2

(anterodorsal part) al
14 L2 (anterodorsal part) — L4H (left part, on swe-apodeme) WN
15a L4F (anterior margin) — L2 (ventral left margin) V2
15b Membranous ventral wall of left complex anterior to L4F — L2 (ventral

left margin) V2, 19
16a L4H (plate-like part in ventral wall of left complex) — anterior ventral

wall of Ive-pouch 73
l6b L4G (anterior margin) and membrane anterior to L4G — dorsal wall

of vla-lobe around genital opening WO, 1
19 Transversely in central dorsal wall of left complex 9910
113a-i Muscles within vla-lobe; mainly diffuse TM, WZ, 73
114c L4H (to the right of nla-process) — hla-hook (dorsally on L3) 0, 122, 13
114d L4H (on nla-process) — hla-hook (ventrally on membranous base 30) WD
115 Ejaculatory duct D next to its opening — L2 (anterior margin) 70
116 L4H (plate-like part in ventral wall of left complex) — L4F (anterior

margin) 13
117 L4H (plate-like part in ventral wall of left complex) and membrane

posterior to it — ventral wall of mla-lobe 13

Figs.74-78: Eurycotis floridana (Blattaria, Blattidae, Polyzosteriinae) — 74: Right phallomere in dorsal
view. — 75: Right phallomere in left-dorsal (somewhat anterior) view. — 76: Right phallomere in left-
ventral view. — 77: Right phallomere in right-ventral view; with transition to left complex. — 78:
Right phallomere in right-ventral view; most elements shown in fig.77 removed. — Scale: Imm.

71

sra

Eurycotis
floridana

W

118

rl
r2

r3
r5
r6

b4a

b4b
sl

s2
s3

s4
s5a

s5b
s5c
s6a
s6b
s6c

s6d
s7
s8

pl (pair)
p2 (pair)

p3 (pair)
p4 (pair)

p5 (pair)
p6 (pair)
p7 (pair)
p9 (pair)

Both insertions on ventral wall of mla-lobe

R3 (right margin) — RIH (anterior dorsal wall of fda-lobe)

R3 — cbe-invagination: RIF (anterior margin), membrane, and R2
(dorsal margin)

RIF (central part, posterior to articulation A3) — R1G (on pia-tooth)
R3 (right margin) — RIF (dorsal margin, on rge-apodeme)

RIF (dorsal margin, on rge-apodeme) — RIH (anterior dorsal wall
of fda-lobe)

Outer wall of Ive-pouch (anterior part next to genital opening) —
tre-tendon
pne-pouch (anterodorsal wall) — tre-tendon

S9 left side (anteriorly on apophysis S9a) — L4H (anteriormost left
ventral wall of left complex)

S9 right side (anteriorly on apophysis S9a) — R3 (anterior margin)
S9 left side (anteriorly on apophysis S9a) — anteriormost ventral wall
of left complex

S9 right side (anteriorly on apophysis S9a) — R3 (left anterior margin)
S9 left side (posterior to apophysis S9a, medially) — anterior left
ventral wall of genital pouch; diffuse

S9 left side (posteriorly and laterally) — posterior left ventral wall of
genital pouch

S9 left side (posteriorly and far medially) — median ventral wall of
genital pouch; small, diffuse

S9 right side (posteriorly on apophysis S9a) — anteriormost right
ventral wall of genital pouch

S9 right side (posteriorly and laterally) — posterior right ventral wall
of genital pouch

S9 right side (posterior to apophysis S9a) — median ventral wall of
genital pouch; small, diffuse

S9 right side (posteriorly and far laterally) — right wall of genital pouch
S9 left side (anteriorly on apophysis S9a) — L2 (anteriorly on Ive-pouch)

S9 right side (most anteriorly on apophysis S9a) — tre-tendon

S9 — Py-sclerite (median part)

S9 — T9 (lateral anterior margin); in most specimens the p2 pass
through eyelets of the vasa deferentia Vd. A pair of muscles having
the same course is also present in segment 8: p2(8).

S9 — rectum (ventral wall)

T9 (lateralmost anterior margin) — membrane far anterior to Pv-sclerite;

ventral insertion of left p4 far to the right
T10 (lateralmost anterior margin) — Pv-sclerite
T9 (lateralmost part) — S9 (lateral margin)

Membrane anterior to Pv-sclerite — anterior margin of paratergite T10p

Membrane anterior to Pv-sclerite (next to p7) — membrane median to
inner end of Pv-sclerite, beneath rectum

13

Io 19

80, 81

80, 81, 82
58, 80

79

Io 0, 79
St) 10, 79

59, 62, 64 70
59, 62, 64, 82

59, 62, 64, 70
59, 62, 64, 82

62, 64

59, 62, 64
62, 64

59, 62, 64, 82
59, 62, 64

62, 64

59, 62, 64
58, 62, 64, 70

de, OZ, 10, 79
58, 62

58, 62
58, 62

58
58
deh (Ol
39

I)

73

ood

Eurycotis
floridana

nd

Figs.79-82: Eurycotis floridana (Blattaria, Blattidae, Polyzosteriinae) — 79: Right phallomere in dorsal
view; with some muscles; some dorsal elements removed. — 80: Right phallomere in dorsal view;
with some muscles; dorsal elements more extensively removed than in fig.79. — 81: Right phallomere
in left-dorsal (somewhat anterior) view; with some muscles; some dorsal elements removed. — 82:
Right phallomere in right-ventral view; with some muscles; ventral wall of genital pouch more
complete than in fig.77; pia-tooth cut open to show muscle r3. — Scale: 1mm.

74

5.7. Tryonicus parvus (Blattaria, Blattidae, Tryonicinae)

Left complex

L4 is a group of three sclerites in the ventral wall and on the left edge of the left complex
(fig.87, 88, 91): L4G occupies the ventral wall of a broad ventral lobe vla (= ventral
phallomere; fig.87, 88, 98). LAK is a broadly horseshoe-shaped sclerite in the left anterior
ventral wall and bears a bulge-like process (nla in fig.87-89, 97, 98). LAN occupies the
posterior left edge of the left complex and has a hinge-like articulation with L4K (A5 in
fig.88, 97, 98). Anterodorsally LAN has a long ribbon-like extension to the anterior (L4d
in fig.88-97). The posterior end of LAN is on a short bulge-like process pda (fig.91, 94,
96) and is to the right connected with the L2-sclerotisation of another process paa (fig.94).

The small sclerite L11 lies in the anteriormost left dorsal wall (fig.88, 91). The right
anterior part of the dorsal wall has some complicated membranous foldings (fig.91-93).
Beneath and posterior to these foldings there are the somewhat cup-shaped L1-sclerite
and, to the right of L1, the dorsoventrally curved L2-sclerite. The plate-like dorsal part
of L1 (fig.94) has a hinge-like articulation with L2 (A2 in fig.94). The posterior parts of
L1 occupy two bulge-like processes (dea in fig.94, 95; only the ventral wall of the right
dca-process is membranous). The ventral part of L1 sclerotises the dorsal wall of a broad
pouch (pne in fig.95). This pne-pouch is strongly deepened in its rightmost part, where
L1 has a ribbon-like extension to the anterior. At the left end of the pne-pouch L1 is in
close contact with the anterior end of the L4d-ribbon (fig.94, 95). The ventral wall of the
pne-pouch is membranous and contains the opening of the phallomere-gland (P in fig.96).

L2 occupies the dorsal wall right-anterior to L1 (fig.94). From here it curves ventrad and
back to the left (fig.95) into a position beneath the pne-pouch; then it extends, becoming
narrower, to the posterior edge of the left complex (fig.97). The posterior end of L2 occu-
pies the paa-process, which is upcurved and somewhat hook-like. For most of its length
L2 extends broadly along the anterior edge of a pouch (Ive in fig.94, 95, 97), which is
dorsoventrally curved — in the same way as L2 itself and as it has been described for L2
and lve of Archiblatta. The edge along the bottom of the Ive-pouch is labelled 7 in fig.92,
94, 96, 97. Ventrally the invagination of the lve-pouch begins at the posterior right margin
of L4K (posterior end of edge 7 in fig.97). Immediately posterior to this point L2 and
L4K are in close contact. L2 is restricted to the inner wall of the Ive-pouch. The outer
Ive-wall is completely membranous and is at the same time the dorsal (or left) wall of the
vla-lobe (fig.91, 92; in the figures vla is pulled to the right). The ejaculatory duct (D in
fig.87, 91, 92) opens far anteriorly into this membrane. In the rightmost dorsal wall of
the vla-lobe there lies a ribbon-like extension of the R2-sclerite of the right phallomere
(R2m in fig.91, 101, 102).

Figs.83,84: Tryonicus parvus (Blattaria, Blattidae, Tryonicinae) — 83: Male postabdomen in dorsal
view; with phallomere complex, subgenital plate, marginal parts of abdominal tergites 9 and 10,
supraanal lobe, subanal lobes, paraprocts, Pv-sclerites, distal part of rectum, and basal parts of cerci.
— 84: Same as in fig.83, after removal of further parts of abdominal tergites 9 and 10 (especially
T10v), parts of paraprocts, and supraanal lobe. Distal part of rectum, basal parts of cerci, and right
subanal lobe cut open. Anterior to transverse line: like in fig.83. — Scale: Imm.

78

DDIIIIIISSSSSSSSTCUUNUN

Oe =

76

.

\O
eo
a)
[0,0]
nH
lot)
Pion!
FL

parvus

Tryonicus

(Blattaria, Blattidae,

Tryonicinae) — 85

Male postabdomen

‚ with

1e¢Ww

dorsal v
phallomere complex,

in

subgenital plate, and

of

l tergite 9.

parts

lateral

mina

86
plate in dorsal view.
Imm.

abdo
— Scale

Subgenital

TT)

tt

Tryonicus parvus

Figs.87-90: Tryonicus parvus (Blattaria, Blattidae, Tryonicinae) — 87: Phallomere complex in ventral
view. — 88: Left complex in left-dorsal view. — 89: Left complex in left-dorsal view; some parts
removed (compare fig.88). — 90: hla-hook and sclerite L3 in left-dorsal view. — Scale: 1mm.

78

Figs.91-94: Tryonicus parvus (Blattaria, Blattidae, Tryonicinae) — 91: Left complex in right-dorsal
view. (In the text this view is designated as dorsal). — 92-94: Left complex in right-dorsal view; with

successive removal of its parts (mainly of dorsal ones); figs.92,93: scale slightly larger; fig.93: detail
from fig.92, some more parts removed. — Scale: 1mm.

37

Tryonicus parvus

Figs.95-98: Tryonicus parvus (Blattaria, Blattidae, Tryonicinae) — Left complex in
with successive removal of its parts (mainly of dorsal ones). — Scale: Imm.

79

right-dorsal view;

80

The large hook hla is evaginated from the left ventral wall of the left complex. The distal
part of hla is sclerotised by L3 (fig.87-90, 97, 98), a large basal part of hla is membranous
(30 in fig.97, 98). By introverting this membrane 30, hla can be retracted rather far into
the phallomere (the retracted state is shown in the figures). L3 is connected with L4K by
a ribbon of weak sclerotisation (L3a in fig.89, 98), which crosses the membrane 30.

Right phallomere

Sclerite R3 occupies the (right-)anterior ventral wall (fig.99-103). Sclerite R2 articulates
with the left posterior end of R3 (A7 in fig.100-102). R2 is a plate of irregular shape,
with a ridge-like elaboration in its left-dorsal part (fig.99, 100, 103). The right-ventral part
of R2 has the extension R2m (fig.101, 102) mentioned above. Posterior to the central part
of R3 the ventral wall of the right phallomere curves dorsad and somewhat anteriad to
form a central invagination (cbe in fig.99-101; compare fig.102 and 104).

Posterior to the right part of R3 sclerite RIF adjoins, and the two sclerites articulate (A3
in fig.99, 100, 102). RIF extends from its central part behind A3 in two directions: The
ventral arm bends left-dorsad (along edge 16, fig.102, 104) and largely occupies the right-
dorsal wall of the cbe-invagination (fig.99, 104). The posterior margin of this arm forms
a ridge (pva in fig.99, 103). The median end of this arm articulates with the left-dorsal
end of R2 (A6 in fig.99, 100, 101). The dorsal arm of RIF extends posterodorsad (fig.99,
100, 102); its dorsal margin folds back to the right and ventrad to form a sclerotised groove
rge (fig.99, 102). Behind A3 and on the ventral arm the sclerotised cuticle is considerably
thickened (cross-sections in fig.104).

The posterior part of the right phallomere is a large dorsal lobe fda (fig.99, 103), whose
dorsal and ventral walls are partly occupied by sclerite R1J. The dorsal anterior end of
R1J articulates with the dorsal arm of RIF (A8 in fig.99), its ventral anterior end is in
close contact with the base of the ventral arm of RIF (A9 in fig.102, 103). The tendon
tre (fig.99) originates from the anteriormost dorsal wall of the fda-lobe. In its right wall
the fda-lobe has a large membranous area (17 in fig.99, 103).

Subgenital plate and posterior abdominal segments

Fig.83, 84 (posterior segments); fig.86 (subgenital plate S9). The ventral part of tergite 10
T10v is moderately extensive. Strip-like Pv-sclerites are present; they are laterally
connected with the paraprocts Pp. The Ca-sclerites are curved ribbons immediately median
to the cercal bases (no distinct bulges present). The Cc-sclerites are dorsoventrally curved
plates on paired shallow outfoldings beneath the anterior margin of T10v and above the
subanal lobes sbl. Cb-sclerites are missing. The articulations A98 and A99 are well-
developed.

Figs.99-104: Tryonicus parvus (Blattaria, Blattidae, Tryonicinae) — 99: Right phallomere in dorsal
view. — 100: Right phallomere in left-dorsal (somewhat anterior) view. — 101: Right phallomere in
left-ventral view. — 102: Right phallomere in right-ventral view. — 103: Right phallomere in dorsal
view; some dorsal elements removed. — 104: Right phallomere in right-ventral view; most elements
shown in fig.102 removed. — Scale: Imm.

81

Tryonicus parvu

82

>

10

capensis

107

angustus

83
5.8. Polyphaga aegyptiaca (Blattaria, Polyphagidae, Polyphaginae)

Left complex

Sclerite L1 is situated in the central dorsal wall; its anterior part lies within a deep pouch
(pne in fig.117). Beneath the pne-pouch there is another very deep and broad pouch (Ive
in fig.118, 122), with the arch-shaped L2-sclerite extending along its edges. The left
posterior dorsal wall contains the sclerotisation of two processes (pda and paa in fig.117,
118) and of an intervening invagination (dte in fig.117, 118), which is firmly connected
with the L2-sclerotisation in the Ive-pouch around the posterior edge of the left complex
(10 in fig.118, 122). This dorsal sclerotisation is composed of parts of L2 (right part) and
of L4 (left part: L4N). Sclerite L8 lies in the right dorsal wall of the left complex (fig.117).
From the left wall of the left complex there protrudes a large hook-process (hla in fig.117)
with its L3-sclerite. The large L4M-sclerite occupies the ventral wall (fig.115, 117, 126).
Another L4-sclerite lies within the ventral base of the hla-hook (L4K in fig.122-124).

The anterior hood-shaped part of L1 largely occupies the walls of the pne-pouch (fig. 111,
118, 119, 120). In the left-dorsal half of the pne-pouch the walls are membranous —
especially in the posterior part, where the phallomere-gland (P in fig.118, 120, 121) opens
from ventrally. The anterior end of L1 is flat and plateau-like. Posteriorly L1 leaves the
pouch and has an arm-like extension on each side (fig.118, 120, 121). The extensions
curve ventrad and approach each other again (fig.121). The membranous cuticle enclosed
by this (open) sclerite-ring forms two cushion-like evaginations (dea in fig.117, 120, 121).

The Ive-pouch (fig.118, 122) spans almost the whole breadth of the left complex. (The
edge along the bottom of the Ive-pouch is labelled 7 in fig.122, 123). L2 occupies the
margins of the dorsal (fig.122) and ventral (fig.123) walls of the Ive-pouch. The left part
of L2 in the dorsal lve-wall broadens posteriorly and bends around the posterior edge of
the left complex (along 10 in fig.118, 119a, 122) into the dorsal wall. Here it continues
into the sclerotisation of the paa- and pda-processes and of the dte-invagination (fig.117-
119b; L2 and LAN, with the border between them somewhere within dte). The pda-process
is finger-like; the paa-process is saucer-shaped and partially encloses the dea-processes
from left-ventrally. The pda-sclerotisation has a tongue-like extension to the left (L4d in
fig.118, 123, 124). The paa-sclerotisation has an arm-like extension to the right (12 in
fig.118, 119a). The right parts of L2 and of the Ive-pouch curve dorsad and back to the
left (fig.118, 122, 123), and along this curvature the Ive-pouch becomes rapidly less deep.

Figs.105-108: 105,106: Ergaula capensis (Blattaria, Polyphagidae, Polyphaginae) — 105: Sclerite L1
in dorsal pouch pne in dorsal view; with some surrounding elements and phallomere-gland P. — 106:
Sclerite L1 in dorsal pouch pne in ventral view; with some surrounding elements and phallomere-
gland P. — 107,108: Tryonicus angustus (Blattaria, Blattidae, Tryonicinae) — 107: Sclerite L1 in dorsal
pouch pne in left-dorsal view; with some surrounding elements and phallomere-gland P. — 108:
Sclerite L1 in dorsal pouch pne in right-ventral view; with some surrounding elements and
phallomere-gland P. — Scale: 1mm.

84

This recurved dorsal part of the Ive-pouch, with its dorsal and ventral walls sclerotised
by L2, approaches L1 and articulates with it (A2 in fig.118, 120, 121). The ventral wall
of the Ive-pouch also shows this dorsoventral curvature, but it is in its posterior part
additionally invaginated to the right, and the invagination is strengthened by an arm-like
extension of L2 (11 in fig.118, 122, 123). The posterior end of this L2-extension is, around
an edge, in close contact with the posterior margin of sclerite L8 (fig.117, 118).

The ventral wall of the Ive-pouch is, except for the L2-sclerotisations along its margins,
membranous; it is at the same time the anterior dorsal wall of the very broad ventral lobe
vla (fig.123). The ejaculatory duct (D in fig.123, 124) opens into the right part of this
membrane. Dorsal to the genital opening there are some membranous outfoldings (goa in
fig.122-124). Sclerite L5 lies in the left dorsal wall of the vla-lobe and is in close contact
with the posterior margin of L2 (fig.123, 124).

The hla-hook (with L3) is evaginated from the left wall of the left complex (fig.117). The
base of hla is rather complicated (fig.122-125a) and contains sclerite L4K in its
posteroventral part. L4K shows a horseshoe-like dorsoventral curvature, with a broad
ventral and a pointed dorsal part. The rightmost part of the left complex is the lobe Iba
with sclerite L7 in its ventral wall (fig.115, 117, 118). The Iba-lobe is distinctly separated
from the vla-lobe.

Right phallomere

Sclerite R3 occupies the anterior (right-)ventral wall (fig.134-137). The right margin of
R3 forms a groove-like apodeme age (fig.134, 137). Posterior to the left part of R3 sclerite
R2 adjoins (fig.135-137), and the two sclerites are fused; a strip of weaker sclerotisation
is probably the suture (A7* in fig.135-137). Along its ventral margin R2 forms a ridge
bearing several processes (fig.136, 137, 141), the largest of which is behind A7*. Posterior
to the central part of R3 the ventral wall of the right phallomere curves dorsad and
somewhat anteriad to form a sclerotised central invagination (cbe in fig.134-136; compare
fig.137 and 138).

— 20,85

Figs.109,110: Polyphaga aegyptiaca (Blattaria, Polyphagidae, Polyphaginae) — 109: Male
postabdomen in dorsal view; with phallomere complex, subgenital plate, marginal parts of abdominal
tergites 9 and 10, subanal lobes, paraprocts, distal part of rectum, basal parts of cerci, and part of
musculature. — 110: Same as in fig.109, after removal of further parts of abdominal tergites 9 and
10 (especially T10v). Distal part of rectum and basal parts of cerci cut open. Another part of
musculature shown. — Scale: 1mm.

<=) OS®)

Figs.111-114: Polyphaga aegyptiaca (Blattaria, Polyphagidae, Polyphaginae) — 111: Male
postabdomen in dorsal view; with phallomere complex, subgenital plate, and lateral parts of abdominal
tergite 9. — 112a,b: Left (a) and right (b) margins of subgenital plate (compare fig.111); with insertions
of muscles p6. — 113: Subgenital plate in dorsal view; with insertion areas of muscles (except p6);
most of dorsal sclerotisation S9d of subgenital plate removed. — 114: Subgenital plate in dorsal view;
anterior part of ventral sclerotisation removed, dorsal sclerotisation S9d complete. — Scale: 1mm.

86

RQ

Ges

SSW

SS

N

zu

pansies

EZ

ER

0 IIE DIGG,

RAGE

N

SOs

87

Figs.115,116: Polyphaga aegyptiaca (Blattaria, Polyphagidae, Polyphaginae) — 115: Phallomere
complex in ventral view. — 116: Phallomere complex in ventral view; with some muscles; ventral
wall of genital pouch more complete than in fig.115, with parts of dorsal sclerotisation S9d of
subgenital plate in its posterior part (compare fig.114). — Scale: 1mm.

88

Posterior to the right part of R3 the large RIM-sclerite adjoins, and the two sclerites
articulate (A3 in fig.134, 135, 137). From its central part behind A3 sclerite RIM extends
to the left, where it bends left-dorsad (along edge 16; fig.137, 138) and then occupies the
right-dorsal wall of the cbe-invagination (fig.134). The left-ventral wall and the top of cbe
are sclerotised by a plate-like part of R2 (fig.134-137); the dorsal margin of this R2-part
is fused to the anterior margin of RIM. A line of weaker sclerotisation (13 in fig.134,
138) is probably the boundary between R2 and RIM.

From its fusion line 13 with R2 and from articulation A3, RIM extends far posteriad as
a dorsoventrally curved plate of irregular shape (fig.134, 137). It largely occupies the
ventral and right walls of the posterior part of the right phallomere (labelled fda and pva
in fig.134-138). The dorsal margin of RIM folds back to the right to form a sclerotised
groove (rge in fig.134, 138, 140); rge is a posterior continuation of the age-groove on
R3 and extends to the posterior edge of the fda-lobe. Dorsal to rge there is an outfolding
to the right, which contains the very weak ribbon-like sclerites RIL (fig.134). At the left
end of the fda-lobe there is a dorsal outfolding to the left, which contains sclerite RIK
in its ventral wall (fig.134). Beneath RIK the left marginal part of RIM forms a
longitudinal ridge projecting to the left (pva, compare fig.134 and 139). The dorsal wall
of the fda-lobe is mostly membranous (fig.134), and most anteriorly the tendon tre has
its origin (fig.134, 135, 139).

Subgenital plate and posterior abdominal segments

Fig.109, 110 (posterior segments); fig.113, 114 (subgenital plate S9). The entire tergite 10
T10, including its ventral part T10v, is divided along its midline. T10v is moderately
extensive. Separate Pv-sclerites are missing; they are assumed to have been incorporated
into the anterior margins of the paraprocts Pp. The sclerites Ca, Cb, and Ce are missing.
The bulges next to the cercal bases the Ca-sclerites lie upon in other species, however,
are present (compare fig.59). The articulations A99 are well-developed. The articulations
A98 are missing: the sclerotisations Ell and T10 are far away from each other. Each
subanal lobe sbl has a small groove (14 in fig.110) beneath the cercal base.

——>} p.89

Figs.117-121: Polyphaga aegyptiaca (Blattaria, Polyphagidae, Polyphaginae) — 117: Left complex in
dorsal view. — 118: Left complex in dorsal view; some parts removed (mainly dorsal ones). — 119a,b:
Dorsal parts of sclerites L4 and L2, separated from remainder of left complex, in dorsal view. — 120:
Sclerite L1 in dorsal pouch pne, separated from remainder of left complex, in dorsal view; with some
surrounding elements and phallomere- gland P. — 121: Sclerite LI in dorsal pouch pne, separated
from remainder of left complex, in ventral view; with some surrounding elements and phallomere-
gland P. — Scale: Imm.

——> p.90

Figs.122-124: Polyphaga aegyptiaca (Blattaria, Polyphagidae, Polyphaginae) — Left complex in dorsal
view; with further successive removal of its parts (mainly of dorsal ones); fig.122: left complex after
removal of the parts shown in fig.119-121. — Scale: 1mm.

89

90

“

WE N

AS
SNS \\
MASS

91

IN

SRRERRRRRWN

Figs.125,126: Polyphaga aegyptiaca (Blattaria, Polyphagidae, Polyphaginae) — 125a,b: Left (a) and

right (b) part of left complex, separated from remainder of left complex, in dorsal view; further parts
removed (mainly dorsal ones; compare fig.124). — 126: Ventral wall of left complex in dorsal view.
— Scale: Imm.

92

Musculature
Muscles Positions of insertions in fig.
12 L1 (plateau-like anterior end of pne-pouch) — L4M (left anterior part) 127, 128, 129, 131
13 L1 (right-ventrally on pne-pouch) — L2 (right anterior part) 24807831
14 L2 (left anterior part) — L4K (dorsal right end, within base of

hla-hook) 297830282
15 L4M (anterior part) — L2 (left part) and membrane to the left of it 130, 133
16a L4M (anterior part) — L2 (right anterior part) 1302133
l6b L4M (posterior to 16a) — ejaculatory duct D next to its opening 132
19 Transversely within right dorsal wall of left complex; ventral parts:

L1 — L2 (areas next to articulation A2) 110727729530
110 L2 (left posterior part) — L4N and L2 on invagination dte 12930
111 L4K (anteroventral part) — L4d (= left part of LAN) Ws, WB
112 L2 (rightmost part) — L8; very short and stout 128-2980
113 Ejaculatory duct D next to its opening (ventral wall) — dorsal wall

of vla-lobe (rightmost part) 152
rl R3 (right margin) — RIM (dorsal margin, on posterior part of

rge-groove), RIL, and membrane in between these sclerotisations 109589
r2 R3 — cbe-invagination: RIM (anterior margin) and R2

(right-dorsal part) 140, 141
r6 R1M (dorsal margin, on rge-groove) — RIK and surrounding

membranes 140
r9 R3 (left-posterior part) — R2 (left-ventral part) 141
b2 Right dorsal wall of left complex (next to L8) — R3

(left-posterior part) 110, 127, 141
b4a L2 (right anterior edge of Ive-pouch) — tre-tendon NOD, WAI, W222,

1302159,

b4b L8 (anterior margin) — tre-tendon 109207289
sl S9 left side (laterally and anteriorly) — membrane anterior to

hla-hook U), WSs, TG. 27)
——> p.93

Figs.127-129: Polyphaga aegyptiaca (Blattaria, Polyphagidae, Polyphaginae) — Left complex in dorsal
view; each figure with some muscles; parts of left complex removed to various extents. — Scale:
Imm.

— p.94

Figs.130-133: Polyphaga aegyptiaca (Blattaria, Polyphagidae, Polyphaginae) — 130: Sclerite L2 in
ventral pouch Ive in dorsal view; with insertion areas (white) of muscles (compare fig.122). — 131:
Sclerite L1 in dorsal pouch pne in ventral view; with muscles 12 and 13 (compare fig.121). — 132,133:
Left complex in dorsal view; each figure with some muscles; parts of left complex removed to various
extents. — Scale: 1mm.

> [a5

Figs.134-138: Polyphaga aegyptiaca (Blattaria, Polyphagidae, Polyphaginae) — 134: Right phallomere
in dorsal view. — 135: Right phallomere in left-dorsal (somewhat anterior) view. — 136: Right
phallomere in left-ventral view. — 137: Right phallomere in right-ventral view. — 138: Right
phallomere in right-ventral view; most elements shown in fig.137 removed. — Scale: 1mm.

93

94

} — u

hb

rm

Zr

95

96

RIL

1

Polyphaga
aegyptiaca

Figs.139-142: Polyphaga aegyptiaca (Blattaria, Polyphagidae, Polyphaginae) — 139,140: Right
phallomere in dorsal view; each figure with some muscles; dorsal elements removed to various
extents. — 141: Right phallomere in left-dorsal (somewhat anterior) view; with some muscles; some
left-dorsal elements removed. — 142: Right phallomere in right-ventral view; with some muscles;
ventral wall of genital pouch more complete than in fig.137. — Scale: Imm.

97

s3 S9 left side (most anteriorly, median to sl) — L4M (anterior margin) 110, 113, 116,
1272133

s4 S9 right side (anteriorly) — R3 (left anterior margin) 110, 113, 116,142

s6 S9 right side (laterally and anteriorly) — R3 (right anterior margin) 110, 113, 116, 142

s8 S9 right side (anteriorly) — tre-tendon 1095113139

s12 S9 right side (most anteriorly, median to s8) — L4M (anterior margin) 109, 113, 116,
127133

pl (pair) S9 — paraproct Pp (anterior margin, corresponding to

Pv-sclerotisation); right muscle divided into two bundles 1095113
p2 (pair) S9 — T9 (lateral anterior margin) 109, 113
p3 (pair) S9 — rectum (ventral wall) 109, 113

p4 (pair) T9 (lateralmost anterior margin) — membrane anterior to paraproct
Pp (far laterally); right muscle inserting on lateral wall of genital

pouch, next to right phallomere; both muscles twisted 109
p5 (pair) T10 (lateralmost anterior margin) — paraproct Pp (lateral anterior
margin, corresponding to the Pv-sclerotisation) 109

p6 (pair) T9 (lateralmost part) — S9 (lateral margin) (left muscle only) and

adjacent membranes; ventral insertion of right muscle extending far

anteriad into right wall of genital pouch, lying immediately anterior

to insertion of p4 109, 112a,b
ps Longitudinally within membrane between Pv-sclerotisations 110

5.9. Cryptocercus punctulatus (Blattaria, Cryptocercidae)

Left complex

L4 is a group of four sclerites: The ribbon-shaped L4N (fig.150; L4d is part of L4N) lies
on a transverse outfolding of the dorsal wall, whose right end is somewhat lobe-like (paa
in fig.150). The ventral wall of this outfolding extends far anteriad, where L4K takes its
position. The small L4P, not present in all specimens, lies at the anteriormost left edge
of the left complex. L4G is a plate in the posterior ventral wall of the ventral lobe vla (=
ventral phallomere; fig.148, 152).

Sclerite L1 lies in the central dorsal wall. Its hood-shaped anterior part occupies most of
the walls of a deep pouch (pne in fig.150-154). The anterior summits of L1 and pne are
expanded and plateau-like, with upcurved margins. The posterior part of L1 leaves the
pne-pouch and has an arm-like extension on each side. These extensions curve ventrad
and join each other again to form a complete sclerite ring (fig.153, 154). The membranous
cuticle enclosed by this ring forms two cushion-like bulges dea, with a small sclerotised
peak between them (18 in fig.153). The phallomere-gland (P in fig.152, 153) opens into
the membrane to the left of L1.

The large sclerite L2 lies ventral to L1 (fig.152). To the posterior L2 becomes narrower,
and then it curves around the posterior edge of the left complex into the dorsal wall
(fig.151, 152). The area of this curvature forms a large bulge (paa in fig.150-152). In the
dorsal wall L2 extends anteriad as far as to the opening of the phallomere-gland. The right
anterior part of L2 lies in the dorsal wall of a pouch (Ive in fig.150-152; the edge along

98

RTIIIID

tocercus
unctulatus

IE

99

the bottom of the pouch is labelled 7). The broad ejaculatory duct (D in fig.150, 151)
opens most anteriorly into this Ive-pouch.

The ventral wall of the Ive-pouch is at the same time an anterior part of the dorsal wall
of the vla-lobe. The ribbon-shaped sclerite L5 (fig.151, 152) lies more posteriorly in the
dorsal vla-wall. The ventral vla-wall is membranous anteriorly and sclerotised by L4G
posteriorly.

The large hook hla (fig.150, 151), whose distal part is sclerotised by L3, is evaginated
from the left wall of the left complex — beneath and somewhat posterior to L4K.

Right phallomere

Sclerite R3 occupies the anterior (right-)ventral wall (fig.160-163). The lateral and anterior
margins of R3 form a weakly sclerotised groove-like apodeme age (fig.160, 163). Along
the right margin of R3 the ventral sclerotisation of age folds to the left (19 in fig.163,
164). Sclerite R2 articulates with the left part of R3 (A7 in fig.161-163). R2 has the shape
of a ridge (fig.161, 162). Posterior to the central part of R3 the ventral wall of the right
phallomere curves dorsad and somewhat anteriad to form a central invagination (cbe in
fig.160-162; compare fig.163 and 164).

Posterior to the right part of R3 sclerite RIF adjoins, and the two sclerites articulate (A3
in fig.160, 161, 163). From its central part behind the A3-articulation RIF extends in two
directions: The ventral arm bends left-dorsad (behind the edge 16 in fig.163, 164) and
extends into the right-dorsal wall of the cbe-invagination (fig.160, 164). This part of RIF
bulges outwards (pva in fig.163, 164) by the cuticle being extensively thickened (cross-
section in fig.164). The median end of this arm articulates with the left-dorsal end of R2
(A6 in fig.160, 164). The dorsal arm of RIF extends posterodorsad (fig.160, 163);
anteriorly its dorsal margin folds back to the right to form a sclerotised groove (rge in
fig.160, 163), which is a posterior continuation of the age-groove on R3.

The posterior part of the right phallomere is a large dorsal lobe fda (fig.160-163, 166),
whose dorsal and ventral walls are partly occupied by sclerite R1J. The dorsal anterior
tip of R1J approaches the dorsal arm of RIF (A8 in fig.160), its ventral anterior tip
approaches the base of the ventral arm of RIF (A9 in fig.163, 166). The tendon tre has
its origin in the anteriormost dorsal wall of fda (fig.160). In its right wall fda has a large
membranous area (17 in fig.160, 166). At the left end of fda there is another small sclerite
RIK (fig.160).

Figs.143a,b: Cryptocercus punctulatus (Blattaria, Cryptocercidae) — a: Male postabdomen in dorsal
view; with phallomere complex, subgenital plate, marginal parts of abdominal tergites 9 and 10,
supraanal lobe, subanal lobes, paraprocts, distal part of rectum, basal parts of cerci, and part of
musculature. Supraanal lobe shown through a window cut into the membrane anterior to ventral
sclerotisation of abdominal tergite 10 T10v. — b: Posterior insertion of muscle p4 at anterior margin
of abdominal tergite 10 T10. Enlarged detail from left part of fig.143a, further parts of abdominal
tergite 9 T9 and anterior part of p4 removed. — Scale: 1mm.

100

Cryptocercus

punctulatus |

Fig.144: Cryptocercus punctulatus (Blattaria, Cryptocercidae) — Male postabdomen as in fig.143a,
after removal of further parts of abdominal tergites 9 and 10 (especially T10v) and supraanal lobe.
Distal part of rectum, basal parts of cerci, dorsal wall of right subanal lobe, and anterior margins of
paraprocts cut open. — Scale: 1mm.

= 2,910

Figs.145-147: Cryptocercus punctulatus (Blattaria, Cryptocercidae) — 145: Male postabdomen in
dorsal view; with phallomere complex, subgenital plate, and lateral parts of abdominal tergite 9. —
146: Left margin of subgenital plate (compare fig.145); with insertion of muscle p6. — 147: Subgenital
plate in dorsal view; with insertion areas of muscles (except p6); dorsal sclerotisation S9d of
subgenital plate complete in the right part but largely removed in the left part. — Scale: 1mm.

— 2 1D, OW

Figs.148,149: Cryptocercus punctulatus (Blattaria, Cryptocercidae) — 148: Phallomere complex in
ventral view. — 149: Phallomere complex in ventral view; with some muscles; ventral wall of genital

pouch more complete than in fig.148. — Scale: 1mm.

101

si

m

unctulatu

102

tocercu

a ee ee —————u

103

Cryptocercus
punctulatus

=}

Fr,

= _..4

Figs.150-154: Cryptocercus punctulatus (Blattaria, Cryptocercidae) — 150: Left complex in dorsal
view. — 151, 152: Left complex in dorsal view; with successive removal of its parts (mainly of dorsal
ones); fig.152: pne-pouch with some adjacent parts completely cut off from the other elements. —
153: Sclerite L1 in dorsal pouch pne in dorsal view; with some surrounding membranes, part of
sclerite L2, and phallomere-gland P. — 154: Sclerite L1 in dorsal pouch pne in ventral view; with
some surrounding membranes and phallomere-gland P. — Scale: Imm.

104

Subgenital plate and posterior abdominal segments

Fig.143a,b, 144 (posterior segments); fig.147 (subgenital plate S9). The whole postabdo-
men is retracted anteriad and completely covered by the heavily sclerotised tergite and
sternite of abdominal segment 7. The tergal and sternal sclerotisations of the postabdomen
are rather weak. The ventral part of tergite 10 T10v is rather extensive. Separate Pv-
sclerites are missing; they are assumed to have been incorporated into the anterior margins
of the paraprocts Pp. The sclerites Ca, Cb, and Ce are missing, and there are no Ca-
bulges (compare fig.59). The articulations A98 are well-developed. A99 are not true
articulations since the contact between paratergite T10p and paraproct Pp is not very close.

Musculature
Muscles Positions of insertions in fig.
11 Membranous posterior dorsal wall of pne-pouch — LAN
(including L4d) and adjacent membranes 5559
12 L1 (plateau-like anterior end of pne-pouch) — L4K 156
13 L1 (right-ventrally on pne- pouch) — L2 (most of anterior half) 158,159
14 L2 (right anterior part) — L4K; very delicate I>, 158
l6b L4G (anterior margin) and membrane anterior to L4G — dorsal wall
of vla-lobe posterior to genital opening (in part on L5) 1594157]
17 Left ventral wall of left complex — left posterior edge of left complex 158
19 Transversely within anterior left dorsal wall of left complex 144, 155
110 L2 (posteriormost part, on paa-process) — membrane left-dorsal
to paa-process 153
113 Ejaculatory duct D next to its opening (dorsal wall) — anterior dorsal
wall of vla-lobe; anterior part of muscle divided into two bundles. 143a, 155
114 L4K and membrane anterior to base of hla-hook — hla-hook
(dorsal anterior margin of L3 and membrane anterior to it) 157
119 Left posterior ventral wall of left complex — hla-hook (ventral
anterior margin of L3 and membrane anterior to it) 156, 158
rl R3 (right margin) — anterior dorsal wall of fda-lobe, in part on
anterior margin of R1J and on base of tre-tendon 143a, 164, 165
r2 R3 — cbe-invagination: RIF (anterior margin and left part),
membrane, and R2 (dorsal margin) 166, 167
r3 RIF (dorsal and central parts, posterior to articulation A3) — R1J
(right margin) 166, 167, 168
r7 R3 (right anterior part) — tre-tendon 143a, 165
rs Both insertions on central part of R3 167
b4a Right dorsal wall of left complex — tre-tendon 143a, 156, 165
b4b Right dorsal wall of left complex — tre-tendon 143a, 165
b4c Central dorsal wall of left complex — tre-tendon 143a, 165

Figs.155-159: Cryptocercus punctulatus (Blattaria, Cryptocercidae) — 155-158: Left complex in dorsal
view; each figure with some muscles; parts of left complex removed to various extents. — 159: Sclerite
L1 in dorsal pouch pne in ventral view; with muscles 11 and 13 (compare fig.154). — Scale: Imm.

105

Ligne mem

Cryptocercus

S

unctulatu

106

Cryptocercus
punctulatus 164

Figs.160-164: Cryptocercus punctulatus (Blattaria, Cryptocercidae) — 160: Right phallomere in dorsal
view. — 161: Right phallomere in left-dorsal (somewhat anterior) view. — 162: Right phallomere in
left-ventral view. — 163: Right phallomere in right-ventral view. — 164: Right phallomere in right-
ventral view; most elements shown in fig.163 removed. — Scale: Imm.

107

166

165
Cryptocercus

unctulatus

168

167

108

s1+3(+7?) S9 left side (most anteriorly; only medially) — ventral basal line Bl

s2+4+6

s4b

s8
s10

pl (pair)

p2 (pair)
p3 (pair)

p4 (pair)
p5 (pair)
p6 (pair)

p7 (pair)

p10 (pair)

of left complex and membrane anterior to base of hla-hook; right
part of muscle inserting also on anterior margin of L2 (= s7?)

S9 right side (anteriorly; medially and laterally) — R3 (anterior
margin) and anterior ventral wall of genital pouch

S9 right side (medially and anteriorly) — R3 (right anterior margin);
present in some specimens only

S9 right side (most anteriorly) — tre-tendon
S9 right side (medially and most anteriorly) — ejaculatory duct D
next to its opening (right wall)

S9 — membrane anterior to paraproct Pp or Pv-sclerotisation;
very broad

S9 — T9 (lateral anterior margin)

S9 — rectum (ventral wall); divided into two groups of fibers on
both sides

T9 (lateralmost anterior margin, also extending onto paratergite
T9p) — T10 (lateralmost anterior margin, also extending onto
paratergite T10p)

T10 (lateral anterior margin) — paraproct Pp (lateral anterior margin,

corresponding to Pv-sclerotisation)

T9 (lateralmost part) — S9 (lateral margin)

Posteriad-directed outfolding anterior to paraproct Pp — membrane
anterior to “articulation” A99 (between paratergite T10p and
paraproct Pp)

Paratergite T10p (anterior margin) — paraproct Pp (lateral anterior
margin, corresponding to Pv-sclerotisation)

143a, 144, 147,
149, 157, 158

144, 147, 149,
168

143a, 144, 147,
149, 168
143a, 147, 165

143a, 147

143a, 147
143a, 147

143a, 147

143a,b

143a

143a, 146

144

143a, 144

5.10. Lamproblatta albipalpus (Blattaria, Blattidae, Lamproblattinae)

Left complex

Sclerite L1 is situated in the right anterior dorsal wall; its anterior part lies within a large
pouch pne (fig.177, 178). Posterior to L1 there is a complicated sclerotisation (L4T and
L2C: fig.177-179, 182) bearing two processes pda and paa. Beneath these elements there
is a large pouch Ive (fig.180) containing the sclerites L2A and L2B in its dorsal wall. The

<—— p.107

Figs.165-168: Cryptocercus punctulatus (Blattaria, Cryptocercidae) — 165: Right phallomere in dorsal
view; with some muscles. — 166: Right phallomere in dorsal view; with some muscles; some dorsal
elements removed. — 167: Right phallomere in left-dorsal (somewhat anterior) view; with some
muscles; some left-dorsal elements removed. — 168: Right phallomere in right-ventral view; with
some muscles; ventral wall of genital pouch more complete than in fig.163; fda-lobe cut open to

show muscle r3. — Scale: 1mm.

109

large hook-process hla (fig.177) with its sclerite L3 protrudes from the left wall. Sclerite
LAK takes a position dorsal to the base of hla (fig.177). LAR is the sclerotisation of the
ventral wall of a broad ventral lobe vla (= ventral phallomere; fig.174, 181).

The pne-pouch (fig.177, 178) is deep but rather flat. The anterior part of L1 occupies the
ventral wall of the pne-pouch and its anterior dorsal wall. The posterior part of L1

169 Lamproblatta albipalpus

Fig.169: Lamproblatta albipalpus (Blattaria, Blattidae, Lamproblattinae) — Male postabdomen in
dorsal view; with phallomere complex, subgenital plate, marginal parts of abdominal tergites 8, 9,
and 10, supraanal lobe, subanal lobes, paraprocts, Pv-sclerites, distal part of rectum, basal parts of
cerci, and part of musculature. — Scale: Imm.

110

170 Lamproblatta albipalpus

Fig.170: Lamproblatta albipalpus (Blattaria, Blattidae, Lamproblattinae) — Male postabdomen as in
fig.169, after removal of further parts of abdominal tergites 9 and 10 (especially T10v), parts of right
paraproct, and supraanal lobe. Distal part of rectum and basal parts of cerci cut open. Another part
of musculature shown. — Scale: Imm.

= ll

Figs.171-173: Lamproblatta albipalpus (Blattaria, Blattidae, Lamproblattinae) — 171: Male
postabdomen in dorsal view; with phallomere complex, subgenital plate, and lateral parts of abdominal
tergite 9. — 172: Left margin of subgenital plate (compare fig.171); with insertion of muscle p6. —
173: Subgenital plate in dorsal view; with insertion areas of muscles (except p6). — Scale: Imm.

111

SSA

S

A Mr

Lamproblatta
albipalpus

112

sclerotises a process (dea in fig.177, 178) and has a plate-like extension to the left, which
is separated from the main part of L1 by a strip of weaker sclerotisation (22 in fig.176,
I):

The complex sclerite posterior to L1 is composed of LAT - the sclerotisation of the spine-
shaped pda — and L2C - the sclerotisation of the cup-shaped paa (fig.176-179). pda is
almost completely sclerotised in its dorsal wall but only basally in its ventral wall (fig.179,
182), and it resembles a hypodermic needle: At its pointed end (26 in fig.182) the cuticle
is invaginated to form a very narrow channel (sbe in fig.182) which runs back through
the whole spine and whose end is expanded and bulb-like (sbe in fig.182, 183; possibly
the reservoir of a gland). The paa-process is completely sclerotised. At its right base the
cuticle is deeply invaginated to form a heavily sclerotised hood-shaped apodeme (boe in
fig.179, 182, 183) which caps the right end of the sbe-bulb (fig.182). The membrane (25
in fig.178, 179, 182) that adjoins this L4T+L2C-sclerite ventrally is somewhat invaginated
anteriad, and here the phallomere-gland opens (P in fig.178, 179). Ventral to and to the
left of this invagination the membrane extends posteriad towards the transverse edge 23,
along which the cuticle bends ventrad and anteriad to continue into the dorsal wall of the
Ive-pouch (fig.180).

The lve-pouch spans almost the whole breadth of the left complex. (The edge along the
bottom of the pouch is labelled 7 in fig.180). Its dorsal wall is largely occupied by the
sclerites L2A (left part) and L2B (right part), which articulate with each other (A4 in
fig.180). In the area around A4 the Ive-pouch has a very deep recess from anteriorly. The
right part of L2B bends dorsad and back to the left along the longitudinal part of edge
23 (compare fig.178 and 180). This dorsal part of L2B articulates with L1 anteriorly (A2
in fig.176, 178, 180); posteriorly it has an extension to the left (24 in fig.176, 178). L2A
extends like an arch along the margins of the left dorsal Ive-wall. Only in the anteriormost
part of the Ive-pouch L2A enters the ventral wall (fig.181). At the left posterior end of
the Ive-pouch the sclerite abruptly narrows, leaves the pouch (sclerotisation now
designated LAS, fig.178, 180, with L4d as its distalmost part), and curves into the dorsal
wall of the left complex (fig.177, 178).

The membranous ventral wall of the Ive-pouch is at the same time the anterior dorsal wall
of the vla-lobe (fig.174, 180, 181). The ejaculatory duct (D in fig.178-181) opens far on
the right into this membrane. Dorsal to the genital opening there is a small membranous
outfolding (goa in fig.177, 179-181). LAR in the ventral wall of the vla-lobe is a transverse
plate with a ribbon-like anterior extension (fig.174, 181). The small sclerite L7 lies in the
anterior right edge of the vla-lobe and is in close contact with a ribbon-like extension of
the R2-sclerite of the right phallomere (R2m in fig.174, 191-193). Another small sclerite
L8 lies in the right dorsal wall of the left complex (fig.176, 177).

Figs.174-176: Lamproblatta albipalpus (Blattaria, Blattidae, Lamproblattinae) — 174: Phallomere
complex in ventral view. — 175: Phallomere complex in ventral view; with some muscles; ventral
wall of genital pouch more complete than in fig.174. — 176: Left complex in right-dorsal view; dorsal
wall of pne-pouch largely removed. — Scale: 1mm.

113

Lamproblatta

Ipus

albipa

114

LUD LE
2000,00,

LAT
via pda

—

Lamproblatta
albipalpus

Figs.177-181: Lamproblatta albipalpus (Blattaria, Blattidae, Lamproblattinae) — 177: Left complex
in dorsal view. — 178-181: Left complex in dorsal view; with successive removal of parts of left
complex (mainly of dorsal ones). — Scale: 1mm.

115

The hla-hook (fig.174, 177) is evaginated from the left wall of the left complex and is
largely sclerotised by L3. Around the base of hla the cuticle is circularly invaginated
(fig.178, 179). Sclerite L4K shows a dorsoventral curvature: it lies mainly in the left dorsal
wall of the left complex, above the base of hla (fig.177), but its left part bends like a
horseshoe ventrad into the invagination around the hla-base (fig.178).

Right phallomere

Sclerite R3 occupies the anterior ventral wall (fig.190-194). In the posterior part of R3
the cuticle is considerably thickened (cross-section in fig.193). Sclerite R2 articulates with
the left posterior margin of R3 (A7 in fig.190-194). R2 forms a large ridge (fig.191, 192,
194), whose left-dorsal part curves dorsad and slightly back to the right (fig.190). The
right-ventral end of R2 has the extension R2m (fig. 191-195). Posterior to the central part
of R3 the ventral wall of the right phallomere curves dorsad to form a narrow, groove-
like central invagination (cbe in fig.190, 191; compare fig.193 and 195).

Posterior to the right part of R3 sclerite RIF adjoins, and the two sclerites articulate (A3
in fig.190-194). From its central part behind the A3-articulation RIF extends in two
directions: The ventral arm bends left-dorsad (at and behind edge 16 in fig.193, 195) and
largely occupies the right-dorsal wall of the cbe-invagination (fig.190). The distal part of
this arm forms a somewhat spoon-shaped process pva (fig. 190-195). At its distal anterior
margin this arm articulates with R2 (A6 in fig.190, 195); at its basal posterior margin it
has a distinct extension (20 in fig.190, 192-195). The dorsal arm of RIF extends
posterodorsad (fig.190, 191) and forms a sclerotised groove (rge in fig.190, 195). The part
of RIF posterior to A3, the extension 20, and the dorsal arm show an extensive thickening
of the cuticle directed to the interior of the phallomere (cross-sections in fig.195).

The posterior part of the right phallomere is a large dorsal lobe fda (fig.190, 194; in the
figures fda is pulled to the right and to the posterior), whose dorsal and ventral walls are
partly occupied by sclerite R1J. The dorsal anterior tip of R1J articulates with the dorsal
arm of RIF (A8 in fig.190), its ventral anterior margin articulates with the extension 20
of RIF (A9 in fig.192-195, 197). Near the A9-articulation the cuticle of R1J is, like that
of extension 20, thickened to the interior, and the articulation is thus very stout and deeply
immersed in the phallomere (fig.193, 195). The right wall of fda has a large membranous
area (17 in fig.190, 193, 194). The posterior edge of fda bears a sclerotised spine (sra in
fig.190).

Subgenital plate and posterior abdominal segments

Fig.169, 170 (posterior segments); fig.173 (subgenital plate S9). Tergite 10 T10 is not
completely divided longitudinally, but around its posterior edge there is a median mem-
branous field (21 in fig.169). The ventral part of tergite 10 T10v is moderately extensive
and is, except for an anterior transverse bridge, also divided by membrane 21. The
paraprocts Pp are divided (by the articulations A97) into a large median part sclerotising
the dorsal wall of the subanal lobe sbl and a small lateral plate-like part. Along the
anteriormost and medianmost dorsal wall of the subanal lobe sbl each paraproct forms a
heavy groove-like apodeme (fig.170; cut through on the right side). The lateral plate of

116

Pp is narrowly connected with the paratergite T10p laterally, and A99* is hence no longer
a true articulation. Strip-like and twisted Pv-sclerites are present; they are laterally
connected with the lateral plates of the paraprocts Pp. The Ca-sclerites are curved ribbons
on rather indistinct bulges immediately median to the cercal bases. The very small Cb-
sclerites lie at the bottom of a small funnel-like invagination. The Cc-sclerites are
dorsoventrally curved plates on a paired shallow outfolding beneath the anterior margin
of T10v and above the subanal lobes sbl. The articulations A98 are well-developed.

Musculature

Muscles Positions of insertions

12 L1 (left-anteriorly on pne-pouch) — membrane anterior to L4K and
to hla-base

13 L1 (ventrally on pne-pouch) — L2A and L2B (area of articulation A4)

15 LAR (anteriormost part) — L2A (anteriormost part)

16a LAR (anteriormost part) — L2B (left posterior part)

l6b Left bundle: L4R (left-posterior part) — dorsal wall of vla-lobe, far
left-posterior to genital opening
Right bundle: L4R (right-posterior part) — dorsal wall of vla-lobe,
posterior to genital opening

19 Anterior dorsal wall of genital pouch — L8

110 L2A (anteriormost part) — L4T and L2C between processes
paa and pda, membranous area 25 (compare fig.182)

111 L4K (posterodorsal part) — L4d (= dorsal part of L4S) and
membrane to the left of it

112 Membrane next to L2B (right ventral wall of lve-pouch) — L8 and
membrane in dorsal wall of pne-pouch

113a Ejaculatory duct D next to its opening — dorsal wall of goa-lobe

113b Ejaculatory duct D next to its opening — dorsal wall of vla-lobe
immediately to the right of genital opening

114 Membrane anterior to L4K and to hla-base — hla-hook (on L3);
muscle divided into an anterior and a posterior bundle inside hla.

120 L2A (leftmost part) — membrane left-posterior to opening of
phallomere-gland P

121 Membrane anterior to L4K — membrane left-posterior to opening
of phallomere-gland P; very delicate

122 Membrane anterior to L4K — hla-hook (basally and dorsally on L3)

123 L4K - hla-hook (basally and ventrally on L3)

124 Membrane posterior to L4K — membrane left-posterior to opening

of phallomere-gland P; very few diffuse fibers

in fig.
184

187

188

188

188, 189

188, 189
170, 184, 185

186
184, 188

184, 185, 186, 188
188

188
184, 185
188
184
184, 185

184, 186

185

Figs.182-185: Lamproblatta albipalpus (Blattaria, Blattidae, Lamproblattinae) — 182: Sclerites L2C
and LAT, processes paa and pda, and phallomere-gland P in dorsal view. — 183: Sclerites L2C and
LAT and process paa in dorsal view; some further parts removed (compare fig.182). — 184,185: Left
complex in dorsal view; each figure with some muscles; parts of left complex removed to various
extents. — Scale: 1mm.

Lamproblatta
albipalpus

~

117

118

KIM

IN
N

N

STRING

RR

— h-

187

ii Lamproblatta

albipalpus

N

NN N
N
NR
NIS
DV
>
»

SS

N
N
N

N
WOW

\
SICKO

ue

188 189

Figs.186-189: Lamproblatta albipalpus (Blattaria, Blattidae, Lamproblattinae) — Left complex in

dorsal view; each figure with some muscles; parts of left complex removed to various extents. —
Scale: Imm.

119

r2 R3 — cbe-invagination: RIF (anterior margin), membrane, and R2

(dorsal margin) 197, 198
r3 RIF (dorsal and central parts, posterior to articulation A3) — R1J

(right margin) 19751935199
r6 RIF (dorsal margin, on rge-groove) — dorsal wall of fda-lobe

(in part on R1J) 196
b2 L8 and membrane to the right of it (= right dorsal wall of

vla-lobe) — membrane ventral to R2 184, 198
sl S9 left side (most anteriorly on apophysis S9a) — membrane anterior

to hla-base ADS SES, IS), Weeds)
s3 S9 left side (anteriorly on apophysis S9a) — L4R (anteriormost part) 170, 173, 175, 188
s4 S9 right side (anteriorly on and median to apophysis S9a) — R3

(left anterior margin) IFO WS. WS, I!
s5a S9 left side (posteriorly and medially) — left ventral wall of genital

pouch IW, 178
s5b S9 left side (posteriorly and quite laterally) — left wall of genital

pouch a, 178, 178
s6 S9 right side (anteriorly on apophysis S9a and laterally) — R3

(right anterior margin) 7027321737199
s12 S9 right side (anteriorly on apophysis S9a) — L4R (anteriormost part) 169, 173, 175, 188
p3 (pair) S9 — rectum (ventral wall) 169, 173

p4 (pair) T9 (lateral anterior margin, also extending onto paratergite T9p) —
membrane far anterior to Pv-sclerite; muscles on both sides divided

into three bundles (except for their ventralmost parts) 169, 170
p5 (pair) T10 (lateralmost anterior margin) — Pv-sclerite 169
p6 (pair) T9 (lateralmost part) — S9 (lateral margin) 169, 172
p7 (pair) Membrane anterior to Pv-sclerite - membrane (far) anterior to

contact A99* between paratergite T10p and paraproct Pp 170

5.11. Anaplecta sp. (Blattaria, Blattellidae, Anaplectinae)

Left complex

Sclerite L4K (fig.205, 208-210) lies in the left wall. Its posterior part partly encloses the
(retracted) hook hla (fig.209, 210) and its sclerite L3. The highly complicated L2-sclerite
is in the center of the left complex (fig.210-215). Its anterior part forms a tube-like
apodeme (lve-apodeme = anterior part of lve-pouch), on which the nla-bulge rests. At the
left base of the Ive-apodeme L2 forms, together with parts of L4 (L4N), a stout sclerite
ring (fig.211, 212) bearing two processes: pda and paa (fig.209, 211, 214). From the right

eS 120

Figs.190-195: Lamproblatta albipalpus (Blattaria, Blattidae, Lamproblattinae) — 190: Right
phallomere in dorsal view. — 191: Right phallomere in left-dorsal (somewhat anterior) view. — 192:
Right phallomere in left-ventral view. — 193: Right phallomere in right-ventral view; membrane 17
largely removed. — 194: Right phallomere in ventral (somewhat posterior) view. — 195: Right
phallomere in right-ventral view; most elements shown in fig.193 removed. — Scale: 1mm.

120

Lamproblatta

S

albipalpu

121

Lamproblatta
albipalpus

Figs.196-199: Lamproblatta albipalpus (Blattaria, Blattidae, Lamproblattinae) — 196: Right
phallomere in dorsal view; with muscle r6; some dorsal elements removed. — 197: Right phallomere
in dorsal view; with some muscles; dorsal elements more extensively removed than in fig.196. — 198:
Right phallomere in left-dorsal (somewhat anterior) view; with some muscles. — 199: Right
phallomere in right-ventral view; with some muscles; ventral wall of genital pouch more complete
than in fig.193; membrane 17 cut open to show muscle r3 (cut through). — Scale: Imm.

122

base of the Ive-apodeme L2 extends to the right to form a curved plate-like sclerotisation
(fig.212-214). The large lobe vla, with sclerite L4G in its ventral wall (fig.205), is the
ventralmost part of the left complex.

L4K consists of a plate-like, dorsoventrally curved posterior and a bulge-like (nla in
fig.206, 209; veiled by membrane in fig.208) anterior part (fig.209, 210), which are only
narrowly connected (fig.205). The dorsal wall to the right of L4K is membranous; its
central part is depressed ventrad and anteriad (pne in fig.208, 209). The hla-hook is
evaginated from the left posterior wall of the left complex (fig.210, 212). The distal half
of hla is sclerotised by L3, which is rather weak except for its distalmost part. The basal
membranous half of the hla-hook (30 in fig.210-212) can be introverted and hla can thus
be deeply retracted into the left complex (this state 1s shown in the figures). To the right
of the hla-base the posterior wall of the left complex folds inwards (fpe in fig.210-213)
— separating the area of hla from the area dominated by L2 (fig.211).

The Ive-apodeme has its entire dorsal wall (fig.210, 211) and the margins of its ventral
wall (fig.211, 225) sclerotised by L2. The sclerotised cuticle is considerably thickened
(cross-sections in fig.211-215). The top of the Ive-apodeme and the nla-process are firmly
connected (two areas of the internal surface of the cuticle adhere to each other). The Ive-
apodeme is the narrow anterior part of a lve-pouch, which posteriorly expands to the right.
At the base of the Ive-apodeme, L2 forks into a left and a right branch (immediately
posterior to the cross-section through lve in fig.213).

The left branch bears a node-like apodeme (29 in fig.212, with a tuft of fine cuticular
threads) and forms the sclerite-ring mentioned above (fig.211-213). This ring is slightly
sunken anteriad into the left complex; the cuticle within the ring is evaginated to form
the processes paa and pda, which are both sclerotised in their ventral walls only (fig.209,
214; the sclerotisations of both are connected with the basal ring). The membrane 31 in
fig.211-214 is the area of contact between the bases of paa and pda (cut through in
fig.214). Apart from L2 (dorsal part of the ring, paa-sclerotisation), L4 also contributes
to this structure (L4N: ventral part of the ring, pda-sclerotisation). The left branch of L2
has another posteriad-directed extension on its ventral side (28 in fig.214, 215) which lies
in the dorsal wall of another process (gta in fig.215, 216). The sclerotised cuticle of the
left L2-branch is thickened in most of its parts (cross-sections in fig.212-215).

The right branch of L2 broadens and extends far to the right, where it curves dorsad
(fig.212-214). Posteriorly this upcurved part extends somewhat back to the left and is
involved in some complicated cuticular foldings (near 32 in fig.211, compare fig.209-213).
Posterodorsal to the right L2-branch there are some additional membranous in- and
outfoldings (fig.210-213).

— 21,128)

Figs.200,201: Anaplecta sp. (Blattaria, Blattellidae, Anaplectinae) — 200: Male postabdomen in dorsal
view; with phallomere complex, subgenital plate, marginal parts of abdominal tergites 9 and 10,
subanal lobes, paraprocts, Pv-sclerites, distal part of rectum, basal parts of cerci, and part of
musculature. — 201: Same as in fig.200, after removal of further parts of abdominal tergites 9 and

10 (especially T10v). Distal part of rectum and basal parts of cerci cut open. Another part of
musculature shown. — Scale: 1mm.

123

7?

200

201

®

naplecta s

124

Anaplecta sp.

re as i mee ware A

Figs.202-204: Anaplecta sp. (Blattaria, Blattellidae, Anaplectinae) — 202: Male postabdomen in dorsal
view; with phallomere complex, subgenital plate, and lateral parts of abdominal tergite 9. — 203: Left

margin of subgenital plate (compare fig.202); with insertion of muscle p6. — 204: Subgenital plate
in dorsal view; with insertion areas of muscles (except p6). — Scale: Imm.

125

205

Anaplecta sp. |

L4G

126

The ejaculatory duct D opens anterior to the right L2-branch (fig.210, 211). In its dorsal
wall there is a small outfolding (goa in fig.212). The ventral wall of the duct continues
posteriad into the dorsal wall of a broad membranous outfolding vfa (fig.211, 215). The
vla-lobe, with sclerite L4G in its ventral wall (fig.205; cut open in fig.206; fig.214-220)
is another broad outfolding ventral to vfa. The right part of the vla-lobe curves dorsad
(fig.208). The right dorsal wall of the vla-lobe has a broad and flat invagination vte
(fig.206, 208, 209, 219) functioning as a tendon (muscle 16b in fig.221).

Two phallomere-gland ducts (P in fig.215, 216) open anterior to the ventral wall of the
gta-process. Between the orifices there is a small infolding (ipe in fig.215-217). Anterior
to the orifices the ventral wall has a broad membranous pouch (vpe in fig.209-214, 216-
218). Anterior to vpe the nla-bulge with its L4K-sclerotisation adjoins (fig.218, 219).
Posterior to nla the anteriormost ventral wall of the genital pouch forms a membranous
pouch ate (fig.205, 208), which functions as a tendon (muscles s3 and 16a in fig.222).

Right phallomere

The triangular R3-sclerite occupies the anterior (right-)ventral wall (fig.226-229). The right
and anterior margins of R3 form a groove-like apodeme (age in fig.226, 229). Sclerite R2
articulates with the left posterior margin of R3 in two points (A7 in fig.227-229). R2
forms a ridge (fig.227, 228) with three bulges. Posterior to the central part of R3 the
ventral wall of the right phallomere curves dorsad and anteriad to form a central
invagination (cbe in fig.226-228; compare fig.229 and 230). Posterior to the right part of
R3 there adjoins the large sclerite RIN, and the two sclerites articulate (A3 in fig.226,
227, 229). RIN occupies most of the posterior part of the right phallomere, the broad
dorsal lobe fda (fig.226, 231). From behind the A3-articulation RIN has a long extension
to the left (34 in fig.226), which lies in the dorsal wall of the cbe-invagination. The left
end of extension 34 turns back to the right like a hook, and it articulates with the left-
dorsal end of R2 (A6 in fig.226, 227, 230). The fda-lobe bears a sclerotised bulge in its
ventral wall (33 in fig.227, 228, 231).

Subgenital plate and posterior abdominal segments

Fig.200, 201 (posterior segments); fig.204 (subgenital plate S9). Within the membrane
extending anteroventrad from the posterior edge of tergite 10 T10 there is a pair of ribbon-
like sclerites (T10v or Cc?). Strip-like Pv-sclerites are present: The right one is well
developed and connected with the paraproct Pp (or with paratergite 10 T10p?) laterally.
The left one is very small and isolated. The articulations A99 are missing (A99* in fig.201:
paratergites T10p and paraprocts Pp have fused). The Ca-sclerites lie on curved bulges
immediately median to the cercal bases. Cb- and Cc-sclerites are missing. The articulations
A98 are well-developed.

<—— p.125

Figs.205-207: Anaplecta sp. (Blattaria, Blattellidae, Anaplectinae) — 205: Phallomere complex in
ventral view. — 206: Phallomere complex in: ventral view; some ventral parts removed (compare
fig.205). — 207: Phallomere complex in ventral view; with some muscles; ventral wall of genital
pouch more complete than in fig.205. — Scale: 1mm.

27

208

Anaplecta sp.

Figs.208-210: Anaplecta sp. (Blattaria, Blattellidae, Anaplectinae) — 208: Left complex in dorsal view.
— 209,210: Left complex in dorsal view; with successive removal of its parts (mainly of dorsal ones).
— Scale: 1mm.

128

211

Anaplecta sp.

Figs.211-213: Anaplecta sp. (Blattaria, Blattellidae, Anaplectinae) — Left complex in dorsal view; with
successive removal of its parts (mainly of dorsal ones); fig.213: hla-hook separated from remainder
of left complex. — Scale: 0.5mm.

129

Figs.214-217: Anaplecta sp. (Blattaria, Blattellidae, Anaplectinae) — Left complex in dorsal view;
with successive removal of its parts (mainly of dorsal ones). — Scale: 0.5mm.

130

218 |

Anaplecta sp.

ay

220 |

ok

Figs.218-220: Anaplecta sp. (Blattaria, Blattellidae, Anaplectinae) — Left complex in dorsal view;
with successive removal of its parts (mainly of dorsal ones). — Scale: 1mm.

131

Figs.221-225: Anaplecta sp. (Blattaria, Blattellidae, Anaplectinae) — 221-224: Left complex in dorsal
view; each figure with some muscles; parts of left complex removed to various extents. — 225:
Anterior part of lve-apodeme in ventral view; with the muscles inserting on it. — Scale: 0.5mm.

182

VI aD

Z

|

|

|
Anaplecta sp. |
5
Figs.226-233: Anaplecta sp. (Blattaria, Blattellidae, Anaplectinae) — 226: Right phallomere in dorsal
view. — 227: Right phallomere in left-dorsal (somewhat anterior) view. — 228: Right phallomere in
left-ventral view. — 229: Right phallomere in right-ventral view; with transition to left complex. —
230: Right phallomere in right-ventral view; most elements shown in fig.229 removed. — 231: Right
phallomere in dorsal view; with some muscles; some dorsal elements removed. — 232: Right
phallomere in left-dorsal (somewhat anterior) view; with some muscles; some dorsal elements
removed. — 233: Right phallomere in right-ventral view; with some muscles; ventral wall of genital
pouch more complete than in fig.229. — Scale: 0.5mm.

0

133

Musculature
Muscles Positions of insertions in fig.
12 Left dorsal wall of left complex (left part of pne-”pouch”) — LAK

(posterior part) 221
13 Anterior central dorsal wall of left complex (anterior part of

pne-"pouch”) — L2 and membrane to the right of L2 (right dorsal

wall of Ive-pouch) AO, 22, 222
15 LAK (anterior part, dorsal base of nla-process) — L2 (on apodeme 29) 222, 223
16a ate-tendon — L2 (right edge of Ive-apodeme) RN, PAM, PPD
16b L4G - vte-tendon in right dorsal wall of vla-lobe 221, 224
19 Transversely within right dorsal wall of left complex 22
110 L2 (anterior left edge of Ive-apodeme) - sclerite ring with L4N and

L2 at common base of processes paa and pda DDP SIED2S
113 Ejaculatory duct D next to its opening (ventral wall) — bottom of

infolding between lobes vfa and vla AD
114 L4K (anterior part, ventral base of nla-process) — hla-hook

(anterior margin of L3) 2223
125 L2 (anterior ventral wall of Ive-apodeme) — ipe-infolding between

openings of phallomere-gland ducts P 224, 225
126 L2 (anterior right edge of Ilve-apodeme) — anterior edge of vpe-

infolding 224, 225
rl R3 (anterior right margin) — RIN (rightmost dorsal wall of

fda-lobe) 200, 231
r2 R3 — cbe-invagination: RIN-extension 34 (anterior margin),

membrane, and R2 (dorsal margin) 2317232
b2 L4G (right margin) — membrane ventral to R2 and R3 224, 232
s2 S9 right side (laterally and anteriorly) — R3 (central anterior margin) 201, 204, 207, 233
s3 S9 left side (medially and anteriorly) — ate-tendon 201, 204, 207, 222
s4 S9 right side (entire anterior margin) — R3 (left margin) 20182047 2077233
s5 S9 left side (laterally and anteriorly) — left wall and left ventral

wall of genital pouch 201, 204, 207
s6 S9 right side (lateral margin) — anteriormost right ventral wall of

genital pouch 201, 204, 207, 233
s7 S9 left side (medially and most anteriorly) — L2 (right edge

of Ive-apodeme) 200, 204, 207,

221225

s10 S9 right side (medially and most anteriorly) — bottom of infolding

between lobes vfa and vla (ventral to ejaculatory duct) 200, 204, 207,

222272

pl (pair) S9 — membrane anterior to (right muscle) or median to

(left muscle) Pv-sclerite; very delicate; left muscle in most cases

divided into two bundles 200, 204
p3 (pair) S9 - rectum (ventral wall) 200, 204

p4 (pair) T9 (lateralmost anterior margin, also extending onto
paratergite T9p) — membrane anterior to (right muscle) or
anteromedian to (left muscle) Pv-sclerite; muscles on both sides
divided into three bundles (except for their ventralmost parts) 200, 201

134

p5 (pair) T10 (lateralmost anterior margin) — on (right muscle) or median
to (left muscle) Pv-sclerite 200
p6 (pair) T9 (lateralmost part) — S9 (lateral margin) 200, 203

5.12. Nahublattella sp. (Blattaria, Blattellidae, Plectopterinae)

All figures are side-reversed and show mirror-images of the original structures. In the subsequent
descriptions and in the homology discussions (chapter 6.) the terms “left” and “right” will also be
exchanged. (This will be practised in other Plectopterinae, too: Euphyllodromia, Supella). The natural
orientation is shown in fig.236b and 239b.

Left complex

Sclerite L1’ lies in the posterodorsal wall (fig.243). The L2’-sclerotisations (L2D’, L2E’)
are in the center of the left complex (fig.242-244). The anterior part of L2D’ forms a
tube-like apodeme (lve-apodeme = anterior part of Ive-pouch), to the right of which the
ejaculatory duct (D in fig.242) opens. At the left base of this apodeme, L2E’ forms,
together with a L4’-sclerotisation (L4N’), a large trifid process via (fig.244, 245). In the
anteriormost ventral wall of the left complex lies sclerite L4V’, which bears the whip-like
process nla (fig.239a, 247). The left posterior edge of the left complex bears a long hook
hla with its sclerite L3’ (fig.242-244). The base of the hook is partly enclosed by the
lateral sclerite L4U’.

The left posterior part of the left complex resembles a bulge whose left wall is sclerotised
by L4U’ (fig.242-244). The hla-hook is evaginated from the posterior wall of this bulge;
its distal half is sclerotised by L3’. The membrane of the basal half (30 in fig.242-244)
can be introverted, and hla can be retracted in the same way as in Anaplecta (all figures
show hla in its retracted state). Right-ventral to the hla-base the posterior wall of the left
complex folds inwards (fpe in fig.242-245): This fpe-fold separates the area of hla from
the other parts of the left complex. The dorsal wall to the right of hla contains a bristle
area (35 in fig.242). Ventral to this area there is a flat pouch invaginated to the anterior
(pne in fig.242, 243). The ventral wall of the pne-pouch is sclerotised by the anterior part
of L1’. The cup-shaped posterior part of L1’ occupies a bulge-like process (dea in fig.243,
244) beset with setae. Beneath the dca-process there is a membranous lobe (cla in fig.244).
Anteroventral to the cla-lobe the cuticle is again evaginated: The posterior end of sclerite
L2D’ forms a bifid process (psa in fig.244-246).

Figs.234,235: Nahublattella sp. (Blattaria, Blattellidae, Plectopterinae) — All figures show mirror-
images of the original preparations. — 234: Male postabdomen in dorsal view; with phallomere
complex, subgenital plate, marginal parts of abdominal tergites 9 and 10, subanal lobes (covered),
paraprocts (covered), Pv-sclerites, distal part of rectum, basal parts of cerci, and part of musculature.
— 235: Same as in fig.234, after removal of further parts of abdominal tergites 9 and 10 (especially
T10v). Distal part of rectum and basal parts of cerci cut open. Another part of musculature shown.
— Scale: Imm.

135

S9s

Nahublattella sp.

235

S9s

S9s

136

The whole central part of the left complex is invaginated anteriad to form a large pouch
(Ive in fig.242-246; the edges along the bottom of this Ive-pouch are labelled 7 in fig.242).
This invagination contains the Ive-apodeme (middle part; with the L2D’-sclerotisation),
the via-process (left side), and the terminal part of the ejaculatory duct D (right side). The
Ive-apodeme is completely sclerotised by L2D’ — except for a membranous stripe (44 in
fig.239a, 243, 245) in its ventral wall. At the base of the Ive-apodeme the Ive-pouch
broadens. Here, L2D’ has a short extension to the left (36 in fig.239a, 243-246) and a
long ventral extension to the posterior (28 in fig.245, 246, which lies in the ventral wall
of the Ive-pouch). The main part of L2D’, however, extends far posteriad within the dorsal
wall of the Ive-pouch (fig.242); it bears a small apodeme (37 in fig.242, 245), and its
posteriormost part sclerotises the psa-process. Along its right margin this main part of
L2D’ folds ventrad and back to the left (towards edge 38 in fig.245; compare fig.244) to
form a shallow sclerite groove. Along edge 38 the cuticle turns to the right again and
continues into the dorsal wall of the ejaculatory duct (D in fig.244, 245). The ejaculatory
duct opens from the right side into the Ive-pouch (fig.242-246). The phallomere-gland (P
in fig.242) opens posterior to the dorsal wall of the ejaculatory duct.

The via-process evaginates posteriad from the left wall of the Ive-pouch (fig.244; in fig.241
via is isolated and shown in ventral view). Distally via branches into three spines paa,
pda, and vsa. via has a basal and a distal sclerite separated by a membranous ring (39 in
fig.241, 244, 245; sclerites not termed separately). The basal sclerite is roughly cylindrical
(fig.244), with a deep recess at its ventral anterior margin (fig.241, 245). Ventrally it bears
the vsa-spine. At its right anterior margin it articulates with extension 36 of sclerite L2D’

—_ — 87

Figs.236-238: Nahublattella sp. (Blattaria, Blattellidae, Plectopterinae) — All figures except 236b show
mirror-images of the original preparations. — 236a: Male postabdomen in dorsal view; with phallomere
complex, subgenital plate, and lateral parts of abdominal tergite 9. — 236b: Same as in fig.236a but
smaller scale and natural orientation. — 237: Subgenital plate in dorsal view; with insertion areas of
muscles (including p6). — 238: Dorsal sclerotisation S9d of subgenital plate; the part of the cuticle
shown has been cut off from the subgenital plate along the line between the points labelled x (compare
x in fig.237). — Scale: Imm.

— 0488

Figs.239-241: Nahublattella sp. (Blattaria, Blattellidae, Plectopterinae) — All figures except 239b show
mirror-images of the original preparations. — 239a: Phallomere complex in ventral view. — 239b:
Same as in fig.239a but smaller scale and natural orientation. — 240: Phallomere complex in ventral
view; with some muscles; ventral wall of genital pouch completely retained (compare fig.239a),
including dorsal sclerotisation S9d of subgenital plate and its muscles. — 241: via-process (including
paa, pda, and vsa, and sclerotisations L2E’ and L4N’) in ventral view. — Scale: Imm.

7.202139

Figs.242-244: Nahublattella sp. (Blattaria, Blattellidae, Plectopterinae) — All figures show mirror-
images of the original preparations. — 242: Left complex in dorsal view. — 243,244: Left complex in
dorsal view; with successive removal of its parts (mainly of dorsal ones); fig.244: hook hla and
adjacent areas as well as retained parts of process dca with sclerite L1’ separated from remainder of
left complex. — Scale: 1mm.

137

Nahublattella sp.

$9a

7

28

138

Nahublattella .

139

> 2% L
pda

Lie

cla

WU a psa

140

(A10 in fig.243-245). At its left anterior margin it has a ribbon-like extension (L4d’ in
fig.241, 242, 244, 245) running posteriad along the left edge of the Ive-pouch (7 in
fig.242). The distal sclerite branches into the sclerotisations of paa and pda. According
to their assumed origin, the right-dorsal parts of the via-sclerotisation (including A10 and
paa) are designated L2E’, the left-ventral parts (including L4d’ and pda) are designated
LAN’. (The boundary between L2E’ and L4N’ is perpendicular to the division into a basal
and a distal sclerite).

The ventral wall of the Ive-pouch (fig.246, including the ventral wall of the ejaculatory
duct D) is, except for the L2D’-extension 28 and the sclerotisations within the Ive-
apodeme, membranous. To the posterior it continues into the dorsal wall of a completely
membranous lobe (vla in fig.245-247, 239a). The ventral wall of vla is part of the ventral
wall of the left complex (fig.239a, 247). Sclerite L4V’ (fig.239a, 247) occupies the
anteriormost ventral wall; its right part extends onto and completely sclerotises the very
long process nla (fig.239a, 247, 248). nla has a broad base but soon narrows to become
whip-shaped.

Right phallomere

Sclerite R3’ occupies the anterior (right-)ventral wall (fig.253-257); its lateral and anterior
margins form a groove-like apodeme (age in fig.253, 256, 257). At its left and right ends
R3’ has extensions to the posterior (40 and 41 in fig.253-257). Posterior to the left part
of R3’ sclerite R2’ adjoins; the two sclerites have a broad articulation (A7 in fig.255,
257). R2’ is a plate of irregular shape, which as a whole slightly bulges posteriad (fig.254-
257). It bears a tooth (42 in fig.254, 255, 260) projecting dorsad and a horseshoe-shaped
bulge with small spines (43 in fig.254, 255, 260; seen from inside in fig.256, 258).
Posterior to the central part of R3’ the ventral wall of the right phallomere curves dorsad
and slightly anteriad to form a narrow, groove-like central invagination (cbe in fig.253-
256; compare fig.257 and 258), whose left-ventral wall is completely sclerotised by R2’.

Sclerite RIN’ broadly articulates with the right posterior part of R3’ (A3 in fig.253, 255,
257). From the A3-articulation RIN’ extends like an arch posteriad, leftward, and anteriad
again; it occupies the margins of a broad dorsal lobe fda (fig.253), which is the posterior
part of the right phallomere. The left anterior part of RIN?’ articulates with the left margin
of R2’ (A6 in fig.253-255, 258); at A6 RIN’ turns sharply back to the right (34 in fig.253),
and its bristled terminal part lies on a bulge in the posterior dorsal wall of the cbe-
invagination (34 in fig.253, 255, 259).

Figs.245-248: Nahublattella sp. (Blattaria, Blattellidae, Plectopterinae) — All figures show mirror-
images of the original preparations. — Left complex in dorsal view; with successive removal of its
parts (mainly of dorsal ones); fig.245: hook hla and its base (with sclerite L4U’) separated from
remainder of left complex and from each other (compare fig.244); fig.248: only sclerite L4V’ and
process nla retained. — Scale: Imm.

141

Nahublattella sp.

142

Subgenital plate and posterior abdominal segments

Fig.234, 235 (posterior segments); fig.237, 238 (subgenital plate S9). The dorsal
sclerotisation S9d of the subgenital plate comprises two isolated sclerites (fig.238), which
are asymmetrical and beset with stout setae. The entire tergite 10 T10, including its ventral
part T10v, is divided along its midline. T10v is rather extensive. Strip-like Pv-sclerites
are present; they are laterally connected with the paraprocts Pp. The Ca-sclerites are very
long and lie on curved bulges immediately median to the cercal bases. Cb- and Cce-sclerites
are missing. The articulations A98 and A99 are well-developed.

Musculature
Muscle Positions of insertions in fig.
11 L1’ (anteriorly on pne-pouch) — L4d’ (= part of L4N’) 249
12 Membranous basal part 30 of hla-hook — L4U’ (dorsal part) 249
13 L1’ (anteriorly on pne-pouch) — L2D’ (posteriormost part, on
apodeme 37) 250
14 L2D’ (posteriormost part, on apodeme 37) — L4U’ (ventral part) 249
15 L4V’ (left posterior margin) — L2D’ (extension 28 in ventral wall
of Ive-pouch) 291
16a L4V’ (right part) - L2D’ (most anteriorly on Ive-apodeme) 240, 250
l6b Ventral wall of vla-lobe — dorsal wall of vla-lobe Sk, ASD
19a L1’ (right wall of pne-pouch) — membrane to the right of pne-pouch 249
19b Transversely in dorsal wall of left complex (including bristle area 35) 249, 261
110 L2D’ (anterior left edge of Ive-apodeme) — left base of via-process
(with LAN’ and L2E’) 250
114 L2D’ (most anteriorly on Ive-apodeme) — hla-hook (anterior margin
of L3’) 240, 249, 250
115 L2D’ (anteriorly on Ive-apodeme) — ejaculatory duct D next to its
opening (dorsal wall) 249
127 Outlet channel of phallomere-gland P — membrane to the right of P 249
128 L1’ (anterior right ventral wall of pne-pouch) — L1’ (ventral anterior
margin) 250
129 Outlet channel of phallomere-gland P — ejaculatory duct D next to its
opening (dorsal wall) 249
130 Longitudinally in ventral wall of left complex 251

— > p.143

Figs.249-252: Nahublattella sp. (Blattaria, Blattellidae, Plectopterinae) — All figures show mirror-
images of the original preparations. — Left complex in dorsal view; each figure with some muscles;
parts of left complex removed to various extents. — Scale: Imm.

pa

Figs.253-258: Nahublattella sp. (Blattaria, Blattellidae, Plectopterinae) — All figures show mirror-
images of the original preparations. — 253: Right phallomere in right-dorsal view. — 254: Right
phallomere in right-dorsal view; dorsal elements largely removed. — 255: Right phallomere in dorsal
(somewhat anterior) view; membrane of cbe-invagination cut open to show parts of R2’ lying beneath
it (42, 43). — 256: Right phallomere in left view. — 257: Right phallomere in ventral view. — 258:
Right phallomere in ventral view; most elements shown in fig.257 removed. — Scale: 0,5mm.

143

1

jr 7)
8
=
<P)
en
Bun‘
8
=
=)
het
8

144

ahublattella sp.

145

Nahublattella sp.

Figs.259-261: Nahublattella sp. (Blattaria, Blattellidae, Plectopterinae) — All figures show mirror-
images of the original preparations. — 259: Right phallomere in right-dorsal view; with some muscles;
some dorsal elements removed; membrane of cbe-invagination cut open to show parts of R2’ lying
beneath it (42, 43). — 260: Right phallomere in dorsal (somewhat anterior) view; with muscle 12;
dorsal elements removed. — 261: Right phallomere in ventral view; with some muscles; ventral wall
of genital pouch much more complete than in fig.257 (compare fig.239a). — Scale: 0,5mm.

146

131 Left ventral wall of left complex — ventral base of via-process

(with LAN’ and L2E’) USN
132 Both insertions in left posterior ventral wall of left complex WS)
133 S9d (right sclerite) — S9d (left sclerite) 240
134 Longitudinally in posterior ventral wall of genital pouch (membrane

in between the two S9d-sclerites) 240
135 Longitudinally in anterior ventral wall of genital pouch 240
rl R3’ (right margin) — membrane anterior to RIN’ (rightmost dorsal

wall of fda-lobe) 2347259
r2 R3’ — cbe-invagination: membrane and R2’ (dorsal margin) I), ASD
r10 RIN’ (left ventral wall of fda-lobe) — left dorsal wall of fda- lobe 259
b5 Anterior dorsal wall of left complex — leftmost part of right

phallomere (to the left of and posterior to RIN’-part 34) 72, 28 Ol
s3 S9 left side (most anteriorly on apophysis S9a) — L4V’ (anterior

margin) 23923120:

2X0, 231

s4 S9 right side (most anteriorly on apophysis S9a) — R3? (anterior left

margin) RD, Ze, ZA), AOI
s5a S9 left side (far posteriorly; laterally and medially) — left ventral

wall of genital pouch 235, 237, 240
s5b S9 left side (far posteriorly and laterally) — left wall of genital

pouch, next to L4U’ JE); 22, 2D ZO
s6a S9 right side (far posteriorly) — right ventral wall of genital pouch,

near right margin of left complex 237, 240, 249
s6b S9 right side (far posteriorly and laterally) — R3’ (anterior right

margin) 235, 237, 240, 261
s6c S9 right side (very far posteriorly and medially) — right ventral wall

of genital pouch; composed of several delicate bundles 237, 240
s7 S9 left side (most anteriorly on apophysis S9a) — L2D’ (most

anteriorly on Ive-apodeme) 234, 237, 240, 249
s10 S9 right side (most anteriorly on apophysis S9a) — ejaculatory duct

D (right wall) 234, 237, 240, 249
p3 (pair) S9 — rectum (ventral wall) DB An a
p4 (pair) T9 (lateral anterior margin) — membrane far anterior to Pv-sclerite

(far medially) 234
p5 (pair) T10 (lateralmost anterior margin) — membrane anterior to Pv-sclerite 234
p6 (pair) T9 (lateralmost part) — S9 (lateral part) My Da
p9 (pair) Membrane anterior to Pv-sclerite — membrane beneath posterior

part of rectum 234

5.13. Parcoblatta lata (Blattaria, Blattellidae, Blattellinae)

Left complex

The left complex is not as complicated as in the previous species and contains only few
sclerotisations. A deep infolding from the posterior (fpe in fig.268-271) divides the left
complex into a left part with the retractable hla-hook and its L3-sclerite and a right part
with the long L2-sclerite.

147

The left part has the shape of a bulge. The hla-hook evaginates from the posterior wall
of the bulge; when retracted, hla lies in the center of the bulge (fig.268, 269; all figures
show hla in the retracted state). The distal half of hla is sclerotised by L3, which becomes

Parcoblatta
lata

89s

Fig.262: Parcoblatta lata (Blattaria, Blattellidae, Blattellinae) — Male postabdomen in dorsal view;
with phallomere complex, subgenital plate, marginal parts of abdominal tergites 9 and 10, supraanal
lobe, subanal lobes, paraprocts, distal part of rectum, basal parts of cerci, and part of musculature.

Supraanal lobe shown through a window cut into ventral sclerotisation of abdominal tergite 10 T10v.
— Scale: 1mm.

148

263

Parcoblatta
lata

$3a

SSS

SS

R

SSB

235 x 2 = =
2)
22)
r

a
RT

=

“ 3S9s

Fig.263: Parcoblatta lata (Blattaria, Blattellidae, Blattellinae) — Male postabdomen as in fig.262, after
removal of further parts of abdominal tergites 9 and 10 (especially T10v) and supraanal lobe. Distal
part of rectum and basal parts of cerci cut open. Another part of musculature shown. — Scale: Imm.

——> p.149

Figs.264,265: Parcoblatta lata (Blattaria, Blattellidae, Blattellinae) — 264: Male postabdomen in
dorsal view; with phallomere complex, subgenital plate, and lateral parts of abdominal tergite 9. —
265: Subgenital plate in dorsal view; with insertion areas of muscles (including p6). Insertion areas
of s5a and s6a shown through two windows cut into dorsal sclerotisation S9d of subgenital plate. —
Scale: Imm.

149

IN

SS

SS

N

Parcoblatta

lata

150

Parcoblatta 266
lata

a

/

fifi}
Uff)

YY)
WY),
hid fg

A,

fr

Figs.266,267: Parcoblatta lata (Blattaria, Blattellidae, Blattellinae) — 266: Phallomere complex in
ventral view. — 267: Phallomere complex in ventral view; with some muscles; ventral wall of genital
pouch more complete than in fig.266, with parts of dorsal sclerotisation S9d of subgenital plate in

its posterior part (compare fig.265). — Scale: Imm.

151

heavier distally. The membranous basal half (30 in fig.268-270) of hla becomes introverted
in the retraction of hla (fig.270, 271). The terminal leftward-bent part of hla has a groove
along its anterior surface (hge in fig.266), whose ventral wall has a distinct notch (45 in
fig.266).

The part to the right of the fpe-infolding has in its center a deep invagination to the anterior
(pouch Ive in fig.268, 269) and a spine to the posterior (via in fig.268); both contain parts

of sclerite L2. The anterior part of the Ive-pouch is a narrow tube-like apodeme (Ive-
apodeme) with a flattened and broadened top. It is completely sclerotised by L2 — except
for a membranous stripe (44 in fig.266) in its right-ventral (more anteriorly) or right (more
posteriorly) wall, which does not reach the top of the apodeme. Roughly in the middle of
L2, the ejaculatory duct joins the lve-pouch from the right (fig.268), and Ive becomes
much broader. At this point, the right edge of the Ive-apodeme (with the membranous
stripe 44) bends anteriad to continue into a dorsal fold of the ejaculatory duct D (fig.268,
269). The ventral main part of the ejaculatory duct extends to beneath the Ive-pouch
(fig.268, 269; cross-section in fig.270) and wraps partly around it from ventrally (fig.270-
272). In the area posterior to the confluence of the lve-apodeme and the dorsal part of the
ejaculatory duct, L2 is a groove-shaped sclerotisation in the left edge of the lve-pouch
(cross-sections in fig.270-272). This groove-shape of L2 extends posteriad as far as to the
posterior end of edge 7 (fig.270, 273), where L2 becomes completely restricted to the
dorsal wall of the Ive-pouch.

Posterior to this point L2 forms the sclerotisation of the via-spine (fig.272-275). via has
a longitudinal groove in its right-dorsal wall (vge in fig.272-275), whose anterior end
deepens to form a small, bulb-like, and heavily sclerotised apodeme (vge, tve in fig.273).
The phallomere-gland P opens to the right of the tve-apodeme. The ventral wall of the
left complex has some outfoldings in the area beneath the ejaculatory duct (47, 48, 49 in
fig.266, 271, 272). In between these outfoldings and posterior to them the ejaculatory duct
opens to the outside, and this area can be regarded as the genital opening. Dorsal to and
to the left of the via-process there is a dorsoventrally curved membranous lobe (vla in
fig.266, 268, 270).

The tendon ate (fig.266, 268, 271) has its origin in the anteriormost ventral wall of the
genital pouch; it is a long and thin invagination of the cuticle with sclerite L4V in its
anterior dorsal wall.

==> 1,1159

Figs.268-270: Parcoblatta lata (Blattaria, Blattellidae, Blattellinae) — 268: Left complex in dorsal
view. — 269,270: Left complex in dorsal view; with successive removal of its parts (mainly of dorsal
ones). — Scale: 1mm.

—) 9.158

Figs.271-275: Parcoblatta lata (Blattaria, Blattellidae, Blattellinae) — 271,272: Left complex in dorsal
view: with successive removal of its parts (mainly of dorsal ones); fig.271: hla-hook separated from
remainder of left complex (compare fig.270). — 273-275: Posteriormost part of sclerite L2 on process
via, phallomere-gland P, and surrounding membranes in dorsal view (scale larger than in fig.272);
with successive removal of parts of the cuticle (mainly of dorsal ones). — Scale: Imm.

152

Parcoblatta

lata

158

si Parcoblatta
EN lata

154

Right phallomere

The long, spatulate R3-sclerite occupies the anterior ventral wall (fig.280-284). Its right
part has a long extension to the posterior (fig.280, 284). The age-groove or -apodeme is
very broad at the anteriormost margins of R3; to the posterior it soon decreases and ends
on both sides (fig.266, 284). Sclerite R2 adjoins posterior to the left part of R3; the two
sclerites are broadly separated by membrane (at A7 in fig.282-284). From posterior to A7,
R2 extends anteriad and then curves to the left (fig.281, 283, 285). Most ventrally R2 has
a strong tooth (fig.283, 284); in its other parts it forms a very low ridge (fig.285). At its
left end R2 is fused to sclerite RIS (at A6* in fig.281, 283, 285). RIS likewise forms a
low ridge (pva in fig.281, 282), and next to its fusion with R2 it has a bulge-like cuticular
thickening to the interior (ewe in fig.282, 283, 285). Posterior to the central part of R3
the ventral wall of the right phallomere curves dorsad and anteriad to form a central
invagination (cbe in fig.280, 281, 283; compare fig.284 and 285) with R2 and RIS in its
left-ventral wall.

Sclerite RIP adjoins posterior to the right part of R3 (fig.281, 284), and the two sclerites
articulate (A3 in fig.281, 282, 284). RIP occupies the ventral wall and the margins of the
dorsal wall of a large lobe fda (fig.281, 284). The left anterior tip of RIP closely
approaches the free end of RIS (fig.281, 282). Above the fda-lobe there is another, smaller
and membranous lobe (dla in fig.280).

Subgenital plate and posterior abdominal segments

Fig.262, 263 (posterior segments); fig.265 (subgenital plate S9). The ventral part of tergite
10 T10v is rather extensive. There are no separate Pv-sclerites; the Pv-sclerotisations are
assumed to be incorporated into the anterior parts of the paraprocts Pp (fig.263; a deep
indentation at the median margin of each paraproct possibly marks the border between Pp
and Pv). The Ca-sclerites lie on curved bulges immediately median to the cercal bases.
Cb- and Ce-sclerites are missing. The articulations A98 and A99 are well-developed. The
asymmetrical subanal lobes sbl are highly elaborated (fig.263): The left sbl bears a small
spine on its posterior edge. The right sbl bears some posteriad-directed bulge-like
processes and an anteriad-directed small hook (50 in fig.263, veiled by membrane).

Musculature
Muscle Positions of insertions in fig.
12 Membranous basal part 30 of hla-hook — posterior left dorsal wall
of left complex 276
14 L2 (posteriormost part, on tve-apodeme) — anterior left dorsal wall
of left complex 276
16a Ventral wall of genital pouch — L2 (anteriormost right edge of
Ive-apodeme) 20 ANT

Figs.276-279: Parcoblatta lata (Blattaria, Blattellidae, Blattellinae) — Left complex in dorsal view;
each figure with some muscles; parts of left complex removed to various extents; fig.279: of muscle
l6b only ventral insertion area shown. — Scale: Imm.

155

SR

SS

SEN

Parcoblatta
lata

Parcoblatta
lata

Figs.280-285: Parcoblatta lata (Blattaria, Blattellidae, Blattellinae) — 280: Right phallomere in dorsal
view. — 281: Right phallomere in dorsal view; some dorsal elements removed. — 282: Right phallomere
in left-dorsal (somewhat anterior) view. — 283: Right phallomere in left-ventral view. — 284: Right
phallomere in right-ventral view. — 285: Right phallomere in right-ventral view; most elements shown
in fig.284 removed. — Scale: Imm.

157

8

Parcoblatta
| lata

Figs.286-288: Parcoblatta lata (Blattaria, Blattellidae, Blattellinae) — 286: Right phallomere in dorsal
view; with some muscles; some dorsal elements removed. — 287: Right phallomere in left-dorsal
(somewhat anterior) view; with muscle r2; some dorsal elements removed. — 288: Right phallomere
in right-ventral view; with some muscles; ventral wall of genital pouch more complete than in fig.284.

— Scale: Imm.

158

16b

114a
114b

130a
130b
136
137a
137b
138
139

140
rl

r2

s3a

s3b

s4

s5a

s5b

s6a

s6b
s7

Central ventral wall of left complex — posterior ventral wall of
Ive-pouch; diffuse

L2 (most anteriorly on Ive-apodeme) — hla-hook (left part of L3)
L2 (most anteriorly on Ive-apodeme) — right part of membranous
basal part 30 of hla-hook

Longitudinally in ventral wall of left complex (on both sides of
fpe-infolding); diffuse

Ventral wall of outfolding 48 — dorsal wall of outfolding 48; diffuse
Longitudinally in membranous basal part 30 of hla-hook
Transversely in anterior right ventral wall of left complex
Longitudinally in posterior right ventral wall of left complex
Some isolated fibers in right ventral wall of left complex
ate-tendon (anterior part with L4V) — ejaculatory duct D

(ventral wall)

Transversely in anterior ventral wall of left complex; diffuse

R3 (right-anteriormost part) - membrane in right anterior dorsal
wall of dla-lobe

R3 — cbe-invagination: RIS (right part), RIP (leftmost anterior
part), and membrane

S9 left side (far anteriorly on apophysis S9a) — left ventral basal
line Bl of left complex
S9 left side (most anteriorly on apophysis S9a) — ate-tendon

S9 right side (anteriorly and posteriorly on apophysis S9a) — R3
(anterior and left margin)

S9 left side (very far posteriorly) — left ventral wall of genital
pouch (very far posteriorly, in part on S9d)

S9 left side (medially) — anterior left ventral wall of genital pouch;
very delicate

S9 right side (main part inserting very far posteriorly, some smaller
bundles more anteriorly) — right ventral wall of genital pouch (very
far posteriorly, in part on anterior margin of S9d)

S9 right side (laterally) — R3 (entire right margin)

S9 left side (most anteriorly on apophysis S9a) — L2 (most
anteriorly on Ive-apodeme)

PRs. 29D
267, PAU, 21/7

267, 276
278
PENS, 21)
276
278
Du
278

ZI
278

262, 286

286, 287

263, 2695 26/27
NGS), ADS), AST,
Dig

263, 265, 267, 288
265, 267
262,269,207 5207)
ASS}, AOS), AST ZO

263, 265, 267, 288

262,265.20) 216

Figs.289-292: 289,290: Parcoblatta lata (Blattaria, Blattellidae, Blattellinae) — 289: Central part of
left complex in dorsal view; with sclerites L2 and L4V, tendon ate, process via, phallomere-gland P,
and ejaculatory duct D; stippled area: insertion area of muscle 14. — 290: Central part of left complex
in dorsal view; with posterior part of sclerite L2, process via, phallomere-gland P, and ejaculatory
duct D; many parts removed compared with fig.289. — 291,292: Blaptica sp. (Blattaria, Blaberidae)
— Mirror-images of the original preparations. — 291: Central part of left complex in dorsal view; with
sclerites L2’ and L4V’, sclerite-group L10’, tendon ate, process via, phallomere-gland P, and
ejaculatory duct D; stippled area: insertion area of muscle 14. — 292: Central part of left complex in
dorsal view; with posterior part of sclerite L2, process via, phallomere-gland P, and ejaculatory duct
D; many parts removed compared with fig.291. — Scale: 1mm.

159

290

Parcoblatta
2 lata

292

1 Blaptica sp.

MOOR

160

s10 S9 right side (most anteriorly on apophysis S9a) — ejaculatory

duct D (right wall) 262, 265, 267, 276
sl4a S9 right side (far anteriorly on apophysis S9a) — right ventral

basal line Bl of left complex AS, ASS), 2ST, 26
s14b S9 right side (far anteriorly on apophysis S9a) — right ventral

basal line Bl of left complex 262; 203920792116
p3 (pair) S9 — rectum (ventral wall) 2029285

p4 (pair) T9 (lateralmost anterior margin, also extending onto paratergite
T9p) — paratergite T10p (anterior margin); muscles on both sides

completely divided into two bundles 262, 263
p5 (pair) T10 (lateralmost anterior margin) — anterior margin of

Py-sclerotisation 262
p6 (pair) T9 (lateralmost part) — S9 (lateral part, also extending onto dorsal

sclerotisation S9d of subgenital plate) 262, 265
p9 (pair) Membrane anterior to Pv-sclerotisation — membrane median to

paraproct Pp, beneath rectum 262

5.14. Blaberus craniifer (Blattaria, Blaberidae)

All figures are side-reversed and show mirror-images of the original structures. In the subsequent
descriptions and in the homology discussions (chapter 6.) the terms “left” and “right” will also be
exchanged. (This will be practised in other Blaberidae, too: Blaptica, Byrsotria, Nauphoeta). The
natural orientation is shown in fig.295b and 297b.

Left complex

The left complex resembles that of Parcoblatta. Again, a deep infolding from the posterior
(fpe in fig.299-302) divides the left complex into a left part with the retractable hla-hook
and its L3’-sclerite and a right part with the long L2’-sclerite.

The left part has the shape of a bulge whose left and ventral walls are occupied by sclerite
L4U’. The hla-hook evaginates from the posterior wall of the bulge; when retracted, hla
lies in the center of the bulge (fig.295a, 299; all figures show hla in a more or less retracted
state). Only a small distal part of hla is sclerotised by L3’. Most of the extensive
membranous basal part (30 in fig.299-302) of hla becomes introverted when hla becomes
fully retracted. hla can be retracted more deeply than in the previous species. (Full
retraction is shown in fig.295a, with the sclerotised part of hla completely veiled by
membrane; in the other figures hla is only partly retracted — to an extent corresponding
to the maximal retraction in Parcoblatta). The leftward-bent terminal part of hla has a
groove along its anterior surface (hge in fig.297a), whose ventral wall has a distinct notch
(45 in fig.297a).

The part to the right of the fpe-infolding has in its center a deep invagination to the anterior
(pouch Ive in fig.299, 300) and a sclerotised process to the posterior (via in fig.299); both
contain parts of sclerite L2’. The anterior part of the lve-pouch is a short tube-like apodeme
(Ive-apodeme) with a flattened and broadened top. Most anteriorly the Ive-apodeme is
sclerotised all around, more posteriorly the right wall is membranous (44 in fig.297a, 299,
300). Roughly one third down from the top of L2’, the ejaculatory duct D joins the Ive-

161

pouch from the right (fig.299), and Ive becomes broader. In the area posterior to this
confluence, L2’ is a groove-shaped sclerotisation in the left edge of the Ive-pouch (cross-
section in fig.301). This groove-shape of L2’ extends posteriad as far as to the posterior
end of edge 7 (fig.301), where L2’ becomes completely restricted to the ventral wall of
the Ive-pouch. The phallomere-gland (P in fig.299-301) opens next to this point. The
posteriormost part of L2’ sclerotises the via-process (fig.299-302).

Blaberus
craniifer

Fig.293: Blaberus craniifer (Blattaria, Blaberidae) — Mirror-image of the original preparations. — Male
postabdomen in dorsal view; with phallomere complex, subgenital plate, marginal parts of abdominal
tergites 9 and 10, supraanal lobe, subanal lobes (covered), paraprocts, distal part of rectum, basal
parts of cerci, and part of musculature. Right part of supraanal lobe shown through a window cut
into ventral sclerotisation of abdominal tergite 10 T10v. — Scale: 2mm.

162

S8s

Blaberus
craniifer

Figs.294,295b: Blaberus craniifer (Blattaria, Blaberidae) — 294: Male postabdomen as in fig.293,
after removal of further parts of abdominal tergites 9 and 10 (especially T10 and T10v) and supraanal
lobe. Distal part of rectum and basal parts of cerci cut open. Another part of musculature shown. —
Mirror-image of the original preparations. — Scale: 2mm. — 295b: Same as in fig.295a (next plate)
but smaller scale and natural orientation (no mirror-image).

163

Blaberus
craniifer

EN

I
SS

N

N

Ye

164

Beneath via, the ventralmost part of the left complex forms a broad ventral lobe vla, whose
edges are sclerotised by the tuberculate L10’. The right anterior end of L10’ is connected
with L2’. The genital opening is more or less right-dorsal to the middle part of sclerite
L2’, though it is hardly possible to define its exact position.

The membranous tendon ate has its origin in the ventral basal line of the left complex
(BI in fig.297a, 302); it is a short and broad invagination of the cuticle. To the left of ate
there is another small membranous invagination (55 in fig.297a, 302).

Right phallomere

Sclerite R3’ occupies the anterior (right-)ventral wall (fig.308-312a). Its right part has a
short extension to the posterior (fig.308, 312a). The age-groove or -apodeme is very broad
at the anteriormost margins of R3’; to the posterior it soon decreases and ends on both
sides (fig.297a, 312a). Sclerite R2’ adjoins posterior to the left part of R3’; the two
sclerites are broadly separated by membrane (at A7 in fig.312a). The ventral anterior tip
of R2’ lies in a small membranous pouch (56 in fig.308, 312a,b, 313). From here R2’
extends left-dorsad and forms a ridge (fig.310, 311). At its left end R2’ is fused to the
large sclerite RIT’ (at A6* in fig.310, 313). Next to its fusion with R2’, RIT’ has a
bulge-like cuticular thickening to the interior (ewe in fig.308-310, 313). Posterior to the
central part of R3’ the ventral wall of the right phallomere curves dorsad and anteriad to
form a rather indistinct central invagination (cbe in fig.308, 310, 313; compare fig.312a
and 313).

The two large sclerites RIT’ and R4’ adjoin posterior to the right part of R3’ (fig.308,
309, 312a). RIT’ is loosely articulated with R3’ (A3 in fig.309, 310, 312a). RIT’ and
R4’ are the sclerotisations of two lobes lying one above the other, which compose the
posterior part of the right phallomere (fda and dla in fig.308, 309, 313). The left end of
the fda-lobe is somewhat pointed (58 in fig.309, 312a), and next to this point RIT’ is
fused to R2’. R4 mainly occupies the dorsal wall of the dla-lobe (fig.308); its right end
curves into the ventral wall of the phallomere (59 in fig.308, 309), where it closely
approaches articulation A3.

The bulged sclerite R5’ lies in the left-ventral part of the right phallomere (fig.312a;
removed from the other elements in fig.311; cut through in fig.309). The right phallomere
can be retracted and protracted, and during this movement R5’ flaps back and forth
(compare fig.312a and 312b).

Subgenital plate and posterior abdominal segments

Fig.293, 294 (posterior segments); fig.296 (subgenital plate S9). The entire tergite 10 T10,
including its ventral part T10v, is divided along its midline. T10v is very extensive; it
has a pair of extensions to the anterior, which bear node-like apodemes (54 in fig.293).

—— p.163

Figs.295a,296: Blaberus craniifer (Blattaria, Blaberidae) — All figures show mirror-images of the
original preparations. — 295a: Male postabdomen in dorsal view; with phallomere complex, subgenital
plate, and lateral parts of abdominal tergite 9. — 296: Subgenital plate in dorsal view; with insertion
areas of muscles (including p6). — Scale: 2mm.

165

Blaberus
craniifer

GG

Figs.297,298: Blaberus craniifer (Blattaria, Blaberidae) — All figures except 297b show mirror-images
of the original preparations. — 297a: Phallomere complex in ventral view. — 297b: Same as in fig.297a
but smaller scale and natural orientation. — 298: Phallomere complex in ventral view; with some
muscles; ventral wall of genital pouch more complete than in fig.297a. — Scale: 1mm.

166

Blaberus craniifer

Figs.299-302: Blaberus craniifer (Blattaria, Blaberidae) — All figures show mirror-images of the
original preparations. — 299: Left complex in dorsal view. — 300-302: Left complex in dorsal view;
with successive removal of its parts (mainly of dorsal ones); fig.302: hook hla separated from
remainder of left complex (compare fig.301). — Scale: Imm.

167

oe

Blaberus craniifer

168

There are no separate Pv-sclerites; the Pv-sclerotisations are assumed to be incorporated
into the anterior parts of the paraprocts Pp (fig.294; a deep indentation at the median
margin of the left paraproct possibly marks the border between Pp and Pv). On their
anterior margins the assumed Pv-sclerotisations bear the anteriad-directed node-like
apodemes 51 (both sides, smaller on the right) and the posteriad-directed apodeme 52 (left
side only). The sclerites Ca, Cb, and Ce are missing. The bulges next to the cercal bases
the Ca-sclerites lie upon in the other species, however, are present (compare fig.263). The
articulations A99 are well-developed; the articulations A98 are missing: The sclerotisations
E11 and T10 are far away from each other. The left subanal lobe sbl bears a finger-like
process (53 in fig.294: mostly veiled by membrane) in its anterior ventral wall.

Musculature
Muscles Positions of insertions in fig.
12 Membranous basal part 30 of hla-hook — L4U’ (dorsal part) 303
14 L2’ (left-posterior part) — L4U’ (anterior part in left edge of left

complex) and membrane anterior to L4U’ 303
l6a Anteriormost ventral wall of left complex and anteriormost ventral

wall of genital pouch — L2’ (anteriormost right edge of

lve-apodeme) 298, 304
16b Central ventral wall of left complex — L2’ in posterior ventral wall

of Ive-pouch; diffuse 305,306
114a L2’ (most anteriorly on Ive-apodeme) — hla-hook (left wall anterior

to L3’) 298, 303, 304
114b L2’ (anterior left wall of Ive-apodeme) — right part of membranous

basal part 30 of hla-hook (insertion area horseshoe-shaped) 298, 303
130 Longitudinally in ventral wall of left complex (only to the right of

fpe-infolding); diffuse 307
136 Longitudinally in membranous basal part 30 of hla-hook 303

<—— p.167

Figs.303-307: Blaberus craniifer (Blattaria, Blaberidae) — All figures show mirror-images of the
original preparations. — Left complex in dorsal view; each figure with some muscles; parts of left
complex removed to various extents; fig.304: hook hla separated from remainder of left complex; of
muscle Il4a only posterior insertion area on hla shown; left picture shows part of the membranous
base of hla (part of membrane 30) together with muscle 146; fig.306: of muscle 16b only ventral
insertion area shown. — Scale: Imm.

= nd

Figs.308-313: Blaberus craniifer (Blattaria, Blaberidae) — All figures show mirror-images of the
original preparations. — 308: Right phallomere in dorsal view. — 309: Right phallomere in dorsal
view; some dorsal elements removed (mainly lobe dla and sclerite R4’). — 310: Right phallomere in
left-dorsal (somewhat anterior) view; sclerite R5’ removed. — 311: Right phallomere in left-ventral
view; sclerite R5’ and surrounding membranes separated from remainder of right phallomere (along
the undulate line between the points labelled x). — 312a: Right phallomere in right-ventral view. —
312b: Left part of right phallomere in right-ventral view; sclerite RS’ flapped to the anterior. — 313:
Right phallomere in right-ventral view; most elements shown in fig.312a removed. — Scale: 1mm.

169

310
fer

Blaberus
cranii

170

137 In right ventral wall of left complex 307
138 In right ventral wall of left complex 307
141 Transversely in dorsal wall of left complex, between fpe-infolding

and opening of phallomere-gland P; diffuse 304
142 L2’ (left anterior ventral wall of Ive-apodeme) — fpe-infolding 304
143 Membranous basal part 30 of hla-hook — membrane to the right

of hla-base 303
144 ate-tendon — anterior right ventral wall of left complex; diffuse 307
145 Longitudinally in rightmost part of left complex 305
146 Longitudinally in membranous basal part 30 of hla-hook, distal

to 136; composed of diffuse fibers running within the insertion

area of Ilda. 304 (left)
rl R3’ (right-anteriormost part) — R4’ in right anterior dorsal wall

of dla-lobe 293,314
r2 R3’ — cbe-invagination: RIT’ (left part) and membrane 314, 316
rlla R4’ (left edge of dla-lobe) — RIT’ (right ventral wall of fda-lobe);

diffuse 314, 315
rllb R4’ (right dorsal wall of dla-lobe) — RIT’ (rightmost ventral wall

of fda-lobe); diffuse 314, 315
r12 R3’ (right-anteriormost part) — R4’ (leftmost part); very delicate 315
r13 R3’ (right-anteriormost part) — membrane posterior to

cwe-thickening; anterior part of r13 running within muscle r2
(compare fig.314), posterior part abruptly leaving r2 and running

to ewe; very delicate 315
r14 R3’ (anterior right margin) — membrane to the right of right
posterior end of R3’; very delicate 315
r15 Membrane posterior to ewe-thickening — RIT’ (right ventral wall
of fda-lobe); diffuse Bilis
r16 R4 (leftmost part) — membrane to the right of left part of R4’;
very delicate, diffuse 315
r17 Longitudinally in ventral wall of genital pouch beneath right
phallomere; several delicate and diffuse bundles Sy)
r18 Membrane to the left of R2’ — membrane to the left of ewe-thickening 294, 314

Fig.315 shows the muscles r12, r13, r14, r15, and r16. However, none of these muscles was
present in all of the investigated specimens, and in none of the specimens these muscles were
present all together.

b6 Membrane to the left of R2’ — ejaculatory duct D (posterior right
dorsal wall) 294, 316

Figs.314-319: 314-317: Blaberus craniifer (Blattaria, Blaberidae) — Figures 314-317 show mirror-
images of the original preparations. — 314,315: Right phallomere in dorsal view; each figure with
some muscles; dorsal elements removed to various extents. — 316: Right phallomere in left-dorsal
(somewhat anterior) view; with some muscles; some dorsal elements removed. — 317: Right
phallomere in right-ventral view; with some muscles; ventral wall of genital pouch more complete
than in fig.312a. — 318: Byrsotria fumigata (Blattaria, Blaberidae) — Mirror-image of the original
preparations. — Right phallomere in dorsal view. — 319: Nyctibora sp. (Blattaria, Blattellidae,
Nyctiborinae) — Right phallomere in dorsal view. — Scale: Imm.

Blaberus
craniifer

Vyctibora
sp.

171

172

b7

s3a

s3b
s4
s5a

s5b
s6a

s6b

s10
s14

pla (pair)

Membrane posterior to R5’ — ejaculatory duct D (posterior right
ventral wall)

S9 left side (far anteriorly on apophysis S9a) — ventral basal line of
left complex, on infolding 55

S9 left side (most anteriorly on apophysis S9a) — ate-tendon

S9 right side (most anteriorly on apophysis S9a) — R3’ (anterior
and left margin)

S9 left side (far posteriorly) — left ventral wall of genital pouch
(far posteriorly); an additional delicate bundle runs more medially
S9 left side (laterally) — left ventral basal line BI of left complex
S9 right side (far posteriorly and laterally) — right ventral wall of
genital pouch

S9 right side (laterally) — R3’ (entire right margin)

S9 right side (on apophysis S9a) — ejaculatory duct D (right wall)
S9 right side (posterior to apophysis S9a) — right ventral basal line
BI of left complex

S9 — anterior margin of Pv-sclerotisation, on apodeme 51; very
delicate

plb (pair) S9 — posterior (!) margin of paraproct Pp, in anteriormost dorsal

p2 (pair)
p3 (pair)
p4 (pair)

p5 (pair)

p6 (pair)
p9 (pair)

wall of subanal lobe sbl

S9 — T9 (lateral anterior margin); very delicate

S9 — rectum (ventral wall)

T9 (lateralmost anterior margin, also extending onto paratergite
T9p) — membrane far anterior to Pv-sclerotisation; muscles on both
sides divided into two bundles in their dorsolateral parts

T10 (lateralmost anterior margin) — anterior margin of
Pv-sclerotisation (left muscle on apodeme 52)

T9 (lateralmost part) — S9 (lateral part)

Membrane anterior to Pv-sclerotisation — membrane median to
anterior margin of Pv-sclerotisation, beneath rectum; very
asymmetrical

5.15. Further species

317

294, 296, 298,
304

294, 296, 298,
304

294, 296, 298, 317

296, 298
294, 296, 298, 304

296, 298, 303
294, 296, 298, 317
239729923805

22I2 IB

295298
2955236
293, 296
2952.96
293, 294
293

293, 296

294

In some further species only certain parts or elements of the phallomere complex have
been investigated. These will be described within the respective sections of the following

chapter 6.

For Blatta orientalis, Deropeltis sp., Periplaneta americana (Blattidae, Blattinae) and
Ergaula capucina (Polyphagidae, Polyphaginae): Sclerites and most muscles of left
complex (no figures for Blattinae; for Ergaula capucina the morphology of the leftmost
part of the left complex is shown in fig.326d, 327d).

For Tryonicus angustus (Blattidae, Tryonicinae) and Ergaula capensis (Polyphagidae,
Polyphaginae): Sclerite L1, pne-pouch (with opening of phallomere-gland), and dca-

Wis

processes (shown in fig.105-108). Only in E. capensis: morphology of right phallomere
(shown schematically in fig.330m).

For Euphyllodromia angustata and Supella longipalpa (Blattellidae, Plectopterinae),
Loboptera decipiens (Blattellidae, Blattellinae), Ectobius sylvestris (Blattellidae,
Ectobiinae), Nyctibora sp. (Blattellidae, Nyctiborinae), and Byrsotria fumigata, Blaptica
sp. and Nauphoeta cinerea (Blaberidae): Central part of left complex with sclerite L2, Ive-
pouch, and via-process (all species; shown in fig.328c,d,f,g,h,1 and 291, 292; no figures
for Byrsotria). Morphology of right phallomere (all species except Ectobius, Loboptera,
Nauphoeta; shown in fig.3300,r and 318, 319; no figures for Supella). Presence and special
condition of some further elements of left complex: hge-groove, notch 45 (elements of
hla-hook, compare fig.266), ate-tendon with its sclerite L4V (no figures).

6. HOMOLOGY RELATIONS AND CHARACTER STATES

In this chapter a homology hypothesis will be elaborated for the phallomere and
postabdominal elements of the investigated species. This hypothesis should be as detailed
as possible, and it should be provided with as many arguments as possible. The following
list gives a survey which elements are discussed in which section. For the first five sections
(left complex I-V) the discussed elements are listed. Which elements are contained in the

T8
T9
T10
X
320 T10v
Ep
Spl
| Re Af
Dw Pv Pp sbi
Bl
Ive L2
D
¢; via
Bl L4 S9d S9s
Vw Vw
SI M sg

S8
Fig.320: Male postabdomen and phallomere complex in median sagitto-longitudinal section. Left
view; anteriore-, posterior. Only the cuticle is shown: Thin lines are membranous, thick lines are
sclerotised cuticle. Styles and paraprocts are shown though they are not visible in a median section.
Abbreviations in 4.7.

174

remaining sections is evident from the headlines. Some elements will be discussed in more

than one section — according to the various aspects of their relative position.

Gall

02.

6.3.

6.4.

6.5.

6.6.
6.7.

Left complex I: Main sclerites L1 and L6 and associated elements (L1, L6 / pne,
dca, loa, afa / 11, 12, 13, 19, 128, b4 / phallomere-gland P)

Left complex II: Main sclerite L2 and associated elements (L2 / Ive, vla, pda, paa,
via, gta, psa / 14, 15, 16, 18, 110, 112, s7 / ejaculatory duct D)

Left complex III: Main sclerites L4 and L10 and associated elements (L4, L10 /
swe, pda, paa, vsa, via, nla, vla / 11, 12, 14, 15, 16, 17, 110, 111, 114, sl, s3, s12)
Left complex IV: Main sclerite L3 and associated elements (L3 / hla, hge, fpe /
114, 119, 122, 123, 136, 146)

Left complex V: Further main sclerites and muscles (L5, L6, L7, L8, L9, L10,
L11 / Iba / 17, 19, 112, 113, b2)

Left complex VI: The position of the phallomere-gland opening

The elements of the right phallomere

Fig.321: Male postabdomen and phallomere complex corresponding to the hypothetical ground-plan
of Blattaria and Mantodea (on pages 175-177).

a)
b)

m)-o)

Postabdomen from segment 9 on in dorsal view.

Postabdomen from segment 9 on after removal of central parts of abdominal tergites 9 and
10; dorsal view. With marginal parts of tergites 9 and 10, phallomere complex, ejaculatory
duct, subgenital plate, paraprocts, epiproct, subanal lobes, supraanal lobe, cerci, and distal part
of rectum.

Detail from left part of fig.321b after removal of some dorsal parts in dorsal view.
Subgenital plate and phallomere complex in dorsal view.

Left complex in dorsal view.

Right phallomere in dorsal view.

Left complex after removal of some dorsal parts in dorsal view.

Right phallomere after removal of some dorsal parts in dorsal view. Ridges pia and pva shown
through a window cut into ventral wall of lobe fda.

Left complex in ventral view.

Subgenital plate in dorsal view. Ventral wall of genital pouch with dorsal sclerotisation S9d
of subgenital plate retained on left side but removed on right side.

Postabdomen from segment 9 on after removal of rectum, supraanal lobe, epiproct, and of
large parts of abdominal tergites 9 and 10, subanal lobes, and dorsal wall of genital pouch;
dorsal view. With lateral parts of tergites 9 and 10, phallomere complex, ejaculatory duct,
subgenital plate, paraprocts, ventral walls of subanal lobes, and basal parts of cerci. Dorsal
lobe fda cut open lengthwise in its leftmost part. All ground-plan muscles of categories p
(peripheral), s (phallomero-sternal), and b (between left complex and right phallomere) shown.
p4 only shown on right side, p6 and p7 only on left side. Dorsal part of p3 (to rectum) removed.
Left complex after removal of some dorsal parts in dorsal view. Each figure with some ground-
plan muscles of category | (intrinsic muscles of left complex).

Right phallomere after removal of some dorsal parts in dorsal view. With the ground-plan
muscles of category r (intrinsic muscles of right phallomere). Ridge pia shown through a
window cut into ventral wall of lobe fda.

Stippled areas are sclerotised. Abbreviations in 4.7. Elements whose presence in the ground-plan of
Blattaria and Mantodea is uncertain are provided with “?”.

175

I

178

1) Polyphaga

aegyptiaca m) Ergaula

capensis

k) Lamproblatta
albipalpus

i) Cryptocercus

punctulatus

Me g) Eurycotis
d) Metallyticus floridana

e) Sphodromantis sp. “ a
Se! violaceus

Mantodea \/ Blattaria

D) Archiblatta hoeveni

c) Chaeteessa b) Mantoida
eaudata schraderi

a) hypothetical ground-plan

179

u) Anaplecta sp.

q) Blaberus
oe Ss
4 06) Nahublattella sp p) Parcoblatta cramiiten
et ” lata _”®f
DA ab m

un

h) Tryonicus
parvus

Fig.322: Left complex, evolution of main sclerites. — Only the sclerites of the left complex are shown
— largely in their natural arrangement. Dorsal views. Species with “S” behind their names have side-
reversed phallomeres, and a mirror-image of the original preparation is shown. The branching black
lines represent the assumed phylogeny. The ground-plan is in some respects unclear (? in fig.322a,
discussions in 6.3.1., 6.4., 6.5.): Presence or absence of sclerites L3 and L5; connection or separation
of the L4-sclerites in the anterior ventral wall.

180

6.8. The muscles connecting the left complex and the right phallomere
6.9. The phallomero-sternal muscles

6.10. The subgenital plate and associated structures

6.11. The peripheral muscles

6.12. The terminal part of the abdomen

6.13. The asymmetry of the phallomere complex

In each of the sections 6.1.-6.4. and 6.7., as a first point, the homologies between Blattaria
and Mantodea will be analysed. This will be done by a discussion of selected species of
both groups which show similarities in the respective elements. Since Mantodea are not
a subgroup of Blattaria, and vice versa, this comparison is an outgroup comparison for
Blattaria as well as for Mantodea. Thus, it can serve (1) to reconstruct features of the
common ground-plan of Blattaria and Mantodea and (2) to determine polarities of
characters within Blattaria and within Mantodea. The ground-plan features will be shortly
summarised within each section as a second point, and the complete ground-plan will be
presented in chapter 7. As the third and fourth points of each section, the homologies, the
special conditions, and, in part, the evolution of the respective elements will be discussed
for Mantodea and then for Blattaria. The different states of the various characters, their
distribution over the taxa, and their polarities will in most cases become clear from these
discussions. An evaluation of the results in terms of evolution and phylogeny will be done
in 7.2.-7.8.. Fig.322-333 show the homology relations of the elements discussed.

6.1. Left complex I: Main sclerites L1 and L6 and associated elements

6.1.1. Comparison between Blattaria and Mantodea

All Mantodea and most Blattaria have a pouch (pne) in the dorsal part of the left complex,

the walls of which are largely sclerotised (sclerite L1). The membranous part of the pne-

wall is on the right side in Mantodea (e.g. fig.44, 45) but left-dorsal or dorsal in Blattaria

(e.g. fig.151). Arguments for the homology of L1 and pne in Mantodea and Blattaria and

indications for the ground-plan morphology of these elements can best be found by

comparing Mantoida (fig.44, 45) with Polyphaga (fig.120, 121), Ergaula (fig.105, 106),

and Cryptocercus (fig.153, 154); other Blattaria can also contribute.

In Mantoida, Polyphaga, Ergaula (both species), and Cryptocercus L1 and pne show

several similarities:

1. The pne-pouch lies in the central dorsal wall of the left complex and is an invagination
to the anterior.

2. The anterior part of L1 (region Lla in fig.323d,i,l,m) occupies most of the pne-wall
and is hood-shaped.

3. The right posterior part of L1 is a distinct arm-like extension (region Lim in
fig.323d,i,1,m).

4. A stout muscle runs from the posterior or central part of L1 to L2 in the dorsal wall
of the Ive-pouch: 13 (fig.50, 128, 158, 159; Ergaula: only E. capucina studied, no
figure).

181

5. Another stout muscle runs from the anterior or central part of L1 to L4-sclerotisations
in or near the left edge of the left complex: 12 (fig.49, 128, 156; Ergaula: only E.
capucina studied, no figure).

Mantoida, Polyphaga, and Cryptocercus have in common that:
6. The phallomere-gland P opens into the membranous part of the pne-wall. (In Ergaula
the opening is beneath the pne-pouch and the dca-processes.)

Mantoida, Polyphaga, and Ergaula have in common that:

7. The extension Lim (fig.323d,l,m) articulates with L2 (A2 in fig.45, 46, 105, 118).
This articulation is rather narrow. (There is no contact between L1 and L2 in
Cryptocercus: A2 is missing.)

Another feature is present only in Mantoida and Cryptocercus:

8. A stout muscle runs from the pne-pouch to L4-sclerotisations in the dorsal wall of
the left complex, the latter insertion being right-dorsal to the 12-insertion: 11 (fig.48,
SD):

Regarding the I2-insertion on pne, Mantoida is more similar to other Blattarian species

(compare feature 5.):

9. In Mantoida (fig.49), Eurycotis (fig.70), and Anaplecta (fig.221) 12 has its right
insertion in the left wall of the pne-pouch. (In Cryptocercus, Ergaula, and Polyphaga
this insertion is on the anterior face of the pne-pouch.)

In Cryptocercus, Polyphaga, and Ergaula L1 has, apart from Lim on the’ right side,

another distinct extension at its left posterior margin (region Lil in fig.323i,l,m). In

Cryptocercus and Ergaula L1l joins L1m ventrally to form a complete sclerite-ring (region

Lir in fig.323i,m). Mantoida has no Lll-extension, but other Mantodea have such an

extension:

10. In Metallyticus (fig.323b) and Chaeteessa (fig.323c) L1 has an extension at its dorsal
margin, which could well be homologous with the L1l of the respective Blattaria. (In
Mantoida, fig.323d, the corresponding area of L1 is designated as a vestigial L11.)

The formation of a sclerite-ring, however, does not seem to be a ground-plan element:
ll. In Mantoida, Chaeteessa, Metallyticus, as well as in Archiblatta (fig.54, 55, 323f) and
Eurycotis (fig.67, 68, 323e) the posterior part of L1 does not form a sclerite-ring.

Many Blattaria and Mantodea have distinct cuticular evaginations behind L1, which are

either membranous or sclerotised by posterior parts of L1: dca (e.g. in fig.153) and loa

(e.g. in fig.45, 54). The exact homology relations can hardly be determined for these

formative elements. Only in some cases homology is evident, e.g. for the paired

membranous cushions of Polyphaga, Cryptocercus, Tryonicus angustus, and probably

Archiblatta (dea in fig.120, 153, 107, 54). It is unclear whether these dca-processes are

elements of the common ground-plan of Blattaria and Mantodea and what their

morphology was like in this ground-plan. As regards the process loa, Mantoida resembles

Archiblatta:

12. At the posterior margin of L1 there is a completely sclerotised, curved and thorned
process (loa in fig.45, 54). Its sclerotisation is connected with L1 in Archiblatta but
articulated with L1 in Mantoida. However, the homology of these processes is not
certain.

182

a) Sphodromantis
sp.

c) Chaeteessa 4) Mantoida
caudata schraderi

b) Metallyticus
violaceus

Lia

323

Lil

Lim

Lir

Fig.323: Left complex, homologous regions of main sclerite L1. - Only L1-sclerotisations are shown.
Dorsal views, only left drawing of fig.323g and right drawing of fig.323n in ventral view. L1 is
divided into the four regions Lla, Lil, Lim, and Lir (definition in 6.1.1.). If L1 is divided into
several sclerites, these are labelled with the capital letters used in the text and in fig.1-319 (e.g. B =
L1B). The part of the sclerite margin which forms articulation A2 with sclerite L2 is indicated by
dashes; if part of this margin is covered by other parts of the sclerite, the dashes are shorter.

183

SZ
Xs

OS

SS

SAS

>

%
DE

dorsal view

®

vIeW

ventral

floridana

e) Eurycotis

Icus

ryon

2) T

Archiblatta
hoeveni

f)

s

parvu

DER

Polyphaga
aegyptiaca

I)

k) Lamproblatta
albipalpu

yptocercus
punctulatus

) Cr

angustus

h) Tryonicus

ventral view

“

VIEW

dorsal

2

&

=

=

whoo

A.

when See

Co wi

74

En

=
=
=
8
oo
Sons
ea)
Ale
E

°

CAPCRSIS

184

The membranous part of the pne-wall takes different positions within pne:

13. It is right-dorsal in Mantoida (fig.44, 45), dorsal in Archiblatta (fig.53, 54), left-dorsal
in Polyphaga, Ergaula, and Cryptocercus (fig.117, 151). The ground-plan position can
be assumed to be somewhere within this span.

This outgroup comparison suggests the similarities 1.-13. to be features of the common
ground-plan of Blattaria and Mantodea (rather uncertain as regards 10.-12.).

Main sclerite L1 is divided into four regions (fig.323):

— Lila (anterior): The large anterior part of L1 within the pne-pouch (hood-shaped in most
species).

— Lim (median): The extension at the median (right) posterior margin of Lla. Lim has
an articulation A2 with L2.

— Lil (lateral): The extension at the lateral (left) posterior margin of Lla.

— Lir (ring): The sclerotisation connecting L1l and L1m ventrally and giving the posterior
part of L1 the shape of a ring. (The ring can be complete or with a short gap.)

6.1.2. The elements in the common ground-plan of Blattaria and Mantodea

The features 1.-13. in 6.1.1. permit a reconstruction of the ground-plan morphology of
Li, pne, and some adjacent elements (fig.321le,g): Region Lla is hood-shaped, is situated
within a deep pouch pne, and has two posterior extensions: regions L1l (not certain) and
Lim (long and distinct). Lil and Lim do not join each other ventrally to form a sclerite-
ring (no region Lir). Lim articulates with L2 (A2). The membranous part of the pne-
wall is more or less dorsal and contains the opening of the phallomere-gland P. At the
posterior margin of L1 there is a thorned loa-processes (not certain). Muscles 11, 12, and
13 are present. 12 inserts on the left wall of the pne-pouch. I1 is dorsal to 12.

6.1.3. Homology relations and character states of the elements in Mantodea

In Chaeteessa (fig.32, 34), Metallyticus (fig.24, 25), and Sphodromantis (fig.10) L1 and
the pne-pouch are similar to Mantoida (fig.44, 45): The anterior part of L1 lies in the
deep pne-pouch and is more or less hood-shaped. The phallomere-gland P (not found in
Metallyticus) opens into the membranous part of pne. A large extension L1m (fig.323a-
d) from the right posterior part of L1 articulates (A2) with the right part of L2.

In contrast to Mantoida, the membranous part of the pne-wall is not right-dorsal but on
the right or right-ventral (Sphodromantis); this rotation (clockwise as viewed from behind)
is regarded as derived.

The articulation area A2 is elaborated differently: In Mantoida A2 is exactly on the edge
1 between the pouches pne and Ive (fig.45, 46). In the other species the L1m-extension
bends around edge 1 into the dorsal wall of the Ive-pouch, and here Lim and L2 lie in
the same plane (fig.10, 11, 25, 26, 34, 35; fig.323a-c). This is assumed to be a derived
condition. The sclerotisation bending around the edge is narrow in Chaeteessa but much
broader in Metallyticus and Sphodromantis. In Metallyticus, the right, bending part of L1m
is partly cut off from the basal part of Lim by the stripe of membrane 2 (fig.25, 323b).
In Sphodromantis this separation is complete (2 in fig.10, 323a), and L1 has divided into

185

two sclerites LIA and LIB. This comparison with Metallyticus reveals that in
Sphodromantis L1B is a right part of the L1m-region, that the articulation labelled A2 in
fig.11 is the true A2, and that the membranous stripe 2 1s a derived feature.

The homology relations of the processes behind the pne-pouch and on the edge 1 between
the pouches pne and Ive (loa, paa, afa in fig.10, 25, 34, 45) are — with the exception of
paa, which is discussed in 6.2.3. — somewhat difficult: The sclerotisation of loa originates
in Sphodromantis (fig.10) and Metallyticus (fig.25) from that part of L1 to the right of
the membranous stripe 2 (fig.323a,b), and the base of loa is posterior to the bending part
of Lim. Thus, homology is assumed for these loa-processes. The base of loa is far to the
right in Sphodromantis, but more to the left, in the ventral wall of the pne-pouch, in
Metallyticus. loa of Mantoida protrudes from the left-dorsal wall of the pne-pouch, but
homology with the loa of the other species seems possible if a shift of loa is assumed —
with the situation in Metallyticus being intermediate. Chaeteessa has no loa-process. In
Sphodromantis the sclerotisation of loa is reduced to a stripe in the dorsal wall (compare
feature 2=1n ©: 1-19).

The part of Lim bending ventrad around edge 1 sclerotises anteriorly the low bulge afa
in Metallyticus (fig.25, 26) and the hammer-shaped afa in Sphodromantis (fig.10, 11);
these afa might be homologous. The membranous lobes of Mantoida and Chaeteessa (afa
in fig.34, 45) might be homologous with the afa of Metallyticus and Sphodromantis (not
with loa), since their bases are anterior (not posterior) to the bending part of L1m. If this
homology is true, in Metallyticus and Sphodromantis the L1m-region has, while becoming
broader, additionally expanded onto the afa-processes.

6.1.4. Homology relations and character states of the elements in Blattaria
Ergaula, Polyphaga, and Cryptocercus

In Ergaula (fig.105, 106, 323m), Polyphaga (fig.120, 121, 3231), and Cryptocercus
(fig.153, 154, 3231) L1 and pne are quite close to the ground-plan but also have some
probably derived features: The anterior end of L1 is plateau-like, and the insertion of 12
has shifted to this plateau (fig.128, 156; compare 9. in 6.1.1.). The extensions Lil and
Lim curve ventrad and approach each other. However, only in Ergaula and Cryptocercus
the extensions unite to form a complete ring; in Polyphaga the ring is open (arrow in
fig.3231). The dea-processes — with their bases encircled by the L1-ring — are very similar
in Cryptocercus and Polyphaga. In Ergaula the morphology of dea is quite different. Only
Cryptocercus has a sclerotised peak (18 in fig.153) in between the dea, and the close
contact between L1m and L2 (A2-articulation) has been lost (fig.151).

Tryonicus angustus and T. parvus

In Tryonicus angustus (fig.107, 108, 323h) L1, pne, and dca are similar to the previous
species: The pne-pouch is very distinct and deep. The opening of the phallomere-gland
has the same position as in Polyphaga and Cryptocercus (compare fig.107 and 120, 153).
L1 articulates with L2 (A2 in fig.107, 108). The shape of the dca resembles Polyphaga
and Cryptocercus (fig.107, 120, 153). The extensions L1l and Lim are distinct (fig.323h)
and form a (open) sclerite ring encircling the dca-processes. The sclerotised peak 18

186

(fig.107, 108) resembles that of Cryptocercus (18 in fig.153), but its sclerotisation is
connected with the L1-ring dorsally and ventrally.

Some features are certainly derived (compared with the previous species and with the
ground-plan): Lia (fig.323h) and pne are flat (not hood-shaped). The L1-ring is not
complete since L1m has a gap between its base on Lla and A2 (arrow in fig.323h; this
situation differs from Polyphaga where the ring has a gap ventrally between L1l and A2:
arrow in fig.3231). A2 has become a broad hinge-like articulation, and the part of Lim
next to A2 is strongly enlarged (compare fig.323h and i,l,m). L1, pne, and dea are —
compared with Ergaula, Polyphaga, and Cryptocercus — rotated 40° (counterclockwise as
seen from behind): The membranous part of the pne-wall (removed in fig.107) is on the
left.

In Tryonicus parvus (fig.94, 95, 323g) L1, pne, and dea are even further rotated, and the
membranous part of the pne-wall is ventral. Compared with Polyphaga or Cryptocercus,
L1 and pne are rotated 120°; compared with e.g. Sphodromantis, where L1 and pne are
rotated in the opposite direction, the angle of rotation is 300°. Therefore, in comparing T.
parvus with the other species, L1 should be viewed from ventrally (fig.323g, left picture).
The anterior part of L1 (Lila in fig.323g) is a flat ribbon in the dorsal wall (rotation!
former ventral wall) of the distinct but narrow anterior part of the pouch pne (fig.95). The
position of the phallomere-gland opening is, having the L1-rotation in mind, exactly the
same as in Cryptocercus or T. angustus. The sclerotisation of the two bulges dca posterior
to Lla can be interpreted (fig.323g) as a complete sclerite-ring composed of the regions
Lim, Lil, and L1r (like in Ergaula and Cryptocercus) and an additional expansion of L1
onto dea. The Lil-arm runs mesad because of the L1-rotation. Lim first extends far
laterad, then it turns to the left, where it forms, like in T. angustus, a large plate and a
broad hinge-like articulation A2 with L2.

Archiblatta, other Blattinae, and Eurycotis

In Archiblatta L1, pne, and dea (fig.53, 54, 323f) can be easily identified: They take a
position in the central dorsal wall of the left complex. The anterior part of L1 (Lla-region
in fig.323f) lies in a pouch pne. At its right margin L1 articulates with L2 (A2 in fig.54).
The dea are membranous cushions at the left-posterior margin of L1 (fig.54); however,
the dea are not very similar to those of e.g. Cryptocercus (fig.153). loa resembles loa of
Mantodea (feature 12. in 6.1.1.). Some features can be regarded as derived: The pne-
pouch is less deep and distinct than in all species discussed before (fig.53, 54). The Lla-
region has become level as in Tryonicus. The phallomere-gland (P in fig.56) opens in the
same position as in Ergaula — beneath the dcea-processes (fig.54-56, 105, 106). (This
situation has certainly been achieved independently in Ergaula and Archiblatta). There are
no distinct arms Lil and Lim (and also no ring-formation or region Li1r). The vestiges
of Lil and Lim can be localised according to their characteristic relative positions
(fig.323): L1l is left-anterior to the dca-cushions; L1m is right-anterior to the dea-cushions
and bears articulation A2.

In other Blattinae (with Deropeltis, Blatta, and Periplaneta studied) L1 is similar to
Archiblatta, but the dea-processes are rather variable, and the pne-pouch is less distinct

187

(as in Eurycotis, see below). The musculature of these species (not studied in Archiblatta)
confirms the assumed homologies for L1 and pne: Like in e.g. Mantoida or Cryptocercus,
there is a stout muscle from Lla to L4-sclerotisations (compare fig.53: L4C) in the left
edge of the left complex (12) and another one to L2 (compare fig.55) in the dorsal wall
of the Ive-pouch (13). Muscle I1 is missing. A derived feature peculiar to Blattinae (and
Eurycotis, fig.70) is the shift of the left insertion of muscle b4b to the anterior summit of
the pne-pouch (discussion in 6.7.1.).

In Eurycotis (fig.65-67) the characteristics of Lla, Lil, Lim, and A2 (fig.323e) and the
position of the phallomere-gland opening (P and edge 6 in fig.54, 55, 67, 68) are quite
the same as in Archiblatta. The pouch-shape of pne, however, is by far less distinct. The
processes posterior to L1 could be dea (as labelled in fig.66, 67) or loa (the right one?).
The insertion of 12 (fig.70) is still on the left part of the pne-pouch but has shifted from
L1 to the adjacent membrane. (The position of the I2-insertion on L4 is the same as e.g.
in Mantoida: discussion in 6.3.1.). Like in Blattinae, muscle I1 is missing. Muscle 13 from
L1 to L2 is represented by three bundles (l3a,b,c in fig.71), which together occupy the
same insertion area as the 13 of Blattinae, and an apomorphic tripartition can be assumed.
Muscle b4b inserts, like in the Blattinae, anteriorly on the pne-pouch (fig.70). The origin
and homology of the sclerites L6A and L6B (fig.66, 322g) only found in Eurycotis is

questionable: new sclerites or derivatives of L1? Homology with sclerite L8 of Ergaula,
Polyphaga (fig.117, 3221,m), and Lamproblatta (fig.177, 322k) is unlikely (different
muscle insertions); homology with L9 of Ergaula (fig.105) is also not very probable.

Lamproblatta

Like in the other species, L1 lies in the dorsal wall of the left complex, its anterior part
Lla (fig.323k) is inside a deep pouch (pne in fig.177), and its right part articulates with
L2 (A2 in fig.178). Furthermore, L1 and pne can be identified by the characteristic muscle
connections with the area of L4 in the left edge of the left complex (12 in fig.184; the
L4-sclerotisations are highly modified, discussion in 6.3.4.) as well as with L2 (13 in
fig.187; the insertion on L2 is far posteriorly). Like in Polyphaga, Ergaula, and Crypto-
cercus, the I2-insertion on L1 has shifted far anteriad. Muscle Il is missing.

As compared with other Blattaria and Mantodea, L1 and pne have shifted right-anteriad.
Most of the anterior part of L1 (Lla in fig.323k) is level, but, in contrast to Blattinae,
Eurycotis, and Tryonicus, there is a reminiscence of the hood-shape since the anteriormost
part of L1 bends into the dorsal wall of pne (fig.177, 178). This dorsal part of Lla may
even be regarded as a vestige of an anterior plateau which has been inclined posteriad.
Sclerite arms (regions Lil and L1m) are not distinct. The part of L1 containing articulation
A2 can be designated as the vestigial L1m-region (fig.323k; that A2 in fig.178 really is
A2 is shown in 6.2.4.). The demarcation of L1l in fig.323k is tentative. For the process
dea (fig.177) the homology with the dea (or loa, fig.54?) of the other species is
questionable. Region L1r is missing (no sclerite ring). The phallomere-gland opens, like
in Archiblatta or Ergaula, into the membrane extending ventrad from the posterior margin
of L1 (P in fig.178); however, parts of L2 and L4 (with the processes paa and pda,
fig.178) have shifted into the interspace between L1 and the opening (compare in 6.6.4.).

188

Anaplecta

In the previous species L1 and pne are in the central dorsal wall. In Anaplecta the
corresponding area is membranous and just somewhat depressed (fig.209). This area is
interpreted as the vestige of a pne-pouch, with L1 completely lost. This assumption is
supported by the muscles 12 and 13 (fig.201, 221), which run to L4-sclerotisations in the
left edge of the left complex (12) and to L2 in the dorsal wall of the Ive-pouch (13). These
are the same connections as in the species discussed before. 11 is, like in some other
species, missing. (The homology of L4 and L2 is discussed in 6.2.4. and 6.3.4). Further
support comes from McKittrick (1964): She identifies in another species of Anaplecta
(“sp. C”) a sclerite L1 (McKittrick’s fig.112), which has the same position as the assumed
pne-vestige in the species I studied.

Nahublattella

L1’ and pne (fig.242, 243) show some characteristic features: They take a dorsal position.
The level anterior part of L1’ (L1a in fig.323n) occupies the ventral wall of a pouch pne.
The posterior part of L1’ completely sclerotises a bulge-like process (dea in fig.243, 244).
This is interpreted as a sclerite-ring (regions L1l, Lim, and LIr in fig.323n) encircling
the dca-process which has spread posteriad over the whole dca (similar to but more
complete than in Tryonicus parvus, compare fig.323g and n). Further arguments for
homology come from the muscles on pne: I1 (fig.249) runs leftward to the sclerite-ribbon
L4d’ (which is probably homologous with L4d of Mantoida, fig.44, and Cryptocercus,
fig.150: discussion in 6.3.4.; compare Il of Mantoida, fig.48, and Cryptocercus, fig.155).
13 (fig.250) runs to L2 in the dorsal wall of the Ive-pouch (compare 13 of e.g. Mantoida,
fig.50, Polyphaga, fig.128, Cryptocercus, fig.158, 159, and Anaplecta, fig.221); the
homology relations of L2 and Ive are discussed in 6.2.). 19a (fig.249) runs to the membrane
to the right of L1’ — as do 19 in Anaplecta (fig.221) and the posterior part of 19 in
Polyphaga (fig.127, 129).

The right insertion of 12 (fig.249) has shifted away from the pne-pouch (discussion in
6.3.4.). Muscle 128, with both insertions on L1’, is peculiar to Nahublattella (fig.250).
Like in Cryptocercus, but in contrast to the other species, L1 and L2 are no longer in
contact (articulation A2 lost).

Parcoblatta and Blaberus

There are no vestiges of L1 and pne. Muscles I1 and 13 have been lost. 12 has shifted in
the same way as in Nahublattella (discussion in 6.3.4.).

6.2. Left complex II: Main sclerite L2 and associated elements

6.2.1. Comparison between Blattaria and Mantodea

In the Mantodean and in several Blattarian species the L2-sclerotisations, the Ive-pouch,
the vla-lobe, the processes paa and pda, and the genital opening show the same principal
arrangement and similar positions relative to pne and L1. The proportions of these

189

elements, however, can be very different. To determine the homology relations between
Blattaria and Mantodea and to reconstruct the ground-plan a comparison between
Mantoida, Polyphaga, Tryonicus, Archiblatta, and Eurycotis is most useful. Some L4-
sclerotisations and the pda-process will be considered in this section, but the homology
discussion of these elements will be completed in 6.3.

The right parts of L2 and of the Ive-pouch are level in Mantodea (fig.11, 26, 34, 46) but
curve dorsad and back to the left in most Blattaria (fig.54, 55, 94, 95, 118, 122). If this
up- and recurved area is extensive, the walls of the Ive-pouch cannot be designated as
dorsal and ventral: The wall corresponding to the dorsal Ive-wall of Mantodea (containing
L2) is the inner Ive-wall; the wall corresponding to the ventral Ive-wall of Mantodea
(mostly membranous) is the outer Ive-wall (compare in 5.5.-5.7.). The respective walls of
Ive are homologous, the absence or presence of a curvature being the only difference. The
vla-lobe usually shows the same curvature (in the figures vla is often pulled to the right),
but the walls of vla will be designated throughout as dorsal and ventral.

Mantoida and Polyphaga have a lot of features in common:

1. The pouch Ive (fig.46, 122) is flat and lies ventral to the pouch pne. However, in
Polyphaga Wve spans almost the whole width of the left complex and is on the whole
very large, whereas in Mantoida Ive is restricted to the right part and much smaller.

2. Sclerite L2 (fig.46, 122) is arch-shaped and extends along the anterior and lateral
edges of the Ive-pouch (edges 7 in fig.46, 122). However, in Mantoida L2 is mainly
restricted to the dorsal Ive-wall (only its leftmost part bends into the ventral wall,
compare fig.46, 47), whereas in Polyphaga L2 bends into the ventral wall all along
the edge 7 (compare 7 and L2 in fig.122 and 123).

3. The left part of L2 leaves the Ive-pouch posteriorly, bends around the posterior edge
of the left complex into the dorsal wall, and sclerotises a process (paa in fig.46, 117).

4. This L2-sclerotisation on paa is on its left connected with the sclerotisation of a
closely adjacent process (pda in 44, 117). However, the shapes of both paa and pda
are quite different in Mantoida and Polyphaga.

5. The right end of L2 articulates with the L1m-region (A2 in fig.45, 46, 118). However,
the right parts of L2 are up- and recurved in Polyphaga but level in Mantoida.

6. A muscle (13 in fig.50, 128) runs from the dorsal wall of Ive to pne (compare in
Os lleie))s

7. A muscle (14 in fig.50, 132) runs from the left edge of Ive to sclerotisations in the
left part of the left complex. However, the latter sclerotisations are very different in
Polyphaga and Mantoida.

8. Ventral to the Ive-pouch there is a broad ventral lobe (vla in fig.47, 123). The anterior
part of the dorsal vla-wall is at the same time the ventral Ive-wall.

9. The dorsal vla-wall is mostly membranous. The ejaculatory duct (D in fig.46, 47, 123,
124) opens into the right anterior part of this membrane.

10. The ventral vla-wall is part of the ventral wall of the left complex (fig.41, 115) and
is largely sclerotised (by L4 or L4M, respectively).

11. Two or three muscles inserting in the anterior ventral wall of the left complex run to
the Ive-pouch: The leftmost one inserts on the left anterior edge of Ive (15 in fig.50,
Ba):

190

12. The anterior right muscle (or the anterior part of the right one in Mantoida) inserts
on the right anterior edge of Ive (anterodorsal part of 16 in fig.50; 16a in fig.133).

13. The posterior right muscle (or the posterior part of the right one in Mantoida) runs
to the ejaculatory duct near its opening (posteroventral part of 16 in fig.52, l6b in
fig.132). Muscle 16 is undivided in Mantoida; in Sphodromantis, however, 16 is divided
in the same way as in Polyphaga (compare 16a and 16b in fig.132, 133 and 16, 18).

All features listed for Polyphaga are also true of Ergaula (both species, muscles only
investigated in E. capucina) — with the exception that the pda-process is missing.

Homology is assumed for all these similarities between Mantoida and Polyphaga and for
all elements given the same name. 1.-13. are regarded as features of the common ground-
plan of Blattaria and Mantodea. Some of the mentioned differences between Polyphaga
and Mantoida are bridged by various other Blattaria, which will be discussed subsequently;
these species, however, are in some features rather different from Mantoida and/or
Polyphaga.

The Ive-pouch and L2 of Tryonicus (fig.95) are, like Ive and L2 of Polyphaga, ventral to
and to the right of the pne-pouch, and the right parts of L2 and Ive curve dorsad and back
to the left (compare fig.95 and 94). In contrast to Polyphaga, this up- and recurved part
is by far more extensive, and it 1s directed anteriad and conceals the pne-pouch from
dorsally. Some features correspond with both Mantoida and Polyphaga: The relative
position of articulation A2 is the same; the dorsal Ive-wall is largely sclerotised by L2;
the left posterior part of L2 leaves the Ive-pouch and provides the sclerotisation of a
process (paa in fig.94, 97), which is to the left connected with the sclerotisation of another
process (pda in fig.94); the outer (= ventral) Ive-wall is membranous; this membrane is
at the same time the dorsal vla-wall (fig.87, 91) and contains the genital opening (D in
fig.91, 92) anteriorly; the ventral vla-wall is part of the ventral wall of the left complex
and is largely sclerotised (by L4G in fig.87).

In some features Tryonicus is more similar to Mantoida than Polyphaga is, and these
similarities are regarded as further features of the common ground-plan of Blattaria and
Mantodea:

14. The Ive-pouch of Tryonicus does not extend as far to the left as in Polyphaga but is
restricted to the right part of the left complex as in Mantoida (compare feature 1.;
fig.46, 97, 122).

15. L2 does not, in contrast to Polyphaga, occupy the margins of the outer = ventral Ive-
wall but is restricted to the inner = dorsal Ive-wall (fig.92, 94-98); this situation ap-
proximates that in Mantoida (compare feature 2.).

16. The shapes of both paa and pda are quite similar in Mantoida and Tryonicus (fig.44,
96; compare feature 4.): both are short and bulge-like, and paa is somewhat upcurved.

As a consequence, some features of Polyphaga (and Ergaula) are assumed to be derived:
(1) the extension of the Ive-pouch almost to the left edge of the left complex; (2) the L2-
sclerotisation in the marginal ventral Ive-wall; (3) the special shapes of paa and pda.

In some other features Tryonicus is certainly derived: (1) L2 is much broader than in
Mantoida and Polyphaga and has lost the arch-shape of the ground-plan since its dorsal
part is directed anteriad (fig.94-97; compare feature 2). (2) Another feature concerns the

191

left parts of the vla-lobe and the Ive-pouch. In all three species the invagination of the
lve-pouch (edge 7 in e.g. fig.96-98) and the free left edge of the vla-lobe (edge 61 in e.g.
fig.87, 98) start at the same point in the ventral wall of the left complex. Anterior to this
point the ventral vla-wall is confluent with the remaining ventral wall of the left complex
(e.g. fig.87). This point takes a different position in the three species. In Tryonicus it is
far anteriorly (fig.87): The left edge 61 of vla extends far anteriad, and the invagination
of the Ive-pouch begins far anteriorly. In Mantoida this point is by far more posteriorly
(7 and 61 in fig.41, 47): The left edge 61 of vla does not extend so far anteriad, the ventral
vla-wall is except for its posteriormost part confluent with the remaining ventral wall of
the left complex, and the invagination of the lve-pouch begins far posteriorly. In Polyphaga
this point is at the posterior edge of the left complex (fig.122, 123): The vla-lobe does
not have a free left edge 61 at all, the ventral vla-wall is completely confluent with the
remaining ventral wall of the left complex (fig.115), and the invagination of the Ive-pouch
begins most posteriorly. The ground-plan state of this feature is assumed to be somewhere
in between the situation of Mantoida and that of Polyphaga:

17. The ventral vla-wall is for most or all of its length confluent with the remaining ventral
wall of the left complex, the left edge 61 of the vla-lobe ends far posteriorly or is
missing, and the invagination of the left(-ventral) part of the Ive-pouch begins far
posteriorly.

Archiblatta likewise has a second pouch (lve in fig.54, 55) in the right part of the left
complex. Like in Polyphaga and Mantoida, the sclerotisation L2 (fig.55) runs like an arch
along the edges of the pouch (7 in fig.55). To the same extent as in Tryonicus, the right
parts of L2 and Ive curve dorsad and back to the left, and in this area L2 is very broad.
Similar to Mantoida (fig.46, 47), L2 is mostly restricted to the inner = dorsal Ive-wall but
bends into the outer = ventral Ive-wall in the posterior left part of Ive (at the posterior
end of edge 7 in fig.55, 56). In some features Archiblatta corresponds with all previous
species: L2 articulates with L1 (A2 in fig.54, compare fig.94, 118, 45, 46). The left
posterior part of L2 leaves the Ive-pouch and runs onto a process (paa in fig.55, 56),
which thus corresponds to the paa of the other species in this aspect of its relative position.

The outer = ventral Ive-wall is the dorsal vla-wall (fig.53; vla is pulled to the right), is
membranous, and receives the ejaculatory duct (D in fig.53). The ventral vla-wall is a
right part of the ventral wall of the left complex and is largely sclerotised (by L4G in
fig.54). Like in Polyphaga only, the vla-lobe does not have a left edge 61 but its ventral
wall is entirely confluent with the remaining ventral wall of the left complex.

Derived features of Archiblatta are (1) that paa is mostly membranous, (2) that there is
no connection between the sclerotisations of paa (L2) and pda (L4), and (3) that paa is
quite far removed from pda (compare fig.44 and 53). The two latter features are probably
correlated with a derived feature of the Ive-pouch: (4) The posteroventral part of lve has
strongly receded to the right (compare edges 7 in fig.122, 46, and 55) and is restricted to
a narrow right part of the left complex. The anteroventral part of Ive is still deeply
invaginated to the left and has become tongue-like.

Eurycotis is similar to Archiblatta but in some features more derived (compare fig.67, 68
and 54, 55): (1) L2 is not an arch but a plate (probably the arms of the arch have fused).

192

(2) The posteroventral part of the Ive-pouch is even more reduced than in Archiblatta
(compare fig.55 and 68). (3) A deep notch (9 in fig.63, 69) separates the right main part
of the vla-lobe from the remaining parts of the left complex. This notch lies within the
vla-lobe and does not correspond to the left edge 61 of vla in Tryonicus (compare fig.63
and 87). Like in the other species, the posterior part of L2 extends onto a process (paa
in fig.67, 68), which might thus be regarded as paa (compare fig.55). That paa is
completely sclerotised is primitve compared with Archiblatta, but, like in Archiblatta, paa
and its L2-sclerotisation have been far removed from pda and its L4-sclerotisation (fig.65).

The muscles of Eurycotis, compared with Polyphaga and Mantoida, confirm the assumed
homologies: Eurycotis also has a muscular connection 13 from L2 to the posterior part of
L1 (l3a,b,c in fig.71, 13 in fig.50, 128; compare in 6.1.) and a muscle 14 from L2 to the
left edge of the left complex (fig.71, 50, 132). The 14 of Eurycotis and Mantoida and their
insertion areas are extremely similar and reveal an additional feature of the common
ground-plan of Blattaria and Mantodea:
18. 14 (fig.50, 71) runs from L2 in the Ive-pouch to the swe-apodeme on L4-
sclerotisations in the left edge of the left complex (sclerites L4 or L4H), where it
inserts immediately ventral to muscle 12 (fig.49, 70) coming from the pne-pouch.

Several muscles of Eurycotis run from the ventral wall of the left complex to the Ive-
pouch: 16b (fig.70, 71) runs to the ejaculatory duct near its opening, like 16b in Polyphaga
(fig.132) and the ventral part of 16 in Mantoida (fig.52). Another muscle (16a in fig.73)
runs to the anterior ventral wall of lve, and homology with either I5 or 16a of Polyphaga
(fig.133) and Mantoida (16a = dorsal part of 16 in fig.50) seems possible. Two other
muscles (l3a,b in fig.72) insert on the ventral left edge of the lve-pouch, somewhat like
15 in Polyphaga (fig.133); however, the ventral insertions of I5a and I5b are far posteriorly.
Thus, for 15a,b and l6a of Eurycotis the homologies are not completely clear, but I suppose
that the relations expressed by the designations are the most probable.

Concerning the common ground-plan of Blattaria and Mantodea, one question remains
open: It is not decidable whether the right parts of L2 and Ive are level (like in Mantodea)
or up- and recurved (like in Blattaria).

The definition of the regions of main sclerite L2 (fig.324) is based on the primitive arch-

shape of L2, which is present in its typical form in Mantoida, Archiblatta, and Polyphaga.

From the left to the right four L2-regions are distinguished:

— L2d (dorsal): The sclerotisation of the process paa.

— L2p (posterior): The part of the L2-arch in the left posterior part of the Ive-pouch.

— L2a (anterior): The part of the L2-arch in the anterior part of the Ive-pouch.

— L2m (median): The part of the L2-arch in the right part of the Ive-pouch. L2m has an
articulation A2 with L1.

— L2v (ventral): This is not defined as a separate region of L2. This term is used (mainly
in fig.324) if large parts of L2 have invaded the ventral wall of the Ive-pouch; these
parts of L2 are not necessarily homologous in the species concerned.

The up- and recurved right parts of L2 of Blattaria belong to the regions L2m and L2a.

193

6.2.2. The elements in the common ground-plan of Blattaria and Mantodea

The features 1.-18. in 6.2.1. permit a reconstruction of the ground-plan morphology of
L2, lve, paa, and vla, and of the genital opening (fig.321e,g): The Ive-pouch lies ventral
to the pne-pouch; it is quite broad but does not reach the left edge of the left complex
(and does not have a recess to the right in its posterior part). L2 is largely restricted to
the dorsal Ive-wall and runs like an arch along the lateral and anterior edges of the Ive-
pouch. The left part (L2p-region) of the arch leaves the pouch posteriorly, and the
posteriormost part (L2d-region) sclerotises the short, somewhat upcurved process paa.
L2d is connected with the sclerotisation of the process pda, and paa and pda are close
to each other. The right end of L2, or the dorsal left end in the case of L2-upcurving
(L2m-region), articulates with L1m (A2). The ventral Ive-wall is mostly membranous and
is at the same time an anterior part of the dorsal vla-wall. The ejaculatory duct D opens
into the right anterior part of this membrane. The ventral vla-wall is largely sclerotised
and almost completely confluent with the ventral wall of the remaining left complex (i.e.
the left edge 61 of vla is missing or does not extend far anteriad). Muscles 13, 14, 15, 16a,
and 16b are present.

6.2.3. Homology relations and character states of the elements in Mantodea

In Chaeteessa, Metallyticus, and Sphodromantis, the main features of L2, Ive, and paa
are like in Mantoida (fig.11, 26, 34, 45, 46): Ive lies ventral to the pne-pouch. L2 is
restricted to the dorsal Ive-wall. The right parts of L2 and Ive are level. Posteriorly L2
leaves the lve-pouch to provide the sclerotisation of a process which is therefore regarded
as paa. The right margin of L2 articulates with L1 (A2; compare in 6.1.3.).

In contrast to Mantoida, Metallyticus and Sphodromantis have separated the sclerotisations
of paa (L2) and pda (L4), and the processes are more distinct from each other and more
prominent (fig.10-12, 23-26, 44-46). In Chaeteessa (fig.31, 32, 34) paa is very prominent,
too, but a process pda is missing. Nevertheless, in this species, too, L2 and L4 are
completely separated in the area concerned.

L2 of Chaeteessa, Metallyticus, and Sphodromantis has become plate- or ribbon-like and
has spread over most of the dorsal Ive-wall. Like in Eurycotis, the arms of the arch (regions

Fig.324: Left complex, homologous regions of main sclerite L2 (on pages 194, 195). — Only L2-
sclerotisations are shown. Dorsal views. L2 is divided into the four regions L2a, L2d, L2m, and
L2p (definition in 6.2.1.); the extensive parts of L2 in the ventral wall of pouch lve in Polyphaga
are labelled L2v (not defined as a region of L2). If L2 is divided into several sclerites, these are
labelled with the capital letters used in the text and in fig.1-319 (e.g. B = L2B). Undulate lines are
cutting lines through sclerotisations (if L2 is fused with parts of other main sclerites, e.g. L4). In
fig.324e,f,g the left drawing shows the complete L2, the right drawing shows L2 after removal of
its dorsal parts. The part of the sclerite margin which forms articulation A2 with sclerite L1 is
indicated by dashes. A4 and A10 are articulations between L2-sclerites. paa and via are processes
occupied at least partly by L2-sclerotisations. The sclerite of Parcoblatta (fig.324n) and Blaberus
(fig.3240) which is termed L2 in the text has to be designated correctly L2D+(L2E+L4N), compare
in 6.2.4.; the L4N-part has been removed from the sclerite.

194

a) Sphodromantis sp.

ER

RR

S

b) Metallyticus

violaceus

d) Mantoida
schraderi

e) Eurycotis
floridana

e) Chaeteessa
caudata

g) Tryonicus
parvus
f} Archiblatta
hoeveni

195

L2d

L2m

L2

a

0) Blaberus
ceranlifer

n) Parcoblatta

=
i)
Saye
S
aS,
Ss oo
my &
ral et
a
= EN
= m SE.
2 mc INN = =
>
aS
Es
oo
coed
8s =
2%
= e
es A
33 es
eo =
w
= a =
aS g
a" =
z =
= «
=
(ee

lata

196

L2p and L2m) have probably fused (fig.324a-c, compare fig.324d). In Chaeteessa and
Metallyticus, as compared with Mantoida, the Ive-pouch has strongly narrowed (compare
fig.26, 34 and 46), and L2 is also narrow. In Sphodromantis the anterior part of the Ive-
pouch is deeply invaginated to the left, and L2 forms a broad transverse tongue (fig.11).
This resembles the situation in Archiblatta (compare the course of edge 7 in fig.11 and
55) — certainly a case of parallel evolution.

As in Mantoida, in Chaeteessa, Metallyticus, and Sphodromantis the membranous ventral
Ive-wall is a (left) anterior part of the dorsal vla-wall (fig.12, 27, 35), the ejaculatory duct
(D in fig.11, 12, 26, 32) opens far anteriorly into the dorsal vla-wall, and the ventral vla-
wall is part of the ventral wall of the left complex and is largely sclerotised (by L4 or
L4A in fig.6, 20, 28). The genital opening lies in Sphodromantis within the Ive-pouch
(like in Mantoida and Polyphaga); in Metallyticus and Chaeteessa it is a bit, or far,
respectively, to the right of the pouch and outside of it.

The homology of Ive and L2 of Mantoida and Sphodromantis is confirmed by the
musculature: 15, l6a, and l6b of Sphodromantis (fig.15, 16, 18) show the same arrangement
as IS and 16 in Mantoida (fig.50, 52). Both species have a stout muscle 13 from L2 to L1
(fig.16, 50). Muscle 14 inserts on the left edge of Ive (fig.15, 50) and has its opposite
insertion (on L4-sclerotisations) immediately to the left of the dorsal 12-insertion.
However, the latter insertions of both 12 and 14 are in Sphodromantis by far more to the
right than in Mantoida. This topic will be taken up again in 6.3.3.. Muscle 18 of
Sphodromantis (fig.16) consists of very few fibers at most (completely missing in some
specimens). Snodgrass (1937) finds this muscle well-developed in Tenodera sinensis
(Mantidae; muscle 13 in Snodgrass’ fig. OD). Hence, 18 of Sphodromantis represents either
a vestige or an early evolutionary stage of this muscle. Since 18 is missing in Mantoida
and all Blattaria it is not assumed to be a muscle of the Mantodean ground-plan.

6.2.4. Homology relations and character states of the elements in Blattaria
Polyphaga, Ergaula, Tryonicus, Archiblatta, and Eurycotis

These species have been sufficiently discussed in 6.2.1..

Lamproblatta

The pouch and the ventral lobe labelled Ive and vla in fig.180 resemble in many respects

the lve and vla of the previous species: The Ive-pouch lies ventral to the pne-pouch. Its

dorsal wall is largely sclerotised (L2A, L2B). Its ventral wall is mostly membranous, is

at the same time the dorsal vla-wall (fig.180, 181), and contains the genital opening (D

in fig.180, 181) in its right part. The ventral wall of vla is part of the ventral wall of the

left complex and is partly sclerotised (L4R in fig.174, 181). However, in Lamproblatta
there are two problems:

— Instead of only one sclerite L2 as in the former species there are two within the Ive-
pouch (L2A and L2B), and around articulation A4 the Ive-pouch has a deep recess
(fig.180). The questions arise if either only the part to the left of the recess (with L2A)
corresponds to the Ive-pouch of the other species, or if the part to the right of the recess
(with L2B) is also a true part of Ive, and whether L2B really is a part of L2 (or of L1:

197

a split off part of the L1m-region like L1B of Sphodromantis, fig.10?). This equals the
question which of the articulations A2 (fig.178) and A4 (fig.180) is the true A2.

— In Mantoida, Polyphaga, and Tryonicus the left posterior part of L2 leaves the pouch
and continues into the sclerotisation of the processes paa and pda (fig.46, 96, 118). In
Lamproblatta, the left posterior part of L2A also has an extension that leaves the pouch
(LAS in fig.178, 180), and right-anterior to this area there is a sclerotisation with two
processes (paa and pda in fig.178, 179), which, however, is completely separated from
L2A and L4S). The question arises if paa and pda of Lamproblatta are homologous
with those of the other species.

These problems can be solved by a comparison of the musculature of Lamproblatta and

the other species — especially Polyphaga.

— Lamproblatta also has a stout muscle from L1 to Ive (13 in fig.187; compare Polyphaga,
13 in fig.128). Its insertion area on Ive is within the recess and also includes parts of
L2B.

— Lamproblatta and Polyphaga have 4 muscles (15, 16a, s3, s12) inserting close to each
other in the anteriormost ventral wall of the left complex (fig.133, 188); homology is
assumed for all of them (s3 and s12 are discussed in 6.9.). In Polyphaga the two
posterior muscles (I5 and 16a) run to L2 at the left or right, respectively, anterior edge
of the Ive-pouch. In Lamproblatta 15 runs to L2A on the left of the recess, 16a runs to
L2B on the right of the recess.

— The two l6b-bundles (fig.189) of Lamproblatta are assumed to be homologous with the
l6b of the other species (Polyphaga: fig.132): The ventral insertion is posterior to that
of l6a. The insertion in the dorsal vla-wall is somewhat different in Polyphaga and
Lamproblatta (next to the genital opening or far posterior to it), but the insertions of
the two bundles of Lamproblatta are similarly situated as the insertion of the one bundle
of l6b of Sphodromantis (fig.18) or Cryptocercus (fig.155, 157, see below).

— Muscle 112 of Lamproblatta and Polyphaga inserts in the right part of the ventral =
outer Ive-wall — very close to the dorsal insertion of 16a (fig.128, 129, 186, 188) — and
runs to a small sclerite in the right dorsal wall of the left complex (L8, homology
discussion in 6.5.). Homology is assumed for these 112. In Lamproblatta 112 inserts to
the right of the recess, on L2B. (L8 and 112 are derived features of Polyphaga, Ergaula,
and Lamproblatta.)

— Muscle 110 of Polyphaga (fig.129) runs from L2 in the left dorsal wall of Ive to the
sclerotisation in between the processes paa and pda. Lamproblatta has a muscle (110
in fig.186) from the same area of Ive to the sclerotisation between paa and pda. Such
a muscle is missing in all species discussed before.

The course of 110 suggests homology for paa and pda of Lamproblatta and Polyphaga.
Since the pda-sclerotisation is part of L4, the discussion of this topic will be continued
in 6.3.4..

The insertions of 13, 16a, and 112 clearly demonstrate that L2B is a true part of L2
(regioning in fig.324i) and that the respective part of the pouch is a true part of Ive. That
the right part of L2B curves dorsad and leftward like the right part of L2 in other Blattaria
and that the genital opening is to the right of the recess (fig.122, 180) supports this

198

a) Sphodromantis sp. c) Chaeteessa
caudata

SS

ISIN

SSO
Sv
SSN
Sv
SSS

N
II
N

b) Metallyticus
violaceus

Fig.325: Left complex, homologous regions of main sclerite L4. — Only L4-sclerotisations are shown.
Dorsal views. L4 is divided into the regions L4a, L4b, L4c, L4d, L4l, L4n, L4v, and L4x (definition
in 6.3.1.). If L4 is divided into several sclerites, these are labelled with the capital letters used in the
text and in fig.1-319 (e.g. B = L4B). The sclerites are mostly shown as they are arranged in the left

199

swe

e) Eurycotis
floridana

f) Archiblatta
hoeveni

Lda

L4b

Ldc

1.4d

141

L4n

14v

L4x

g) Tryonicus
parvus

; (contd.): complex; only in Sphodromantis (fig.325a) and Metallyticus (fig.325b) the dorsal sclerites
L4B are shifted to the left (thin arrows). Undulate lines are cutting lines through sclerotisations (if
| L4 is fused with parts of other main sclerites, e.g. L2). Al and A5 are articulations between L4-
4 sclerites. pda, via, nla, sla, vsa, tve, and swe are formative elements occupied at least partly by L4-

200

h) Cryptocercus
punctulatus

Lda

Y
i) Lamproblatta J Yj L4b
albipalpus
L4e
L4d

L4l

L4n

LAV

N L4x

s3 si2

N N N NS
NN

N

k) Polyphaga
aegyptiaca

325

(contd.): sclerotisations. Broad arrows represent muscles inserted on or near L4-sclerotisations; 15 is
certainly homologous in most species; if homology with these 15 is questionable, the muscle is labelled
15? (in Anaplecta and Nahublattella 15? are homologous). In Mantodea the presence of region L4c
is questionable (? in fig.325a-d). The sclerite of Parcoblatta (fig.325n) and Blaberus (fig.3250)

et

pda n) Parcoblatta lata

I) Anaplecta sp.

o) Blaberus craniifer

325

m) Nahublattella sp.

201

(contd.): which is termed L2 in the text has to be designated correctly L2D+(L2E+L4N), compare

in 6.2.4.; only the L4N-part is shown in the figures.

202

interpretation. Thus, the articulation between L2B and L1 (A2 in fig.178) is homologous
with A2 of the other species. In correlation with the right-anteriad shift of L1 (compare
in 6.1.4.), A2 of Lamproblatta is far anteriorly. The articulation A4 and the recess within
the Ive-pouch are derived features of Lamproblatta.

Another derived feature of Lamproblatta is the lack of muscle 14. Another derived feature
common to Lamproblatta and Polyphaga (and Ergaula) — in addition to L8 and 112 — is
that the Ive-pouch and the vla-lobe extend nearly to the left edge of the left complex
(compare edges 7 in fig.122 and 180).

Cryptocercus

The elements L2, Ive, paa, and via (fig.151, 152) can be clearly identified by many
features corresponding with the other species: L2 lies ventral to the pne-pouch and is
connected with it by a stout muscle (13 in fig.158, 159). The posterior part of L2 bends
into the dorsal wall of the left complex, and the area of bending forms a bulge-like process
(paa and L2d in fig.151, 152, 324h; compare Tryonicus, fig.95, 97, 324g). This dorsal
part of L2 extends anteriad as far as to the opening of the phallomere-gland (P in fig.151,
152) — like in Polyphaga (fig.118, 120). The right anterior part of L2 occupies the dorsal
wall of a pouch-like invagination (Ive in fig.150-152), which, however, is restricted to the
anteriormost part of the left complex. The ejaculatory duct (D in fig.150, 151; compare
Mantoida, fig.46) opens into this lve-pouch. The ventral Ive-wall is at the same time part
of the dorsal vla-wall (fig.150-152), and the ventral vla-wall is partly sclerotised (L4G in
fig.148, 152). Muscle 110 (fig.155) runs from L2 to the membrane to the left of paa. This
membranous area is in between the processes paa and pda (pda, fig.150, is discussed in
6.3.4.), and thus the left insertion of this 110 is like that of the 110 of Polyphaga (fig.129)
and Lamproblatta (fig.186). However, the right insertion is by far more posteriorly, and
the homology of these 110 is not certain. Muscle 14 (fig.155, 158) runs from the anterior
part of L2 to L4K in the left part of the left complex like 14 of Polyphaga (fig.132; 14
of Cryptocercus is strongly reduced; L4K is discussed in 6.3.4.).

Cryptocercus has some features which are, compared with the ground-plan, clearly
derived: (1) L2 is, like in most Mantodea and Eurycotis, more plate-like, though the
primitive arch is still recognisable (compare L2-regions in fig.324h and d,f,k). (2) The
right part of L2 is, like in Mantodea, not upcurved (fig.151, 152). However, since the
contact between L2 and LI (articulation A2) has been lost, this is not interpreted as a
primitive situation — as suggested by the outgroup comparison with Mantodea — but as a
reduction of the right part of L2. Consequently, the right-dorsal part of the Ive-pouch,
which contains the upcurved part of L2 in other Blattaria, has been strongly reduced. (3)
The vla-lobe is separated from the remaining ventral wall of the left complex as far as to
the anterior margin of the left complex (edge 61 in fig.148). Accordingly, the invagination
of the Ive-pouch begins very far anteriorly (see left end of edge 7 in fig.150, 151), and
the left-ventral part of lve has been strongly reduced. This is an extreme modification of
the situation in Tryonicus (edge 61 in fig.87 and edge 7 in fig.97). (4) Of the muscles
from the ventral wall of the left complex to the Ive-pouch and the ejaculatory duct only
one is present (l6b in fig.155). According to its insertions (anterior margin of L4G,

203

membrane posterior to genital opening), it is likely to be the homologue of 16b of Eurycotis
(fig.71), Sphodromantis (fig.18), and the other species. 15 and l6a have been lost; this
might be a consequence of the extreme reduction of the Ive-pouch.

Anaplecta

The homology of the elements designated L2, Ive, vla, paa, and pda in Anaplecta with
the respective elements of the other species is suggested by the following features: The
whole area labelled Ive (fig.210-213), whose anterior part narrows to form the Ive-
apodeme, is a large invagination to the anterior, which lies beneath the pne-"pouch”
(fig.209). Ive and pne are connected by a stout muscle (13 in fig.201, 222, 50, 128). L2
is mainly restricted to the dorsal wall of Ive. Anteriorly, however, L2 also occupies the
margins of the ventral Ive-wall (fig.211, 225) — similar to Polyphaga (fig.123) and
Lamproblatta (fig.181). That a phallomero-sternal muscle inserts on the Ive-pouch
resembles Eurycotis (s7 in fig.58, 200). That L2 forks at the base of the Ive-apodeme is
regarded as a vestige of the primitive arch-shape (compare fıg.3241 and d,k); this is
confirmed by the morphology of the two branches of the fork: The right branch is upcurved
at its right margin (fig.211-213), like the right part of L2 in other Blattaria. The cuticular
area containing this part of L2 can therefore be regarded as the right dorsal (= inner) wall
of the Ive-pouch. Anterior to this right L2-part there opens, like in the other species (e.g.
Mantoida, fig.46), the ejaculatory duct (D in fig.211). The left branch of L2 continues
into a sclerotisation bearing two processes (paa in fig.211, pda in fig.214). This is the
same situation as at the left end of the L2-arch of Mantoida, Polyphaga, and Tryonicus.
Muscle 110 runs, like 110 of Polyphaga and Lamproblatta (fig.222, 129, 186), from the
sclerotisation of paa and pda to L2 in the left-dorsal Ive-wall. The relative positions of
the vla-lobe (fig.205, 218-220) and its sclerite L4G (fig.205) in the ventral wall of the
left complex are especially similar to those of vla and L4G of Tryonicus (fig.87, 205),
with vla having a left edge (61 in fig.205) reaching far anteriad (farther than in Tryonicus,
fig.87, but not as far as in Cryptocercus, fig.148).

In contrast to all other species, the edge of the Ive-pouch is — except in the area of the
lve-apodeme — not continuous throughout (compare edges 7 in fig.55, 122, 180, 211, 212)
but interrupted by some apomorphic membranous foldings (fig.212-219): e.g. outfolding
via, infolding vpe (fig.214, 215, 217). For that reason it is difficult to determine the
homologies of the muscles of this area with the 15- and 16-muscles of the other species.
l6b (fig.224) is probably homologous with l6b of e.g. Sphodromantis, Eurycotis,
Cryptocercus, and Lamproblatta (fig.18, 71, 155, 189): All these 16b run from the sclerite
plates in the ventral vla-wall, or from their vicinity, to the dorsal vla-wall. In Anaplecta,
however, the dorsal insertion is not immediately behind the genital opening but is separated
from it by the outfolding vfa (compare fig.223 and 224). vfa is therefore assumed to be
evaginated from the anteriormost dorsal wall of vla and the ventral wall of the ejaculatory
duct. This assumption is supported by two other muscles: s10 inserts on the ejaculatory
duct in Nahublattella (fig.249) and Parcoblatta (fig.276) but on the dorsal base of vla in
Anaplecta (fig.222). 113 of Polyphaga (fig.132), Cryptocercus (fig.155), and Eurycotis
(113h in fig.72) runs from the ejaculatory duct to the dorsal vla-wall posterior to it. 113

204

of Anaplecta (fig.222) also inserts on the ejaculatory duct but bridges the vfa-outfolding
on its way to its insertion on the dorsal base of vla (discussion of 113 in 6.5.). Muscle
16a of Anaplecta (fig.222) resembles 16a of Polyphaga and Lamproblatta (fig.133, 188)
in inserting ventrally behind s3 and dorsally at the right anterior edge of the Ive-pouch.
Whether muscle 15 is homologous with 15 of the other species (fig.133, 188, 223) is
questionable: The insertion on L2 is similar in Anaplecta and e.g. Polyphaga; the anterior
insertion (on L4), however, is situated quite differently in these two species. Homology
is also unclear for the muscles 125 and 126 (fig.224).

As compared with the previous species, Anaplecta has some important derived features:
(1) The anterior part of the Ive-pouch is a tube-like Ive-apodeme. (2) Edge 7 is interrupted
by vfa and vpe. (3) The common sclerotisation of paa and pda is stout and ring-shaped
at its base. (4) Muscle 14 from L2 to left parts of L4 has been lost (like in Lamproblatta).

Nahublattella

The part of the left complex comprising the large pouch Ive (fig.242), the opening of the
ejaculatory duct (D in fig.242), the processes, via, vsa, paa, and pda (fig.244, 245), and
the sclerotisations L2D’, L2E’, and L4N’ show a lot of similarities with the elements of
Anaplecta discussed before:

All these elements lie in the center of the left complex and (antero-)ventral to the pne-
pouch. The anterior part of the Ive-pouch (see edges 7 in fig.242) is a tube-like Ive-
apodeme, whose dorsal wall is completely sclerotised, and whose ventral wall contains a
membranous stripe (44 in fig.206, 212, 239a, 245). Muscle s7 runs from the Ive-apodeme
to the subgenital plate (fig.200, 234).

At the base of the apodeme, L2D’ is somewhat forked (fig.243, 324m), like L2 in
Anaplecta (fig.212, 3241): Extension 36 is the left branch, the posterior main part of L2D’
is the right one. At the left branch there adjoins a ring-shaped sclerotisation lying at the
base of some processes (paa and pda in Anaplecta, fig.211-214; via with vsa, paa, and
pda in Nahublattella, fig.244, 245). A stout muscle 110 (fig.222, 250) runs from the Ive-
apodeme to the left part of this sclerite-ring. However, in Nahublattella the sclerite ring
has become separated from the rest of L2 (L2D’) by an articulation (A10 in fig.244). At
the base of the left branch the L2-sclerotisation bends into the ventral Ive-wall and forms
a posterior extension (28 in fig.215, 245). However, in Anaplecta the cuticle around
extension 28 forms a process (gta in fig.215), which is missing in Nahublattella.

The right branch of L2 or L2D’, respectively, extends rightward in Anaplecta but more
posteriad in Nahublattella. The relation between this sclerotisation and the dorsal wall of
the ejaculatory duct (D in fig.245, 246) is the same in the two species (fig.211, 245), but
only in Nahublattella the right-anterior margin of this L2-part folds narrowly back to the
left (towards edge 38 in fig.245).

Two further muscles inserting on L2D’ of Nahublattella correspond with muscles of the
other species: 13 running to the pne-pouch (fig.250, compare e.g. fig.50, 71, 128, 221),
and 14 running to L4-sclerotisations in the left part of the left complex (fig.249, compare
e.g. fig.50, 71, 129; missing in Anaplecta; homology discussion of L4 in 6.3.4.).

205

d) Ergaula
capucina

S c) Polyphaga
SIS aegyptiaca

b) Lamproblatta

albipalpus a) Cryptocercus

punctulatus 3 2 6

L2
: L3
Fig.326: Left complex of Lamproblatta, Cryptocercus, Polyphaga,
and Ergaula, homology relations of the sclerotisations in the left L4K

part. — Only cuticular elements of the left part of the left complex
are shown. Dorsal views. Scale Imm. The sclerites are patterned
differently according to their homology relations. Ive, hla, and vla
are formative elements. The position of region L4d is given.
Undulate lines are cutting lines. The branching black lines represent
the assumed phylogeny.

LAN, L4S (with L4d)

L4G, L4M, L4R

LS

206

The membranous lobe vla (fig.239a, 245-247) has similar features as the vla-lobe of e.g.
Sphodromantis (fig.6,12), Lamproblatta (fig.174, 180), Cryptocercus (fig.148, 151), and
Eurycotis (fig.63, 66): Its ventral wall is part of the ventral wall of the left complex. Its
dorsal and its ventral walls are connected by a stout muscle I6b (fig.251, 252, 18, 71, 155,
188, 189). (These two features are also true of Anaplecta, fig.205, 218, 224). Its dorsal
wall is at the same time the ventral wall of the Ive-pouch. (This is not true of Anaplecta
because of the membrane foldings between lve and vla, especially vfa).

Like the other species (with the exception of Cryptocercus), Nahublattella has muscles
from the anterior ventral wall of the left complex to the L2-sclerotisations: 15 is certainly
homologous with 15 of Anaplecta (similar posterior insertion on the left branch of L2; the
homology of the anterior insertion is discussed in 6.3.4.), but, as in Anaplecta, homology
with the 15 of the other species is questionable. Muscle 16a (fig.250) could be homologous
with 16a of Anaplecta (and the other species); however, the insertion on the Ive-apodeme
is by far more anteriorly (fig.222, 250), and the insertion in the ventral wall is on
sclerotisation. Alternatively, homology with 126 (fig.224) of Anaplecta seems possible.

Nahublattella shows some important derived features as compared with Anaplecta and, at
least in the case of (2)-(6), all other previously discussed species: (1) The right branch of
L2 (posterior part of L2D’) is by far narrower (compare fig.213 and 245). Moreover, the
whole right posterior dorsal part of the left complex — that part with the right L2-branch
in its ventral wall (Anaplecta: fig.211-213) — is strongly reduced to form just the bifid
psa-process (fig.245; compare fig.328a and b). (2) L2 has divided into L2D’ and L2E’
by articulation A10. (3) The sclerotisation at the common base of the processes paa and
pda, which is ring-shaped in Anaplecta (fig.211-213), has lengthened to form a cylinder
(fig.244). Hence, the processes paa and pda (and vsa) are now only the distal branches
of a larger evagination, which has been defined as a “new” process via (paa, pda, and
vsa are subordinate parts of via). The homologies of the single processes of Anaplecta
and Nahublattella are hardly determinable, but in my view the relations expressed by the
designations are the most probable. In accordance with Anaplecta, the sclerotisation of
via is assumed to comprise a L2-part (L2E’, roughly the L2d’-region; dorsally on via
and near articulation A10; fig.324l,m) and L4N (ventrally on via and near the insertion
of 110; fig.3251,m; discussion of LAN in 6.3.4.). (4) This sclerotisation of via is divided
into a basal and a distal sclerite (39 in fig.241, 244). (5) There is no sclerotisation in the
ventral wall of the vla-lobe (compare fig.205 and 239a). (6) The main muscle of the hla-
hook (114 in fig.249) has its anterior insertion on the Ive-apodeme (discussion in 6.4.).

Parcoblatta, Blaberus, and other Blattellidae and Blaberidae

The L2-sclerotisations, the pouch Ive, the processes via, paa, pda, and psa, the lobe vla,
and the muscles s7, 14, and 110 have been studied not only in Parcoblatta and Blaberus
but also in Supella, Euphyllodromia, Loboptera, Ectobius and Nyctibora (Blattellidae),
Nauphoeta and Blaptica (Blaberidae) (muscles not studied in Ectobius). Morphology and
homology of these elements are shown in fig.328. The morphology of all these species is
derived from a situation similar to Nahublattella.

d) Ergaula
capucina

c) Polyphaga
aegy ptiaca

2 aproblaite
albipalpus

Fig.327: Left complex of Lamproblatta, Cryptocercus, Polyphaga, and
Ergaula, homology relations of the muscles in the left part. — Cuticular
elements are shown as in fig.326, but some parts are removed. Dorsal
views. Scale Imm. All sclerotisations are patterned in the same manner.
The muscles 14, 15, 111, and 114, if present, are shown and patterned
differently according to their homology relations. 14 is always cut
through. The ventral insertion of 15 is shown only in Lamproblatta.
Undulate lines are cutting lines. The branching black lines represent the
assumed phylogeny.

207

a) Cryptocercus
punctulatus

297

Sclerotisation

J

== il

208

Fig.328: Left complex, evolution of

sz a) Anaplecta sp. 3 2 8 main sclerite L2 and sclerite LAN in
nn) Blattellidae and Blaberidae. — The
central part of the left complex is
shown, with pouch lve, terminal parts of
ejaculatory duct (D) and phallomere-
gland (P), processes paa, pda, via, and
psa, ventral lobe vla, and tendon tve.
Dorsal views. Patterned areas are
sclerotised. Undulate lines are cutting
lines through the cuticle. L2D and L2E
are separate L2-sclerites, A10 is the
articulation between them. Broad
arrows represent the muscles 14, 15, 110,
and s7 (not investigated in Ectobius,

fig.328g). Curved arrows in Parcoblatta
(fig.328e) and Blaberus (fig.328k) show
the direction of rotations. X and Y are
special elements of Loboptera.
(Detailed information in 6.2.4.). Species
with “S” behind their names have side-

at reversed phallomeres, and a mirror-
Sclerotisation of processes pda (= posterior

part of region L4l = main part of sclerite ugs Ou (ns Sram Preparztion 3
LAN) and paa (= region L2d = left posterior shown. The branching black lines
part of sclerite L2 or L2E), or of process via represent the assumed phylogeny.

The sclerotisations are patterned as follows:

Regions L2p and L2a

Region L2m Sclerite L10 (only in Blaberus, fig.328k)

A c) Supella d) Euphyllodromia
b) Nahublattella sp. longipalpa® . 4 via angustata®
(Plectopterinae) (Plectopterinae) (Plectopterinae)

209

k) Blaberus
eraniifer®
(Blaberidae)

i) Nauphoeta
cinerea®
(Blaberidae)

h) Nyctibora sp.
(Nyctiborinae)

g) Ectobius
sylvestris
(Ectobiinae)

f) Leboptera
decipiens
(Blattellinae)

e) Parcoblatta
lata
(Blattellinae)

210

Pouch Ive and its L2-sclerotisation The elements designated lve and L2 have
a lot of features in common with lve and L2 of Nahublattella and/or Anaplecta: Ive is a
deep anteriad-directed invagination in the center of the left complex which is partly
sclerotised (by L2; fig.210, 242, 268, 299, 328). The anterior part of this Ive-pouch is a
tube-like Ive-apodeme with a membranous stripe in its ventral or right wall (44 in fig.206,
239a, 245, 266, 297a). A muscle s7 runs from the Ive-apodeme to the left half of the
subgenital plate (fig.221, 249, 276, 328); in Blaberidae, s7 is missing in Blaberus but
present in Nauphoeta. A muscle 110 runs from the Ive-apodeme to the base of the Ive-
pouch; however, the positions of the 110-insertions are not exactly the same in the various
species, and in some species 110 is missing (fig.328d,e,f,k; discussion below). The
ejaculatory duct joins the Ive-pouch from the right side at the base of the Ive-apodeme
(like in Nahublattella; D in fig.242, 268, 299, 328). The genital opening is to the right of
the apodeme — only in Blaberus its position is more dorsal (fig.328k), and only in
Parcoblatta its position is more ventral (fig.328e). Only in Ectobius the ejaculatory duct
opens far to the right of the apodeme (fig.328g). Where the ejaculatory duct joins the lve-
pouch, Ive strongly widens, like in Nahublattella, and the right posterior part of Ive is
membranous (fig.328b-k; both is not true of Ectobius, fig.328g). The phallomere-gland
opens into the posteriormost dorsal Ive-wall (like in Nahublattella, P in fig.328; the
phallomere-gland has been lost in Supella, fig.328c, and Ectobius, fig.328g). A muscle 14
runs from the L2-sclerotisation (fig.249, 276, 303, 328) to the left wall of the left complex,
where it is attached to L4-sclerotisations if present (discussion of L4 in 6.3.4.). 14 was
not found in Loboptera.

The remaining muscles of this area have been investigated only in Parcoblatta and
Blaberus. Two further muscles having homologues in Nahublattella insert anteriorly on
the Ive-apodeme: 114 or I14a,b (fig.249, 276, 303) run to the hla-hook (discussion of 114

in 6.4.3.). 16a (fig.250, 277, 304) runs posteroventrad. In Blaberus and Nahublattella the
posterior l6a-insertion is still in the anteriormost ventral wall of the left complex, but in
Parcoblatta it has shifted far posteriad to the ventral wall of the genital pouch (fig.267).
In Blaberus and Parcoblatta 16a has strongly enlarged. 115 is restricted to Nahublattella
(fig.249); 142 is restricted to Blaberus (fig.304). Parcoblatta and Blaberus have lost muscle
13 (from Ive to pne, compare Nahublattella, fig.250).

Process via and its L2- and L4-sclerotisations All species except Loboptera
(fig.328f) and Ectobius (fig.328g) have a sclerotised process behind the Ive-pouch (via in
fig.328), whose shape and size varies, and whose sclerotisation can be connected with
(fig.328d,e,k) or separated from (fig.328c,h,i) the L2-sclerotisation in the Ive-pouch. The
question arises whether these processes are homologous with the via-process (fig.244) or
with the psa-process (fig.245) of Nahublattella, and whether this homology relation is the
same in all species.

Some similarities strongly suggest that via of Nyctibora is homologous with via of
Nahublattella: The sclerotisation at the right base of via articulates (A10 in fig.328b,h)
with the left posterior end of the L2-sclerite occupying the Ive-apodeme. The basal
sclerotisation of via forms a complete cylinder. A stout muscle runs from the Ive-apodeme

21

to the left base of via (110 in fig.250, 328b,h). Thus, the bipartition of L2 (by A10: L2D
and L2E) is assumed to be homologous in Nyctibora and Nahublattella, and for via of
Nyctibora the same composition of L2E and LAN is assumed as for via of Nahublattella
(further details on L4N in 6.3.4.). In some features Nyctibora is more derived than
Nahublattella: (1) The right posterior branch of L2D (fig.244, 245; compare fig.328b and
h) is reduced to a vestige, and the process psa is completely missing. (In Nyctibora the
vestige can be identified by the insertion of 14, which is much closer to articulation A10
than in Nahublattella: fig.328b,h). (2) The posterior insertion of 110 is upon a long
cuticular tendon (tve in fig.328h). (3) The via-process is no longer forked, and paa and
pda (and vsa?) must be fused or partly reduced.

Nauphoeta (fig.3281) strongly resembles Nyctibora: The via-process, the articulation A10,
the insertions of I4 and 110, the tve-tendon, and the phallomere-gland are arranged in the
same way (compare fig.328h and 1). However, the basal sclerotisation of via is no longer
a complete cylinder and does not reach the base of the tve-tendon.

Supella (fig.328c) is similar to Nauphoeta, but some features are different: The phallomere-
gland and the tve-tendon are missing. The sclerotisation of via has expanded anteriad
along the right margin of L2D’, and articulation A10 is therefore long and hinge-like. The
right insertion of 14 is in the usual position but has shifted from L2D’ to the adjacent
membrane. The posterior insertion of 110 has shifted to the right; its position can be
explained by a clockwise (as seen from behind) rotation of the via-process along its
longitudinal axis (similar to Parcoblatta, see below). The anterior 110-insertion on L2D’
is by far more posteriorly than in the other Blattellidae and Blaberidae having a 110;
however, in Anaplecta, fig.222, and Nahublattella, fig.250, the 110-insertion also extends
far posteriad.

Alternatively, one could assume that in Supella the process is not via but psa (compare
fig.328b and c) and that via is missing. However, no muscle in any of the Dictyopteran
species studied here would then have the same course as 110 of Supella (from the anterior
part of L2 to its right part), and articulation A10 of Supella would likewise have no
homologue at least in Blattellidae and Blaberidae. Therefore, and since the respective area
is quite similar in Supella and Nauphoeta, the process is more likely to be via.

Euphyllodromia (fig.328d), Parcoblatta (fig.328e), and Blaberus (fig.328k) have, in
contrast to Nahublattella, Nyctibora, Nauphoeta, and Supella, the sclerotisation of via
firmly connected with the L2-sclerotisation of the Ive-pouch (like in Anaplecta and in the
ground-plan), and muscle 110 is missing (however, 142 of Blaberus, fig.304, might possibly
be a 110 with its posterior insertion shifted far to the left). The basal sclerotisation of via
is a complete cylinder (fig.328d, 273, 274, 300, 302). Muscle 14 is present (fig.328d,e,k).

In Parcoblatta the via-process and the surrounding area have undergone a rotation
(clockwise as seen from behind; lower curved arrow in fig.328e). This can be recognised
by the following features: (1) The contact between the lumina of the via-process and of
the rest of the left complex (fig.328e, 273, 274) is dorsal to the connection of the
sclerotisations of via and lve, not to the left of this connection as e.g. in Nyctibora
(fig.328h) and Nahublattella (fig.328b). (2) The right part of the Ive-pouch, including the
distal part of the ejaculatory duct, has partly wrapped around the L2-sclerite (again,

212

clockwise as seen from behind; upper curved arrow in fig.328e). (3) Posteriorly the
sclerotisation on the left edge of the lve-pouch bends more and more into the dorsal Ive-
wall (compare fig.271 and 272). (4) The genital opening has been rotated in the same way
and is now in the right ventral wall of the left complex (in between the lobes 47, 48, 49
in fig.266, 271).

In contrast, the area of via has been rotated counterclockwise in Blaberus (as seen from
behind; curved arrows in fig.328k): The contact between the lumina of the via-process
and of the rest of the left complex is situated ventral to the connection of the sclerotisations
of via and lve (fig.328k, 300, 302). Posteriorly the sclerotisation on the left edge of the
Ive-pouch bends more and more into the ventral Ive-wall (compare fig.299 and 300). In
contrast to the other species (fig.328b,c,d,h,1), the genital opening is not exactly on the
right side of the Ive-pouch but more in its dorsal wall.

The tve-tendon is missing in Euphyllodromia and Blaberus. In Parcoblatta the
invagination anteriorly on the vge-groove (vge, tve in fig.273) has exactly the same
position as the tve-tendon in Nyctibora (fig.328e,h): At the base of via, opposite to where
the sclerotisations of via and Ive are connected. In Parcoblatta the right insertion of 14
has shifted to tve, and this might be the reason for the retention of tve despite the loss
of muscle 110.

The via of Nyctibora and Nauphoeta are clearly homologous with via of Nahublattella;
the via-morphology of Supella and the remaining species can be derived from that of
Nyctibora and Nauphoeta. Therefore, for all species shown in fig.328 it is assumed that
the processes designated via are homologous. The presence of two sclerites L2D (in Ive)
and L2E+L4N (on via) is probably plesiomorphic. (Exact argumentation in 7.5.; the
interpretation results from the situation in Nahublattella). In the species having these two
sclerites fused, the resulting sclerite would have to be named correctly L2D+(L2E+L4N).
I will simply designate it L2.

In most of the species shown in fig.328b-k, L2 or L2D occupy the entire left edge of the
Ive-pouch and the adjacent margins of the dorsal and ventral Ive-walls (cross-section like
in fig.270 or 301). This groove shape of L2 or L2D extends posteriad as far as to the
base of via (articulation A10, if present). This is the case in Nahublattella, where, however,
A10 is far anteriorly (fig.328b), and close to A10 there is a kink to the left (edge 7 at 36
in fig.242). Between A10 and the kink, L2D’ bears the extension 28 (fig.245) into the
ventral Ive-wall, which has a homologue in Anaplecta (28 in fig.216). In Supella, Parco-
blatta, Nyctibora, Nauphoeta, and Blaberus (fig.328c,e,h,1,k) L2 is also groove-shaped,
but A10 or the via-base are by far more posteriorly, and there is no kink (except for a
hint of one in Nyctibora) and no extension 28.

In Euphyllodromia (fig.328d) and Loboptera (fig.328f) the sclerotisation of the Ive-pouch
is — except for the anteriormost part — restricted to the dorsal wall (and not groove-shaped),
and the membranous left edge of the Ive-pouch is extensively invaginated. For a correct
interpretation of these invaginations (origin, homology in Euphyllodromia and Loboptera?)
further investigations are necessary. Only Loboptera has a bulge (X in fig.328f; thickened
cuticle?) in the ventral wall of this invagination, which bears a sclerotised whip-like
process (Y in fig.328f). (Since there is no via-process at the posterior end of L2, these

213

X- and Y-structures could possibly be via, which then would be in a rather primitive
position; compare Nahublattella, fig.328b).

Ventral lobe vla Parcoblatta and Blaberus have retained a distinct vla-lobe. Like in
Nahublattella, however, the L4-plate in the ventral vla-wall has been lost (compare e.g.
L4G in Anaplecta, fig.205).

The vla-lobe of Blaberus (fig.297a, 299: beneath the via-process) can be identified as the
true vla by some of its features in common with Nahublattella and other species: The
dorsal vla-wall continues anteriad into the ventral Ive-wall (fig.12, 181, 246, 300). The
ventral vla-wall is part of the ventral wall of the left complex (fig.6, 174, 239a, 297a).
The dorsal and ventral vla-walls are connected by a very stout muscle (l6b in fig.18, 188,
189, 251, 252, 305, 306). In contrast to all other species, however, Blaberus has the dorsal
insertion of 16b on the L2-sclerotisation. This is a consequence of the rotation of via and
of the posterior parts of L2 described above, by which extensive parts of L2 must have
shifted into the insertion area of 16b.

Parcoblatta has a similar vla-lobe (fig.268-270), which, however, lies to the left and dorsal
to the via-process. This location corresponds to the rotation of via and of the genital
opening described above, in which vla has been involved, too. Muscle l6b (fig.278, 279)
is in the same position as in Blaberus; its dorsal insertion, however, is on the membranous
parts of the lve-pouch wrapped around the L2-sclerite; this situation is, again, an effect
of the rotation.

In the other species (fig.328c,d,f,g,h,i; l6b not investigated) there is no distinct lobe vla.
By comparing the relative positions of the phallomere elements adjacent to the vla-lobe
in Blaberus and Parcoblatta, however, a membranous area that is assumed to be the last
vestige of vla (fig.328c,d,h,i) can be determined. Only in Loboptera (fig.328f) and
Ectobius (fig.328g) this is not possible because of extensive reductions in this part of the
left complex.

A sclerotisation L10 on vla is only present in Blaberus (fig.299) and some other Blaberidae
(e.g. Blaptica, fig.291). L10 is discussed in 6.3.4..

The muscles of this area have been investigated only in Parcoblatta and Blaberus. Muscle
15 of Anaplecta (fig.223) and Nahublattella (fig.251) is probably missing in Parcoblatta
and Blaberus or might possibly be incorporated into muscle 1l6b. Nahublattella (fig.251,
252), Parcoblatta (fig.277-279), and Blaberus (fig.305, 307) have some muscles in the
ventral wall of the left complex (130, 131, 132, 137, 138, 140, 144, 145), most of which are
rather diffuse. These can be homologised only in part, and the most probable homologies
are expressed by the designations.

6.3. Left complex III: Main sclerites L4 and L10 and associated elements

6.3.1. Comparison between Blattaria and Mantodea

The homology relations between Blattaria and Mantodea and the common ground-plan
can be best deduced from a comparison between Mantoida, Archiblatta, Eurycotis, and

214

Tryonicus. Features of Periplaneta will also be discussed (no figures). In Mantoida L4 is
one large sclerite, in the Blattarian species L4 is a group of sclerites.

L4 of Archiblatta (fig.53-57) and L4 of Periplaneta are very similar: 5 sclerites L4C,
L4D, L4E, L4F, and L4G in the same arrangement. Eurycotis (fig.65-69) has three
sclerites L4H, L4F, and L4G.

That both L4F and L4G are homologous in the three species is evident from the identical
positions and similar outlines of the sclerites and from the fact that in Eurycotis as well
as in Periplaneta the muscles 15 (fig.72) and l6b (fig.70, 71) insert on them. A special
feature of Eurycotis is the mla-lobe (fig.63, 68, 69).

L4H of Eurycotis (fig.65-68, 325e) is composed of three parts which are homologous with
L4C, L4D, and L4E of Archiblatta (fig.53, 57, 325f) and Periplaneta: (1) The left,
crescent-shaped part of L4H corresponds to L4C. It lies in the left edge and in the
anteriormost ventral wall, an apodeme swe runs along it, and its posteriormost part
occupies a process pda. In Eurycotis and Periplaneta the muscles 12 and 14 (fig.70, 71)
insert on the posterior part of swe and run to the pouches pne and Ive. (2) The left part
of the L4H-plate in the anterior ventral wall is homologous with L4D. It takes a position
left-posterior to the right-anterior end of swe and bears a node-like process nla. Eurycotis
and Periplaneta have a stout muscle from this sclerotisation to the hla-hook (114c in
fig.72). (3) The right part of the L4H-plate in the anterior ventral wall is homologous with
L4E. It takes a position in between the right-anterior end of swe and L4F. Eurycotis and
Periplaneta have a muscle from this sclerotisation to the anterior part of the lve-pouch
(16a in fig.73).

The definition of the regions of main sclerite L4 is mainly based on the condition of L4
— as several isolated sclerites — in Archiblatta (compare fig.53-57 and fig.325f). The choice
of Archiblatta as the type of reference is made for practical reasons and has nothing to
do with an assumption of a primitive state. The positions of muscle insertions on the
various L4-regions (not studied in Archiblatta) are taken from Periplaneta and Eurycotis
— in accordance with the homology relations to Archiblatta discussed above. For
Archiblatta and Eurycotis the regioning is shown in fig.325e,f.

— L4l (lateral): The sclerotisation homologous with sclerite L4C of Archiblatta, minus its
dorsal extension to the right (L4d, see below). L4l bears the swe-apodeme and
sclerotises the pda-process posteriorly. On L4l there are the left insertions of the
muscles 12 and 14, which run to the pouches pne and Ive.

— L4d (dorsal): The sclerotisation homologous with the dorsal, rightward directed
extension of sclerite L4C of Archiblatta.

— LA4n (node): The sclerotisation homologous with sclerite L4D of Archiblatta. L4n bears
the evagination nla. On L4n there is the anterior insertion of muscle 114, which runs
to the hla-hook.

— L4e (central): The sclerotisation homologous with the sclerites L4E and L4F of
Archiblatta.

— L4v (ventral): The sclerotisation homologous with sclerite L4G in the ventral wall of
the vla-lobe of Archiblatta. On L4v there is the ventral insertion of muscle l6b, which
runs to the Ive-pouch.

215

Eurycotis has no distinct L4d-region; the demarcation of L4d in fig.325e is tentative. In
Archiblatta (fig.325f) and Periplaneta L4d is very distinct.

Three other L4-regions are not present in Archiblatta but are apomorphic sclerotisations

of certain subgroups:

— L4a (anterior), L4x: Sclerotisations of Lamproblatta and Polyphaga (and Ergaula)
which have developed by an expansion of the sclerites in the ventral wall of the vla-
lobe (definition in 6.3.4.).

— L4b (between): A new sclerotisation in the ventral wall of the left complex of
Chaeteessa, Sphodromantis, and Metallyticus (definition in 6.3.3.).

The left part of L4 of Mantoida (fig.44, 45) and the muscles inserting on it are very similar
to the regions L4l and L4d of Archiblatta, Periplaneta, and Eurycotis; homology is
assumed for the following similarities and elements (fig.325d,e,f):

1. Both the left part of L4 (Mantoida) and L4C (Archiblatta) occupy the whole left edge
of the left complex and the anteriormost ventral wall (fig.44, 45, 53, 54).

2. An apodeme swe extends along most of this sclerotisation (fig.44, 45, 53, 54);
anteriorly swe is massive and beam-like, posteriorly it is groove-like.

3. The posteriormost part of both L4 and LAC occupies a process (pda in fig.44, 53).
However, only in Mantoida the sclerotisation of pda is connected with the L2-
sclerotisation of the paa-process (compare in 6.2.1.).

4. Both L4 and L4C have a distinct dorsal extension to the right (L4d in fig.44, 53).

5. Muscle 12 inserts on the swe-apodeme in the posterior half of L4 (Mantoida), or L4C
(Periplaneta), or L4H (Eurycotis) (fig.49, 70). 12 runs to the pne-pouch and inserts
on L1 (Mantoida and Periplaneta) or in the membrane to the left of L1 (Eurycotis).

6. Muscle 14 inserts on swe ventral to 12 (fig.50, 71) and runs to the Ive-pouch.

- 7. Muscle sl (fig.48, 70), which comes from the left apophysis of the subgenital plate

(fig.37, 59), inserts on that part of L4, L4C, or L4H in the anteriormost ventral wall.

The right part of L4 of Mantoida has some features in common with the regions L4v and

L4c of Archiblatta, Periplaneta, and Eurycotis (fig.325d,e,f):

8. The posterior right part of L4 (fig.41, 47) has the same position in the ventral wall
of the vla-lobe as the L4G-sclerites (= L4v-region) have in Eurycotis (fig.63, 66) and
Archiblatta (fig.54).

9. The anterior right part of L4 is, like the anterior L4c-region in Archiblatta (L4E in
fig.57) and Eurycotis, situated between the right anterior end of the L4l-region and
the anterior end of the L4v-region.

10. A muscle running to the ventral wall of the ejaculatory duct inserts on or near the
right part of L4 or on LAG, respectively: the posteroventral part of 16 in Mantoida
(fig.52), l6b in Eurycotis (fig.70, 71) (compare in 6.2.1.).

These features 1.-10. suggest the homology relations shown in fig.325d,e,f. In Mantoida,
the regions L4l, L4d, and L4v can be unambiguously identified, and the similarities 1.-
8. and 10. can be regarded as features of the common ground-plan of Blattaria and
Mantodea. L4c might be contained in the anterior right part of L4 (feature 9.), but this is
not certain since in Mantoida the extension of the L4v-region to the anterior cannot be
determined and the L4c-region is only indicated by its relative position in between L4v

216

and L4l (? in fig.325d). Hence, 9. is an uncertain ground-plan feature. In Mantoida there
is no indication for the presence of a L4n-region, and a nla-process is missing.

Evidence from Chaeteessa (complete discussion in 6.3.3.) suggests that the L4n-region is

also an element of the common ground-plan of Blattaria and Mantodea and that the lack

of L4n in Mantoida 1s derived:

11. The heavier sclerotised transverse bridge in the anterior ventral wall of the left
complex of Chaeteessa might, according to its very similar position, well be
homologous with the L4n-region of Eurycotis (compare fig.31 and 65, 69; fig.325c,e).

Tryonicus has some features in common with Archiblatta and Mantoida which suggest the
homology relations shown in fig.325d,f,g. The two sclerites L4K and LAN (fig.85, 97)
together form a broad ribbon in the left wall of the left complex, which takes, like L4C
in Archiblatta (fig.54), a position left-dorsal to the base of the hla-hook. Most of L4K
and LAN is therefore assumed to represent the L4l-region. The nla-process (fig.97) on
L4K corresponds to the nla of Archiblatta (fig.56) and Eurycotis (fig.68) in its shape, in
its location in the anterior ventral wall, and in its position relative to the other L4-
sclerotisations and to the hla-base. Thus, the part of L4K on nla is regarded as the L4n-
region. That part of L4K which anterior to nla extends to the right (fig.95) has the same
relative position as the right-anterior part of the L4C-crescent of Archiblatta (fig.53, 55)
and is hence assumed to belong to the L4l-region. The ribbon-like extension of LAN (L4d
in fig.88-95) corresponds in its shape and relative position with the L4d-region of
Mantoida (fig.45) and especially Archiblatta (fig.53). The posteriormost part of L4N
occupies a process (pda in fig.91) like in Mantoida and Archiblatta (pda in fig.44, 53)
and can be regarded as part of the L4l-region. Like in Mantoida but in contrast to

Fig.329: Left complex, subdivisions of regions L4l and L4d into individual sclerites in Blattaria and
Mantodea. — Region L4l (discussion in 6.3.) is in its primitive condition undivided and connected
with region L2d posteriorly (between processes paa and pda). In many species L4l is connected
with region L4n anteriorly (ground-plan condition unclear). The various subdivisions of L4l in the
subgroups of Blattaria and Mantodea and the hypothetic directions of evolutionary transformation are
shown. The various types — or further derivations of them — are present in the species listed. It is
intended to emphasise principal similarities and differences in the subdivision of L4l. For
comparability, all elements are left in their most primitive condition (like in fig.329a) — except for
the subdivisions of L4l and of the included parts of L2. Inner views from the right side; dorsal—,
ventral, anterior?, posteriorJ.

The following structures are shown:

— The sclerite regions L4l (white) and L4d (light pattern) and parts of L2 (dark pattern; mainly
region L2d on process paa). The right-anterior undulate line (1 in fig.329a) represents the removal
of region L4n. The right-posterior undulate line (2 in fig.329a) represents the removal of the
remainder of L2. L4d is always shown in its most primitive position, orientation, and shape, even
if these have changed or if L4d has been lost.

— The dividing lines which cause a division of these sclerotisations into individual sclerites. Dividing
lines along which the respective sclerites are still in close contact are labelled A+Number
(articulation, e.g. A5; like in the text); if the respective sclerites are farther away from each other
the name of the respective articulation is put in brackets (e.g. (A5)).

217

f} Nahublattella 3
(Parcoblatta) h) Cryptocercus

g) Lamproblatta

c) Metallyticus
Sphodromantis (Blaberus)

b) Chaeteessa

3 2 9 a) Ground-plan

Mantoida

e) Tryonicus
Anaplecta
Polyphaga
Ergaula

d) Archiblatta
Eurycotis

— The individual L4-sclerites produced by the division of L4l and L4d. These are labelled with the
capital letters used in the text and in fig.1-319 (e.g. B = L4B).

— The apodeme swe. swe is always shown in its most primitive condition and position (like in
fig.329a), but only the parts of swe retained in the respective species are patterned and labelled.

— The insertion areas of the muscles 12 and 14, which are also always shown in their most primitive
condition and position (like in fig.329a).

— The processes pda (with its L4l-sclerotisation) and paa (with its L2d-sclerotisation).

The various types can be derived from each other in the way indicated by the arrows. Mantoida

(fig.329a) conforms completely with the ground-plan. Parcoblatta and Blaberus can be derived from
Nahublattella but differ in some respects (e.g. secondary fusion at A10).

218

Archiblatta, paa is bulge-shaped and its sclerotisation is connected with that of the paa-
process (fig.96). Sclerite L4G (fig.92, 325g) resembles L4G of Archiblatta (fig.325f) in
its outline and its position in the ventral wall of the vla-lobe and probably represents the
L4v-region; however, it cannot be excluded that, additionally, sclerotisations of the L4c-
region are contained in this L4G (in its left and anterior parts).

As compared with Archiblatta, Eurycotis, and Mantoida, some features of Tryonicus can
be regarded as derived: The L4l-sclerotisation of Tryonicus is (1) broader and (2) divided
into two sclerites (L4K, L4N; fig.325d,f,g) by the articulation A5 (fig.88, 97; compare
fig.329a,d,e). (3) The swe-apodeme is completely missing. (4) L4d is directed more
anteriad.

On the other hand, some features of Tryonicus can contribute to the common ground-plan

of Blattaria and Mantodea:

12. As already stated in 6.2.1., feature 4., the connection of the sclerotisations of pda
(L4l) and paa (L2d) present in Tryonicus and Mantoida is a ground-plan state. The
separation of these sclerotisations in Archiblatta and Eurycotis is derived (compare
feature 3.).

13. In Tryonicus and in Mantoida no sclerotisations can be unambiguously assigned to
the L4c-region, and a sclerite corresponding to L4F of Archiblatta and Eurycotis is
definitely missing. Thus, L4F, and possibly the whole L4c-region, can be regarded
as a derived element of Archiblatta and Eurycotis.

Concerning the common ground-plan of Blattaria and Mantodea, four questions remain
open: (1) It cannot be decided if there is a L4c-region. (No sclerotisations undoubtedly
homologous with the sclerotisations defined as L4c in Archiblatta, Eurycotis, and Peri-
planeta have been identified in any other species). (2) It cannot be determined whether
the L4l-region is connected with or separated from the L4v-region (or L4c-region, if
present) in the anterior ventral wall of the left complex (? in fig.32le), since there is
always a connection in Mantodea but never in Blattaria. (3) It cannot be decided if a nla-
process is present (but compare in 7.5. (M), (N)). (4) In Tryonicus, Eurycotis, and, if the
assumption in 11. is true, in Chaeteessa the regions L4l and L4n are firmly connected.
Hence, the separation of L4l and L4n in Archiblatta (sclerites L4C and L4D) might be
regarded as apomorphic. However, the position of the connection between L4n and L4l
is rather different in Tryonicus (to the left of nla, fig.96, 97) and Eurycotis (to the right
of nla, fig.67, 68), and these connections might be non-homologous. Thus, it seems better
to regard the respective ground-plan state of L4n (connected with L4l or not) as
unresolved.

6.3.2. The elements in the common ground-plan of Blattaria and Mantodea

The features 1.-13. in 6.3.1. permit the reconstruction of many ground-plan features of
L4, pda, vla, and some adjacent elements: (fig.321e,g,i): L4 is composed of sclerotisations
in the left edge and in the anterior and right ventral wall of the left complex. The L4l-
region is located in the left edge and in the anteriormost ventral wall. The swe-apodeme
runs along most of L4l. swe is massive and beam-like anteriorly and groove-like
posteriorly. There is a distinct dorsal extension L4d directed to the right and possibly

219

slightly anteriad. The posteriormost part of L4l completely sclerotises a bulge-like process
pda. The sclerotisation of pda is connected with the sclerotisation of paa (region L2d).
The right posterior part of L4 (region L4v) lies in the ventral vla-wall. The presence of
the L4c-region is questionable, but a sclerite L4F is certainly missing. The L4n-region is
present. The presence of the nla-process is unclear. The muscles 12, 14, and 16b are present.
12 and 14 have their L4-insertions close together on the swe-apodeme in the left edge of
the left complex.

6.3.3. Homology relations and character states of the elements in Mantodea

In Chaeteessa (fig.28), Metallyticus (fig.20), and Sphodromantis (fig.6), the ventral wall
of the left complex is completely sclerotised, not only along its margins as in Mantoida
(fig.41). In Chaeteessa, however, the marginal ventral sclerotisation is distinctly heavier
and is assumed to correspond to L4 of Mantoida (fig.325a- d) — composed of the ground-
plan regions L4l and L4v (and possibly L4c). The anterior transverse bridge of heavier
sclerotisation present in Chaeteessa is probably also a ground-plan element (L4n-region).
The weaker sclerotisation of the remaining ventral wall is new and is defined as a further
region of L4 (fig.325a-c):

— L4b (between): The sclerotisation of the ventral wall of the left complex between the

ground-plan regions L4, L4n, and L4v (and possibly L4e).

In Metallyticus and Sphodromantis this L4b-sclerotisation is further derived in being as
heavy as the ground-plan regions of L4. The presence of L4n is not assessable for this
uniformity of the ventral sclerotisation (in fig.325a,b the interpretation is done in
accordance with Chaeteessa). The presence of a region L4c is in Chaeteessa, Metallyticus,
and Sphodromantis as uncertain as in Mantoida (? in fig.325a-d).

As compared with Mantoida (fig.44) or Archiblatta (fig.53), in Chaeteessa (fig.31),
Metallyticus (fig.21), and Sphodromantis (fig.9) the L4-sclerotisation in the dorsal wall of
the left complex has expanded to the right: In Chaeteessa and Sphodromantis L4 occupies
most of the dorsal wall, in Metallyticus it is restricted to the anterior part. By this expansion
L4 now covers the external opening of the pne-pouch from dorsally. Possibly in correlation
with this expansion the pne-pouch has rotated to the right (clockwise as seen from behind;
compare in 6.1.3.). These shifts are very obvious in Chaeteessa and Sphodromantis but
less distinct in Metallyticus.

The muscle insertions on the L4l-region of Mantoida (fig.48-52, 325d) and on the dorsal
part of L4 (L4B) of Sphodromantis (fig.15-17, 325a) also demonstrate these shifts: The
muscles Il (to L1 anteriorly in the pne-pouch), 12 (to L1 more posteriorly in the pne-
pouch), 14 (to L2), and 17 (to the left posterior ventral wall of the left complex) are certainly
homologous in the two species, but in Sphodromantis all insertions on L4 have shifted
far to the right. These insertions also show that the extensive dorsal L4-sclerotisations of
Sphodromantis (L4B) have not been produced by an expansion of the L4d-region
(Mantoida: fig.44) but of the L4l-region (fig.325a): 12, 14, and 17 of Mantoida are not
inserted on L4d but on L4l. At the most a small right-anterior part of L4B of
Sphodromantis (posterior to the Il-insertion, fig.17) might be regarded as representing L4d
(fig.325a). Whether this distribution of L4l and L4d in the dorsal wall is also true of

220

Chaeteessa and Metallyticus is unclear (no data for the musculature); in fig.325b,c L4l
and L4d are demarcated in accordance with the situation in Sphodromantis. In any case,
in Chaeteessa, Metallyticus, and Sphodromantis the L4d-region is no longer distinct from
the L4l-region.

The pda-process of Metallyticus (fig.20, 23-26) is in its shape and in its position relative
to the paa-process similar to pda of Mantoida (fig.44-46) and is likewise sclerotised by
L4. Homology is assumed for the pda of the two species. However, in Metallyticus the
sclerotisations of pda and paa are separated, and the processes themselves are more
distinct from each other and by far longer. These two features also apply to Sphodromantis
and Mantis: In Mantis (no figure) pda is shovel-shaped and far on the left side as in
Metallyticus. In Sphodromantis (fig.9-12), certainly a close relative of Mantis, pda is long
and slender and has shifted to the right. Thus, despite the different morphology of pda in
Mantoida and Sphodromantis, these evolutionary stages suggest homology. In Chaeteessa
the pda-process has been completely lost. (The one posterior process of Chaeteessa, fig.28,
has proved to be paa; compare in 6.2.3.).

Only in Sphodromantis and Metallyticus (and Mantis, which will not be further considered)
the dorsal and ventral parts of L4 have become separated by an articulation (Al in fig.6,
10, 20, 24; sclerites L4A, L4B). The dividing line runs within the L4l-region (fig.325a,b).
This is evident from the positions of the involved sclerotisations (compare fig.325a,b and
c,d) and from the muscle insertions: In Sphodromantis (fig.325a) sl inserts on L4A, but
12, 14, and 17 insert on L4B, and all these insertions belong to L4l (compare Mantoida,
fig.325d). This division of L4l reminds one of the L4l-division in Tryonicus (by A5 in
fig.88, 97, 325g). However, the courses of the dividing lines are different: The pda-
sclerotisation, for example, is part of the posterodorsal plate (L4N) in Tryonicus but part
of the ventral plate (L4A) in Metallyticus and Sphodromantis (compare fig.329c and e).
Thus, the articulations Al and A5 are certainly not homologous, and the division of L4l
is a case of parallel evolution.

The swe-apodeme is well-developed in Mantoida and Archiblatta (fig.45, 53). In
Chaeteessa swe has been completely lost. Metallyticus has retained a vestige of Swe on
the left margin of the ventral L4A (fig.24). Sphodromantis has a vestige on the left margin
of the dorsal L4B (fig.10,11). This suggests that swe has been cut through by the division
into L4A and L4B and confirms that the L4l-region participates in both L4A and L4B.

6.3.4. Homology relations and character states of the elements in Blattaria
Archiblatta, Eurycotis, and Tryonicus

These species have been sufficiently discussed in 6.3.1.

Cryptocercus, Lamproblatta, Polyphaga, Ergaula, and Anaplecta

In Tryonicus (fig.325g) the LAl-region is divided by articulation A5: The anterior parts of
L4l form, together with L4n, the L4K-sclerite; the posterior parts of L4l (with the pda-
sclerotisation) form, together with L4d, the L4N-sclerite. The connection of the
sclerotisations of pda (L4l) and paa (L2d, fig.324g) is retained. The swe-apodeme has

221

been lost. The nla-process is well-developed. The L4v-region is a plate in the ventral wall
of the vla-lobe, the L4G-sclerite.

From this situation the morphology of the five species in the heading and the remaining
Blattellidae and Blaberidae can be derived. In all these species, however, L4K and L4N
are no longer articulated but far away from each other. In some species L4K or L4N
undergo further divisions. In Cryptocercus, Polyphaga, Ergaula, and Lamproblatta the
anterior sclerite L4K has been strongly reduced (fig.150, 124, 177). The sclerotisations
of paa and pda remain in most species connected, and this connection often becomes
very close. The L4v-region may retain its shape and position, or it becomes enlarged
(Lamproblatta, Polyphaga, Ergaula) or lost (Nahublattella, Parcoblatta, Blaberus).

As a first point, the evolution of the L4N-sclerite and the processes pda and paa of the
species in the heading will be discussed; then the L4K-sclerite with the nla-process and,
at last, the sclerite in the ventral vla-wall will be considered. paa, pda, and vla have in
part already been discussed in 6.2.4, but a discussion of these elements is only complete
by considering them in context with the other L4-sclerotisations. This will be done in this
section.

L4N-sclerite, processes pda and paa The homology of L4N of Tryonicus
(fig.94-97), Polyphaga (fig.117, 118), Ergaula, Cryptocercus (fig.150, 151), and Anaplecta
(fig.210-215) — and of its derivatives L4S and LAT in Lamproblatta (fig.177-180) — can
best be deduced from a comparison of the prominent substructures. In Tryonicus these
are: (1) The pda-sclerotisation (posteriormost L4l-region), including its close vicinity to
and connection with the paa-sclerotisation (L2d-region). (2) The dorsal extension to the
anterior (L4d-region).

Further evidence comes from the musculature, which, however, has not been studied in
Tryonicus. The homology relations and the resulting regioning of these sclerotisations into
L4 and L4d are shown in fig.325g,h,i,k,1.

The pda and paa of Tryonicus, Lamproblatta, Polyphaga, Ergaula, and Anaplecta have

some features in common, none of which, however, is realised in all these species. But

the whole of the similarities is sufficient to regard the pda and paa of all species as
homologous.

— pda and paa are two processes with their sclerotisations firmly connected. (The
sclerotisation of pda is designated as L4N or L4T, that of paa as L2 or L2C). Ergaula,
however, has lost the right process paa.

— The position of pda and paa on the left complex is dorsal, far posterior, and far to the
left — near the left end of the Ilve-pouch. In Lamproblatta, however, their position relative
to lve is more to the right (fig.179, 180).

— The common sclerotisation of paa and pda is, to the right (Anaplecta, fig.211) or to
the ventral side (Polyphaga, fig.118, 122; Ergaula; Tryonicus, fig.97), firmly connected
with the left end of the L2-sclerotisation in the Ive-pouch. Lamproblatta, however, has
lost this connection (fig.178, 179).

— On the common sclerotisation of paa and pda there inserts a muscle coming from the
left part of the Ive-pouch (110 in fig.129, 186, 222; compare in 6.2.4.; not analysed in
Tryonicus).

DL

— In Polyphaga and Lamproblatta the left process pda is long and pointed, the right
process paa is somewhat saucer- or cup-shaped.
— In Polyphaga and Ergaula the pda-processes are nearly identical.

Polyphaga, Ergaula, and Lamproblatta (but not Anaplecta) have sclerotisations probably

homologous with the L4d-region of Tryonicus:

— In Tryonicus the one end of LAN (to the right of pda) is connected with L2 occupying
paa and the Ive-pouch, and its other end has the extension L4d (fig.96), which is
directed anteriad.

— In Polyphaga and Ergaula LAN is also connected with L2, and its opposite end has an
extension, which, however, is directed to the left (L4d in fig.118).

— In Lamproblatta the sclerotisation L4S (fig.178, 180) is connected with L2 at the left
end of the Ive-pouch. The distal part of L4S resembles L4d of Polyphaga — with the
difference that it is not connected with the sclerotisation of pda and paa (compare
fig.3251 and k). I assume that in Lamproblatta a dividing line has formed which has
separated the following sclerotisations from each other (fig.329e,g): Posterior to the line
is the common sclerotisation of pda and paa (composed of L4T, a part of the former
L4N-sclerite, and L2C, a part of the former L2-sclerite). Anterior to the line are (1)
the part of the L2-sclerotisation at the left posterior end of the Ive-pouch and (2) the
other part of the former L4N-sclerite (L4S) which maintains the connection with L2 at
its one end and has the extension L4d at its other end (fig.178).

— The homology of the extensions L4d in Lamproblatta, Polyphaga, and Ergaula is
confirmed by muscle 111, which inserts on or near L4d and runs to sclerite LAK (fig.128,
188, 327)8

— Nahublattella (complete discussion below) also has an extension similar to L4d of
Tryonicus or Polyphaga (L4d’ in fig.242); it extends, like L4d of Polyphaga, from the
common sclerotisation of paa and pda to the left. In Anaplecta L4d has been lost.

L4d is directed to the right in Mantoida (fig.44), right-anteriad in Archiblatta (fig.53), and
anteriad in Tryonicus (fig.94). In Polyphaga and Ergaula (and Nahublattella) LAd has
even further rotated (counterclockwise as seen from above) and is directed to the left.
From such a position, L4d of Lamproblatta has additionally rotated 90° (clockwise as
seen from behind) and shows a dorsoventral orientation (fig.178).

LAN of Cryptocercus (fig.150) is assumed to be homologous with the L4N of the other
species and to have the same orientation as in Polyphaga and Ergaula (compare fig.117
and 150, 325h and k): Its left part is L4d, its right part is the pda-sclerotisation (fig.325h).
This is suggested by the following features:

— LAN lies, like in the previous species, in the left dorsal wall of the left complex (fig. 150).

— The right part of L4N lies, like the right part of L4N of Polyphaga and the right-
posterior part of LAN of Tryonicus, on a process (pda in fig.150, 118, 96).

— pda is, like in Tryonicus, Mantoida, and Polyphaga, situated to the left of (and
somewhat dorsal to) the paa-process (fig.150, 96, 44, 117).

— Within the angle formed by the sclerites L4N and L2 (fig.150) there is the base of the
hla-hook. Tryonicus shows the same relations (fig.97).

— In Cryptocercus and Mantoida, the muscles I1 (fig.48, 155) and 12 (fig.49, 156) run

223

from the pne-pouch to the left and dorsad (homology discussion in 6.1.1.). The insertion
of 11 is in Cryptocercus on LAN (including L4d), in Mantoida on and to the right of
L4d.

— In Cryptocercus and Mantoida, muscle 19 (fig.49, 155) runs transversely within the
dorsal wall of the left complex (homology discussion in 6.5.). Its left insertion is in
Cryptocercus anterior to the left end of the assumed L4d-part of LAN, in Mantoida near
the right end of L4d; regarding the orientation of L4N assumed for Cryptocercus, these
ends of the sclerites would be homologous. In Cryptocercus, however, 19 does not extend
as far to the right as in Mantoida.

Some features of Cryptocercus are derived: (1) The sclerotisation of pda is largely reduced
(fig.150, 329h), and, possibly as a consequence of this, (2) the sclerotisations of pda (L4)
and paa (L2) have separated (like in Chaeteessa, Metallyticus, Sphodromantis, Archiblatta,
and Eurycotis).

L4K-sclerite, process nla LAK of Anaplecta (fig.209, 212) is probably

homologous with L4K of Tryonicus (fig.98); similarities are:

— The position in the left and left-ventral walls of the left complex.

— The anterior part is on a bulge-like evagination (nla in fig.97, 212). Like in Tryonicus,
this sclerotisation is regarded as the L4n-region (fig.325g,l).

— The posterior part is plate-like, with a broadly truncate posterior margin, and this part
partially encloses the retracted hla-hook from the left side (fig.85, 202). Like in
Tryonicus, this sclerotisation is regarded as an anterior part of the L4l-region
(fig.325g,)).

A comparison of the muscles of Anaplecta and Eurycotis confirms these assignments to

L4n and L4l:

— That in both species the anterior insertion of the main muscle of the hla-hook (114¢ or
114; fig.73, 222) is on or near the sclerotisation of the nla-bulge shows the homology
of these nla. The nla-sclerotisation of Eurycotis is, by definition, the L4n-region
(fig.325e,]).

— In Eurycotis and Mantoida the muscles 12 (fig.49, 70) and 14 (fig.50, 71) insert close
to each other on the L4l-region and run to the pouches pne and Ive, respectively.
Anaplecta also has a muscle 12 from the posterior part of L4K to the pne-’pouch”
(fig.221, compare in 6.1.4.); muscle 14 is missing. Nahublattella (complete discussion
below), however, has both 12 and 14 (to the pouches pne and Ive), and their left insertions
are on a sclerotisation homologous with the posterior part of L4K of Anaplecta (L4U’
in fig.249).

The narrow sclerotisation which extends in Tryonicus from the anteriormost part of L4K
to the right (anterior to nla, fig.96, 97, 325g) is missing in Anaplecta (fig.325l). Since
this sclerotisation probably corresponds to the anteriormost L4l-region of Archiblatta and
Mantoida (bearing the anterior part of the swe-apodeme; fig.44, 53, 325d,f), this feature
of Anaplecta is regarded as derived.

The course of the dividing line through the L4l-region of Anaplecta (separating L4K and
LAN) can be deduced from the positions of the I2- and l4-insertions in Eurycotis,

224

Mantoida, Anaplecta, and Nahublattella and from the distribution of the other
substructures: Anterior to the line (on L4K) there are the insertion of 12 (and 14 in
Nahublattella) — and hence parts of the L4l-region — the L4n-region, the nla-process, and
the I14-insertion. Posterior to the line (on L4N) are the pda-process — and hence posterior
parts of the L4l-region — and the I10-insertion (and the L4d’-extension in Nahublattella).
The course of the line is shown in fig.329e. The distribution of all cuticular elements
present is the same as in L4K and LAN of Tryonicus, and the dividing lines of Anaplecta
and Tryonicus (and Nahublattella) are strongly suggested to be homologous.

L4K of Cryptocercus (fig.150, 151) is homologous with L4K of Tryonicus and Anaplecta

and is likewise composed of the anterior L4l-region and of the L4n-region. Both L4l and

L4n, however, are strongly reduced. These relations are, firstly, suggested by similarities

in the cuticular elements of Tryonicus and Cryptocercus:

— LAK of Cryptocercus has the same position like the left-dorsal half of L4K of Tryonicus:
left-dorsal to the base of the hla-hook (compare fig.85, 97 and 145, 151). In Tryonicus
this sclerotisation has been regarded as an anterior part of the L4l-region (fig.325g,h).
Sclerite L4P of Cryptocercus (fig.151) probably corresponds to that part of L4K of
Tryonicus immediately anterior to the hla-base. The right-ventral half of L4K of
Tryonicus (fig.325g) with the anteriormost L4l-region (the anterior extension to the
right) and the L4n-region (nla-sclerotisation) has been, like the nla-process itself, lost
in Cryptocercus.

Secondly, the same relations result from a comparison of the muscle insertions of

Cryptocercus and other species:

— The 12 and 14 of Cryptocercus (fig.155, 156), Mantoida (fig.49, 50), and Eurycotis
(fig.70, 71) run from the pouches pne and Ive to the leftmost part of the left complex,
where their insertions are close to each other. Homology can be assumed. The left
insertions are on the L4l-region in Eurycotis and Mantoida (fig.325d,e), and on L4K
in Cryptocercus. The contribution of the L4l-region to L4K of Cryptocercus is thus
confirmed (fig.325h). This can be only an anterior part of L4l since the posterior part
(with the pda-sclerotisation) is included in sclerite L4N. In Anaplecta L4K also bears
the left 12-insertion (fig.221), and in Nahublattella the homologue of the posterior part
of L4K (L4U’) bears the left 12- and 14-insertions (fig.249).

— The 114 of Cryptocercus (fig.157), Eurycotis (fig.72), and Anaplecta (fig.222) run from
the anterior left wall of the left complex to a large hook (hla) and are certainly
homologous (discussion in 6.4.). Cryptocercus (fig.157) and Eurycotis (fig.70) have
phallomero-sternal muscles s1+3 or sl inserting immediately anterior to 114: the left
part of s1+3 (= sl) is probably homologous with sl of Eurycotis (sl is missing in
Anaplecta; discussion in 6.9.).

— In Eurycotis and Anaplecta 114 inserts on the L4n-region (on the nla-process, fig.72,
73, 222), and in Eurycotis sl inserts at the border between L4n and the anterior L4l
(fig.73, 325e). In Cryptocercus part of 114 inserts on L4K; this suggests that the L4n-
region also contributes to L4K. The larger part of the 114-insertion and the entire s1+3-
insertion, however, are on membrane (ventral and anterior to L4K); this suggests that
the L4n-region as well as the anteroventral part of the L4l-region (corresponding to the

225

anterior extension to the right on L4K of Tryonicus) are strongly reduced. Probably as
a consequence, the nla-process is missing.

Thus, L4K of Cryptocercus is composed of anterior parts of L4l (with the insertions of
12 and 14) and a highly reduced L4n (with part of the insertion area of 114). LAN of
Cryptocercus is made of the posterior part of the L4l-region (with the reduced pda-
sclerotisation) and of L4d. The distribution of all elements present is the same as in L4K
and L4N of Anaplecta and Tryonicus, and homology can be assumed for L4K, for L4N,
and for the dividing line between them (through LA). The reduction of the L4n-region
and the loss of the nla-process are derived features of Cryptocercus. The anteroventral
part of L4l has also been lost in Anaplecta.

L4K of Lamproblatta (fig.177, 178) resembles L4K of Cryptocercus (fig.150, 151):
— The sclerites take the same position dorsal to the base of the hla-hook.
— A process nla on or near L4K is missing.

Anterior to L4K in Lamproblatta or on and anterior to L4K in Cryptocercus there insert

some muscles having the same course, and the insertions on or near L4K show the same

positions relative to each other:

— A muscle to the subgenital plate (sl in fig.185; left part of s1+3 in fig.157).

— A muscle to sclerite L3 on the hla-hook (114 in fig.184, 157). The 114-insertion is partly
on LAK in Cryptocercus but completely on membrane in Lamproblatta.

— A muscle to the pne-pouch (12 in fig.184, 156). The 12-insertion is on L4K in
Cryptocercus but on the membrane anterior to L4K in Lamproblatta.

Therefore, L4K of Lamproblatta and Cryptocercus are assumed to be homologous and to
have the same composition: anterior part of L4l, vestiges of L4n. However, since in
Lamproblatta the insertions of 12, 114, and sl are exclusively on membrane and muscle
14 has been lost (compare fig.155), the muscles do not yield any direct evidence for the
presence of the regions L4l and L4n and for the distribution of L4l and L4n within sclerite
L4K. The distribution can only be deduced from a comparison with Cryptocercus, as it
is done in fig.325h,i. That the 114-insertion is completely on membrane could be
interpreted as a further reduction of the L4n-region as compared with Cryptocercus. That
the I2-insertion is anterior to L4K (not on LAK as in Cryptocercus) is interpreted as a
shift of this insertion to the anterior, not as a reduction of the respective L4l-sclerotisation
(comparison with Polyphaga, see below).

L4K of Polyphaga is situated not in the dorsal but in the posteroventral part of the hla-
base (fig.122-124; compare fig.151, 178). It is assumed to be homologous with L4K of
Lamproblatta and Cryptocercus and to have shifted and rotated (clockwise as seen from
the left) ventrad around the posterior part of the hla-base. This is suggested by the
following features:

— In Polyphaga and Lamproblatta L4K is broadly horseshoe-shaped and curves into the
base of the hla-hook. (According to the assumed shift and rotation in Polyphaga — the
latter is almost 180° — the orientation of the sclerite is in Lamproblatta and Polyphaga
opposite). In Cryptocercus this curvature of L4K is missing.

— In Polyphaga and Cryptocercus L4K bears the insertion of a muscle coming from the
left-anterior part of L2 (14 in fig.132, 155). 14 is missing in Lamproblatta.

226

— In Polyphaga and Lamproblatta L4K bears the insertion of a muscle coming from the
sclerotisation L4d (or from the adjacent membrane; 111 in fig.128, 184). Taking the
assumed rotation of L4K in Polyphaga into account, the insertion on L4K is in exactly
the same position. I11 is missing in Cryptocercus and all other species and is a derived
feature of Polyphaga, Ergaula, and Lamproblatta.

The insertion of muscle sl (fig.127) has retained the same position as in Lamproblatta
(fig.185) and Cryptocercus (left part of s1+3 in fig.158, 159): on the basal line anterior
to the hla-base. The hla-muscle 114, present in all other Blattaria studied (discussion in
6.4.), is missing in Polyphaga, and the hla-hook and its sclerite L3 are bare of muscles.
The function of 114 has probably been taken over by the very stout 14, which does not
insert on L3 but on the dorsal part of L4K situated within the hla-base.

The muscles 12 are certainly homologous in Polyphaga, Lamproblatta, Cryptocercus
(fig.128, 184, 156), Mantoida, Eurycotis, and Anaplecta (fig.49, 70, 221; discussion in
6.1.). The ground-plan positions of the 12-insertions are shown by the three latter species:
right insertion in the left wall of the pne-pouch; left insertion roughly in the middle of
the left edge of the left complex. In Cryptocercus, Lamproblatta, and Polyphaga, as a first
point, the right 12-insertion has shifted anteriad to the top of the pne-pouch (compare in
6.1.). As a second point, the left insertion also shows a gradual shift to the anterior and
takes a position (1) more anteriorly than in the ground-plan but still on the L4l-
sclerotisation (L4K) in Cryptocercus, (2) even more anteriorly and anterior to the L4l-
sclerotisation (L4K) in Lamproblatta, and (3) still more anteriorly, and ventrally, but again
on sclerotisation (L4M) in Polyphaga. The various stages of this 12-shift are regarded as
synapomorphies of the species concerned. The I2-insertion is assumed to have shifted away
from the L4l-region (Lamproblatta, Polyphaga) and to have later reached a position on
another sclerotisation formed by an enlargement of the ventral sclerotisation of the vla-
lobe (Polyphaga; this aspect is discussed below). Hence, contrary to the definition of L4l
in 6.3.1., the sclerotisation bearing the I2-insertion in Polyphaga is not assigned to L4l
since the fact that the shifted I2 inserts on sclerotisation is not the result of a concomitant
shift or expansion of LAl.

In Polyphaga the contribution of the L4l-region to L4K can be directly deduced from the
l4-insertion on LAK. For the presence of L4n, however, there is, like in Lamproblatta, no
direct evidence (the nla-process and muscle 114 are missing). L4K is hence assumed to
be mainly made of anterior parts of L4l, with little (like in Cryptocercus) or no contribution
from L4n.

The situation in Ergaula capucina (fig.326d, 327d) can be derived from that in Polyphaga
(fig.326c, 327c): L4K is likewise ventral to the hla-base but has shifted even further
anteriad. The dorsal part of L4K, which bends into the hla-base, is distinctly shorter
(compare edges X in fig.326c and d) and fused to the ventral anterior margin of sclerite
L3 (along edge X and more anteriorly). A muscle coming from the same part of L2 as 14
in Polyphaga, which is certainly homologous with this 14, inserts on this compound sclerite
(mainly along edge X: 14 in fig.327d). Muscle 111 has the same insertions as in Polyphaga
and Lamproblatta (fig.327b,c,d) but is much stouter. The muscles I2 and sl insert like in
Polyphaga.

22]

The morphology of L4K and 14 of Ergaula could easily be mistaken as corresponding
with the situation in Blattellidae (Anaplecta excluded) and Blaberidae: In the latter groups
the main muscle of the hla-hook (114 in fig.249, 276, 303) runs from the anteriormost
part of L2 to sclerite L3. The situation in Anaplecta (fig.222) suggests that this hla-muscle
is a true 114 whose anterior insertion has been translocated from the L4n-region (with
nla) to L2 (discussion in 6.4.3.). Looking at Ergaula only, the “hla-muscle” (14) with its
course from L2 anteriorly in the Ive-pouch to the “base of L3” (= L4K) could easily be
misinterpreted as the “114”, with the “translocation” of its anterior insertion to L2 being
a synapomorphy of Ergaula and the respective Blattellidae and Blaberidae. However, the
situations in Polyphaga and Lamproblatta clearly show that in Ergaula the muscle is 14
(not 114), the sclerite is L4K (not the basal part of L3), and the similarity with Blattellidae
is a case of convergence.

Ventral sclerite plate Sclerite L4G in the ventral wall of the vla-lobe of Eurycotis
and Archiblatta is, by definition, the region L4v (fig.325e,f). L4G of Tryonicus (fig.325g)
probably also corresponds exactly to the L4v-region, but it cannot be excluded that parts
of the L4c-region are contained in the sclerite (compare in 6.3.1.). In the latter case, L4G
of Tryonicus and L4G of Eurycotis and Archiblatta would be only partly homologous.
The ventral plate of Anaplecta (L4G in fig.205) lies similarly in the ventral wall of the
vla-lobe like L4G of Tryonicus (fig.87), and in both species L4G is, apart from the ventral
parts of sclerite L4K, the only sclerotisation in the ventral wall of the left complex. This
indicates that L4G of Tryonicus and Anaplecta are strictly homologous. The ventral plates
of Cryptocercus (LAG in fig.148), Lamproblatta (L4R in fig.174), and Polyphaga (L4M
in fig.115) also lie in the ventral vla-wall and can be assumed to be at least in part ho-
mologous with each other and with the L4G-plates of the other species. These plates,
however, are rather different in their relative sizes, and the homology relations should be
analysed in detail.

Some evidence for the exact homology relations comes from the muscles 12, 15, l6a, 16b

(homology discussion in 6.1. and 6.2.), and s3 (homology discussion in 6.9.). In Sphodro-

mantis, Mantoida, Polyphaga, and Lamproblatta s3, 12, 15, 16a, and 16b can be

homologised one by one (with the exception that in Mantoida 16a and l6b have fused).

In Cryptocercus homology is clear for s3 (right part of s1+3), 12, and l6b; 15 and 16a have

been lost. In Eurycotis and Anaplecta homology is also clear for s3, 12, and l6b; as regards

15 and 16a, homology with the 15 and 16a of the other species is questionable. The relations

between the ventral plates and the insertions of l6b, s3, and 12 are different in the various

species:

— In Eurycotis, Cryptocercus, and Anaplecta only 16b (fig.70, 157, 224) inserts, at least
in part, on the ventral plate (L4G in fig.63, 148, 205). s3 (fig.70, 157, 222) inserts in
the membrane left-anterior to L4G. The positions of these insertions may suggest (but
do not prove) that the L4G of Anaplecta and Cryptocercus are strictly homologous with
L4G of Eurycotis (i.e. only L4v but no parts of L4c or of other sclerotisations are
included; fig.325e,h,l). The same might be assumed for Tryonicus since its L4G is
similar to L4G of Anaplecta (fig.325g).

228

— Lamproblatta and Polyphaga (and Ergaula) differ from the previous species (fig.132,
133, 188, 189): (1) Not only l6b but also s3 inserts on the ventral plate (L4R in fig.174;
L4M in fig.115). This is certainly a derived feature. (2) 15 and 16a also insert on the
ventral plate, but since the homology relations with the respective muscles of the
previous species are uncertain, this feature is not interpretable (no L4c-region is included
in fig.325i,k, but its absence is questionable). (3) A special muscle s12 from the right
half of the subgenital plate runs to the ventral plate and inserts immediately to the right
of s3. The presence of $12 is also a derived feature.

— Polyphaga (and Ergaula) shows an additional derived feature already mentioned above:
Muscle 12 inserts on the ventral plate L4M (fig.128).

The derived condition that, in Lamproblatta and Polyphaga (and Ergaula), the insertions

of some muscles are now on the sclerotisation of the ventral plate (at least s3 in

Lamproblatta and s3 and 12 in Polyphaga) is interpreted as an expansion of this plate, and

the sclerotisations bearing these insertions are defined as new regions of L4:

— L4a (anterior): The sclerotisation of the insertion area of s3. (The s3-insertion has not
changed its position.)

— L4x : The sclerotisation of the insertion area of 12. (The 12-insertion has shifted ventrad.)

According to this interpretation, L4M and L4R are not strictly homologous with each
other and with the L4G of the other species: L4R evolved from L4G by expansion (new
region L4a), and L4M evolved from LAR by an additional expansion (new region L4x).
In Ergaula a small anterior part of L4M (with the insertions of s3 and $12) has split off
to form a sclerite of its own (compare fıg.3221 and m).

Nahublattella

The homologue of sclerite L4K of Anaplecta (fig.209) has divided into two sclerites L4U’
and L4V’ (fig.242). L4U’ resembles the posterior part of LAK: It has the same position
on the left edge of the left complex, the same position relative to the hla-hook, and a
similar shape (curved plate). L4V’ resembles the anterior part of L4K: It lies in the
anteriormost ventral wall of the left complex and forms a process (nla in fig.242, 248).
The homology of the nla-processes of the two species is, regarding their different shape,
debatable.

These relations are supported by the muscles: L4U’ bears the insertions of 12 and 14
(fig.249). 12 runs to the basalmost part of the hla-hook (membrane 30). 12 of Anaplecta
(fig.221) runs to the pne-pouch next to the hla-base (30 in fig.210, 211). I assume
homology for the 12 of the two species and a slight shift of the right insertion in
Nahublattella. 14 of Nahublattella inserts immediately ventral to 12 and runs to the lve-
pouch, exactly like 14 of e.g. Eurycotis (fig.70, 71) and Cryptocercus (fig.155, 156). As
mentioned above, 14 has been lost in Anaplecta. L4V’ bears the insertions of 15, 16a, and
s3 (fig.250, 251). 15 has its posterior insertion like 15 of Anaplecta (fig.223) at the left
base of the lve-apodeme, and homology is highly probable for these 15; that the anterior
insertion is on the anterior part of L4K in Anaplecta confirms the homology between this
part of L4K and L4V’ (and, maybe, the homology of the nla-processes, too). The
insertions of s3 and 16a in Anaplecta, however, are on the ate-tendon to the right of the

229

sclerotisation (fig.222). (Homology is quite certain for s3 but not for 16a; discussion in
6.9. and 6.2.4.). That the dorsal insertion of s3 is on a sclerotisation is not comparable
with the situation in Lamproblatta and Polyphaga: In the latter species the sclerotisation
concerned is an expansion of the ventral plate (L4a-region); in Nahublattella the
sclerotisation with the s3-insertion is an expansion of the former L4K-sclerite. (The
respective area of the sclerite could be defined as a new sclerite region, but this is omitted).

Thus, L4U’ is assumed to consist of the same parts of the L4l-region as the posterior part
of L4K in Anaplecta. L4V’ roughly corresponds to the L4n-region (fig.325m, compare
fig.3251); however, the line dividing the two sclerites does certainly not exactly correspond
to the border between L4l and LAn; this is only the case — by definition - in Archiblatta.

The homology relations of the processes paa and pda of Anaplecta and via, paa, pda,
and vsa of Nahublattella (fig.241, 244) and of their sclerotisations have been discussed
in 6.2.4.. L4N’ of Nahublattella is probably the left-ventral sclerotisation of the via-
process (including pda and vsa; fig.325m). The ribbon-like extension L4d’ at the left base
of via (fig.244, 250) closely resembles L4d of Polyphaga and Tryonicus (fig.94, 97, 118,
129) in its position relative to paa and pda and their sclerotisation and to the 110-insertion.
Like in Polyphaga and Cryptocercus, L4d’ is directed to the left. Like in Mantoida and
Cryptocercus, L4d’ has a muscle running to the pne-pouch (II in fig.48, 155, 249).
Homology is assumed for the L4d and I1 of all species. In Nahublattella the whole area
containing via, pda, paa, and L4d’ is sunken anteriad into the left complex and has
become the left part of an expanded Ive-pouch, and L4d’ lies in the left edge of this
enlarged lve-pouch and runs posteriad (L4d’ is, so to speak, invaginated). This is in
contrast to all other species; only Anaplecta shows a slight anteriad invagination of the
paa+pda-sclerotisation (but L4d has been lost). The separation of the paa+pda-
sclerotisation from the L2-sclerotisations in the Ive-pouch reminds of Lamproblatta
(fig.177-179), but the division of the sclerotisations is different and non-homologous: L4d
is connected with the paa+pda-sclerotisation in Nahublattella but with the sclerotisation
of the Ive-pouch in Lamproblatta (compare fig.329f and g).

The identification of the vla-lobe (fig.245-247) was done in in 6.2.4.. That there is no
sclerite plate in its ventral wall (L4v-region; compare L4G of Anaplecta, fig.224) is a
derived feature.

Parcoblatta, Blaberus, and other Blattelidae and Blaberidae

Sclerite L4U’ of Blaberus has the same shape, relative position, and muscle insertions as
L4U’ of Nahublattella (fig.242, 249, 299, 303). In both species 14 runs to sclerite L2, and
12 runs to the hla-base (30 in fig.249, 303). In Parcoblatta, the morphology of the left
part of the left complex (compare fig.268-270 and 299-301) and the arrangement of 12
and 14 (compare fig.276 and 303) are nearly the same as in Blaberus; however, sclerite
L4U has been lost. In Nyctibora LAU is present and very similar to L4U’ of Blaberus.

Blaberus and Parcoblatta both have a tendon-like invagination (ate in fig.268, 271, 302)
near the ventral basal line of the left complex. ate is also present, and in the same position,
in other Blattellidae and Blaberidae (investigated species: those listed in 5.15.; Blaptica:
fig.291). The homology of these ate-tendons is confirmed by the insertion of a phallomero-

230

sternal muscle (s3b, studied in Parcoblatta, Blaberus, and Blaptica; fig.277, 304;

homology discussion of s3b in 6.9.) and by the presence of a sclerite in the dorsal wall

of the tendon (L4V or L4V’, which, however, is present only in Parcoblatta, Nyctibora,

and Blaptica, fig.289, 291).

The evolutionary origin of tendon ate and sclerite L4V of these species is unclear. The

terms used express the possible homologies with structures being in similar positions in

Anaplecta (ate in fig.212) and Nahublattella (L4V’ in fig.244):

— ate and, if present, L4V resemble both ate of Anaplecta and L4V’ of Nahublattella in
bearing the insertion of at least part of (1) muscle s3 (s3b of Parcoblatta and Blaberus,
fig.277, 304) and (2) muscle l6a (only Blaberus, fig.304). However, it is impossible
that both homologies — of ate and L4V - are true in a strict sense since in Anaplecta
the ate-tendon and the sclerotisation homologous with L4V’ of Nahublattella (anterior
L4K) are located side by side.

— As a combined hypothesis accepting a partial homology of the ate-tendons and a strict
homology of the L4V-sclerites, it might be assumed that in the more derived Blattellidae
and Blaberidae, as compared with Anaplecta, the cuticular area forming the ate-tendon
has expanded basad and that by this process L4V has become integrated into the tendon.
Nahublattella could be an intermediate, with the anteriormost ventral part of the left
complex being a very broad ate-"tendon”, and with L4V’ integrated into this “tendon”.
In the other Blattellidae and Blaberidae this anterior part with L4V’ must then be
assumed to have become very narrow, and L4V’ has become smaller. If this is true, the
ate-tendon of Anaplecta would be homologous with the distalmost part of the ate-tendon
(anterior to L4V, if present) of e.g. Supella, Euphyllodromia, Parcoblatta, Nyctibora,
Blaptica, and Blaberus.

— However, the lack of a sclerotisation within the ate of Supella, Euphyllodromia, and
other species might suggest that L4V of Parcoblatta, Nyctibora, and Blaptica is a new
element not homologous with L4V’ of Nahublattella. If this is true, ate of Anaplecta
could be strictly homologous with ate of the more derived Blattellidae and Blaberidae.

These questions concerning ate and L4V cannot be settled here.

The sclerotisation of the via-process has been assumed, in accordance with the situation
in Nahublattella, to be composed of L4N and L2E (posterior L4l-region and L2d-region:
fig.325m,n,o and 324m,n,o; discussion in 6.2.4.). Since the primary processes paa and
pda are no longer distinguishable in these via-processes (fig.328c-k), the exact arrange-
ment of L4N and L2E is less clear than in Nahublattella. In determining the position of
the L4N- and L2E-sclerotisations on via of Parcoblatta and Blaberus one must consider
the rotation of the via-process. An extension corresponding to L4d’ of Nahublattella is
missing in all species (compare fig.328b and c-k), and the L4d-region is assumed to have
been lost like in Anaplecta (fig.325l,m,n,o and 3241,m,n,o).

At least Parcoblatta, Nyctibora, Blaptica, Nauphoeta, and Blaberus (the other species not
investigated) lack, like Nahublattella, a sclerite plate in the ventral wall of the vla-lobe
(fig.266, 268, 297): The L4v-region has been lost. Sclerite L10’ of Blaberus (fig.299) and
the small sclerites L10’ of Blaptica (fig.291) are not assumed to be descendants of L4v
but new sclerotisations having evolved within Blaberidae. In the blaberid Nauphoeta L10’
is missing.

231

6.4. Left complex IV: Main sclerite L3 and associated elements

6.4.1. Comparison between Blattaria and Mantodea

In Archiblatta (fig.53-55) the hla-hook is an evagination of the anterior left ventral wall
of the left complex, and its base is immediately beneath the arched anterior part of the
L4C-sclerite = L4l-region. Mantoida has no process in the corresponding part of the
ventral wall (fig.45, 46), and the neighboring processes paa and pda have proved to be
homologous with paa and pda of Archiblatta and other Blattaria. The elements of the left
complexes of Chaeteessa, Metallyticus, and Sphodromantis — including the processes paa
and pda — have all been homologised with elements of Mantoida. Thus, none of the
Mantodean species studied has a homologue of the hla-hook; hla, and also its sclerite L3
and its main muscle 114, are restricted to Blattaria.

6.4.2. The elements in the common ground-plan of Blattaria and Mantodea

Since hla is present in all Blattaria (discussion in 6.4.3.) but absent in all Mantodea, its
presence in the common ground-plan cannot be reliably decided. However, a comparison
of the copulation habits of Blattaria and Mantodea might indicate that the lack of hla and
L3 in Mantodea is a derived feature.

In Blattaria copulation has several successive phases (data from Scudder 1971, who refers
to Gupta 1947): In Periplaneta, in phase (1), the male places itself in front of the female,
with its rear end facing the female. Then the female climbs upon the back of the male,
both animals facing the same direction. In this phase the hla-hook of the male makes the
first contact of the genital regions: It seizes the terminal lobes of the female subgenital
plate (Scudder: “initial seizing”). (2) This connection being established, the male rotates
ca. 180° in the horizontal plane (clockwise as seen from above). (3) After this rotation
the animals are again in a line, with their rear ends still in contact. Now other phallomere
elements establish a firmer contact — mainly the seizing apparatus formed by the posterior
part of the male’s right phallomere (Scudder: “final holding”). Scudder describes a several-
phase process with similar positions for some subgroups of Ensifera. But, of course, the
connection of male and female genitalia is established by completely different structures.
Nevertheless, it seems plausible that a copulation procedure with a sequence of these
positions might be plesiomorphic for a higher taxon including at least Orthoptera and
Dictyoptera.

Mantodea have a different copulation procedure, which Scudder regards as apomorphic:
The male mounts the female (often by jumping) and then clings to the female thorax with
its grasping legs. Holding this position, the male curves its terminal abdomen to the right
and pushes it into the female genital pouch from laterally (e.g. Kumar 1973). Together
with the modified fore legs, the very special feeding habits of Mantodea (lurking predators)
are certainly derived. It might be plausible that changes in behaviour correlated with these
new feeding habits might have caused changes in the copulation procedure. (So to speak,
it is no longer advisable for the male to place itself in front of the female in the way
Blattaria do).

232

Thus, the outgroup comparison with Ensifera as well as biological properties of Mantodea
suggest that the copulation procedure of Blattaria is plesiomorphic and that of Mantodea
apomorphic: Phase (1), in which Blattaria make use of their hla-hook, can be regarded as
secondarily lost in Mantodea. Additionally, since the phallomeres of Mantodea and e.g.
Periplaneta are rather similar in their morphology (and completely different from those
of Ensifera), it might be assumed that the way the Mantodean phallomeres functioned
before the copulation procedure has changed was similar to that of the Blattarian
phallomeres, and that a hla-hook was present for initial seizing. Though these ideas are
highly speculative, it is at least plausible that hla and the associated elements L3 and 114
were present in the common ground-plan of Blattaria and Mantodea and have been lost
in Mantodea. The same might also be true of the nla-process, which is present in many
Blattaria (fig.69, 98, 212) but never in Mantodea. nla probably has the function to stiffen
the sclerotisation at and near the 114-insertion, and if hla and 114 are lost an additional
loss of nla could be expected.

6.4.3. Homology relations and character states of the elements in Blattaria

The hla-hook is present in all Blattaria. The homology of all these hla is suggested by
their position in the leftmost part of the left complex, by their similar shape, and by the
presence of a special sclerite L3 occupying the distal part of hla (L3, however, can be
very different in its extension). Apart from these superficial features, additional similarities
between certain species confirm this homology assumption. The most important question
in this context is whether the main muscles of the hla-hooks (called 114 in most species)
are homologous.

Archiblatta, Periplaneta, and Eurycotis

The homology of L3, hla, and 114 of these species is quite evident. (1) The hla-base takes
the same relative position: right-ventral to the L4l-region, left-posterior to the L4n-region
with the nla-process, and left-anterior to sclerite L4F (fig.54, 56, 66, 67). (2) L3 occupies
the entire hla except for the basalmost part (30 in fig.65-67). (3) The tip of hla is two-
pointed (fig.53, 65). (4) In Periplaneta and Eurycotis the main muscle of hla (114c in
fig.72) comes from the L4n-region on and near the nla-process and inserts immediately
behind s1 (fig.70). However, only Eurycotis has one accessory hla-muscle 114d (fig.73)
— possibly a subdivision of 114c.

Cryptocercus and Lamproblatta

The hla-base has a similar position relative to the insertions of 12 and s1 (fig.156, 157,
184, 185) as in Eurycotis (fig.70), and the anterior insertion of the main muscle of hla
(114 in fig.157, 184, 185) is likewise immediately behind the s1-insertion (fig.157, 158,
184, 185). Thus, homology can be assumed for the hla, L3, and 114 of these three species
(homology discussion of sl in 6.9.). Cryptocercus has one accessory hla-muscle 119
(fig.156); Lamproblatta has two, 122 and 123 (fig.184-186). These accessory muscles and
114d of Eurycotis all have different insertions, and homology relations are not assumed.
In Cryptocercus — as compared with the previous species, Polyphaga, and Ergaula (see
below) — the base of hla is more posteriorly, and hla is shorter (fig.151).

233

Polyphaga and Ergaula

The homology of hla with hla of Lamproblatta and Cryptocercus is suggested mainly by
the similar position of the hla-base posterior to the sl-insertion (fig.127, 157, 185) and
by the similar relations between the hla-base and sclerite L4K (discussion in 6.3.4.). A
muscle inserting directly on hla or L3 (114) is missing; the very stout 14 has probably
taken over the function of 114 (discussion in 6.3.4.).

Tryonicus

The hla-base has the same relative position as in Eurycotis: right-ventral to the L4l-region
and left-posterior to the nla-process (sclerites L4K and LAN in fig.97). hla and L3 of the
two species are certainly homologous. Tryonicus, however, shows three special features
as compared with the species discussed so far: (1) The hla-base is distinctly more
posteriorly (compare fig.87, 97 and 63, 67). However, this is also true of Cryptocercus.
(2) The introversible membranous basal part of hla (30 in fig.97) is by far more extensive,
and, consequently, hla can be retracted more deeply into the left complex. (3) The basal
margin of L3 is connected with L4 (L4K) by the sclerite ribbon L3a (fig.89,98). This last
feature is restricted to Tryonicus.

Anaplecta, Nahublattella, Parcoblatta, and Blaberus

The two first-mentioned peculiarities of Tryonicus are more pronounced: The hla-base is
at the posterior edge of the left complex, and the membranous basal part of hla (30 in
fig.210, 242, 269, 300) is so extensive that hla can be retracted into the left complex
except for its distalmost part only (fig.210, 242, 269) or even completely (fig.295a). (These
two features have also been investigated and found in all other Blattellidae and Blaberidae
listed in 5.15.). Another feature common to these 4 species is the membranous infolding
fpe separating the left part of the left complex (with the hla-base) from the right part
(fig.210, 243, 268, 299). These similarities clearly suggest the homology of hla and L3
in the 4 species. With Tryonicus as an intermediate, homology can also be assumed with
hla and L3 of the previous species.

Additionally, the homology of hla and L3 in Anaplecta and Eurycotis is more directly
suggested by the anterior insertion of the hla-muscle 114 or I14c,d, which is, in both
species, on and near the nla-process (fig.72, 73, 222). In Nahublattella, Parcoblatta, and
Blaberus, however, the anterior insertion of the main muscle of hla (114 or 114a,b in
fig.249, 276, 303) is on L2D’ or L2, on top of the lve-apodeme (L2a-region). In Anaplecta,
interestingly enough, the top of the Ive-apodeme and the nla-process are firmly connected
with each other (fig.222). This might suggest that all Blattellidae and Blaberidae have
gone through an evolutionary stage showing this connection, and that, at that time, muscle
114 has shifted from L4n to L2a. Homology is assumed for all hla-muscles 114. (The shift
of 114 will be disscussed in a functional context in 7.5.).

Of these 4 species only Parcoblatta and Blaberus show the following features: (1) 114 is
divided into two bundles (l14a and 114b in fig.276, 303; the division in Eurycotis
mentioned above is clearly not homologous with this division). (2) There is a muscle
within the membranous basal part 30 of hla (136 in fig.276, 303). (3) The distal part of

234

hla has a groove hge with a notch 45 in its ventral wall (fig.266, 297a). In the species
studied only in part (listed in 5.15.), the hge-groove and the notch 45 are distinctly present
in Supella, Euphyllodromia, Loboptera, Byrsotria, and Blaptica; Nyctibora has only hge
but no notch 45; in Ectobius and Nauphoeta the hge-groove is quite indistinct, and the
notch 45 is missing. (114 and 136 have not been investigated in these species). Muscle 146
is peculiar to Blaberus (fig.304, left part).

6.5. Left complex V: Further main sclerites and muscles

Some Blattaria and Mantodea have small sclerites in the dorsal wall of the vla-lobe, which
I have designated L5. LS of Metallyticus (fig.26, 27) and Cryptocercus (fig.151, 155) is
posterior to the genital opening. L5 of Periplaneta (no figure) lies more anteriorly, within
the terminal part of the ejaculatory duct. L5 of Polyphaga (fig.123, 124) is far to the left
of the genital opening. L5 of Ergaula is situated like in Polyphaga but is tranversely
orientated and approaches the genital opening more closely (fig.322m). In Anaplecta and
Nahublattella, the extension 28 of the L2- or L2D’-sclerite (fig.214, 215, 245) takes a
very similar position relative to the other parts of L2 and to the genital opening as L5 of
Polyphaga (fig.123) and might be homologous with it. The sclerites L10’ of Blaberus and
Blaptica (fig.291, 300) lie either in the dorsal vla-wall (Blaptica) or along the posterior
edge of the vla-lobe (Blaberus); whether they show any kind of homology relation with
the L5 of the other species 1s unclear, and improbable in my view. Sclerites in the dorsal
vla-wall are missing in Mantoida, Chaeteessa, Sphodromantis, Archiblatta, Eurycotis, Try-
onicus, Lamproblatta, and Parcoblatta. It cannot be decided whether L5 is a ground-plan
element of Blattaria and Mantodea and has been lost several times, or whether such
sclerites have developed several times independently.

Sclerite L7 is present only in Polyphaga, Ergaula, and Lamproblatta. These L7 (fig.115,
174) take the same relative position between the sclerite plate of the vla-lobe (L4M, L4R)
and the right phallomere and are therefore assumed to be homologous. L7 is regarded as
an element of the left complex since in a specimen of Polyphaga with its external genitalia
consisting of two right phallomeres only there was no trace of L7 (compare in 3.1.). Only
in Polyphaga and Ergaula the area containing L7 is elaborated as a special lobe-like
evagination (Iba in fig.115; in Ergaula the morphology is the same, but L7 and Iba are
larger). The Iba-lobe is assumedly homologous with the rightmost part of the vla-lobe of
the other species. (If this is true, not the vla-lobe of Polyphaga and Ergaula alone but the
vla- and Iba-lobes together are the strict homologue of the vla-lobe of the other species.
That Iba is not alone the homologue of the vla of the other species and that L7 is not the
homologue of the L4G-plates is clearly shown by the muscles 15, l6a, and 16b, compare
in 6.2.1.. L7 and Iba are bare of muscles).

Sclerite L8 is likewise restricted to Polyphaga, Ergaula, and Lamproblatta (L8 =
neoformation N of Grandcolas & Deleporte 1992). These L8 take the same position in
the right dorsal wall of the left complex, but they differ somewhat in their position relative
to the pne-pouch (fig.117, in Ergaula similar; fig.177). However, it must be considered
that in Lamproblatta, as compared with Polyphaga, the pne-pouch has shifted right-

235

anteriad (compare in 6.1.4.). The homology of the L8-sclerites is also strongly supported
by the insertions of three muscles (112, b2, 19) in their immediate vicinity: 112 (fig.128,
129, 186, 188; discussion in 6.2.4.) runs to the right ventral (or outer) wall of the Ive-
pouch, with its insertion close to that of 16a (fig.133, 188). b2 (fig.127, 184; discussion
in 6.8.) runs to the ventral part of the right phallomere, where the insertion, however, has
a slightly different position in Polyphaga and Ergaula on the one hand (R3, fig.141) and
Lamproblatta on the other (membrane next to R2, fig.198). 19 (fig.127, 184, 170;
discussion below) runs to the left dorsal wall of the left complex. L8 and the three muscles
are assumed to be homologous, and L8 and 112 are regarded as derived features of these
species.

Ergaula and Eurycotis have sclerites in the dorsal wall of the pne-pouch (L9 in fig.322m;
L6A and L6B in fig.66), but L9 and L6 are probably not homologous. Sclerite L11 (fig.91)
is peculiar to Tryonicus.

Many species have transverse muscles within the dorsal wall of the left complex, which
have been termed 19: Mantoida (fig.49), Eurycotis (fig.70), Polyphaga (fig.127, 129),
Cryptocercus (fig.155), Lamproblatta (fig.170, 185), Anaplecta (fig.221), Nahublattella
(19a and 19b in fig.249), Sphodromantis (fig.17; “b4, 19?” might be the homologue of
either 19 or b4a and b4b of Mantoida: compare in 6.7.3.). However, the exact position
and the extension of these 19 can be rather different. The homology of 19 of Lamproblatta
and Polyphaga (and Ergaula) is highly probable since the right insertion is on or near
sclerite L8 and close to the insertions of 112 and b2. In Polyphaga (and Ergaula) as well
as in Anaplecta and Nahublattella 19 has its left insertion, at least in part, on the right
wall of the pne-pouch. This relation between pne and 49 is assumed to have been lost in
Lamproblatta by the right-anteriad shift of pne and L1 (compare in 6.1.4.). 19 of Eurycotis
has a similar position like the dorsal part of 19 of Polyphaga. In Cryptocercus 19 is far on
the left; that its left insertion is next to the L4d-region (left part of sclerite L4N in fig.155)
‚and close to the Il-insertion resembles the situation in Mantoida (fig.48, 49; compare in
6.3.4.), but this close relation between 19 and L4d is in contrast to Polyphaga (fig.127,
128). On the other hand, 19 of Mantoida is farther to the right than 19 of Cryptocercus,
and its overall position is similar to that of 19 of Eurycotis and the dorsal part of 19 of
Polyphaga. In my view, these similarities are sufficient to assume homology for all these
19-muscles and to regard 19 as an element of the common ground-plan of Blattaria and
Mantodea. In the evolution of 19, some shifts might have occurred, or different parts of
19 might have been reduced or enlarged in the various species.

Some Blattaria have muscles from the ejaculatory duct D to that part of the dorsal wall
of the vla-lobe posterior to the genital opening; these have been termed 113: Polyphaga
(fig.132), Cryptocercus (fig.155), Lamproblatta (fig.188), Anaplecta (fig.222), Eurycotis
(113h in fig.72). Homology is tentatively assumed for them though their positions are
somewhat different. In Anaplecta 113 bridges the base of the vfa-outfolding (an outfolding
from the anteriormost dorsal wall of vla, compare in 6.2.4.). In Eurycotis some other
diffuse muscles within the vla-lobe have been assigned to 113 (113a,b,c,d,e,f,g,i in fig.71-
73); these could be new muscles, or some of them might be split off parts of the true 113
(Archiblatta, Blatta, Periplaneta, and Deropeltis not investigated). In Mantodea no 113-

236

muscles have been found; however, muscle b3 of Sphodromantis (fig.15) has its right
insertion not far from the dorsal vla-wall and might be a shifted 113. Hence, it is not clear
if 113 is present in the common ground-plan of Blattaria and Mantodea.

Mantoida and Cryptocercus have a longitudinal muscle in the posterior left ventral wall
of the left complex (17 in fig.52, 158). Since the position is very similar these 17 could
well be homologous. However, the respective part of the left complex is very different in
the two species (presence or absence of the hla-hook, highly modified L4-sclerotisations
in Cryptocercus), and it is not possible to compare the relative position of 17 in the two
species. Therefore, the homology of these muscles must be regarded as highly
questionable. 17 of Sphodromantis (fig.15) is certainly homologous with 17 of Mantoida
but has undergone a shift (compare in 6.3.3.). Nahublattella, Parcoblatta, and Blaberus
also have longitudinal muscles in the ventral wall of the left complex (130 in fig.251, 307;
130a,b in fig.278, 279); these 130 are assumed to be homologous, but since they take a
rather different position homology with 17 of Cryptocercus is not assumed.

6.6. Left complex VI: The position of the phallomere-gland opening

The opening of the phallomere-gland P certainly has its primitive position within the mem-
branous part of the pne-wall (discussion in 6.1.1.). It opens far anteriorly into this
membrane in Mantoida (fig.45), Chaeteessa (fig.32), and Sphodromantis (fig.10), and far
posteriorly and on the left side in Cryptocercus (fig.153, 154), Polyphaga (fig.120, 121),
Tryonicus angustus (fig.107, 108), and — considering the rotation of the pne-pouch —
Tryonicus parvus (fig.95, 96).

In Ergaula capensis the opening has, as compared with Polyphaga, shifted only a short
distance; by this shift, however, it has reached a position left-ventral to the dea-processes
and outside the pne-wall (compare fig.106 and 121). In Eurycotis (fig.67, 68) and
Archiblatta (fig.54-56) the opening is likewise ventral to the dca-processes and is assumed
to have undergone a similar shift. In Lamproblatta the opening has the same position
relative to the posterior margin of L1 (fig.177, 178) as in the previous three species but
is farther away from L1, and the processes paa and pda take their position between the
opening and the posterior margin of L1 (fig.178). paa and pda have, as compared with
e.g. Polyphaga (fig.118), shifted to the right (relative to the left posterior end of the Ive-
pouch; compare in 6.3.4.) and are assumed to have intruded into the interspace between
L1 and the phallomere-gland opening.

In Nahublattella the opening has a similar position relative to sclerite L1 and the dca-
process as in e.g. Ergaula (fig.243, 244, 328b) but has shifted far anteriad within the
membrane ventral to dea and lies in the posterior right dorsal wall of the Ive-pouch —
posterior to the dorsal wall of the ejaculatory duct D. The muscles 127 and 129 (fig.249)
are derived features of Nahublattella. In Parcoblatta (fig.270, 328e), Blaberus (fig.300,
328k), Euphyllodromia (fig.328d), Nyctibora (fig.328h), and Nauphoeta (fig.3281) the
opening has a similar position as in Nahublattella but is slightly more to the left and close
to sclerite L2 or L2D.

237

The situation in Anaplecta is difficult to interpret. This concerns the presence of two outlet
ducts with their openings close to each other (P in fig.216), the position of these two
openings, and the presence of a muscle 125 (fig.224) inserting between them. Taking a
situation like in Nahublattella as a starting point, the position of the openings could be
explained by the assumption of a further shift to the left within the dorsal wall of the
ejaculatory duct, and then ventrad to beneath the Ive-apodeme. However, the preceding
shift assumed for Nahublattella would have hardly been possible in Anaplecta since
between the membrane posteroventral to the pne-pouch (fig.209) and the posterior dorsal
wall of the ejaculatory duct there are still extensive right parts of L2 (fig.211) “blocking”
this shift. The openings of Anaplecta are in one respect similarly situated as in
Lamproblatta: more or less ventral to the processes paa and pda (compare fig.210 and
178). The position relative to the Ive-pouch, however, is completely different: dorsal to
Ive in Lamproblatta, ventral to lve in Anaplecta. Possibly, the outlet ducts of Anaplecta
are new organs. In this case, for the remaining Blattellidae and for Blaberidae the
possibility has to be considered that their glands and outlet ducts are homologous with
those of Anaplecta (or one of them) and not with those of the other species. As a point
possibly interesting in this context, the spermathecae of the female genitalia have also
been replaced by completely new organs in Blattellidae and Blaberidae (McKittrick 1964).

6.7. The elements of the right phallomere

6.7.1. Comparison between Blattaria and Mantodea

The homology relations and the ground-plan of the elements of the right phallomere can
best be deduced from a comparison of Eurycotis, Chaeteessa, and Mantoida.

The cuticular elements of the right phallomeres of Eurycotis and Chaeteessa show the

following similarities:

1. A sclerite R3 occupies the anteriormost ventral wall of the right phallomere (fig.28,
VD):

2. At least the right and the right anterior margins of R3 form a groove-like apodeme
age (fig.28, 77).

3. The right posterior end of R3 articulates (A3 in fig.28, 77) with more posterior
sclerites (Eurycotis: RIEF in fig.77; Chaeteessa: R1B in fig.28).

4. The anterior part of both RIF and R1B extends to the left and reaches an edge (16
in fig.28, 77) along which it bends sharply dorsad.

5. Then this sclerotisation arches dorsad and then to the left. The arching in a dorsal
direction is extensive in Eurycotis; in Chaeteessa it is less pronounced and the
sclerotisation extends mainly to the left.

6. The posterior margin of this sclerotisation forms a posteriad-directed ridge (pva in
fig.28, 78).

7. To the left of (Chaeteessa) or left-ventral to (Eurycotis) this pva-ridge the right
phallomere has a large central invagination (cbe in fig.29a, 31, 77, 78; the whole of
cbe is a part of the ventral wall of the right phallomere).

238

8. The posterior part of the right phallomere is composed of a dorsal lobe (fda in fig.31,
74) and a ventral tooth or ridge (pia in fig.28, 29a, 77, 78). fda and pia are connected
along the right edge of the right phallomere, and they diverge to the left like the two
halves of an opened book. In Eurycotis pia is as large as fda (and two-pointed); in
Chaeteessa pia is much smaller than fda.

9. The dorsal wall of the fda-lobe is sclerotised (Eurycotis: R1H in fig.74; Chaeteessa:
R1A in fig.31).

10. The posteroventral part of both RIF and R1B (fig.28, 77; posterior to edge 16) extends
onto the pia-tooth. However, the sclerotisations of the anterior and of the posterior
parts of pia are connected in Chaeteessa (R1B) but separated in Eurycotis (RIF
anteriorly and R1G posteriorly, which articulate in A9).

11. A large membranous area (17 in fig.28, 77) is present at the posterior right edge of
the right phallomere, between the right margins of the sclerotisations of fda and pia.

Homology is assumed for all these similarities, for all elements given the same name, and
for the compared sclerotisations taking the same relative positions. 1.-11. are assumed to
be features of the common ground-plan of Blattarıa and Mantodea.

A further similarity between Chaeteessa and Eurycotis is that the sclerotisation adjoining
articulation A3 posteriorly (RIB in fig.28; RIF in fig.74, 77) is separated from the dorsal
sclerotisation of the fda-lobe (RIA in fig.28, 31, 32, RIH in fig.74, 77) by membrane (4
in fig.28, 32, A8 in fig.74). The dividing lines 4 and A8, however, are probably non-
homologous (discussion below).

There are also some essential differences between Eurycotis and Chaeteessa: (1) The
separation or connection of the anterior and posterior sclerotisations of pia (compare
feature 10.). (2) Only Eurycotis has a sclerite R2 (fig.77), which articulates with R3 (A7
in fig.75, 77) and RIF (A6 in fig.75). (3) In Eurycotis the cbe-invagination has a summit
in the center of the right phallomere and becomes shallower to the left of this summit
(where R2 adjoins; fig.75, 78); in Chaeteessa the cbe-invagination becomes continuously
deeper to the left (fig.29a, 31). (4) Only Eurycotis has a tre-tendon in the anteriormost
dorsal wall of the fda-lobe (fig.74). (5) Only Eurycotis has the sclerotisations of pia (R1G)
and fda (RIH) connected with each other posterior to the membranous area 17 (by a
narrow sclerite bridge; fig.77, 78).

As regards the right phallomeres of Mantoida and Chaeteessa, homology is quite evident
for most elements: Sclerite R3 has the same shape and position and a similar age-apodeme
(compare fig.28, 29a and 41, 43). The right posterior end of R3 articulates (A3 in fig.28,
41) with the sclerotisation adjoining posteriorly (RIE or RIB). However, only in Mantoida
the groove-shape of the sclerotisation extends from R3 (age) far beyond A3 onto the
posterior sclerite RIE (fig.41, 43). The posterior part of the right phallomere is, like in
Chaeteessa, composed of a large dorsal lobe (fda in fig.44) with a sclerotised dorsal wall
(R1E in fig.44) and a ventral tooth (pia in fig.41, 43) with dorsal and ventral sclerotisations
(RIE in fig.41, 43). However, in Mantoida the sclerotisations in the dorsal fda-wall and
those on pia are interconnected anteriorly by a broad sclerite bridge (RIE in fig.41, 44;
no membranous stripe 4 as in Chaeteessa, fig.28). Behind this bridge there is, like in
Chaeteessa, a large membranous area (17 in fig.41). In the ventral wall of the right

239

phallomere, to the left of A3 and anterior to pia, Mantoida has likewise a tooth-like
evagination (pva in fig.41, 28). Its sclerotisation, however, is isolated (RID in fig.41);
this is in contrast to both Chaeteessa and Eurycotis (fig.28, 77, 78) and is assumed to be
a derived feature. The edge 16 of Chaeteessa and Eurycotis (fig.28, 77) has also been
lost. The large central invagination cbe to the left of the pva-tooth resembles cbe of
Chaeteessa (fig.43, 29a).

Taking the homology hypotheses assumed so far as a basis, the muscles of the right
phallomere are rather similar in Mantoida and Eurycotis, and the assumed homologies of
the cuticular elements (1.-11.) are confirmed:

12. Some phallomero-sternal muscles insert along the anterior margin of R3 (s2 and s4
in fig.42, 82; homology discussion in 6.9.).

13. The s2-insertion on R3 extends to the right as far as to a keel-apodeme on the age-
apodeme (3 in fig.41, 42, 44 and 74, 77, 82). (Keel 3 is missing in Chaeteessa.)

14. Muscle rl (fig.48, 79) inserts on the right part of R3, immediately to the right of the
s2-insertion and the keel 3, and runs to the dorsal wall of the fda-lobe.

15. Muscle r2 (fig.49, 80) runs from R3 to the cbe-invagination (compare fig.44, 74 and
49, 80). The right part of the posterior r2-insertion is on the R1-sclerotisation that
forms the pva-tooth more posteriorly (fig.50, 80).

16. Muscle r3 (fig.49, 50, 80) runs from the right wall of the right phallomere to the left
where it inserts mainly in the dorsal wall of the pia-tooth. The rest of the left r3-
insertion is on the ventral fda-wall in Mantoida (compare fig.49 and 50) but in the
ventral pia-wall in Eurycotis (compare fig.80 and 82).

These muscles and the keel 3 are assumed to be homologous and to be features of the
common ground-plan of Blattaria and Mantodea. Muscle r4 is only present in Mantoida
(fig.49), the muscles r5 (fig.80) and r6 (fig.79) only in Eurycotis.

Furthermore, Mantoida and Eurycotis have in common that (1) the age-apodeme extends
as far as to articulation A3 (fig.41, 44, 74) and that (2) even the sclerotisation posterior
to A3 is groove-shaped (fig.41; rge in fig.74, 77). Both is not the case in Chaeteessa
(fig.28). (1) is assumed to be a feature of the ground-plan of Blattaria and Mantodea:
17. The age-apodeme reaches articulation A3.

As regards (2), however, the grooves posterior to A3 take different positions relative to
the right r3-insertion (ventral to r3 in Mantoida, dorsal to r3 in Eurycotis) and are regarded
as non-homologous.

Main sclerite R1 is differently divided in the species discussed so far; the questions arise
(1) which of these divisions are homologous and (2) when have these divisions evolved.

The separation of the pva-sclerotisation (sclerite RID, fig.41) in Mantoida is certainly
apomorphic (compare above).

Chaeteessa and Eurycotis have the dividing lines 4 (fig.28, 32) and A8 (fig.74) in a similar
position. Sphodromantis has a dividing line (4 in fig.6, 14) in the same position as 4 of
Chaeteessa, which is not membranous but only weaker sclerotised than the sclerites RIA
and R1B. This weak stripe 4 of Sphodromantis and the membranous stripe 4 of Chaeteessa
are assumed to be homologous. Muscle r3 of Sphodromantis (fig.16, 19) has the same

240

course as r3 of Mantoida (fig.49, 50). The right insertion of r3 is posterodorsal to stripe
4 (compare fig.14 and 16, 19). In Eurycotis, however, the right insertion of r3 is
anteroventral to articulation A8 (compare fig.74 and 80). Thus, homology is highly
improbable for the dividing lines 4 and A8. The dividing line 4 is thus missing not only
in Mantoida but also in Eurycotis, and it is not a feature of the common ground-plan of
Blattaria and Mantodea but a derived feature of a Mantodean subgroup containing at least
Chaeteessa and Sphodromantis.

The question remains whether the articulations A8 and A9, both missing in Mantodea,
could be elements of the common ground-plan of Blattaria and Mantodea. In Eurycotis,
A8, A9, and the sclerite bridge between R1G and RIH (behind membrane 17 in fig.77,
78) are assumed to be functionally correlated: The posterior part of the right phallomere
— composed of fda and pia — can perform a swinging or flapping movement, with A8 and
A9 defining the axis. During this movement the membrane 17 is folded and stretched
again, and the sclerite bridge may have the function to stabilise RIG and R1H against
each other. Muscle r3 (fig.80) moves the flap posterolaterad; rl and r6 (and possibly s8
on the tre-tendon; fig.79) pull it anteromediad. In Mantoida and Chaeteessa nothing
suggests that such a flap-mechanism has ever been present. Thus, from this functional
point of view, A8, A9, and the posterior bridge are probably derived features of Eurycotis
(and other Blattaria, see in 6.7.6.). However, this view is debatable: In the copulation of
Periplaneta the hla-hook has its function in the “initial seizing” and the flap-mechanism
in the “final holding”. Since the copulation habits of Mantodea are derived, the flap-
mechanism could be in the same way completely obliterated as the hla-hook and some
correlated elements possibly are on the left side (compare in 6.4.2.). On the other hand,
the right phallomeres of Mantoida and Chaeteessa (fig.28, 41) also seem to have the ability
to grasp (mainly by the pia- and pva-teeth), and the final holding could well have been
performed by other structures different from the flap-mechanism in the common ground-
plan of Blattaria and Mantodea. Thus, it is improbable but cannot be completely excluded
that A8, A9, and the posterior bridge are elements of the common ground-plan of Blattaria
and Mantodea.

As a result, the articulations A8 and A9 separating the sclerites RIF, R1G, and RIH are
probably derived elements of Blattaria. The dividing line between RIE and RID
(Mantoida) as well as the dividing line 4 between RIA and RIB (Chaeteessa and
Sphodromantis) are certainly derived features of Mantodean subgroups. If these hypotheses
are true,

18. Ri is an undivided sclerite in the common ground-plan of Blattaria and Mantodea.

However, if A8 and A9 should prove to be elements of this ground-plan, Ri would
have to be regarded as tripartite - composed of RIF, R1G, and RIH like in Eurycotis.
(When R1 will subsequently be assumed to be undivided in this ground-plan, this
must be seen with these reservations in terms of A8 and AQ).

A further difference between Eurycotis (and some other Blattaria) and all Mantodea studied
is the presence or absence of the tre-tendon and muscle s8 and the different condition of
the muscles b4.

241

Muscles connecting dorsal parts of the left complex and of the right phallomere have been
termed b4. Mantoida, Eurycotis, and Polyphaga have two such muscles, b4a and b4b
(fig.36, 48, 58, 109); Cryptocercus has three, b4a, b4b, and b4c (fig.143a). The Blattarian
species have the right insertions of all b4-muscles on the tre-tendon, deeply immersed in
the body, and the homology of the b4-group as a whole is rather certain. Mantoida has
the right insertions of both b4a and b4b on the left dorsal anterior margin of the fda-lobe
(fig.48; the b4b-insertion is not shown; it is immediately posterior to the b4a-insertion).
Since the external origin of tre is at the dorsal anterior margin of fda, the right insertions
of the b4-muscles take the same relative position in Blattaria and Mantoida. The left
insertions, however, take rather different positions, and some shifts must have taken place:
The left b4a-insertion is in Eurycotis and Polyphaga on the utmost right part of the Ive-
pouch (fig.70, 129, 130), in Mantoida on an infolding to the right of the Ive-pouch (fig.48,
compare fig.46). These positions are quite similar. The left b4b-insertion is in Polyphaga
(fig.127) in the anterior right dorsal wall of the left complex, far right-dorsal to the pne-
pouch; in Eurycotis (fig.70) it is on the top of the pne-pouch; in Mantoida the position
is intermediate — dorsal to the pne-pouch, but next to its base (fig.48).

On the basis of these relations, it is in my view acceptable to regard the b4a and b4b of
Mantoida, Eurycotis, and Polyphaga as strictly homologous and to assume homology
between these b4-muscles as a whole and the b4-group of Cryptocercus. The immersion
of the right insertions (by tre) can possibly be regarded as the derived condition and as
an autapomorphy of Blattaria or of a Blattarian subgroup. The same might be assumed
for the cooperating s8-muscle (homology discussion in 6.9.).

19. Muscles b4a and b4b are present.

20. Muscle s8 and the tre-tendon are probably absent.

For the simple sclerites R2 and R3 there is no necessity for defining regions. The

complicated main sclerite R1 will be divided into regions, which mainly (and arbitrarily)

correspond to the division into individual sclerites in Eurycotis, and which are defined as

follows (fig.33le, 332e):

— Rld (dorsal): The sclerotisation homologous with sclerite RIH of Eurycotis (fig.74) on
the fda-lobe. On R1d there are the posterior insertions of the muscles rl and r6 (fig.80).

— Riv (ventral): The sclerotisation homologous with sclerite RIG of Eurycotis (fig.77,
78) on the posterior part of the pia-tooth. On Riv there is the left insertion of muscle
r3 (fig.80, 82).

— RIt (tooth): The sclerotisation homologous with that part of sclerite RIF of Eurycotis
which extends dorsad from the edge 16 (fig.77, 78) and lies in the right-dorsal wall of
the cbe-invagination. Along its posterior margin R1t forms the ridge or tooth pva. On

Rit there is the insertion of the right-dorsal part of muscle r2 (fig.80, 81).

— Rle (central): The sclerotisation homologous with the remainder of sclerite RIF of
Eurycotis, which adjoins sclerite R3 posteriorly and articulates with it (A3), which
extends onto the anterior part of the pia-tooth, and which forms the groove rge along
its dorsal margin. Rle is situated centrally between the other R1-regions: The border
to Rid is articulation A8; the border to R1v is articulation A9; the border to Rit is
edge 16. On Ric there is the right insertion of r3 (fig.80, 82), the anterior insertion of
r6 (fig.79; on rge), and the posterior insertion of r5 (fig.80; on rge).

242

In fig.331c and 332c this regioning of R1 is transferred to Chaeteessa — according to the
homology relations assumed above (features 1.-11.). In Chaeteessa (and Sphodromantis)
the Rlc-region extends far dorsad into the RIA-sclerite (fig.331c); this results from the
position of the right r3-insertion in Sphodromantis (compare fig.16 and 331a,c). The
regioning of R1 of Mantoida is shown in fig.331d, 332d.

6.7.2. The elements in the common ground-plan of Blattaria and Mantodea

The features 1.-20. in 6.7.1. permit the reconstruction of many ground-plan features of the
right phallomere (fig.321f,h): R3 is a curved plate in the anteriormost ventral wall. The
right posterior end of R3 articulates with the Rlc-region (A3). At least the right and right
anterior margins of R3 form a groove-like age-apodeme, which reaches articulation A3
(but the groove does not exceed A3). RI is (probably) an undivided sclerite, with all its
regions firmly connected. Along edge 16 the regions Rle and Rit are sharply angled to
each other. Rit forms a posteriad-directed tooth or ridge pva at its posterior margin. A
large central invagination cbe is situated to the left of or left-ventral to pva. The dorsal
lobe fda and the ventral tooth pia are distinct; they are connected along the right edge of
the right phallomere and diverge to the left. Their walls are largely sclerotised by R1
(regions Rld and Riv). In the posterior right edge of the right phallomere there is a
membranous area 17. The parts of RI in the dorsal wall of fda and those on pia are
interconnected anterior to membrane 17. Posterior to membrane 17 there is no dorsoventral
connection (like in Mantodea) or, at most, a very narrow one (like in Eurycotis). The
muscles rl, r2, r3, s2, and s4 are present. The insertions of s2 and rl are separated by
the keel-apodeme 3. The tre-tendon and the articulations A8 and A9 are probably missing.
It is unclear if R2, the articulations A6 and A7, and the muscles r4, r5, and r6 are present
or not.

6.7.3. Homology relations and character states of the elements in Mantodea

The R3-sclerites of Chaeteessa (fig.28), Mantoida (fig.41), Metallyticus (fig.20), and
Sphodromantis (fig.6) are very similar. The age-apodeme is always deeper in its left part,
where it is more or less plate-like (this is least distinct in Chaeteessa). In Metallyticus and
Sphodromantis this deepening of age is very abrupt. Only in Sphodromantis this left part
of age has developed a curvature to the posterior and back to the right (fig.6, 8). Only in
Chaeteessa the left marginal part of R3 bends dorsad into the cbe-wall (fig.29a, 32). Two
other derived features of Chaeteessa are that the utmost right-posterior part of age and
the keel-apodeme 3 have been lost (fig.28). In the other species the groove-like shape of
the sclerotisation even exceeds A3 (distinct in Mantoida, fig.41, 43, and Sphodromantis,
fig.6, 8; less distinct in Metallyticus, fig.20, 21); the keel 3 has been retained (fig.6, 13,
20, 23, 41, 44). At least in Sphodromantis and Mantoida keel 3 separates the insertion
areas of s2 and rl (fig.15, 48). The apodeme are (fig.6, 8) is a derived feature of
Sphodromantis.

The cbe-invagination becomes in all four species continuously deeper to the left (fig.6, 8,
20, 21, 28, 29a, 41, 43), and a sclerite R2 is always missing (compare Eurycotis, fig.77,
78; but see below: Metallyticus).

243

The posterior part of the right phallomere is in all species composed of a large dorsal lobe
fda (fig.13, 23, 31, 44) and a smaller, leftward projecting ventral tooth pia (fig.6, 20, 28,
41). In Sphodromantis pia has become very small by a reduction of its posterior part.
Left-anterior to pia there is always another tooth-like process pva (fig.6, 20, 28, 41).

The R1-sclerotisations are very similar in Chaeteessa and Sphodromantis (fig.6, 28): The
sclerotisation posterior to articulation A3 is connected with the pva-sclerotisation and with
the pia-sclerotisation (sclerites R1B) but more or less separated (by 4 in fig.6, 28) from
the dorsal fda-sclerotisation (sclerites RIA). The dividing line 4 is also present in
Metallyticus (fig.20), but RIA has expanded far into the right ventral wall and occupies
the ventral wall of pia (fig.20, 21). Only Mantoida shows the plesiomorphic state with

the dividing line 4 missing. In both Metallyticus and Mantoida the pva-sclerotisation has
been separated from the R1-sclerotisation posterior to A3 (sclerites RID in fig.20, 41).
That in Mantoida these elements are really pva and parts of RI (and not R2) is shown
by the posterior insertion of r2, whose right part inserts on that sclerite (fig.49, 50;
compare fig.16, 19). (R2 of Eurycotis bears the left part of the r2-insertion; compare
feature 15. in 6.7.1.). In Metallyticus, however, R1D adjoins the left posterior end of R3
in a similar way as R2 in Eurycotis (fig.20, 77) and could really be the homologue of R2.
But regarding the situations in the other Mantodea, it is certainly more probable that the
tooth is the true pva and RID the respective part of R1. A definite decision might come
from an investigation of muscle r2.

The membranous area 17 has retained its primitive condition only in Chaeteessa and
Mantoida (fig.28, 41). In Metallyticus it has been largely reduced by the expansion of
RIA onto pia (fig.20). In Sphodromantis it has lost its boundary to the membranous ventral
fda-wall by the reduction of the posterior part of the pia-tooth and its sclerotisation (fig.6).

The regioning of R1 of Chaeteessa (6.7.1.) can be transferred to Mantoida, Metallyticus,
and Sphodromantis (fig.331a-d and 332a-d). One minor problem is the exact course of the
boundary between the regions Ric and RIt, since the edge 16 is distinct only in
Chaeteessa (fig.28). In Mantoida and Metallyticus sclerite R1D is tentatively equated with
the Rit-region, but it probably does not exactly correspond to Rit as defined in Eurycotis.

The muscles of Sphodromantis (fig.15, 16, 19) are very similar to those of Mantoida
(fig.48, 49, 50), but r4 is much stouter. Of the muscles connecting the right phallomere
and the left complex dorsally (b4a and b4b in Mantoida, fig.48) at most one is retained
(b4, 19? in fig.17), but its homology with b4 of Mantoida is questionable since both its
insertions are on the left complex. This muscle of Sphodromantis could also be
homologous with 19 of Mantoida (fig.49; compare in 6.5.).

6.7.4. Homology relations and character states of the elements in Blattaria I: The
anteroventral elements

In the following discussions in 6.7.4., 6.7.5., and 6.7.6., data of some species are included
whose right phallomeres have been studied only in part: Archiblatta, Ergaula,
Euphyllodromia, Nyctibora, Byrsotria (fig.330f,m,o,r, 318, 319), and Supella. Archiblatta
resembles Eurycotis (fig.330g); Ergaula resembles Polyphaga (fig.3301); Nyctibora and
Byrsotria resemble Blaberus (fig.330s). For a discussion of the right phallomeres it is

244

useful to consider first the anteroventral elements (R2, R3, cbe-invagination; 6.7.4.) and
the tre-tendon (6.7.5.), whose homology relations are quite evident. Then the posterodorsal
parts will be discussed (R1, pva-tooth, fda-lobe, pia-tooth; 6.7.6.).

The sclerites R2 and R3 and the cbe-invagination can, as compared with Eurycotis, easily
be identified in all Blattaria: R3 (fig.77, 102, 137, 163, 193, 229, 257, 284, 312a) is always
a plate in the anteriormost (right-)ventral wall of the right phallomere, and parts of its
anterior and lateral margins nearly always form a groove-like age-apodeme. The right and
left posterior margins of R3 articulate with more posterior sclerites: A3 (between R3 and
R1) is always distinct. A7 (between R3 and R2) is in most cases also a true articulation;
sometimes, however, R2 and R3 are fused in this place (Polyphaga, Ergaula: A7*), or
R2 and R3 are more distant from each other and no longer articulated (Parcoblatta,
Blaberus; the term A7 is still used to designate the homology of the respective areas).
Posterior to the central part of R3 (between A3 and A7 or A7*), the ventral wall of the
right phallomere always bends dorsad and (more or less strongly) anteriad to form a central
invagination (cbe in fig.78, 104, 138, 164, 195, 230, 258, 285, 313). cbe is variable in its
extension and distinctness and is sclerotised to a varied extent. cbe always has its summit
in the center of the right phallomere and a descent in the left-ventral direction (missing
in Mantodea), at whose base that sclerite adjoins which also approaches R3 in A7 or A7*
(R2 in fig.75, 100, 135, 161, 191, 227, 255, 282, 310). R2 has either as a whole the shape
of a ridge, often projected into teeth or bulges (fig.75, 76, 135, 1365 TOR GZS eso:
227, 228, 281, 285, 287, 310, 311), or R2 is more plate-like but likewise beset with tooth-
like evaginations (fig.99, 100, 102, 254-256). All species with the muscles studied have
a stout r2 from R3 to the cbe-invagination (fig.80, 81, 140, 141, 166, 167, 197, 198, 231,
232, 259, 260, 286, 287, 314, 316), and some phallomero-sternal muscles insert at the
anterior margin of R3 (fig.82, 142, 168, 199, 233, 261, 288, 317; homology discussion in
6.9.). According to these corresponding relative positions and similarities in shape,
homology is assumed for the sclerites R2 and R3, the articulations A3 and A7 (or A7*),
the cbe-invaginations, the age-apodemes, and the r2-muscles of all species.

Fig.330: Right phallomere, evolution of main sclerites. — The sclerotisations of the right phallomere
are shown. The view is roughly cranial but the phallomeres are unfolded. For correct orientation
compare fig.1-319.

Of sclerite R3 (white) only the posterior part is shown (anterior part cut off along undulate line). The
other sclerotisations are shown completely and patterned differently. The individual sclerites of Ri
are labelled with the capital letters used in the text and in fig.1-319 (e.g. F = RIF). X (part of sclerite
R2) and Y (part of region Rt) are sclerotisations occupying the cbe-invagination (compare in 6.7.4.).
Articulations between sclerites are labelled with the numbers used in the text and in fig.1-319 (e.g.
6 = A6). If an articulation has been lost by fusion of the respective sclerites, the point of fusion is
labelled by adding * to the name of the lost articulation (e.g. 6* = A6*). tre and cwe are formative
elements. cl represents a certain dividing line between R1-sclerotisations (compare in 6.7.6.). Species
with “S” behind their names have side-reversed phallomeres, and a mirror-image of the original
preparation is shown. The branching black lines represent the assumed phylogeny. The ground-plan
is in some respects unclear (? in fig.330a; discussion in 6.7.1.): Presence of sclerite R2 and of
articulations A6 and A7.

245

zB GT aL tusAIOY LYHLIGQIIAY (| BuEpLioy syoskiny (8

eyEpnes epioueyy (q 0 ¢ jesnayjodéy (8

Bssvajovy) (9

9
4

AAAI

: em /|\_ So er
eu joe =

snared snsmo0Ary (Y

9
be

u Fu ‚erejsndur
‘ds syugwoapoyds (a SNIIVIOIA ve ernuoapoyjäydny (0

smayAgerap (pP

8s}

ey

eI

IR

I
es)

sisuades
vypnedig (w

: S:
ey
eu z

nenn
BER
ET
we,

"ds wrogyoAn (a BIEL eppegosaeg (b

4}

2
a

esendÄdse edeydäjog (I sadyediqye enejgoadue] (A

246

The age-apodeme is rather variable in its extension: In Polyphaga (fig.137), Ergaula,
Anaplecta (fig.229), Archiblatta, and Eurycotis (fig.77), it is restricted to the right and the
right anterior margins of R3. In Cryptocercus (fig.163) and Nahublattella (fig.257) it
extends along the whole anterior and lateral margins of R3. In Parcoblatta (fig.284),
Nyctibora, Byrsotria, and Blaberus (fig.312a) it is restricted to the anterior marginal areas
of R3. Only in Tryonicus (fig.102) and Lamproblatta (fig.193) age has been lost.

R2 is an isolated sclerite in Eurycotis (fig.75, 76), Tryonicus (fig.100, 101), Cryptocercus
(fig.161, 162), Lamproblatta (fig.190, 191), Anaplecta (fig.227, 228), Nahublattella
(fig.254, 255), and Supella. In all these species, the right-ventral end of R2 articulates
distinctly with R3 (A7), the left-dorsal end of R2 articulates distinctly with R1-
sclerotisations in the dorsal wall of the cbe-invagination (A6), and R2 is restricted to the
ventral base or, at least, to the ventral wall of the cbe-invagination.

In Euphyllodromia (fig.3300), Parcoblatta (fig.282, 283, 330q), Nyctibora (fig.330r),
Blaberus (fig.310, 311, 330s), and Byrsotria R2 is fused to R1-sclerotisations (RIS or
RIT) in the area corresponding to the A6-articulation of the other species. The point of
fusion is A6*, with the ewe-thickening (fig.282, 310) in its immediate vicinity. This topic
will be discussed in 6.7.6.. In Parcoblatta and Blaberus R2 is in close vicinity to R3 in
the area corresponding to the A7-articulation of the other species (A7 in fig.284, 312a)
but is not articulated with R3. In Euphyllodromia and Byrsotria (fig.318) A7 is a distinct
articulation. In Nyctibora A7 is distinct and hinge-like (fig.319).

In Polyphaga (fig.135-137, 3301) and Ergaula (fig.330m) R2 and R3 are clearly
identifiable by their shapes (R3 is a broad curved plate, R2 forms a dental ridge) and by
their positions relative to each other, to cbe, and to the r2-insertions (fig.141; compare
Cryptocercus, fig.161, 167, and Eurycotis, fig.75, 81; r2 not investigated in Ergaula).
However, in both species R2 shows two peculiarities: (1) R2 is fused to R3. (2) R2 has
spread over the cbe-invagination (sclerotisation X in fig.330l,m) and is broadly fused to
R1-sclerotisations in the dorsal cbe-wall; hence, the whole of cbe is sclerotised (fig.134,
135). (1): In Polyphaga the stripe of weaker sclerotisation A7* (fig.135, 137) takes the
same position as articulation A7 in other Blattaria and is assumed to be the line of fusion
between R2 and R3 (and a vestige of A7). In Ergaula R2 and R3 are fused without any
vestige of A7 (no weak line), and the border is not exactly determinable. Moreover, R2
of Ergaula has become so broad that R3 is for most of its breadth confluent with R2
(compare fig.3301 and m). (2): In Polyphaga the sclerotisation of cbe has a weak line (13
in fig.134, 138) and an adjacent notch within the sclerite margin; these structures are
assumed to mark the border between R2 and R1, and the sclerotisation in the ventral wall
and on the summit of cbe is assumedly part of R2 (X in fig.3301). R1 is restricted to the
posterior dorsal wall of cbe. In Ergaula R1 and R2 are firmly connected (no weak line);
the interpretation of the cbe-sclerotisation is done in accordance with Polyphaga (X in
fig.330m is part of R2).

Archiblatta also has the whole cbe-invagination sclerotised (compare fig.74 and 75 of
Eurycotis; the sclerotisation concerned is Y in fig.330f, compare fig.330g), but, in contrast
to Polyphaga, the cbe-sclerotisation is slightly weaker near the ventral base of the cbe-
invagination; in the corresponding area of Eurycotis R2 has its dorsal margin and the

247

membranous ventral wall of cbe adjoins (fig.75, 76). Therefore, in Archiblatta the
sclerotisation of cbe is assumed to have developed by an expansion of RI (Rit-region,
fig.330f). The situation in Eurycotis, with R1 occupying a large part of cbe (fig.75;
compare e.g. fig.99, 160), can be regarded as a primitive stage of such a development. As
a result, homology is assumed for the cbe-sclerotisations of Polyphaga and Ergaula
(mainly part of R2, X in fig.3301,m), but the cbe-sclerotisation of Archiblatta (mainly part
of region Rit, Y in fig.330f) is not homologous with these. In Archiblatta R2 and the
heavier sclerotised dorsal parts of R1 are still distinctly articulated (A6) in the left-dorsal
wall of the cbe-invagination (like in Eurycotis, A6 in fig.75). In Polyphaga and Ergaula
this articulation is missing (A6* in fig.135, 137).

The shape of R2 is rather variable. Details are shown in the figures. Some peculiar features
are: In Parcoblatta (fig.285) R2 is strongly curved. In Nahublattella R2 bears the
conspicuous elements 42 and 43 (fig.254). Only in Tryonicus and Lamproblatta R2 has
an extension posterior to articulation A7 (R2m in fig.102, 91 and 174, 193), which lies
in the rightmost part of the vla-lobe (left complex), and which has in Lamproblatta a close
contact with sclerite L7.

Only Blaberus (fig.311, 312a), Byrsotria, and Nyctibora (fig.319) have a peculiar sclerite
RS ventral to R2 and A7. The R5 of the three species take exactly the same relative
position and are certainly homologous.

As regards the muscles, r9 is specific to Polyphaga (fig.141; Ergaula not studied), and
r8 is specific to Cryptocercus (fig.167). Since the posterior insertions take completely
different positions, a homology of r8 and r9 is most unlikely.

6.7.5. Homology relations and character states of the elements in Blattaria II: The
tre-tendon

A tre-tendon is present in Archiblatta (fig.330f), Eurycotis (fig.74, 330g), Tryonicus
(fig.99, 330h), Ergaula (fig.330m), Polyphaga (fig.134, 3301), and Cryptocercus (fig.160,
3301). Homology is ascertained by the similar position of the tre-base in the anterior dorsal
wall of the right phallomere, by a muscle from the right half of the sugenital plate (s8),
and by two or three muscles from the dorsal part of the left complex (b4-group; fig.79,
139, 165; muscles not studied in Tryonicus). A muscle from tre to R3 is specific to
Cryptocercus (r7 in fig.165). In Lamproblatta as well as in Anaplecta, Nahublattella,
Parcoblatta, Blaberus, and all other Blattellidae and Blaberidae studied (fig.330n-s) tre,
s8, and b4 are missing.

6.7.6. Homology relations and character states of the elements in Blattaria III: The
posterodorsal elements

The elements discussed here are those dorsal and posterior to the summit of the cbe-
invagination and posterior to articulation A3 (compare fig.321f,h). In Eurycotis (fig.74-
77), for example, this part of the right phallomere is composed of the dorsal wall of cbe,
the ridge pva, the dorsal lobe fda, and the ventral tooth pia, and it contains the
sclerotisations comprised in R1: three sclerites RIF, R1G, and RIH. This part of the right
phallomere has undergone very complicated evolutionary changes.

ee

ER

b) Metallyticus

violaceus c) Chaeteessa
caudata 33 1

d) Mantoida schraderi e) Eurycotis floridana f) Tryonicus parvus

Fig.331: Right phallomere, homology of main sclerites and homologous regions of main sclerite R1
(dorsal views). — The cuticular elements of the right phallomere are shown, but some membranous
parts are removed. Patterned areas are sclerotised, white areas are (except for sclerite R3)
membranous. Undulate lines are cutting lines. R3 is separated from the remainder of the right
phallomere and shifted anteriad. R5 of Blaberus is not shown. The individual sclerites of R1 are
labelled with the capital letters used in the text and in fig.1-319 (e.g. F = RIF). Articulations between
sclerites are labelled with the numbers used in the text and in fig.1-319 (e.g. 6 = A6). Articulations
A3 and A7 are not labelled (see fig.332). If an articulation has been lost by fusion of the respective
sclerites, the point of fusion is labelled by adding * to the name of the lost articulation (e.g. 6* =
A6*). tre and cwe are formative elements.

249

R3

g) Lamproblatta N
albipalpus N

i) Polyphaga

h) Cryptocercus aegyptiaca

punctulatus

k) Anaplecta sp.

I) Nahublattella sp.

n) Blaberus
eraniifer

m) Parcoblatta
lata

331

250

Eurycotis and Archiblatta

For Eurycotis this area has been fully discussed in 6.7.1.. In Archiblatta, like in Eurycotis
and in all Mantodea, the fda-lobe and the pia-tooth are both very distinct. Sclerite RIF
(fig.330f, regions Rle and Rit) closely resembles RIF of Eurycotis. The sclerotisations
of the Rid- and Rlv-regions are more complicated than in Eurycotis (compare fig.330f
and g) but similarly structured in a dorsal (RIH = Rid) and a ventral (RIG = Rly)
sclerite. The sclerite bridge connecting RIH and R1G in Eurycotis (behind membrane 17
in fig.77; fig.330g) has a short gap in Archiblatta; instead, there is a ribbon-like connection
between RIH and RIG across the ventral wall of the fda-lobe (compare fig.330f and g).

Tryonicus, Cryptocercus, and Lamproblatta

In these species (fig.99-104, 160-164, 190-195) the posterodorsal part of the right

phallomere has only two sclerites RIF and R1J. RIF corresponds to RIF of Eurycotis

(fig.74-78). The area posterior to RIF contains R1J, which is a fusion product of RIH

and R1G of Eurycotis, and is an undivided lobe (fda, no ventral tooth pia present). The

regioning of R1 is shown in fig.33le,f,g,h and 332e,f,g,h.

RIF (fig.102, 163, 193) is in these species, like in Eurycotis, somewhat horseshoe-shaped

(open to the left, with a dorsal and a ventral arm), and along RIF there are the following

structures in common, which are all regarded as homologous:

— The ventromedian end of RIF articulates with R2 (A6 in fig.75, 100, 160, 164, 190).

— The ventral arm lies in the dorsal wall of the cbe-invagination (fig.74, 99, 160, 190).
It bears a ridge (pva in fig.80, 99, 164, 190, 196), which is formed by cuticular
evagination in Eurycotis, Tryonicus, and Lamproblatta, and by cuticular thickening in
Cryptocercus (cross-section in fig.164).

— At the base of this ventral arm the posterior margin of RIEF articulates with sclerite R1J
(A9 in fig.102, 103, 190, 193), or the sclerites are at least in close vicinity (A9 in
fig.163, 166). This corresponds to the position of articulation A9 of Eurycotis (compare
fig.77 and 78). Special features of Lamproblatta are the extension 20 of that part of
RIF bearing A9 and the immersion of the whole articulation.

— From A9 RIF extends to articulation A3; then it curves into the dorsal wall of the right
phallomere.

— At its dorsomedian end RIF has another articulation with RLJ (A8 in fig.99, 190) or,
at least, closely approaches R1J (A8 in fig.160). This corresponds to the position of
articulation A8 of Eurycotis (fig.74).

— This dorsal arm of RIF has, like in Eurycotis, a sclerotised groove at its dorsal margin,
between the articulations A3 and A8 (rge in fig.74, 77, 99, 102, 160, 163, 190, 193).

R1J (fig.99, 102-104, 160-164, 166, 190, 192-195, 197) bears both the articulations A8
(like RIH or region Rid in Eurycotis) and A9 (like R1G or region Riv in Eurycotis)
and is therefore regarded as a compound sclerite Rld+R1v (fig.331f,g,h, 332f,g,h). Thus,
in contrast to the situation in Mantodea, Eurycotis, and Archiblatta, there is now a very
broad connection between the Rid- and Rlv-regions posterior to membrane 17, and this
is clearly a derived feature. For Cryptocercus, Lamproblatta, and Eurycotis these relations

251

are confirmed by a comparison of the muscles, since R1J bears insertions which are in

Eurycotis either on R1H or on R1G:

— Muscle r3 of Cryptocercus and Lamproblatta (fig.166, 197) runs from that part of RIF
posterior to articulation A3 to the right margin of R1J. It is assumed to be homologous
with r3 of Eurycotis (fig.80), which inserts on R1G.

— Muscle r6 of Lamproblatta (fig.196) runs from the rge-groove to the dorsal wall of the
right phallomere, like r6 of Eurycotis (fig.79). The left insertion is partly on R1J in
Lamproblatta and on R1H in Eurycotis. Such a muscle is missing in Cryptocercus.

— Muscle rl of Cryptocercus (fig.165) runs from the age-apodeme on R3 to the dorsal
wall of the right phallomere, like rl of Eurycotis (fig.79). The posterior insertion is
partly on R1J in Cryptocercus and on R1H in Eurycotis. Such a muscle is missing in
Lamproblatta.

The fda-lobe of Tryonicus (fig.99, 102-104), Cryptocercus (fig.160-164, 166), and
Lamproblatta (fig.190-195, 197) largely corresponds to fda of Eurycotis. However, parts
of its ventral wall assumedly correspond to the pia-walls of Eurycotis (after having been
leveled). Thus, the fda-lobes of these species are not strictly homologous with fda of
Eurycotis. The levelling of pia is also a derived feature.

At least Lamproblatta has a similar flap-mechanism as Eurycotis (with fda being the flap
and the stout A8 and A9 defining the axis of movement). To what extent this is also
practised in Tryonicus and Cryptocercus is questionable since the articulations A8 and A9
are by far less distinct.

Polyphaga and Ergaula

In Polyphaga the posterodorsal part of the right phallomere contains the large sclerite

R1M and the smaller sclerites RIK and RIL (fig.134). The regions Ric, Rit, Rid, and

Riv can be identified and demarcated by their characteristic features (fig.3311, 3321), but

some points remain unclear.

— RIM articulates with R3 (A3 in fig.137) and forms a rge-groove on its dorsal margin
(from A3 to the posterior: fig.134, 137, 140), and rge bears the insertion of a stout
muscle (r6 in fig.140). These features resemble the Rlc-region of Eurycotis (fig.33le,i,
332e,1) and the other species. In contrast to the other species, the rge-groove extends
much farther posteriad (compare fig.74, 99, 160, 190).

— To the left of A3 RIM bends around an edge (16 in fig.137) to occupy the dorsal wall
of the cbe-invagination. The right part of muscle r2, coming from R3, inserts at the
anterior margin of this part of RIM (fig.140). More to the left this part of RIM forms
a ridge (pva in fig.139, 137, 138). These features resemble the Rit-region of Eurycotis
(fig.33le,i, 332e,i) and the other species. In contrast to the other species, the pva-ridge
is not transversely but longitudinally orientated (compare pva in fig.139 and 80, 99,
197). However, the shape of pva of Polyphaga is not so different from pva of
Lamproblatta (compare fig.139 and 197), if a lengthening of pva along the longitudinal
axis and a shortening along the transverse axis is assumed for Polyphaga.

— Corresponding to the probable lengthening of pva and rge to the posterior, it is assumed

252

b) Metallyticus
violaceus

f) Tryonicus parvus

e) Eurycotis floridana

d) Mantoida schraderi 332

Fig.332: Right phallomere, homology of main sclerites and homologous regions of main sclerite Ri
(ventral views). — The cuticular elements of the right phallomere are shown, but some membranous
parts are removed. Patterned areas are sclerotised, white areas are (except for sclerites R3 and R5)
membranous. Undulate lines are cutting lines. R3 and, in Blaberus, R5 are separated from the
remainder of the right phallomere and shifted anteriad. The individual sclerites of R1 are labelled
with the capital letters used in the text and in fig.1-319 (e.g. F = RIF). Articulations between sclerites
are labelled with the numbers used in the text and in fig.1-319 (e.g. 6 = A6). If an articulation has
been lost by fusion of the respective sclerites, the point of fusion is labelled by adding * to the name
of the lost articulation (e.g. 6* = A6*). The articulation points of A3 and A7 are connected by arrows.
cwe is a formative element.

253

R3

g) Lamproblatta \%
albipalpus \

i) Polyphaga

h) Cryptocercus aegyptiaca
punctulatus

k) Anaplecta sp.

R3

n) Blaberus
craniifer

m) Parcoblatta
lata

332

I) Nahublattella sp.

254

that in Polyphaga the regions Rit and Rlc have considerably expanded posteriad and
make up most of RIM (fig.3311, 3321).

— Muscle rl (fig.139) has its anterior insertion on the right margin of R3, like rl and r5
of Eurycotis. The dorsal part of rl has its posterior insertion in the dorsal wall of the
right phallomere (on the RIL-sclerites), like rl of Eurycotis (fig.79). This suggests that
the two RIL-sclerites are part of the Rid-region (fig.3311). The ventral part of rl has
its posterior insertion on the rge-groove (fig.134, 139), similar to r5 of Eurycotis (fig.80)
but more posteriorly. Thus, rl of Polyphaga is certainly homologous with rl of
Eurycotis but possibly also includes the homologue of r5. rl of Cryptocercus closely
resembles rl of Polyphaga, but a contribution of a r5-part is less probable since no
fibers insert on rge or RIF (fig.165).

— Eurycotis (fig.79) and Lamproblatta (fig.196) have the left insertion of r6 in the dorsal
wall of the fda-lobe and, by definition, in the Rid-region. r6 of Polyphaga has a very
similar course; its left insertion is on sclerite RIK, which is therefore assumed to belong
to the Rid-region (fig.3311). Thus, the Rid-sclerotisations of Polyphaga have become
rather fragmented (3 sclerites) and far removed from each other (as the insertions of rl
and r6 are). However, RIK and RIL could also be new elements not homologous with
sclerotisations of other Blattaria. RIK of Cryptocercus (fig.160) could well be
homologous with RIK of Polyphaga. However, since r6 is missing in Cryptocercus,
the somewhat similar position of the sclerites is the only indication for homology.

— Further parts of the regions Rld and Riv might be included in the posterior part of
R1M. Compared with Cryptocercus (fig.160, 163) or Lamproblatta (fig.190, 193), this
would correspond to a fusion of RIF and R1J across the membrane 17 and the
articulations A8 and A9. This is possibly indicated by the complete loss of muscle r3,
which in Cryptocercus (fig.166) and Lamproblatta (fig.196, 197) moves RIF and R1J
upon each other: The loss of r3 could be the consequence of such a fusion. In fig.3301,
3311, and 3321 Rid and Rlv are shown according to this assumption.

The posterior part of the right phallomere is assumed to be composed of the fda-lobe and
of the pva-ridge (fig.136-138). Like in Lamproblatta, Cryptocercus, and Tryonicus, the
ventral tooth pia has been lost, and its leveled vestiges are assumed to be contained in
the ventral wall of fda.

RIM of Ergaula is very similar to that of Polyphaga but narrower (compare fig.3301 and
m). RIL and RIK, however, are missing. The regioning of RIM is assumed to be the
same as in Polyphaga.

Anaplecta

The posterodorsal part of the right phallomere is, like in Tryonicus, Cryptocercus, and

Lamproblatta, an undivided lobe (fda in fig.226-230; no pia-tooth present). In contrast to

these species (with RIF and R1J), however, there is only one sclerite present (RIN in

fig.226-230), which somewhat resembles RIM of Polyphaga (fig.134, 137). The regioning

of RIN is assumed to be as follows (fig.331k, 332k):

— The part of RIN immediately posterior to articulation A3 (fig.229) is the Rlc-region.
However, the rge-groove is missing.

255

— The extension 34 of RIN, which to the left of A3 bends into the cbe-invagination
(compare fig.229 and 230; fig.226), exactly corresponds with the R1t-region of the other
species by its relative position, by its articulation with R2 (A6 in fig.226, 227, 230),
and by bearing the insertion of the right part of muscle r2 (fig.231) (compare fig.331k
and 331f,h). Rit of Anaplecta is somewhat thickened to the outside (pva) but does not
form a true ridge.

— The posterior main part of RIN takes the same position as the R1J-sclerite in Tryonicus,

Cryptocercus, and Lamproblatta and is probably composed of the regions Rid and Riv.

RIN is assumed to have developed by a fusion of the former RIF and R1J across the
membrane 17 and the articulations A8 and A9 (compare fig.160, 163 and 226, 229), as
it has also been assumed for RIM of Polyphaga (compare fig.33lı and k, 3321 and k).
Moreover, like in Polyphaga, muscle r3 has been lost. In contrast to Polyphaga, however,
the Rit-region retains the same degree of independence and the same transverse
orientation as it has in e.g. Eurycotis and Tryonicus, and there are no free sclerites RIL
and RIK. Thus, it is not clear if RIN of Anaplecta and RIM of Polyphaga are strictly
homologous and if the fusion of the former RIF and R1J and the loss of r3 are
homologous in the two species. (Therefore the sclerites are given different names).

The muscle connecting R3 and RIN (rl in fig.231) could be homologous with rl or r5
or both muscles of Eurycotis (fig.79, 80) and with the rl of Polyphaga (fig.139) and the
other species. Like in Polyphaga, the muscle will be named rl in Anaplecta (and in the
other Blattellidae and Blaberidae discussed below).

Nahublattella

The posterodorsal part of the right phallomere is, like in Anaplecta, an undivided lobe
(fda in fig.253, 256: no ventral tooth pia) with one sclerite (RIN’), but the Rit’-
sclerotisation seems to be missing (compare 34 in fig.226). However, similar to the left
end of 34 in Anaplecta, the left end of RIN’ (34 in fig.253) articulates with R2’ (A6 in
fig.254, 255, 226) and curves back to the right like a hook. Therefore, the Rit’-region is
assumed to have fused to the main part of RIN’ lying posterior to it (fig.3311). Apart from
this difference, RIN’ of Nahublattella is regioned in the same way as RIN of Anaplecta
(compare fig.331k and 1, 332k and 1). A peculiar feature of Nahublattella is the hinge-like
shape of articulation A3 (fig.253, 257). Muscle rl is certainly homologous with rl of
Anaplecta (fig.231, 259). Muscle r10 (fig.259) is specific to Nahublattella.

Supella

The posterodorsal part of the right phallomere is again an undivided lobe fda with one
large sclerite RIN’, but RIN’ has expanded over the whole fda-lobe and over the whole
dorsal wall of the cbe-invagination. The R1t’-region must have been firmly integrated into
this sclerotisation. Articulation A6, indicating the left end of the Rit’-region, is distinct
and, like in Nahublattella, on the summit of cbe (compare fig.253, 254). A hook-like or
curved sclerotisation near A6, however, is not present. Supella resembles Nahublattella in
probably having R1t’ completely integrated into RIN’, but because of the large expansion
of RIN’ in Supella the situations in the two species are hardly comparable.

256

Parcoblatta, Blaberus, and other Blattellidae and Blaberidae

The morphology of the posterodorsal part of the right phallomere of Parcoblatta and
Blaberus is in some repects very different from Anaplecta and Nahublattella. Concerned
are two areas, which will be discussed separately: (1) the Rit-region and (2) the dorsal
lobe fda. The essence of the changes having taken place can be understood by considering
the morphology of some more blattellid and blaberid species included in this investigation.

The Rit-region Within Blattellidae and Blaberidae the Rit-region (with pva)
undergoes some changes which also involve R2. These developments are shown in
fig.330n-s.

Anaplecta has Rit (fig.330n, 34 in fig.226) in the same relative position as e.g. Tryonicus
(fig 330h): situated in the dorsal wall of cbe, connected with Rlc to the right, articulated
with R2 to the left (A6). In contrast to Tryonicus, the left end of R1t shows the hook-
like curvature, which is a derived feature.

Euphyllodromia has a similar ribbon-like sclerotisation in the dorsal wall of cbe (fig.3300),
which by its position can be identified as the Rit’-region. The right end of this Rit’
approaches, like in Anaplecta, the Rl1c’-sclerotisation immediately behind articulation A3
but is narrowly separated from Ric’ by membrane (at cl in fig.3300). The left end of R1t’
shows, like in Anaplecta, a hook-like curvature, but this curved part 1s swollen to the
interior of the phallomere by extensive thickening of the cuticle (cwe in fig.3300).
Moreover, the left end of R1t’ is not articulated with R2’ but fused to it (at 6* in fig.3300).
Thus, the former sclerite RIN’ has divided (at cl) into two new sclerites: R1S’ (Rit’-
region, now firmly connected with R2’) and RIP? (rest of the former RIN’). The
separation of RI1t’ from Ric’, its fusion to R2’, and the cwe-thickening are derived
features.

Nyctibora shows the same situation (fig.330r), but Rit (sclerite R1S) and the rest of Ri
(sclerite RIP) are slightly farther removed from each other. (i.e. the two points of division,
called cl again, are farther away from each other). The cwe-thickening and its curvature
are very distinct (fig.319).

In Parcoblatta, Blaberus, and Byrsotria the fusion of R1t and R2 and the ewe-thickening
are very similar to Nyctibora (fig.282, 283, 285 and 309, 310, 313), and cwe marks the
border between R2 and RIt (with cwe belonging to Rit). However, the condition of the
right end of Rit varies: In Parcoblatta (fig.330q, 281, 282) this end of the R1t-region
(sclerite R1S) is still free. It has been far removed from its previous point of contact with
Ric (sclerite RIP) (or, in other words, R1t has been shortened; compare the cl-points in
fig.330q and r). Instead, it has approached the opposite end of sclerite R1P.

In Blaberus (fig.330s, 309) and Byrsotria (fig.318) the Rit’-region is firmly connected
with the rest of R1’. From the phylogenetic context, discussed later in 7.3., it follows that
this is due to a secondary fusion of the sclerites RIS and R1P and does not correspond
to the primary connection of these sclerotisations within the RIN-sclerite of Anaplecta,
Nahublattella, and Supella (fig.226, 253, 330n,0,p). Therefore, the resulting sclerite,
though having the same composition as RIN, is named differently: RIT’. (The
sclerotisations contained within RIN and RIT’ are homologous throughout but the

2]

sclerites themselves are not). Whether the fusion of RIS and RIP to form RIT’ had as
its starting point a similar situation as in Nyctibora, or if it was preceded by a shortening
of RIS like in Parcoblatta, is unclear. (In the regioning of RIT’ in fig.330s the former
situation has been assumed, compare fig.330r).

The Rit-morphology of all these species also shows that the complete incorporation of
Rit’ into sclerite RIN’ in Nahublattella and, in a different way, in Supella is in both cases
a special derivation, and that the situations in Euphyllodromia, Parcoblatta, Nyctibora,
and Blaberus are derived from a situation similar to Anaplecta (fig.330n), with Rit
connected with Ric only at its right end.

The pva-ridge on RIt is very low in Euphyllodromia, Parcoblatta (fig.282), and Nyctibora
(fig.319) and has been completely lost in Blaberus and Byrsotria.

The situation in Nahublattella could be interpreted in another way: That part of RIN’
which near A6 curves back to the right (right part of 34 in fig.253) could alone be the
Rit’-region, which is shortened like in Parcoblatta and, by this, far away from Ric’ with
its right end. According to this (improbable) interpretation, the R1’-morphology of
Nahublattella would be likewise much more primitive than in Euphyllodromia,
Parcoblatta, Nyctibora, and Blaberus: There would be no fusion between R1t’ and R2’,
and cwe would be missing. Instead, some features would have to be regarded as derived
peculiarities of Nahublattella: a fusion between the left end of Rit’ and the left end of
the posterior R1N’-sclerotisation (next to articulation A6); a reduction of the hook-
curvature at the left end of R1t’ (in the same area); an extreme shortening of Rit’ (which
in any case would be a parallelism as compared with Parcoblatta). In my view, the
interpretation of RIN’ of Nahublattella made above is by far more probable.

The dorsal lobe fda In Parcoblatta (fig.280, 281), Nyctibora (fig.319), Byrsotria
(fig.318), and Blaberus (fig.308, 309) the posterodorsal part of the right phallomere is not
an undivided lobe as in Anaplecta (fda in fig.226), Nahublattella (fda in fig.253), Supella,
and Euphyllodromia, but it is, from posteriorly, divided into two lobes lying one above
the other: dla (dorsally) and fda (ventrally).

Sclerite RIP of Parcoblatta resembles RIN of Anaplecta: Both sclerites articulate with
R3 (A3 in fig.226, 229, 281, 284), have a similar shape, and largely occupy the walls of
a posterior lobe (fda in fig.226, 281). Homology is assumed for RIP and RIN — minus
the Rit-region of RIN (compare fig.330n and q). Consequently, the ventral lobe fda of
Parcoblatta is assumedly the homologue of fda of Anaplecta. Apart from r2 (fig.286),
the right phallomere of both Parcoblatta and Anaplecta has only one further muscle (rl
in fig.231, 286), which has the same course and is assumed to be homologous. The
posterior insertion of rl is in the anteriormost dorsal wall of fda in Anaplecta but in the
anteriormost dorsal wall of dla in Parcoblatta. Thus, it can be assumed that the dla-lobe
is a new outfolding originating from the anterior dorsal wall of the formerly undivided
fda. Hence, fda of Parcoblatta is not strictly homologous with the fda of Anaplecta and
the other species. (Moreover, like in e.g. Anaplecta, the ventral fda-wall of Parcoblatta
probably still contains the leveled vestiges of the pia-walls. Thus, the homology between

258

the fda of Parcoblatta and the fda of e.g. Eurycotis, which has a well-developed pia-
tooth, is not strict in even two respects).

In Nyctibora (fig.319) sclerite RIP and the fda- and dla-lobes take the same relative
positions as in Parcoblatta (fig.280, 281). Additionally, however, there is a sclerite in the
dorsal wall of dla (R4 in fig.319, 330r). In Byrsotria (fig.318) the situation is essentially
the same as in Nyctibora, but sclerite R4’ is in two respects more derived: (1) It is
expanded to the right and in contact with sclerite RIT’ (59 in fig.318). (2) The left end
of R4’ (60 in fig.318) bends around the left edge of dla into its left ventral wall. In
Blaberus (fig.308) the situation is like in Byrsotria, but R4 is even further expanded to
the right and curves into the ventral wall of the right phallomere (59 in fig.308, 309) where
it closely approaches articulation A3. Moreover, the dla-lobe is extremely enlarged as
compared with the fda-lobe (compare fig.308 and 318). In Nyctibora, Byrsotria, and
Blaberus (fig.314) the posterior insertion of rl is, like in Parcoblatta, in the anteriormost
dorsal dla-wall, but it is also on sclerite R4. (According to the insertion of rl, R4 would
have to be classified, by definition, as a Rld-sclerotisation. But since R4 is most probably
a new sclerite and not a split off part of R1, the designation R4 is preferred).

The right phallomere of Blaberus has, in contrast to Parcoblatta and Anaplecta, not only
the plesiomorphic muscles rl and r2, but also some further, certainly new muscles. Two
of them run from R4’ to the right ventral wall of fda (rlla and rl1b in fig.314). One
muscle having the same course is also present in Nyctibora. The other muscles (r12-r18
in fig.314-317) have been found only in Blaberus (Byrsotria not studied); however, no
specimen of Blaberus had all these muscles.

6.8. The muscles connecting the left complex and the right phallomere

The b4-muscles have been discussed in 6.7.1., 6.7.3., and 6.7.5., muscle b3 of
Sphodromantis in 6.5.

b2-muscles are present in Sphodromantis (fig.15), Mantoida (fig.49), Polyphaga (fig.110,
127, 141), Lamproblatta (fig.184, 198), and Anaplecta (fig.224, 232). The position of the
right insertion is quite similar in all species: on the left part of R3 in Sphodromantis and
Mantoida, on the left margin of R3 in Polyphaga, next to the left margin of R3 and R2
in Anaplecta, and next to the left margin of R2 in Lamproblatta. That the b2 of
Lamproblatta and Polyphaga are homologous despite the somewhat different position of
the right insertion is strongly suggested by the very similar position of the left insertion:
in the right dorsal wall of the left complex, next to L8 and to the insertions of 19 and 112.
In Anaplecta, Mantoida, and Sphodromantis the left insertion is rather different from that
of Polyphaga and Lamproblatta: It is on the wall of the vla-lobe — near the right edge of
via in Mantoida and Anaplecta, and in the left anterior ventral wall of vla in Sphodro-
mantis. The b2 are certainly homologous in Polyphaga and Lamproblatta on the one hand
and in Mantoida and Sphodromantis on the other. Whether the b2 of these two groupings
and of Anaplecta are homologous and whether b2 is a muscle of the common ground-
plan of Blattaria and Mantodea or a derived element of these groupings is questionable.

259

bl-muscles are only present in Mantodea: Sphodromantis (fig.15, 16), Mantoida (fig.43).
Muscle b5 of Nahublattella (fig.235, 259, 261) and the muscles b6 (fig.294, 316) and b7
(fig.317) of Blaberus are certainly apomorphic muscles of these species.

6.9. The phallomero-sternal muscles

Phallomero-sternal muscles inserted on the right half of the subgenital plate have even
numbers, those inserted on the left half have odd numbers. The insertions on the subgenital
plate are designated as ventral, those on the phallomere complex or on the walls of the
genital pouch as dorsal. The anterior margin of the subgenital plate is the whole margin
between the right and left contacts between the subgenital plate and the paratergites of
segment 9 (compare fig.333a). Fig.333 gives an overview and a homology hypothesis for
the phallomero-sternal musculature of the species studied.

Mantoida has 6 phallomero-sternal muscles s1-s6, whose ventral and dorsal insertions are
arranged almost symmetrically (fig.40, 333b). In my view, this is the most primitive
situation within the species studied, and it is taken as the basis for the terminology. sl
and s2, s3 and s4, as well as s5 and s6 are assumed to be pairs of primitively symmetrical
muscles. Thus, three pairs of primary muscles will be distinguished, which in the
primitive situation have the following basic arrangement:

— sl and s2 insert on the median anterior margin of the subgenital plate and run to the
ventral basal line of the left complex or right phallomere, where they insert more
laterally.

— s3 and s4 insert on the median anterior part of the subgenital plate, closely behind s1
and s2, and run to the ventral basal line of the left complex or right phallomere where
they insert more medially than s1 and s2.

— s5 and s6 insert on the lateral anterior margin of the subgenital plate and run to the
lateral or lateroventral walls of the genital pouch.

Within this basic arrangement, the dorsal insertion of s2 and s4 is on the anterior margin
of sclerite R3, that of sl is on anteriormost parts of L4-sclerotisations (anterior L4l-
region), and that of s3 is on a membranous part of the basal line.

The evolution of the phallomero-sternal musculature comprises two kinds of development:

(1) The morphology of the 6 primary muscles changes by subdivision, fusion, shift, or

loss. The exact homology relations are often difficult to determine. (2) New secondary

muscles develop, each with a characteristic course by which the homology relations can

in most cases easily be determined. These secondary muscles are:

— s7 inserts on the median anterior part of the left side of the subgenital plate and runs
to the anterior part of the Ive-pouch.

— s8 inserts on the median anterior part of the right side of the subgenital plate and runs
to the top of the tre-tendon.

— s10 inserts on the median anterior part of the right side of the subgenital plate and runs
to the ejaculatory duct or to an area near its opening.

— s12 inserts on the median (or more lateral) anterior part of the right side of the subgenital

260

ol

subgenital plate

Sg
ventral insertion
*— muscle

dorsal insertion

T9p

ventral wall of genital pouch

t
: a ER ;
lve fo]

left complex right phallomere a) General scheme

u ee m nn mm

114 tre

Fig.333: Phallomero-sternal muscles, homology relations. — 333a gives an overview of the mode of
representation and the symbols and abbreviations used. 333b-1 (following pages) show the
morphology of the various species.

The anterior half of each figure shows the subgenital plate, with
— ventral insertions of muscles pl, p2, and p3 (4,®, and W).
— ventral insertions of all phallomero-sternal muscles.

— paratergites of abdominal segment 9 T9p.

The posterior half of each figure shows the phallomere complex, with

— ventral basal line Bl = border between ventral + lateral walls of phallomere complex (below BI)
and ventral + lateral walls of genital pouch (above Bl).

— border between left complex and right phallomere = median vertical broken line.

— muscle 114 (to hook hla; base of arrow = anterior insertion area).

— pouch lve (with L2-sclerotisations).

— ejaculatory duct D.

— tendon ate.

— tendon tre.

— sclerite R3.

— dorsal insertions of all phallomero-sternal muscles (@).

— if the dorsal insertion area of a muscle on or near the basal line of the left complex is sclerotised
this is shown by a stippled field around the insertion.

The shape of the subgenital plate is generalised, only the shape of the anterior margin (with apophyses
S9a) corresponds with the special condition in the respective species. The numbers in the ventral
insertion areas of the phallomero-sternal muscles are the numbers used in the text and in fig.1-319
to designate the muscles (e.g. 2 = s2, 5a = s5a). Species with “S” behind their names have the
phallomeres and the subgenital plate side-reversed, and a mirror-image of the original preparation is
shown. Dorsal views. Further information in 6.9..

1
f) Cryptocercus punctulatus

g) Polyphaga aegyptiaca

261

1
1
i
1
|
1
'
ı

h) Anaplecta sp.

k) Parcoblatta lata 1) Blaberus craniiferS

plate and runs to the ventral basal line of the left complex where it inserts immediately
to the right of s3.

— s14 inserts on the median anterior part of the right side of the subgenital plate and runs
to the ventral wall of the genital pouch beneath the rightmost part of the left complex.

Mantoida and Sphodromantis

Mantoida (fig.40, 42, 333b) conforms with the basic arrangement, except that the dorsal
insertion of s3 is posterior (not median) to that of sl. Sphodromantis (fig.5, 7, 333c) is
also close to the basic arrangement, but the right side has four muscles instead of three.

263

The arrangement of the insertion areas on the subgenital plate (compare s4 in fig.40 and
s4a and s4b in fig.5) and on the phallomere complex (s4a and s4b insert medially like s4
in Mantoida, compare fig.7 and 42) suggests that this is due to a division of s4. Both
species lack secondary muscles.

Eurycotis

sl, s2, s3, and s4 conform with the basic arrangement (fig.62, 64, 333d), but s2 is rather
weak. s5 and s6 have divided into three (s5a,b,c) or four (s6a,b,c,d) bundles, whose
insertions occupy a large area in the anterior half of the subgenital plate. However, most
of these bundles are very diffuse and indistinctly bordered to each other. Moreover, it is
not clear if really all these muscles are derivatives of s5 and s6 or if some of them are
new. Of the secondary muscles s7 and s8 (fig.58) are present.

Anaplecta

The left side has three muscles (fig.204, 207, 333h): The secondary s7 (fig.200) runs, like
in Eurycotis, to the lve-pouch. s5 can be identified by the lateral position of its ventral
insertion and by its dorsal insertion on the lateroventral wall of the genital-pouch. The
third muscle (named s3) might be sl or s3. It has its ventral insertion posterior to s7 like
s3 of Eurycotis (sl of Eurycotis inserts anterior to $7, compare fig.333d and h) and has
its dorsal insertion immediately anterior to 16a (fig.222) like s3 in Eurycotis (fig.70, 73),
Polyphaga, and Lamproblatta (discussion below; fig.133, 188). This muscle of Anaplecta
is therefore regarded as s3; sl is missing.

On the right side all three primary muscles are present (S2, s4, s6 in fig.204, 207, 333h).
Their ventral insertions resemble the basic arrangement (s2 and s4 on the median anterior
margin; s2 anterior to s4; s6 far laterally), but the s4-insertion is very broad, and s2 is
strongly reduced. The insertions on R3 are side by side: s4 on the left, s2 in a small central
area, $6 on the right. Thus, in contrast to the basic arrangement, the dorsal insertion of s6
has expanded to the anterior right margin of R3 (but still occupies parts of the membranous
ventral wall of the genital-pouch as well). The same is true of s6a of Eurycotis (fig.64,
333d), whose assignment to s6 is thus confirmed by the morphology of s6 of Anaplecta.
Noticeably, just sl and s2 have been reduced — two muscles which probably are a pair.

Apart from s7 there is another secondary muscle: s10, which, however, is not directly
inserted on the ejaculatory duct but more ventrally on the infolding between the lobes vla
and vfa (fig.221, 222; in fig.333h this infolding is symbolised by a transverse line beneath
the ejaculatory duct; compare discussion in 6.2.4.).

Lamproblatta and Polyphaga

On the left side, s1 and s3 of Lamproblatta (fig.173, 175, 333e) conform with the basic
arrangement, but the dorsal insertion of sl is rather far on the left (compare Eurycotis,
fig.70). In Polyphaga (fig.113, 116, 333g) both insertions of s3 and the dorsal insertion
of sl exactly correspond with Lamproblatta (sl: membrane anterior to the hla-hook,
fig.127, 185), but the ventral insertion of s1 has shifted leftward, too, and is not anterior
to but to the left of s3. Homology can certainly be assumed for sl as well as for s3. The

264

leftward shift of the dorsal sl-insertion, and that of the ventral sl-insertion in Polyphaga,
might be correlated with a parallel shift of the hla-base from the left ventral wall (fig.65)
to the left edge of the left complex (fig.117, 177). Lamproblatta has a lateral muscle (s5b)
showing the s5-insertions of the basic arrangement. Whether another muscle having a
much more median position (s5a) is a true part of s5 or a new muscle is unclear. In Po-
lyphaga s5 is completely missing — possibly in consequence of the leftward shift of s1.

On the right side, the s4 of Polyphaga and Lamproblatta conform with the basic
arrangement and are easily identified. To the right of s4 both species have another muscle
(named s6) running to the anterior margin of R3, which might be s2 or s6. Its ventral
insertion is lateral (not posterior) to s4 (fig.113, 173) and extends far laterad. This suggests
it to be s6. Its dorsal insertion on the right anterior margin of R3 only, not on the genital
pouch, however, might suggest it to be s2. But since the dorsal insertion of s6 of Anaplecta
and Eurycotis has expanded to the anterior margin of R3, and since s2 is small in Eurycotis
and very small in Anaplecta, these muscles of Lamproblatta and Polyphaga are interpreted
as $6; s2 is assumed to have been lost. The restriction of the dorsal s6-insertion to the
anterior margin of R3 (no longer in the wall of the genital pouch) would be a derived
feature of Polyphaga and Lamproblatta. (If the muscle should be s2, the laterad expansion
of its ventral insertion would be a derived feature).

Of the secondary muscles s12 is present (fig.133, 188): The dorsal and ventral insertions
are exactly the same in both species, and, in addition, the dorsal insertion is surrounded
in the same way by s3, 15, and 16a. s12 is peculiar to Lamproblatta and Polyphaga (and
Ergaula) and is assumedly a synapomorphy of these species. s8 is present in Polyphaga
(and Ergaula) but missing in Lamproblatta.

Cryptocercus

The left side (fig.147, 149, 333f) has only one muscle s1+3, with its ventral insertion
along the median (not the lateral) anterior margin of the subgenital plate and its dorsal
insertion only on and near the basal line (not in the more posterior walls of the genital
pouch, fig.158). By this extension of the insertion areas s5 is unlikely to contribute to this
muscle and is, like in Polyphaga, assumed to have been lost. sl+3 seems to be composed
of the previous sl and s3 since its dorsal insertion includes the areas occupied by sl and
s3 in Polyphaga and Lamproblatta (s1: anterior to the hla-base, compare fig.157 and 127,
185; s3: farther on the right, compare fig.158 and 133, 188). The dorsal insertion of s1+3
reaches with some fibers sclerite L2 in the lve-pouch; this insertion is typical for s7, which
may therefore also contribute to s1+3 (fig.333f).

The right side has one large primary muscle, s2+4+6, and, in some specimens only, a very
small one, s4b. The ventral insertion area of the large muscle includes, in contrast to $1+3
on the left side, the lateral anterior margin of the subgenital plate (fig.147, 333f). Thus,
this muscle assumedly contains not only the median s2 and s4 but also the lateral s6. This
is also suggested by the fact that the dorsal insertion (fig.168) extends far into the ventral
wall of the genital pouch (compare s6a of Eurycotis, fig.64). In this latter feature
Cryptocercus would be more primitive than Polyphaga and Lamproblatta, whose s6-
insertion is restricted to the anterior margin of R3. To what extent each s2 and s4 contribute

265

to the large muscle is unclear; s2 might also be missing like in Lamproblatta and
Polyphaga (as shown in fig.333f). The small s4b can, according to its ventral insertion
posterior to the large muscle (fig.147), only be a split off part of s4 or a new muscle but
certainly not s2.

It is difficult to interpret this highly peculiar condition of the primary muscles. At least,
the outgroup comparison between the other Blattaria and Mantoida strongly suggests that
it is not primitive for Dictyoptera. Possibly, extensive fusions of muscles have taken place.
However, there is still another possible explanation: It could be due to a retention of a
nymph-like situation in adult morphology (a neotenic trait), with the differentiation of the
single primary muscles not yet completed. This question could possibly be settled by an
investigation of the ontogeny of the phallomero-sternal muscles in other Blattaria.

Two secondary muscles are present: s8 to the tre-tendon and s10 to the ejaculatory duct
near its opening.

Nahublattella

Nahublattella (fig.237, 240, 3331) closely resembles Anaplecta, but the ventral insertions
of the median (on the apophyses S9a) and of the lateral (more posteriorly on the subgenital
plate) muscles are extremely far away from each other.

The secondary muscles s7 and s10 are easily identified by their insertions (fig.249). s3
can, like in Anaplecta, be identified by its ventral insertion posterior to s7 (compare
fig.333h and 1). sl is missing. The muscle from the right apophysis S9a to the anterior
margin of R3 is probably s4, not s2. (Since in Anaplecta s2 is reduced, a loss of s2 seems
for Nahublattella more probable than a loss of s4).

The lateral primary muscles s5 and s6 have undergone a division (or new muscles have
been added). The dorsal as well as the ventral insertion areas of s5a and s5b together have
the same extension as those of s5 of Anaplecta (compare fig.204 and 237, fig.207 and
240), and a division can readily be assumed. The same correspondence is found for the
ventral insertion areas of s6a and s6b or s6, respectively (compare fig.204 and 237).
Dorsally s6b of Nahublattella inserts on the anterior margin of R3 like s6 of Anaplecta;
the dorsal insertion of s6a, however, is completely different (compare fig.207 and 240).
Thus, for s6a the derivation from s6 is not certain. The question is the same for the very
delicate muscles s6c (fig.237, 240).

Parcoblatta and Blaberus

The phallomero-sternal musculature (fig.265, 267, 333k and 296, 298, 3331) can be derived
from Nahublattella but also shows some differences.

In Parcoblatta, the secondary muscles s7 and $10 as well as the primary muscle s4 conform
with Nahublattella. s3 also inserts like in Nahublattella but is divided into two bundles
s3a and s3b (compare fig.237, 240 and 265, 267). Dorsally s3b inserts on the ate-tendon,
s3a to the left of ate (fig.267). sl and — with the same reservations as in Nahublattella —
s2 are missing. Concerning these muscles, the situation in Blaberus is the same except
that s7 is missing. (However, s7 is present in its typical position in Nauphoeta, another
member of Blaberidae; fig.3281).

266

The lateral primary muscles of the left side (s5) are in Blaberus divided into three bundles
(s5sa — 2 bundles — and s5b in fig.296, 298). s5a and s5b resemble s5a and s5b of
Nahublattella in their dorsal insertions: s5b near the anterior margin of L4U’ (fig.250,
304); s5a in the left ventral wall of the genital pouch (fig.240, 298). The ventral insertions
are similar, too, but in Blaberus s5a has shifted posteriad (fig.237, 296). The division of
s5 into s5a and s5b is assumedly homologous in Nahublattella and Blaberus. The smaller
median muscle of Blaberus, also named s5a, is regarded as a median subdivision of s3a.
In Parcoblatta s5 shows a similar and certainly homologous division: The ventral and
dorsal insertions of s5b are similar to Blaberus (fig.265, 267, 296, 298), but the dorsal
one is somewhat more posteriorly. The ventral and dorsal insertions of s5a are, as
compared with Blaberus, by far more posteriorly.

The lateral primary muscles of the right side (s6) are in Blaberus present as two bundles
(s6a and s6b in fig.296, 298), which are certainly homologous with s6a and s6b of
Nahublattella: The dorsal insertions of s6a and s6b take the same positions as in
Nahublattella (fig.240, 298), but s6b has considerably expanded posteriad and now
occupies the whole right margin of R3’. As regards the ventral insertions, s6a has, as
compared with s6b, shifted far posteriad (like s5a on the left side!). sb of Parcoblatta
shows the same division: The dorsal and ventral insertions of s6b are situated like in
Blaberus (fig.265, 296), and the dorsal insertion likewise occupies the whole right margin
of R3. The dorsal and ventral insertions of s6a are, as compared with Blaberus, shifted
even farther posteriad (like those of s5a on the left side!).

Only Parcoblatta and Blaberus have the secondary muscle s14, which is divided into two
bundles in Parcoblatta (s14a,b). The ventral insertion is closely behind s4 (fig.265, 296,
333k,l). The dorsal insertion is immediately anterior to s6a in Blaberus, but, corresponding
to the posteriad shift of s6a, far anterior to s6a in Parcoblatta.

6.10. The subgenital plate and associated structures

The subgenital plate is poor in complex structures, and hardly any character is valuable
for the phylogenetic analysis in the frame of this study.

In most species the subgenital plate is asymmetrical, but the degree of asymmetry varies,
and various parts are concerned: mainly the styli S9s and the dorsal sclerotisation S9d in
Anaplecta (fig.204); the posterior edge of the plate in Polyphaga (fig.113, 114); mainly
S9d in Sphodromantis (fig.5), Lamproblatta (fig.173), Parcoblatta (fig.265), and Blaberus
(fig.296); S9d and the apophyses S9a in Nahublattella (fig.237, 238); the apophyses S9a
and the lateral and posterior edges in Mantoida (fig.40); the whole subgenital plate in
Metallyticus (fig.22). Only in Chaeteessa (fig.30), Eurycotis (fig.62), Tryonicus (fig.86),
and Cryptocercus (fig.147) the subgenital plate is symmetrical or nearly so.

The apophyses S9a can be very different in their length and distinctness (compare
Chaeteessa, fig.30, and Parcoblatta, fig.265). Cryptocercus has no apophyses at all
(fig.147). As explained in 3.1., the areas designated as apophyses S9d are not in all species
homologous in a strict sense, but short apophyses may be homologous with only the
anterior parts of long apophyses. Such relations are obvious if Parcoblatta (fig.265) is

267

compared with Blaberus (fig.296): In Parcoblatta the recess between the apophyses
extends far beyond the insertions of sl4a,b and nearly reaches the level of the p3-
insertions. In Blaberus the recess does not even reach s14, much less p3. Thus, either the
recess has deepened in Parcoblatta, or the posterior parts of the apophyses have fused
with each other in Blaberus. To compare more distantly related species in this regard,
however, is hardly possible.

The area designated as the dorsal sclerotisation S9d is rather variable and is certainly not
strictly homologous in all species. In Metallyticus (fig.22), Eurycotis (fig.62), Tryonicus
(fig.86), Polyphaga (fig.114), Lamproblatta (fig.173), Anaplecta (fig.204), Parcoblatta
(fig.265), and Blaberus (fig.296) S9d is restricted to the more posterior part or even to
the margins of the dorsal wall of the subgenital plate (= posterior ventral wall of genital
pouch), and it is firmly connected with the ventral sclerotisation around the lateral and
posterior edges of the plate. In Sphodromantis (fig.5) and Cryptocercus (fig.147) S9d is
also continuous with the ventral sclerotisation but extends by far more to the anterior. In
Mantoida (fig.40), Chaeteessa (fig.30), and Nahublattella (fig.237, 238) S9d also extends
far anteriad but is isolated from the ventral sclerotisation. In Nahublattella S9d is highly
elaborated (division, bristles, muscles 133 and 134; fig.238, 240).

In several species, some patterns in the sclerotisation of the subgenital plate remind one
of its presumable composition (true sternite, two coxites; compare in 3.1.) and might
therefore be regarded as primitive. In Eurycotis (fig.62), Polyphaga (fig.113), Cryptocercus
(fig.147), and Nahublattella (fig.237) an anterior (sternite?) and a posterior (transversely
fused coxites?) sclerotisation are separated by a membranous field — except for a
lengthwise connection of the lateralmost parts. In Parcoblatta (fig.265), Blaberus (fig.296),
and Lamproblatta (fig.173; the anterior sclerotisation is very narrow) the field is no longer
membranous but still distinctly weaker sclerotised than the other parts. In Tryonicus
(fig.86) the sclerotisation is weaker in the anterior third, but there is no heavier
sclerotisation along the anterior margin. In Mantoida such a zoning of the plate is only
slightly indicated (not shown in fig.40). In the other Mantodea (fig.5, 22, 30) and in
Anaplecta (fig.204) the sclerotisation of the subgenital plate is uniform. An interesting
feature of Cryptocercus (fig.147) is that in the posterior part of the plate the lateral areas
are distinctly heavier sclerotised than the median area; the transverse fusion of the coxites
is probably not complete.

6.11. The peripheral muscles

The muscles pl, p2, and p3 are, if present, always inserted close to each other on the
anterior margin of the subgenital plate, between the median and the lateral primary
phallomero-sternal muscles (fig.333b-l; fig.5, 40, 62, 113, 147, 173, 204, 237, 265, 296).

The pl-muscles are ventral muscles of segment 9. Their posterior insertions are on or
close to the Pv-sclerites or, if separate Pv-sclerites are absent, on the anterior margin of
the paraprocts Pp. The pl are rather weak (Sphodromantis, fig.1; Mantoida, fig.36;
Polyphaga, fig.109) or even consist of very few fibers only (Eurycotis, fig.58; Anaplecta,

268

fig.200; Blaberus, fig.293), or they are completely missing (Lamproblatta, Nahublattella,
Parcoblatta). In Polyphaga, Anaplecta, and some specimens of Eurycotis the pl are
divided into two bundles on one or on both sides.

In Blaberus the pl are also divided into two bundles on each side; the one bundle shows
the usual insertion on the anterior margin of the paraproct, but the other bundle inserts on
its posterior margin (fig.293), and this is certainly a derived feature. According to
McKittrick (1964), the posterior insertions of the ventral muscles of segment 9 take the
same positions in the females of most Blaberidae, too.

In Cryptocercus the pl (fig.143a) are extremely broad. This is unlikely to be a primitive
state since the pl are by far narrower in all other Blattaria as well as in Mantoida and
Sphodromantis. In last-instar nymphs of Blaberus and Eurycotis the pl are by far broader
than in the respective adults (though not as broad as in Cryptocercus). Hence, the
broadness of pl of Cryptocercus might be a neotenic trait.

The p2-muscles are dorsoventral muscles of segment 9. They are either very delicate
(Eurycotis, fig.58; Polyphaga, fig.109; Cryptocercus, fig.143a; Blaberus, fig.293) or
completely missing (Mantoida, Sphodromantis, Lamproblatta, Anaplecta, Nahublattella,
Parcoblatta).

A special feature of Eurycotis is that the p2 as well as the serially homologous muscles
of abdominal segment 8 pass through two pairs of eyelets in the vasa deferentia (as shown
in fig.58: Vd, p2, and p2(8)). Snodgrass (1937) finds such eyelets, in the same
arrangement, also in Blatta orientalis (Blattinae); Pipa (1988), fig.7, describes eyelets for
Periplaneta americana. | have additionally investigated the vasa deferentia of Deropeltis,
Periplaneta, Parcoblatta, and Blaberus: There are no traces of eyelets in Blaberus. Eyelets
or vestiges of them have been found in Periplaneta, Deropeltis, and Parcoblatta, but either
the passage is more or less narrowed, or there is only a thickening of the vas deferens
without any passage. The degree of eyelet reduction can be rather different in the four
places (often asymmetrical; this was also the case in some specimens of Eurycotis) and
in different specimens of a species. If passages were present in these species, these were
never passed through by muscles (though very thin p2 were often present). Pipa (1988),
however, finds the p2 passing through the eyelets in Periplaneta (S-9 in his fig.7).

In last-instar nymphs of Eurycotis and Blaberus p2 and p2(8) are by far stronger than in
the adults of the same species, and they all run through eyelets in the vasa deferentia. The
eyelets and their penetration by p2 and p2(8) are assumed to be nymphal features, which
in the adults can be retained to rather various extents (even within a single species). The
same seems to be true of the muscles themselves. A far-reaching retention of these
structures in the adult is thus regarded as a neotenic trait.

The p3-muscles (rectal muscles) are present in all species and have a similar fan-shape
throughout. In Cryptocercus they are divided into two fans on each side.

The p4-muscles have their anterior insertions always far laterally on the anterior margin
of tergite 9 T9. In many species they additionally extend onto the paratergites T9p
(Mantoida, fig.36, 37; Cryptocercus, fig.143a; Lamproblatta, fig.169, 170; Anaplecta,
fig.200, 201; Parcoblatta, fig.262, 263; Blaberus, fig.293, 294). In the latter case, except

269

in Cryptocercus, the p4 are divided into several bundles: throughout their length in
Mantoida and Parcoblatta; only anteriorly in Lamproblatta, Anaplecta, and Blaberus.

The posterior insertions take rather different positions: on the lateral anterior margin of
tergite 10 T10 (Cryptocercus, fig.143a,b), or on the anterior margin of the paratergites 10
T10p (Sphodromantis, fig.1; Mantoida, fig.36, 37; Parcoblatta, fig.262, 263), or in the
membrane median to T10p (Anaplecta, fig.200); in this latter case they can be far
anteriorly (Lamproblatta, fig.169, 170; Polyphaga, fig.109; Blaberus, fig.294) or
extremely far medially (left p4 of Eurycotis, fig.58; Nahublattella, fig.235).

In Cryptocercus the insertion on tergite 9 and that on tergite 10 take the same relative
position (compare fig.143a and b), and p4 is clearly a dorsal muscle of segment 9. The
p4 of the other species are assumed to be the same dorsal muscles, and the posterior
insertion is assumed to have undergone a ventromediad and anteriad shift which is
variously pronounced in the different species. The homology of these p4 is suggested by
the constant position of the anterior insertion and by the following fact: In last-instar
nymphs of Eurycotis and Blaberus the posterior p4-insertion is by far more laterally than
in the adults; that means, it shifts mediad during late ontogeny. In the various species, the
final position of the posterior p4-insertion in the adult might depend on the extent to which
the adult character state prevails against the nymphal state. A dorsolateral position (like
in Cryptocercus, fig.143b) is probably a neotenic trait. However, in some species this could
also be a primitive feature.

In Periplaneta americana, whose posterior p4-insertions have a similar ventromedian
position as in Eurycotis (compare fig.58), the innervation of p4 is known (Pipa 1988): It
is accomplished by a nerve-branch (the common base of 4Alc and 4A1d in Pipa) which
innervates, apart from p4 (359, 360 in Pipa), the various groups of dorsal muscles (M and
MDMS9 in Pipa). This is consistent with the assumption that even those p4 having their
posterior insertions far medially are true, though modified, dorsal muscles.

The p5-muscles are dorsoventral muscles of segment 10. Dorsally they always insert on
the lateral anterior margin of tergite 10 T10. Their ventral insertions are on or near the
Pv-sclerites (Eurycotis, fig.58; Lamproblatta, fig.169; Anaplecta, fig.200; Nahublattella,
fig.235) or, if separate Pv-sclerites are missing, on the anterior margin of the paraprocts
Pp (Sphodromantis, fig.1; Polyphaga, fig.109; Cryptocercus, fig.143a; Parcoblatta,
fig.262; Blaberus, fig.293). Hence, the position of the p5-insertions (Pv-sclerites or
anterior margin of paraprocts Pp) differs in the same way as in the pl-muscles. The
insertions of pl and p5 suggest that in those species without separate Pv-sclerites the Pv-
sclerotisations have become incorporated into the anterior part of the paraprocts. (This
part of the paraprocts is then labelled Pv in the figures). Moreover, the insertions of pl
and p5 (ventral muscles of segment 9, dorsoventral muscles of segment 10) suggest that
the Pv-sclerites (or the Pv-parts of the paraprocts) are sternal sclerotisations of abdominal
segment 10. However, this question cannot be finally settled here.

The p6-muscles are dorsoventral muscles of segment 9. The dorsal insertion is always far
laterally on tergite 9 T9. The ventral insertion is either close to the line of contact between
the lateral margin of the subgenital plate and the paratergite of segment 9 T9p (Eurycotis,
fig.69; Cryptocercus, fig.146; Lamproblatta, fig.172; Anaplecta, fig.203; left muscle of

270

Polyphaga, fig.112a) or slightly posterior to this area (Mantoida, fig.39; Sphodromantis,
fig.4). In Polyphaga the ventral insertion has expanded into the lateral wall of the genital
pouch (left muscle, fig.112a) or has completely shifted to this area (right muscle, fig.112b).
In Nahublattella (fig.237), Parcoblatta (fig.265), and Blaberus (fig.296) the p6 insert
distinctly more medially on the subgenital plate.

The p7-muscles have their anterior insertions far medially in the membrane anterior to
paraprocts and Pv-sclerotisations; their posterior insertions are far laterally where the
paratergites 10 T10p meet the paraprocts Pp (articulations A99; lateral to the posterior
pl-insertions). p7 is well-developed in Mantoida (fig.37), Sphodromantis (fig.2),
Lamproblatta (fig.170), and Cryptocercus (fig.144). In Lamproblatta the posterior (or
lateral) insertion of the left p7 has distinctly shifted anteriad. In Eurycotis p7 is represented
by only very few fibers (fig.59). In the other species no p7 have been found.

The muscles p8 and p9 will not be discussed: Their homology relations are uncertain since
they are not clearly distinguishable from other muscles of the anal region. The muscles
p10 of Cryptocercus (fig.144) are probably subdivisions of the p5-muscles.

The muscles pl-p7 are certainly present in the common ground-plan of Blattaria and
Mantodea. p3 and p5 are very uniform in the species studied. The differences in the
morphology of pl, p2, and p4 are assumed to be of limited value for a phylogenetic
analysis, because these differences probably depend on the extent to which nymphal
features are retained in the adult. As regards p6, the mediad shift of the ventral insertions
could be a synapomorphy of the species concerned.

6.12. The terminal part of the abdomen

The homology relations of most elements of this area are quite evident and need no
discussion. The homologies concerning the supraanal lobe spl, the epiproct Ep, and the
tergite 10 T10 are discussed in 3.1. The homology between the Pv-sclerites and the anterior
part of the paraprocts Pp (in species without separate Pv-sclerites) is discussed in 6.11..
There are hardly any features valuable for a phylogenetic analysis, but the following
features are worth mentioning and might gain some more value in future investigations
including more species.

The area where the paraproct Pp, the Pv-sclerite, and the paratergite 10 T10p meet each
other shows in several species some peculiarities. However, the ground-plan condition of
this area is in most respects uncertain since Mantodea seem to have this area highly
modified — similar to but certainly independently of certain Blattaria — and since the
outgroup comparison with other Ectognatha suffers from the uncertainty of homology
relations. Hence, the plesiomorphic or apomorphic nature of these peculiarities is
debatable. A rather primitive condition might be assumed to be represented in e.g.
Eurycotis (fig.59): The lateral tip of Pp articulates (A99) with the ventromedian tip of
T10p laterally and is in close vicinity to the lateral end of a completely free Pv anteriorly.
If the Pv-sclerites really represent the medially divided sternite 10 (compare in 6.11.),
their complete isolation could be plesiomorphic. Tryonicus (fig.83), whose Pv-sclerites are

271

fused to the paraprocts laterally, would then have a more derived state of this character.
In the species having no separate Pv-sclerites by fusion to Pp (e.g. Mantoida, fig.37,
Polyphaga, fig.110) this character would be even more derived. The fusion of the
paraprocts Pp and the paratergites T10p in Lamproblatta (partial; fig.170), Anaplecta,
Mantoida, and Sphodromantis (complete; fig.201, 37, 2), corresponding to a partial
(Lamproblatta) or complete loss of A99, is certainly a derived feature. The presence of
two articulations per side is peculiar to Lamproblatta (A97 and A99 in fig.169, 170). The
interpretation in this species is done in accordance with Eurycotis (fig.59) and Tryonicus
(fig.83): The lateral articulation is the true A99; A97 is assumed to be a new articulation
within the paraproct Pp.

In several species tergite 10 T10 has undergone a complete longitudinal division by a
median stripe of membrane (Polyphaga, fig.109; Nahublattella, fig.234; Blaberus, fig.293)
— certainly a case of threefold parallel evolution. The membranous area 21 of Lamproblatta
(fig.169) might represent an early stage of such a division.

The articulation A98 between the cercal base and tergite 10 has been lost only in
Polyphaga and Blaberus — certainly another case of parallel evolution.

The ventral sclerotisation of tergite 10 T10v is only in Anaplecta separated from the dorsal
main part of T10 (fig.200). However, the T10v-sclerites of Anaplecta could also be
homologous with the Ce-sclerites of the other species (compare fig.200 and e.g. 58).

The various paired sclerites median to the cercal base (Ca, Cb, Cc) are certainly
homologous in the way expressed by the designations. All three pairs are present only in
(some) Blattaria but not in Mantoida and Sphodromantis (Chaeteessa and Metallyticus not
investigated). Sclerites median to the cercal base are also present in e.g. Caelifera
(Snodgrass 1935, fig.7), but whether there is any kind of homology with the Blattarian
sclerites 1s unknown. It is therefore also unclear if some or all of these sclerites are
elements of the Dictyopteran ground-plan or derived features of Blattarian subgroups or
of Blattaria as a whole. Ca-sclerites are present in Eurycotis (fig.58, 59), Tryonicus (fig.83,
84), Lamproblatta (fig.169, 170), Anaplecta (fig.200), Nahublattella (fig.234, 235), and
Parcoblatta (fig.262, 263), and they are crescent-shaped in most species. Except in
Tryonicus and Lamproblatta the Ca extend along distinct curved Ca-bulges. Cc-sclerites
are present in Eurycotis (fig.58, 59), Tryonicus (fig.83, 84), Lamproblatta (fig.169, 170),
and possibly Anaplecta (fig.200: T10v?). Cb-sclerites are peculiar to Lamproblatta
(fig.169, 170). In Polyphaga, Cryptocercus, and Blaberus all three pairs are missing, but
in Polyphaga and Blaberus at least the Ca-bulges are distinct.

A distinct supraanal lobe spl has been found in Mantodea (fig.1, 36) and in Eurycotis,

Tryonicus, Cryptocercus, Lamproblatta, Parcoblatta, and Blaberus (fig.58, 83, 143a, 169,
262, 293). An epiproct Ep is present in Mantodea (fig.1, 36) but never in Blattaria.

6.13. The asymmetry of the phallomere complex

The right phallomeres of the Mantodean species, especially Chaeteessa, and of Eurycotis
are very similar in the arrangement of the sclerotisations (R1, R2), the formative elements
(invagination cbe, lobe fda, tooth pia, ridge pva, apodeme age), the main muscles (rl,

272

r2, r3, s2, s4), and some morphological details (keel 3, edge 16). The right phallomeres
of all other Blattaria studied can be derived from that of Eurycotis without any problems;
especially the area comprising sclerites R2 and R3, invagination cbe, and muscle r2 is
very similar in all species. Therefore, homology is assumed for all these right phallomeres.
This assumption also includes those species with the right phallomere situated on the left
side (Nahublattella, Supella, Euphyllodromia, Byrsotria, and Blaberus investigated in this
paper): The right phallomeres of these species can be integrated into the homology
hypothesis without any problems, and the right phallomeres of Blaberus and Byrsotria
(situated on the left side) and the right phallomere of Nyctibora (situated on the right side)
are nearly identical. (The only principal difference is the fusion of RIP and RIS to form
RIT in the two blaberid species).

The left complexes of Mantoida and of Archiblatta and Eurycotis are quite similar in the
principal arrangement of the sclerotisations (L1, L2, L4), the formative elements (e.g.
pouches Ive and pne, ventral lobe vla, apodeme swe, processes paa and pda), the main
muscles (12, 13, 14, 16, 19, sl, s3), the genital opening, and some morphological details
(L4d-region). Most of the morphological gaps between these species are bridged by other
Blattaria, e.g. Tryonicus (shape of paa and pda and relation between them), Polyphaga
(shape of sclerites L1 and L2, position of phallomere-gland opening), or Cryptocercus
(muscle I1). The left complexes of the other Blattaria (e.g. Parcoblatta) can be extremely
different from those of Archiblatta and Eurycotis, but the morphology of each species can,
if several other species are included in the comparison, be traced back to the basic pattern.
Therefore, homology is assumed for all these left complexes. This assumption likewise
includes those species with the left complex situated on the right side (Nahublattella,
Supella, Euphyllodromia, Blaptica, Nauphoeta, and Blaberus investigated in this paper):
The left complexes of these species can be integrated into the homology hypothesis without
any problems. The left complex of Blaberus (situated on the right side) and the left
complex of Parcoblatta (situated on the left side) are very similar; concerned are the
principal arrangement and shape of most cuticular elements, the course of most muscles,
as well as many details (ate-tendon, hge-groove, notch 45). Differences between Blaberus
and Parcoblatta are in most cases bridged by other species of Blattellidae and Blaberidae:
Loboptera and Nyctibora (orientation as in Parcoblatta) have, like Blaberus, a sclerite
L4U, which is missing in Parcoblatta. Nauphoeta (orientation as in Blaberus) has, like
Parcoblatta, a muscle s7, which is missing in Blaberus. The lve-apodeme and the via-
process of Nyctibora (orientation as in Parcoblatta) and Nauphoeta (orientation as in
Blaberus) are very similar and do not show the strong differences as present between
Parcoblatta and Blaberus (which are due to the differently directed rotation of this area).

From the homology of the right phallomeres and from that of the left complexes it follows
that the asymmetry of the whole phallomere complex is homologous in all species studied.
Thus, the asymmetry of the phallomere complex is a feature of the common ground-plan
of Blattaria and Mantodea (and maybe Isoptera). Moreover, from a comparison of the
ground-plan morphologies of the left complex (fig.32le,g) and of the right phallomere
(fig.321f,h) it follows that the asymmetry of the phallomere complex was in the common
ground-plan of Blattaria and Mantodea already as extreme and of the same very special
kind as in the extant species. For the investigated members of Plectopterinae

273

(Nahublattella, Euphyllodromia, Supella) and Blaberidae (Nauphoeta, Byrsotria, Blaptica,
Blaberus) it is thus evident that the phallomere complex has undergone a change of its
left-right-asymmetry (like a mirror-image): This hypothesis first proposed by Bohn (1987)
is strongly supported, and many new arguments are now available. According to e.g. Bohn
(1987), the phallomere complex of most species of Ectobius (Ectobiinae) is normally
orientated, but some species show the same orientation as Blaberidae. The latter species,
as compared with the former, clearly show that a reversal of the left-right-asymmetry must
be regarded as a possible evolutionary pathway.

Vestiges of a bilateral symmetry or side-homologies within the phallomere complex are
revealed in only very few respects: (1) The primary phallomero-sternal muscles sl and s2
are assumed to be side-homologous (e.g. fig.37), and side-homology might also be
assumed for the areas of their dorsal insertions. In the primitive case (compare in 6.9.)
these are the anterior L4l-region on the left side and the anterior margin of R3 on the
right side. These two areas have additionally in common that the margin of the
sclerotisation is more or less groove- or beam-like (apodemes swe or age). (2) The
transverse phallomere muscles (b-muscles) might be assumed to have primitively a
symmetrical course, and their left and right insertion areas might be side-homologous. The
situation in Mantoida might be interpreted in this way: The insertions of muscle bl (fig.43)
are next to those of the side-homologous s3 (left side) and s4 (right side), and the resulting
side-homology would again concern the (median) anterior margins of R3 and L4. (3) The
dorsal transverse muscles b4a and b4b (fig.48, 58, 109) have their right insertions close
to each other, but the left insertions are quite distant from each other. From their course
it can be at most deduced that there is some kind of side-homology between the
dorsomedian parts of the left complex and of the right phallomere. As regards the re-
maining parts of the right phallomere and of the left complex, there are in no species any
similarities in the positions, in the special shapes, or in the muscular connections of
elements which show similar spatial interrelationships on both sides. Hence, no further
side-homologies can be reliably assumed.

The primary phallic lobes of nymphal Blattaria and Mantodea are certainly homologous
with those of the other Ectognatha (in a more or less strict sense). In most other Ectognatha
(also in the most primitive: Archaeognatha, Zygentoma) the external genitalia developing
from these phallic lobes are bilaterally symmetrical. Hence, the male external genitalia
have certainly been bilaterally symmetrical in some early members of the common stem-
group of Blattaria and Mantodea. However, from this it cannot be concluded that there
must be extensive vestiges of this bilateral symmetry in the sclerotisations, in the formative
elements, or in the musculature of the phallomere complex of the extant species (or of
the later members of the common stem-group): The sclerotisations, muscles, and formative
elements present in the common ground-plan of Blattaria and Mantodea (or more or less
extensive parts of them) might have evolved later — at a time when the extreme asymmetry
had already established. Nothing is known about homology relations between these
Dictyopteran phallomere elements and the elements of the male genitalia in other insect
groups, and hence there is no information about which elements have already been present
when the Dictyoptera branched off from their (unknown; Kristensen 1995) sister-group.

274

7. THE GROUND-PLAN AND THE EVOLUTION OF THE PHALLOMERE
COMPLEX AND THE PHYLOGENY OF BLATTARIA AND MANTODEA

In the sections of chapter 6 many features of the common ground-plan of Blattaria and
Mantodea have been reconstructed. This ground-plan will be given completely in 7.1.

In 7.2. and 7.3. the evolution of the phallomere complex will be described for Mantodea
and for Blattaria. This will be done in accordance with a phylogenetic hypothesis which
results as the most parsimonious solution from the distribution of the phallomere character
states analysed in chapter 6. In this description, all derived character states present in the
various subgroups will be listed, and these derivations are regarded as autapomorphies of
the respective subgroups. For each autapomorphy the section of chapter 6. in which the
respective feature has been discussed will be given. Most of the phallomere characters are
consistent with each other in the distribution of their states over the subgroups defined in
7.2. and 7.3., and this phylogenetic hypothesis is thus highly supported. Some derived
character states which appear as autapomorphies of single species in the frame of the
sample of species included in this investigation and which are uninformative in the present
analysis will also be mentioned, since in later investigations they might be detected in
other species, too, and might then serve as synapomorphies and help in integrating further
species into this phylogenetic hypothesis.

A survey of all assumed aut/synapomorphies is given in 7.4. — together with a phylogenetic
tree (diagram 1) showing the most parsimonious solution.

For some characters the polarity of the states does not become unambiguously clear from
the discussions and informations given in chapter 6, and the respective interpretations
given in 7.1.-7.4. are not yet sufficiently substantiated. The evolution of these characters
and the polarity of their states will be discussed in 7.5. The single topics will be designated
with letters and referred to in 7.2. and 7.3.

For some characters there will, despite the previous discussions, remain some doubt in
terms of polarity. In some other characters whose polarity is rather clear the distribution
of the character states over the species is in some way inconsistent with the phylogenetic
hypothesis in 7.2. and 7.3. The respective (possibly or clearly) derived character states
and the groupings they would suggest will be listed in 7.6. All these inconsistencies will
be also mentioned in 7.2. and 7.3.

In the following discussions, assemblages of species regarded as holophyletic are called
“subgroups” and numbered according to their hierarchy. Assemblages not regarded as
holophyletic are called “groupings” and are designated with capital letters for cross
reference. The character states which are assumed to be autapomorphies of subgroups, and
also the character states whose role as possible autapomorphies is discussed but regarded
as improbable, are, for easy reference, numbered (bold printed and put in brackets, e.g.
(45)).

275

7.1. The common ground-plan of the phallomere complex of
Blattaria and Mantodea

Fig.32la-p shows the cuticular elements and the muscles of the phallomere complex and

of the other parts of the male postabdomen as they have assumedly been in this ground-

plan. For some characters, however, the ground-plan state could not be resolved, since

— the respective elements are (probably) present in the ground-plan of Blattaria but
completely absent in Mantodea: presence or absence of hla-hook, nla-process, dca-
processes, tre-tendon, sclerites L3 and R2, articulations A6, A7, A8, and A9, muscles
114, r6, s7, and s8.

— the respective elements are present in the ground-plan of Mantodea but completely
absent in Blattaria: presence or absence of muscles r4 and bl.

— the condition of the respective elements is different in the ground-plan of Mantodea and
in that of Blattaria: connection or separation of the L4-sclerotisations L4v/L4c, L4l,
and L4n in the anterior ventral wall of the left complex; presence or absence of the
curvature (dorsad and back to the left) of the right parts of L2 and Ive; connection or
separation of the Rl-regions Rid and R1v posterior to the membranous area 17.

— the homology relations between Blattaria and Mantodea or within Blattaria are
questionable: presence or absence of loa-process, L5-sclerite, L4c-region, muscles 17,
113/b3, and b2.

In fig.321 the elements or properties concerned are omitted or supplied with question-
marks. As regards all the data in 7.1., compare in 6.1.2., 6.2.2., 6.3.2., 6.4.2., 6.5., 6.7.2.,
8.10.9) and 7.5.

Cuticular elements

Left complex

L1 is an undivided sclerite in the central dorsal wall of the left complex. A large anterior
part of L1 (L1a-region) is situated in a deep and distinct pne-pouch and is hood-shaped
(but not plateau-like anteriorly). The right posterior part of L1 is a distinct arm-like
extension (Lim-region). There is possibly another arm-like extension formed by the left
posterior part of L1 (Lil-region). Lil and Lim do not join each other ventrally to form
a sclerite-ring (no region LIr). The membranous part of the pne-wall has a roughly dorsal
position and receives the opening of the phallomere-gland P. A completely sclerotised loa-
process is probably present. Whether there are dca-processes is not decidable (omitted in
fig.321).

The pouch Ive lies ventral to the pne-pouch. L2 is an undivided arch-shaped sclerite which
extends along the edges 7 of the lve-pouch and is (almost) completely restricted to its
dorsal wall. Whether the right parts of Ive and L2 are level or curved dorsad and back to
the left is not decidable (curvature omitted in fig.321). The right end of L2, or its dorsal
left end if the right parts of L2 are up- and recurved, (L2m-region) shows a narrow (not
hinge-like) articulation A2 with Lim. The left end of L2 (L2p-region) leaves the Ive-
pouch posteriorly, bends into the dorsal wall, and forms the sclerotisation of the paa-
process (L2d-region). paa is completely sclerotised, short, and somewhat upcurved. The

276

ventral wall of the lve-pouch is mostly membranous and is at the same time the left anterior
part of the dorsal wall of the vla-lobe. The ejaculatory duct D opens into the right anterior
part of the dorsal vla-wall. Both the presence of a small sclerite L5 in the dorsal vla-wall
and, if present, its exact position are questionable.

The ventral wall of vla is part of the ventral wall of the left complex and is sclerotised
by the right posterior part of L4 (L4v-region). The ventral wall of vla is for most or all
of its length confluent with the remaining ventral wall of the left complex (i.e. only most
posteriorly vla has a free left edge 61, or edge 61 is missing). The lve-pouch and the vla-
lobe are rather broad but do not reach the left edge of the left complex.

The L4-sclerotisations as a whole form an arch (open posteriorly) in the ventral wall and
at the left edge of the left complex. The L4-regions L4l, L4d, L4n, and L4v (and possibly
L4c) are present. The ventral wall between these regions is membranous. The morphology
of the L4l- and L4d-regions is like in Mantoida and Archiblatta: LAI is undivided and
extends over the left anterior margin and the whole left edge of the left complex. In the
dorsal wall LA is restricted to the left margin. The posteriormost part of L4l sclerotises
a short bulge-like process pda, which takes a position immediately to the left of the paa-
process and whose sclerotisation is connected with the L2d-sclerotisation of paa. The swe-
apodeme extends over most of the length of L4l. In its anterior part swe is beam-shaped
by cuticular thickening, in its posterior part it is groove-like. L4d is distinctly prominent
from the outline of L4l and directed to the right (and possibly slightly anteriad). The L4n-
region is present; whether it is connected with or separated from the L4l-region cannot
be decided. The nla-process on L4n is possibly also present (according to its functional
correlation with 114 and hla; discussion in 7.5. (M), (N)). Whether the L4c-region is
present is not decidable. (If it should be present, it is certainly firmly connected with the
L4v-region right-posterior to it. At least, there is no separate sclerite L4F). Whether the
L4v-region (or the L4c-region, if present) is connected with or separated from the anterior
end of the L4l-region is unclear. The hla-hook and its L3-sclerite are probably present.
If this is true, the ground-plan condition of hla and L3 can be assumed to correspond with
the ground-plan situation in Blattaria (compare in 7.3.).

Right phallomere

Sclerite R3 lies more or less transversely in the anteriormost ventral wall. At least its right
margin and the right part of its anterior margin form a groove- (or somewhat beam-) like
apodeme age, which reaches the A3-articulation. The age-groove bears a keel 3. The right
posterior end of R3 has an articulation A3 with the Ric-region. Posterior to R3 the ventral
wall of the right phallomere curves dorsad and somewhat anteriad to form a large
invagination cbe.

The posterior part of the right phallomere is composed of a dorsal lobe fda and of a ventral
tooth pia. fda and pia are confluent along the right edge of the right phallomere and
diverge to the left. R1 is probably an undivided sclerite (or, with less probability, it is
divided into three sclerites RIF, R1G, and R1H, corresponding to the regions Ric+RIt,
Rlv, and Rid, by the articulations A8 and A9). RI occupies the area behind the A3-
articulation (Ric-region), part of the right-dorsal wall of the cbe-invagination (RIt-

DT

region), the dorsal (and possibly part of the ventral) wall of fda (Rid-region), and the
dorsal and ventral walls of pia (Rlv-region). The regions Ric and Rit form a distinct
angle along the edge 16. Along its posterior margin Rit has a posteriad-directed ridge
pva. At and near the posterior right edge of the right phallomere there is a membranous
area 17. The dorsal and ventral parts of R1 are (probably) connected anterior to this area
but separated (or only narrowly connected) posterior to it.

The presence of sclerite R2 and of its articulations A6 and A7 is questionable. The
membranous stripe 4 and probably the articulations A8 and A9 and the tre-tendon are
missing.

Musculature
Muscles certainly present in the ground-plan (all shown in fig.3211-p)

l1: from the dorsal wall of the pne-pouch to the L4d-region (6.1.1., 6.3.4.).

12: from sclerite L1 in the left wall of the pne-pouch to the swe-apodeme on the L4l-
tesion (0:1. 1.,:6.3.1.).

13: from sclerite L1 in the posterior ventral wall of the pne-pouch to the L2a-region in
the anterior dorsal wall of the Ive-pouch (6.1.1., 6.2.1.).

14: from the L2a-region anteriorly in the lve-pouch to the swe-apodeme on the L4l-region
(insertion on swe ventral to 12) (6.2.1., 6.3.1.).

15: from the anterior ventral wall of the left complex to the left(-anterior) edge of the Ive-
pouch (6.2.1.).

16a: from the anterior ventral wall of the left complex to the right(-anterior) edge of the
Ive-pouch (6.2.1.).

l6b: from the ventral wall of the left complex to the membranous ventral wall of the Ive-
pouch at or near the genital opening (ventral insertion posterior to 16a) (6.2.1.).

19: transversely within the dorsal wall of the left complex (6.5.).

b4a, b4b: connect the dorsal parts of the left complex and of the right phallomere; left
insertions in the right marginal area of the Ive-pouch (b4a) or in the dorsal wall dorsal to
the pne-pouch (b4b); right insertions on the dorsal anterior margin of the fda-lobe (on
the tre-tendon, if it is present in the ground-plan) (6.7.1.).

rl: from the right margin of sclerite R3 (to the right of keel 3) to the Rid-region in the
dorsal wall of the fda-lobe (6.7.1.).

r2: from sclerite R3 to the cbe-invagination and the Rlt-region (and to sclerite R2, if it
is present in the ground-plan) (6.7.1.).

r3: from the Ric-region to the Riv-region in the dorsal wall of the pia-tooth (6.7.1.).
sl: from the left median anterior margin of the subgenital plate to the lateral ventral basal
line of the left complex (on the L4l-region) (6.9.).

s2: from the right median anterior margin of the subgenital plate to the lateral ventral basal
line of the right phallomere (on sclerite R3, to the left of keel 3) (6.9.).

s3: from the left median anterior part of the subgenital plate (posterior to sl) to the median
ventral basal line of the left complex (on membrane) (6.9.).

s4: from the right median anterior part of the subgenital plate (posterior to s2) to the
median ventral basal line of the right phallomere (on sclerite R3) (6.9.).

278

s5: from the left anterior margin of the subgenital plate to the left(-ventral) wall of the
genital-pouch (6.9.).
s6: from the right anterior margin of the subgenital plate to the right(-ventral) wall of the
genital-pouch (6.9.).

Muscles possibly present in the ground-plan (except for 114 omitted from fig.3211-p)

17: longitudinally within the posterior left ventral wall of the left complex (6.5.).

113, b3: from the ejaculatory duct D to the dorsal wall of the vla-lobe posterior to or to
the right of the genital opening (6.5.).

114: from the L4n-region on or near the nla-process to sclerite L3 and the hla-hook
(6.4.2.).

r4: from the Rid-region in the left dorsal wall of the fda-lobe to the left ventral wall of
the fda-lobe (6.7.1., 6.7.3.).

r6: from the Ric-region to the Rld-region (6.7.1., 6.7.6.).

bi: from the left margin of sclerite R3 to the anteriormost ventral wall of the left complex
(6.8.).

b2: from the left margin of sclerite R3 to an area next to the right end of the Ive-pouch
(6.8.).

s7: from the left anterior margin of the subgenital plate to the Ive-pouch (6.2.4., 6.9.).
s8: from the right anterior margin of the subgenital plate to the tre-tendon (6.7.1., 6.9.).

Asymmetry

From the features of the ground-plans of the left complex and the right phallomere it
follows that the very special kind of extreme asymmetry — corresponding to the extreme
differences between these two ground-plans — is also a feature of the ground-plan.

7.2. The evolution of the phallomere complex and the phylogeny in Mantodea
(= subgroup 1.)

The ground-plan of the phallomere complex of Mantodea

The features listed subsequently can be ascribed to the ground-plan of Mantodea since
they are true either of all investigated species (cuticular elements) or at least of
Sphodromantis and Mantoida (muscles), which are representatives of the two basal sister-
groups (subgroups 1.1. and 1.2., see below). For all these features it is not clear whether
they belong to the common ground-plan of Blattaria and Mantodea or whether they are
autapomorphic for Mantodea.

The following elements are absent: sclerotisations L3, R2; formative elements dca-
processes, hla-hook, nla-process, tre-tendon; articulations A6, A7, A8, A9; muscles 114,
r6, s7, s8. The L4-sclerotisations L4v/L4c, L4l, and L4n are firmly connected with each
other in the anterior ventral wall. The right parts of L2 and Ive are level. The loa-process
is present. The membranous part of the pne-wall is right-dorsal. The regions Rid and
Riv are not connected with each other posterior to the membranous area 17. The muscles
r4 and bl are present.

279

Subgroup 1.1.: Mantoida

The phallomere complex of Mantoida is rather close to the Mantodean ground-plan but
also has some derived features: On the left complex, the L4n-region has been lost (6.3.1.,
6.3.3.). On the right phallomere, the Rit-region (with the pva-tooth) has been separated
from the Ric-region (6.7.1., 6.7.3.; like in Metallyticus: compare (G) in 7.5. and grouping
B (123) in 7.6.).

Subgroup 1.2.: Chaeteessa + (Metallyticus + Sphodromantis)

There are some conspicuous autapomorphies on the left complex: The ventral wall has
developed the L4b-region which occupies all the interspaces between the primary L4-
regions L4v/L4c, L4n, and L4l and makes the ventral wall completely, or nearly so,
sclerotised (6.3.3.; in the primitive case L4b is distinctly weaker than the primary L4-
regions). The swe-apodeme has been lost or reduced to vestiges (6.3.3.). In the dorsal wall
the L4l-region has strongly expanded to the right (6.3.3.). The L4d-region, if present at
all, is no longer prominent from the outline of the L4l-region (6.3.3.). Possibly in
correlation with this expansion of L4l, the pne-pouch has rotated (clockwise as seen from
behind), and the membranous part of its wall is somewhat more on its right side (6.1.3.).
The distal part of the right posterior extension Lim of sclerite L1 curves into the dorsal
wall of the Ive-pouch, and Lim and L2 are therefore in the same plane in the area of
articulation A2 (6.1.3.). The Ive-pouch has become distinctly narrower (6.2.3.). Sclerite
L2 in its dorsal wall has lost its primitive arch-shape (probably by a fusion of the arms
of the arch, L2p and L2m) and is now ribbon- or plate-like (6.2.3.; compare (B) in 7.5.).
At least the paa-process has distinctly lengthened (6.2.3.). Whether the pda-process has
lengthened is not assessable since pda has been lost in Chaeteessa. The sclerotisations of
pda and paa (or, to apply this character state also to Chaeteessa, the L4- and L2-
sclerotisations of the corresponding area) have been separated from each other (6.2.3.;
compare (A) in 7.5.).

On the right phallomere, the Ric-region has been divided by the membranous stripe 4
separating the sclerites RIA and R1B (RIA and RIC in Metallyticus) (6.7.1., 6.7.3.).

Subgroup 1.2.1.: Chaeteessa

The left complex has a membranous pouch pbe between the pne- and Ive-pouches (fig.34).
The pda-process has been lost (6.2.3., 6.3.3.). The loa-process has also been lost (assuming
that at least the loa of Mantoida and Metallyticus + Sphodromantis are homologous;
6.1.3.). The vla-lobe has considerably broadened as compared with the narrow Ive-pouch
(fig.32), and the genital opening has come into a position far to the right of the Ive-pouch
(6.2.3.). On the right phallomere, sclerite R3 bends to the right along its left margin (6.7.3.,
fig.32). The keel-apodeme 3 has been lost (6.7.3.). The utmost right-posterior part of the
age-groove (near articulation A3) has been reduced (6.7.3.).

Subgroup 1.2.2.: Metallyticus + Sphodromantis

Most of the autapomorphies are on the left complex: The dorsal and ventral parts of L4
have been separated by articulation Al at the left edge of the left complex, which divides

280

the L4l-region (sclerites L4A and L4B; 6.3.3.). The ventral wall is uniformly sclerotised
since the L4b-region has become as heavy as the primary L4-regions (6.3.3.). The L1-
extension Lim, which curves into the dorsal lve-wall, as well as articulation A2 have
become much broader (6.1.3.). The curving part of Lim now also sclerotises the afa-
process on the anterior part of the edge 1 between the pne- and Ive-pouches (6.1.3.; this
autapomorphy is uncertain since the homology of afa with the elements called afa in
Mantoida and Chaeteessa is not certain, and since afa of Metallyticus is nothing but a
shallow bulge). Sclerite L1 is, at least in its posterior part, divided by a stripe of membrane
2 within the L1m-region (6.1.3.). On the right phallomere, the deepening of the left part
of the age-groove is very abrupt — certainly also a derived condition (6.7.3.).

Subgroup 1.2.2.1.: Metallyticus

The restriction of L4 (L4B-sclerite) to the anterior part of the dorsal wall seems to be an
autapomorphy since in both Chaeteessa and Sphodromantis L4 or LAB occupies the whole
dorsal wall (6.3.3.). The phallomere-gland P has probably been lost (at least, it was not
found; 6.1.3.). The Rit-region (with the pva-tooth) has separated from the Ric-region
(6.7.1., 6.7.3.; like in Mantoida: compare (G) in 7.5. and grouping B (123) in 7.6.). The
Rid-region has expanded into the ventral wall of the pia-tooth and has largely ousted the
membranous area 17 (6.7.3.).

Subgroup 1.2.2.2.: Sphodromantis

There are several autapomorphies: The Lim-extension has become extremely broad
(fig.323a; 6.1.3.). Sclerite L1 is now completely divided by the membranous stripe 2
(sclerites L1A and L1B; 6.1.3.). The afa-process is highly elaborated (fig.10; 6.1.3.). The
membranous part of the pne-wall with the phallomere-gland opening has undergone a
further rotation to the ventral side of the pne-pouch (6.1.3.). The sclerotisation of the loa-
process has been strongly reduced (6.1.3.). The tongue-like deepening of the anterior part
of the Ive-pouch and of L2 to the left is certainly also a derived feature since lve and L2
are narrow in both Metallyticus and Chaeteessa (6.2.3.). On the right phallomere, the
crescent-like curvature and the extreme deepening of the left part of the age-apodeme and
the resulting pouch 5 in the ventral wall of the genital pouch are derived features (fig.6;
6.7.3.). The posterior part of the pia-tooth and its Riv-sclerotisation have been reduced
(0:72.39):

LaGreca (1954) investigated the phallomeres of species of Amorphoscelididae,
Eremiaphilidae, Hymenopodidae, Mantidae, and Empusidae. Not many of the characters
dealt with in this chapter are recognisable in the figures of LaGreca or discussed in his
text, but at least three features are present in all these species: (1) L4 always occupies the
whole ventral wall of the left complex: the L4b-sclerotisation is present. (2) The
sclerotisations of paa and pda are always separated from each other. (3) L4 is always
divided into a dorsal (L4B) and a ventral (L4A) sclerite: the articulation Al is present.
(1) and (2) suggest that these families all belong to subgroup 1.2.; (3) additionally suggests
that they all belong to subgroup 1.2.2.

281

7.3. The evolution of the phallomere complex and the phylogeny in Blattaria
(= subgroup 2.)

The ground-plan of the phallomere complex of Blattaria

The features listed subsequently can be ascribed to the ground-plan of Blattaria since they
are true either of all investigated species or at least of representatives of the two basal
sister-groups (subgroups 2.1. and 2.2., see below). For all these features, except for the
last-mentioned (reduction of muscle s2), it is not clear whether they belong to the common
ground-plan of Blattaria and Mantodea or whether they are autapomorphic for Blattaria.

The following elements or properties are assumed to be present in the ground-plan since
they are found in all Blattaria: sclerotisations L3, R2; formative element hla-hook. The
L4v-region (in close connection with the L4c-region, if L4c is present) is separated from
the regions L4l and L4n. The regions Rid and Riv are connected with each other
(narrowly in the primitive case) posterior to the membranous area 17.

The following elements or properties occur only in part of the investigated species but are
assumed to be present in the ground-plan since they are found in representatives of the
two basal sister-groups: formative elements dca-processes, nla-process, tre-tendon, rge-
groove; articulations A6, A7, A8, A9; muscles 114, r6, s7, and s8. The right parts of L2
and of the Ive-pouch curve dorsad and back to the left. The base of the hla-hook is in the
left anterior ventral wall of the left complex, and the introversible membranous basal part
30 of hla is narrow. The dea are two membranous cushions posterior to L1.

Another ground-plan feature of Blattaria might be that muscle s2 is distinctly thinner than
its left counterpart s1 (6.9.); this situation is distinct in Eurycotis (fig.62: s2 weak),
Polyphaga, and Lamproblatta (fig.113, 173: s2 lost). In Cryptocercus the condition of s2
is not assessable; in Anaplecta, Nahublattella, Parcoblatta, and Blaberus s2 is also weak
or absent, but, in addition, sl has been completely lost. The situation in the Mantodean
ground-plan, with sl and s2 of similar stoutness (Mantoida, fig.37, 40), is regarded as
more primitive than in all Blattaria since both muscles as well as their symmetry are
preserved. An asymmetry in the stoutness of sl and s2 is assumed to be an autapomorphy
of Blattaria.

Subgroup 2.1.: Archiblatta + Eurycotis (and Periplaneta, Blatta, Deropeltis)

Periplaneta, Deropeltis, and Blatta have been studied only in part, but at least all the
derived features listed subsequently are also present in these species. The muscles have
not been studied in Archiblatta.

Most of the autapomorphies are on the left complex: The anteroventral part of the Ive-
pouch is, like a tongue, deeply invaginated to the left (6.2.1.). The posteroventral part of
the Ive-pouch has strongly receded to the right (or is even more reduced: Eurycotis; 6.2.1.).
Correlated with the latter feature is that the paa-process has shifted far to the right and
is far away from the pda-process (6.2.1.). The sclerotisations of pda (L4l-region) and paa
(L2d-region) have, possibly again in correlation with the previous feature, been separated
from each other (like in some other subgroups: compare (A) in 7.5.; 6.2.1.). The L4c-
region is highly elaborated (or L4c is as a whole an autapomorphy of this subgroup), and

282

there is a distinct L4F-sclerite (6.2.1., 6.3.1.). The ventral insertions of the 15-muscles (15a
and I5b in Eurycotis) have shifted posteriad and take a position on L4F (6.3.1.; 15a,b
could also be new muscles, but this would be a derived condition, too). The left insertion
of muscle b4b has shifted to the top of the pne-pouch (6.1.4., 6.7.1.). The anterior part
of sclerite L1 (Lla-region) has been leveled, and muscle Il has been lost (6.1.4.; these
two derived features, however, are also present in other subgroups of Blattaria: compare
grouping M (24) and grouping R (25) in 7.6.). On the right phallomere, the Rit-region
has enlarged and occupies most of the cbe-invagination including its anterior wall (6.7.4.).

Subgroup 2.1.1.: Archiblatta

On the left complex, the paa-process has lost most of its sclerotisation (L2d-region;
6.2.1.). On the right phallomere, the Rit-region has expanded to sclerotise the entire cbe-
invagination and has developed a broad connection with sclerite R2 (6.7.4.). Whether the
condition that the anterior L4c-region forms an isolated sclerite L4E is an apomorphy of
Archiblatta (and Periplaneta, Blatta, Deropeltis) or the plesiomorphic state within
subgroup 2.1. is not decidable (6.3.1.).

Subgroup 2.1.2.: Eurycotis

As compared with Archiblatta, the cuticular elements of the left complex show many
derived features: The L4d-region, if present at all, is no longer prominent from the outline
of the L4l-region (6.3.1.). The posteroventral part of the Ive-pouch is extremely reduced
(6.2.1.). Sclerite L2 has lost the arch-shape and is plate-like: the arms of the arch, L2m
and L2p, are assumed to have fused (6.2.1.; like in the Mantodean subgroup 1.2.; compare
(B) in 7.5.). Within the vla-lobe there is a deep incision 9 (6.2.1.). The mla-lobe covering
L4F has evaginated from the ventral wall (6.3.1., fig.63, 69). The pne-pouch has become
less distinct (6.1.4.). The left-dorsal wall of pne contains two probably new sclerites
L6A,B (6.1.4., 6.5.).

As compared with Periplaneta, Blatta, and Deropeltis, the muscles also show some derived
features: The right insertion of 12 has shifted from sclerite L1 to the adjacent membrane
(6.1.4.). Muscle 13 has divided into three bundles I3a,b,c (6.1.4., 6.2.1.).

The following derived features have not been investigated in the other species of subgroup
2.1., and it is unclear whether they are autapomorphies of Eurycotis or of a larger
holophyletic group within 2.1.: 115 (fig.70) and some muscles in the mla- and vla-lobes
(113-group except for 113h, 6.5.; 116, 117, 118; fig.71-73) are new. The presence of a
separate muscle r5 might also be derived (fig.80; this could be a new muscle or a split
off part of rl).

Subgroup 2.2.: Tryonicus + (Cryptocercus + (Lamproblatta + (Ergaula + Polyphaga)))
+ (Anaplecta + (Nahublattella + (Supella + (Euphyllodromia + (Parcoblatta + (Nyctibora
+ (Blaberus + Nauphoeta + Blaptica + Byrsotria}))))))

The most important synapomorphy of all these species is that the L4l-region reveals the
same division (at A5 or (A5) in fig.329e-h): As far as the various substructures are
preserved, and as far as no further subdivisions of the L4l-region have occurred, the

283

anteroventral part of L4I bears the left insertions of 12 and 14 and forms a sclerite (L4K)
together with the L4n-region or its vestiges. The posterodorsal part contains the pda-
sclerotisation and, at its connection with the paa-sclerotisation (L2d-region), the posterior
110-insertion and forms a sclerite (L4N) together with the L4d-region. The L4l-, L4d-,
and L4n-morphology of all species comprised in this subgroup follows this description
(muscles not known in Tryonicus) or can be derived from this situation (6.3.1., 6.3.4.).
(This division is completely different from the division of the L4l-region in the Mantodean
subgroup 1.2.2. where the I2- and 14-insertions are together with L4d on the dorsal L4B
and the pda-sclerotisations together with the L4n-region on the ventral L4A; fig.329c).

Additionally, the swe-apodeme has been completely lost (6.3.1., 6.3.4.). (swe has also been
reduced in the Mantodean subgroup 1.2. The loss might be correlated with the division
of L4l in both groups, for which region there is now no longer any need to be stiffened.
For Chaeteessa, however, this explanation does not fit).

In its ground-plan subgroup 2.2. probably possesses a muscle 110 from the Ive-pouch to
the common sclerotisation of paa and pda. (However, this feature is not investigated in
Tryonicus, and homology is not certain for 110 of Cryptocercus. In some members of
subgroup 2.2.3.2.2.2. 110 is missing, but this is certainly a secondary loss, compare (R)
in 7.5). 110 might be a posterior part of the ground-plan muscle 14, which might have
divided together with the L4l-region (its left insertion area); in this case, the similar
division of 14 and the shift of the posterior part of its left insertion to the paa- and pda-
sclerotisation would be an autapomorphy of this subgroup.

The L4d-sclerotisation has rotated (counterclockwise as seen from above): In Tryonicus
L4d is directed anteriad; in the other species L4d is directed to the left, or, after a further
rotation, dorsad (Lamproblatta), or it has been lost (6.3.4.).

On the right phallomere, the pia-tooth has been lost (6.7.6.). The regions Rid and Riv
have developed a broad connection at the posterior edge of the fda-lobe (i.e. the former
sclerites RIG and RIH have broadly fused to form R1J; 6.7.6.). (In some more derived
species R1J has additionally fused with RIF, the sclerites RIM or RIN being the results).

A possible autapomorphy is the extreme reduction of muscle s2 (more than in the Blattarian
ground-plan and in subgroup 2.1.); this feature, however, has not been investigated in
Tryonicus and is not assessable in Cryptocercus (6.9.). Another possible autapomorphy is
the sclerite-ring formed by the posterior part of L1 (by the regions LH, Lim, and LIr;
Culm os compare (FP) in7:3.).

For this subgroup 2.2. there are two possibilities for the next subordinate sister-group
relation; both are supported by derived character states or possible autapomorphies. Hence,
there is a trichotomy not resolvable with the present state of knowledge. Alternative B,
followed in fig.322 and 330, might be more probable.

Alternative A: Holophyly of Subgroup 2.2.1. + Subgroup 2.2.3. is supported by two
derived character states of the hla-hook: The introversible membranous basal part 30 of
hla has become more extensive (hla can therefore be retracted more deeply into the left
complex; 6.4.3.). The base of hla has shifted posteriad (6.4.3.). These two features are
possibly intercorrelated (compare (M), (N) in 7.5.). A posteriad shift of the hla-base,
however, is also present in Cryptocercus (fig.151; compare in 7.7.).

284

Alternative B: Holophyly of Subgroup 2.2.2. + Subgroup 2.2.3. is supported by three
derived character states of the regions L4l and L4d: The anterior and posterior parts of
the L4l-region (in the primitive case included in the L4K- and L4N-sclerites) are still
hinged to each other ın Tryonicus (articulation A5) but always far removed from each
other in the species of these subgroups (6.3.4.). The utmost right-anterior part of the L4l-
region, which in Tryonicus extends rightward anterior to the L4n-region, has been lost
(6.3.4.). The L4d-ribbon has further rotated (counterclockwise as seen from above) and
is now directed to the left. (In Lamproblatta L4d has additionally rotated into a
dorsoventral orientation; in subgroup 2.2.3. L4d has been preserved only in Nahublattella;
6.3.4).

Subgroup 2.2.1.: Tryonicus (parvus and angustus)

Synapomorphies of the two species of Tryonicus investigated in this paper are the rotation
of the pne-pouch (counterclockwise as seen from behind; extreme in T. parvus; 6.1.4.),
the enlargement and plate-like condition of the L1m-region, and the consequently hinge-
like condition of articulation A2 (6.1.4.). The anterior part of L1 has become level (6.1.4.;
like in the subgroups 2.1. and 2.2.3.; compare grouping M (24) in 7.6.).

Derived features of 7: parvus (characters not investigated in T. angustus) are the extension
R2m of sclerite R2 and the loss of the age-apodeme (6.7.4.; both features also in Lam-
problatta: compare grouping G (129) and (130) in 7.6.). The sclerite bridge L3a which
connects the L4n-region and L3 and crosses the hla-base 30 is probably also derived
(6.4.3.).

Subgroup 2.2.2.: Cryptocercus + (Lamproblatta + (Ergaula + Polyphaga))

On the left complex, sclerite L4K has been reduced in a specific way (6.3.4.): The parts
of L4K which in the ground-plan of subgroup 2.2. (as in Tryonicus) take a position right-
ventral to the hla-base have been lost. This concerns mainly the L4n-region, and the
anterior insertion of 114 (muscle lost in Polyphaga and Ergaula) is at least mostly on
membrane. (In Polyphaga and Ergaula this reduced L4K has shifted to the ventral side
of the hla-base). Probably in consequence of this L4n-reduction, the nla-process has been
lost (6.3.4.).

Muscle 12 shows a shift to the anterior: This concerns a gradual anteriad shift of the left
insertion (6.3.4.; least distinct in Cryptocercus: insertion still on sclerite L4K, 1.e. on region
L4l) and a complete anteriad shift of the right insertion from the left wall of the pne-
pouch to its top (6.1.4.). That the anterior face of the pne-pouch, i.e. of sclerite L1, has
become plateau-like (6.1.4.; distinct in Cryptocercus, Polyphaga, and Ergaula; plateau
vestigial in Lamproblatta) is possibly correlated with the shift of 12 and is also assumed
to be an autapomorphy of this subgroup. That this feature of pne is obsolete in Lampro-
blatta is assumed to be a secondarily derived condition, possibly correlated with the
apomorphic right-anteriad shift of pne by which the insertion angle of I2 on pne has
become very acute.

Another possible autapomorphy might be the complete loss of muscle s2, which feature,
however, is not assessable in Cryptocercus (6.9.). (The loss of s2 in subgroup 2.2.3.2. is
certainly a case of parallel evolution since s2 is present in Anaplecta.)

285

Subgroup 2.2.2.1.: Cryptocercus

The left complex has several derived features: The right part of L2, which is upcurved in
other Blattaria, has been reduced; in consequence, the respective right-dorsal part of the
Ive-pouch has been reduced, and the contact between L2 and L1 (articulation A2) has
been lost (6.1.4., 6.2.4.). The left edge 61 of the vla-lobe has expanded almost to the
anterior margin of the left complex; in consequence, the left-ventral part of the lve-pouch
has also been extremely reduced (6.2.4.; a similar derived state is present in Tryonicus
and in Anaplecta: compare grouping C (124) in 7.6.). The pda-sclerotisation (posterior
L4l-region) has been reduced and also separated from the paa-sclerotisation (L2d-region)
(6.2.4., 6.3.4.; the separation of L4l and L2d has also developed in the subgroups 1.2.
and 2.1.: compare (A) in 7.5.). New muscles of Cryptocercus are 119 (6.4.3.), r7 (6.7.5.),
r8 (6.7.4.), and possibly 17 (6.5., 7.1.).

There are some derived features in the phallomero-sternal and peripheral musculature: s1
and s3, and probably s7, are fused (6.9.). s4 and s6, and s2 if present, are fused (6.9.).
The pl are extremely broad. The posterior insertions of the p4 are far laterally (6.11.).
All these seemingly primitive features are assumed to be neotenic traits; that they are not
primitive but derived results from the outgroup comparison with Mantodea. s5 has been
lost (6.9.; like in Polyphaga and Ergaula: compare grouping H (131) in 7.6.).

Subgroup 2.2.2.2.: Lamproblatta + (Ergaula + Polyphaga)

There are many autapomorphies on the left complex: L8 is a new sclerite in the right
dorsal wall, with the insertions of 112, 19, and b2 upon it or in its immediate vicinity (6.5.).
L7 is a new sclerite on the right part of the vla-lobe (or on the Iba-lobe, which is a right
part of vla; 6.5.). The muscles I11 (6.3.4.), 112 (6.2.4., 6.5.), and s12 (6.2.4., 6.3.4., 6.9.)
are also peculiar to this subgroup. The L4-plate in the ventral vla-wall has expanded to
include the dorsal insertion of muscle s3 (new region L4a, larger sclerites L4R and L4M;
6.3.4.). The left insertion of 12 has shifted further anteriad and away from sclerite L4K
(or region L4l; 6.3.4.). The Ive-pouch has expanded almost to the left edge of the left
complex (6.2.1., 6.2.4.).

Other derived features could possibly also be autapomorphies of this subgroup: If s2-parts
should be included in s2+4+6 of Cryptocercus, s2 would have been lost in subgroup
2.2.2.2. (6.9.). The dorsal insertion of muscle s6 has become restricted to the anterior
margin of sclerite R3 (6.9.). (If the muscle named s6 should be s2, the loss of s6 and the
laterad expansion of the ventral s2-insertion would be the autapomorphies). If s7 really is
a ground-plan muscle of Blattaria (compare (L) in 7.5.), and if vestiges of s7 are included
in $1+3 of Cryptocercus, the complete loss of s7 would also be an autapomorphy of
subgroup 2.2.2.2. (6.9.). Muscle 11 has been lost. (However, Il has certainly been lost
several times: at least also in the subgroups 2.1., 2.2.3.1., and 2.2.3.2.2.; compare grouping
R (25) in 7.6.)

Subgroup 2.2.2.2.1.: Lamproblatta

There are many derived features on the left complex. L2 has divided into three sclerites:
L2A in the left part of the Ive-pouch, L2B in the right part of the Ive-pouch, and L2C

286

on the paa-process (6.2.4.). LAN has divided into two sclerites: L4S containing the L4d-
region, L4T on the pda-process (6.3.4.). (The sclerotisations of paa and pda remain
connected). Around articulation A4 (between L2A and L2B) the Ive-pouch has developed
a deep recess (6.2.4.). The L4d-region has rotated into a dorsoventral orientation (6.3.4.).
The processes paa and pda have elaborated the cuticular invaginations boe and sbe
(fig.182). The muscles 120, 121, 122, 123, and 124 have evolved (fig.184-188). Muscle 14
has been lost (like in Anaplecta: compare grouping K (133) in 7.6.).

On the right phallomere, the tre-tendon and its muscles b4 and s8 have been lost (6.7.5.:
like in subgroup 2.2.3.: compare (I) in 7.5. and grouping E (73) in 7.6.). The age-groove
has been lost (6.7.4.), and sclerite R2 bears an extension R2m to the left (6.7.4.; both
features also in Tryonicus: compare grouping G (129) and (130) in 7.6.). Muscle rl has
been lost (6.7.6.).

Subgroup 2.2.2.2.2.: Ergaula + Polyphaga

There are many autapomorphies on the left complex: Sclerite L4K has shifted to the
posteroventral part of the hla-base (6.3.4.). Muscle 114 has been lost and functionally
replaced by 14 (6.3.4., 6.4.3.). The L4-plate in the ventral vla-wall has undergone a further
expansion (new region L4x, larger sclerite L4M) and includes now the left insertion area
of 12 (6.3.4.). Additionally, this I2-insertion has shifted further anteriad and also ventrad
(6.3.4.). The cuticular area around sclerite L7 has been elaborated as a new lobe Iba which
represents the rightmost part of the vla-lobe (6.5.).

On the right phallomere, the sclerites R2 and R3 have fused (and articulation A7 has been
lost; 6.7.4.). The large sclerite RIM has developed, either by a posteriad expansion of the
former RIF alone or, more probably, by an additional fusion of RIF and parts of the
former R1J (6.7.6.; with the loss of the membranous area 17 and of the articulations A8
and A9 as a result; compare (H) in 7.5.). Probably in correlation with this feature (in its
latter interpretation) muscle r3 has been lost (6.7.6.). (The fusion of RIF and R1J and
the loss of r3 have also been achieved in subgroup 2.2.3.: compare grouping F (128) and
(64) in 7.6.). The rge-groove on the Ric-region has distinctly expanded posteriad (6.7.6.).
The pva-ridge on the R1t-region has achieved a longitudinal orientation and has likewise
expanded posteriad (6.7.6.). (That means, within the RIM-sclerite, as compared with the
RIF-sclerites, the regions Rle and Rit have expanded posteriad). Sclerite R2 has
expanded to occupy most of the cbe-invagination and is connected with Rit in the dorsal
wall of cbe (6.7.4.). The articulation A6 between R2 and RIt has been lost (6.7.4.). (Since
a fusion of Rit and R2 does not necessarily result in a loss of A6 — compare Archiblatta
in 6.7.4. — each of the two latter features is regarded as an autapomorphy of its own.)

Subgroup 2.2.2.2.2.1.: Polyphaga
The only derived feature known as compared with Ergaula is the ventral gap in the sclerite

ring formed by the posterior part of L1 (6.1.4.). r9 is a new muscle (6.7.4.: Ergaula not
investigated).

287

Subgroup 2.2.2.2.2.2.: Ergaula (capensis and capucina)

On the left complex, the anteriormost part of L4M has split off to form an isolated sclerite
(with the insertions of s3 and s12; 6.3.4.). Sclerite L4K has shifted somewhat farther
anteriad (6.3.4.). The dorsal part of L4K within the hla-base has shortened and fused to
the ventral anterior margin of sclerite L3 (6.3.4.). Muscle I11 has distinctly enlarged
(6.3.4.; investigated only in E. capucina). The paa-process has been lost (6.3.4.). On the
right phallomere, R2 has broadened, and R3 is now for most of its breadth confluent with
R2 (6.7.4.). The weak lines A7* and 13, representing the fusion lines between R2 and
R3 or R2 and RIt, respectively, in Polyphaga, have been lost (6.7.4.).

Subgroup 2.2.3.: Anaplecta + (Nahublattella + (Supella + (Euphyllodromia +
(Parcoblatta + (Nyctibora + (Blaberus + Nauphoeta + Blaptica + Byrsotria))))))

All these species belong to Blattellidae and Blaberidae sensu McKittrick (1964). In
Anaplecta, Nahublattella, Parcoblatta, and Blaberus the whole phallomere complex has
been investigated, including its muscles. In the other species only certain parts or elements
have been studied, or their presence or absence has been checked (mainly the elements
listed in 5.15.). It will be exactly specified which derived features are known to be present
in which of these species. Ectobius and Loboptera will not be considered in the following
analysis since too few features have been investigated to correctly assess and assign these
species, which are probably highly modified in their phallomere morphology.

At least the following apomorphies are present in all species comprised in this subgroup:
On the left complex, the introversible membranous basal part 30 of the hla-hook has
become very extensive, and hla can be almost completely retracted (6.4.3.). The hla-base
has shifted to the left posterior edge of the left complex (6.4.3.). (These two features are
possibly intercorrelated; compare (M), (N) in 7.5). The left part of the left complex, which
contains the hla-base, has been separated from the parts more to the right by the fpe-
infolding (6.4.3.). The anterior part of the Ive-pouch has been elaborated as a tube-like
Ive-apodeme (6.2.4.). The common sclerotisation of the processes paa and pda has become
stout and ring-shaped in its basal part (6.2.4.). (The resulting very close relation of paa
and pda and their sclerotisations might be the basis for the formation of the via-process
with an elongated common basal part of paa and pda in subgroup 2.2.3.2.).

On the right phallomere, the tre-tendon and its muscles b4 and s8 have been lost (like in
Lamproblatta: compare (I) in 7.5. and grouping E (73) in 7.6.; 6.7.5.). Sclerite RIN has
developed by a fusion of the former RIF and R1J (6.7.6.; the loss of the membranous
area 17 and of the articulations A8 and A9 are concomitant derivations; all regions of R1
are now included in one sclerite, like in the common ground-plan of Blattaria and
Mantodea). The loss of muscle r3 is probably correlated with this feature (6.7.6.). (The
fusion of RIF and R1J and the loss of r3 have also been achieved in subgroup 2.2.2.2.2.:
compare grouping F (128) and (64) in 7.6. and (H) in 7.5.). The rge-groove on the Rlic-
region has been lost (6.7.6.; compare (J) in 7.5.). The median end of the Rit-region has
developed a hook-like curvature (6.7.6.). (This feature is absent in Supella; it is assumed
to be rendered unrecognisable by the extreme expansion of sclerite RIN’. In subgroup
2.2.3.2.2.2. this curved area forms the cwe-thickening).

288

The following features have been investigated only in Anaplecta, Nahublattella,
Parcoblatta, and Blaberus: Muscle sl has been lost (6.9.). The muscles termed 15 in
Anaplecta and Nahublattella are possibly new muscles, or they are true 15 with the anterior
insertion shifted to the L4n-region; in any case, one of these character states is probably
apomorphic for subgroup 2.2.3. (6.2.4., 6.3.4.). (In Parcoblatta and Blaberus 15 has been
lost or integrated into 16b).

Subgroup 2.2.3.1.: Anaplecta

On the left complex, sclerite L1 and the dca-processes have been lost (like in all or many
species of subgroup 2.2.3.2.2., compare grouping L (109) and (110) in 7.6.; according to
McKittrick (1964), however, L1 is present in another species of Anaplecta; 6.1.4.). The
pne-pouch is therefore completely membranous; it has been reduced to a shallow
depression in the central dorsal wall. (pne has been lost completely in all or many species
of subgroup 2.2.3.2.2., compare (111) in 7.4. and grouping L (134) in 7.6.; 6.1.4.). Muscle
14 has been lost (6.2.4., 6.3.4.; like in Lamproblatta: compare grouping K (133) in 7.6.).
Some membranous foldings have developed in the area of the Ive-pouch, e.g. vfa and vpe
(6.2.4.). The gta-process (6.2.4., fig.215) and the vte-tendon (fig.208) have evolved. The
L4d-region has been lost (like in subgroup 2.2.3.2.2., compare grouping L (95) in 7.6.).
Muscle I1 has been lost (like in some other subgroups: compare groupings L and R (25)
in 7.6.). 125 and 126 are probably muscles peculiar to Anaplecta. (However, 126 might be
homologous with l6a of Nahublattella, Parcoblatta, and Blaberus; 6.3.4.).

The firm connection between the nla-process and the top of the Ive-apodeme might be
autapomorphic for Anaplecta. According to its possible correlation with the translocation
of the anterior 114-insertion from nla to the top of Ive, however, this feature could also
be an autapomorphy of the whole subgroup 2.2.3., lost again at the base of subgroup
2.2.3.2. (compare (M), (N) in 7.5.; 6.4.3.). That there are two phallomere gland openings
in a rather peculiar position — possibly new organs — could also be either a derived feature
of Anaplecta alone or an autapomorphy of subgroup 2.2.3., with a loss of one opening
and a deplacement of the other at the base of subgroup 2.2.3.2. (6.6.).

Subgroup 2.2.3.2.: Nahublattella + (Supella + (Euphyllodromia + (Parcoblatta +
(Nyctibora + (Blaberus + Nauphoeta + Blaptica + Byrsotria)))))

At least the following apomorphies are present in all species listed: The anterior insertion
of muscle 114 has been translocated from the L4n-region on the nla-process to the L2a-
region on top of the Ive-apodeme (6.4.3.; or, if the homology of the 114-muscles should
not be true, a new muscle from L2a to L3 has developed). The size of the vla-lobe has
distinctly decreased (6.2.4., 6.3.4.). The common base of the paa- and pda-processes has
been elongated and forms, together with paa and pda, the via-process (6.2.4., 6.3.4.). The
right posterior dorsal part of the left complex — the part dorsal to the right half of the Ive-
pouch — has been reduced (6.2.4.). The division of L2 into L2D (within the Ive-pouch)
and L2E (together with LAN on the via-process) by articulation A10 can also be regarded
as an autapomorphy of this group since this situation is present in both the subordinate

289

sister groups 2.2.3.2.1. and 2.2.3.2.2. (The absence of this division and of A10 is regarded
as a secondary loss having occurred several times within subgroup 2.2.3.2.2.: compare
(Q) in 7.5.).

The following derived features are also assumed to be autapomorphies of subgroup 2.2.3.2.
but have been investigated in sufficient detail only in Nahublattella, Parcoblatta, and
Blaberus: On the left complex, L4K has divided into the sclerites L4U and L4V — with
the nla-process probably still present on L4V’ of Nahublattella. (In Blaberus and possibly
also in Parcoblatta L4V has been completely lost; 6.3.4.). Sclerite L4G (L4v-region) in
the ventral vla-wall has been lost (6.3.4.). The right insertion of muscle 12 has shifted to
the membranous basal part 30 of the hla-hook (6.3.4.). Muscle 130 has developed, having
a longitudinal course in the ventral wall of the left complex (6.5.). Muscle s2 has been
lost (6.9.; like in subgroup 2.2.2. or 2.2.2.2.: compare above). The muscles s5 and s6 have
divided into s5a and s5b or s6a and s6b, respectively (or, new muscles s5a and s6a have
developed; 6.9.). The p4-insertions on the subgenital plate have shifted mesad (6.11.).

Subgroup 2.2.3.2.1.: Nahublattella

On the left complex, the sclerotisation L2E’+L4N’ of the via-process has been divided
transversely by the membrane ring 39 (6.2.4.). The nla-process, if actually nla, has become
whip-shaped (6.3.4.). The central dorsal wall contains a bristle area 35 (fig.242). The
muscles 127, 128, 129, 131, 132, and 135 have developed (fig.240, 249-252). On the right
phallomere, sclerite R2 has become plate-like, bearing the highly elaborated evaginations
42 and 43 (6.7.4.). The Rit-region has probably completely fused to the rest of sclerite
RIN’ (6.7.6.). There is a new muscle r10 (6.7.6.). The dorsal sclerotisation S9d of the
subgenital plate is highly elaborated, divided, and provided with the new muscles 133 and
134 (6.10.).

+ Nauphoeta + Blaptica + Byrsotria))))

At least the following apomorphies are present in all species listed (with the exceptions
mentioned): On the left complex, the hla-hook has evolved a hge-groove with a notch 45
in its ventral wall (6.4.3.). (In Nyctibora and Nauphoeta hge is not that distinct, and the
notch 45 is missing; this is probably due to secondary reduction). The ate-tendon has
become very long and narrow (6.3.4.). (That ate is rather short in Blaberus, Blaptica, and
Byrsotria is interpreted as a secondary reduction since ate is very long and thin in
Nauphoeta, and since these four species form the holophyletic subgroup 2.2.3.2.2.2.2.2.2.).
Sclerite L4V (essentially the L4n-region) has been reduced to a small sclerite in the dorsal
wall of the ate-tendon or has been completely lost (or, L4V has been generally lost, and
the sclerite within ate, present only in Parcoblatta, Nyctibora, and Blaptica, is a new one;
6.3.4.). The nla-process has been lost (6.3.4.). The via-process is no longer branched, i.e.
paa and pda are no longer distinct (6.2.4.). The right posterior branch of L2, whose distal
part sclerotises the psa-process in Nahublattella, has been completely lost (6.2.4.). The
L4d-region has been lost (6.3.4.; like in Anaplecta: compare grouping L (95) in 7.6.). The

290

ventral extension 28 of L2, present in Anaplecta and Nahublattella, is lacking; this is also
assumed to be a derived feature (6.2.4.; compare the possible homology of 28 and LS:
OS»):

The following apomorphies have been investigated only ın Parcoblatta and Blaberus; they
might be autapomorphies of the whole subgroup 2.2.3.2.2. or of any subordinate subgroup
containing at least Parcoblatta and Blaberus: The age-groove has been restricted to the
anterior part of sclerite R3, i.e. the age-part along the posterior right margin of R3 has
been lost (6.7.4.). Muscle 114 has divided into two bundles 114a and 114b (6.4.3.; this
division is completely different from the division of 114 in Eurycotis). Muscle 13 has been
lost (6.1.4., 6.2.4.). The muscles 136, 137, and 138 have developed (fig.276-278, 303, 307;
6.4.3.). Muscle 15 has been lost or integrated into l6b (6.2.4.). Muscle 16a has distinctly
enlarged (6.2.4.). Muscle s3 has divided into two bundles s3a and s3b (6.9.). The muscles
s14 or sl4a,b are new (6.9.). The dorsal insertion of muscle s6b has expanded posteriad
to occupy the entire right margin of sclerite R3 (6.9.). The ventral insertions of s5a and
s6a have shifted posteriad (6.9.). Some further derived features of Parcoblatta and
Blaberus are also present in Anaplecta but not in Nahublattella (compare grouping L (109),
(110), (134), and (25) in 7.6.): Sclerite L1, the dca-process(es), and muscle Il have been
lost. The pne-pouch, which has become indistinct in Anaplecta, has been completely lost
(6.1.4.).

Subgroup 2.2.3.2.2.1.: Supella

Sclerite RIN’ has extremely expanded to occupy the entire dorsal wall of the cbe-
invagination (6.7.6.). The indistinctness of the Rit’-region and the lack of a hook-like
curvature at the median end of Rit’ (present in the ground-plan of subgroup 2.2.3.) are
assumed to be results of this expansion.

Subgroup 2.2.3.2.2.2.: Euphyllodromia + (Parcoblatta + (Nyctibora + (Blaberus +
Nauphoeta + Blaptica + Byrsotria)))

All species listed share some very important derived features on the right phallomere
(6.7.6.): The median ends of the Rit-region and of sclerite R2 have fused: loss of
articulation A6. (A6 is still an articulation in Supella). The hook-curvature at the median
end of Rit has been elaborated as the cwe-thickening. At its lateral end, Rit has been
separated from the R1c-region: resulting sclerites RIP and RIS. (In Supella Ric and Rit
are still connected with each other. In Blaberus, Nauphoeta, Blaptica, and Byrsotria Rit
and Ric are also connected: sclerite RIT’; this situation is interpreted as a secondary
fusion of these regions and as a synapomorphy of these species, which view is suggested
by the assumed autapomorphies of the subgroups 2.2.3.2.2.2.2. and 2.2.3.2.2.2.2.2.).

Subgroup 2.2.3.2.2.2.1.: Euphyllodromia

No derived features restricted to this species have so far been found in the phallomere
complex. According to the phylogenetic hypothesis presented here, the following derived
features appear as autapomorphies: The L2-sclerotisations of the via-process and of the
Ive-pouch have fused secondarily (i.e. articulation A10 has been lost). Muscle 110 has

291

been lost. (Both features also in some other species of subgroup 2.2.3.2.2.2.; compare (Q),
(R) in 7.5.). The membranous left wall of the Ive-pouch has deeply invaginated to the left
(also in Loboptera; 6.2.4.).

Subgroup 2.2.3.2.2.2.2.: Parcoblatta + (Nyctibora + (Blaberus + Nauphoeta + Blaptica
+ Byrsotria))

Only two possible autapomorphies are present in the phallomere complex: First, the
presence of a dla-lobe (6.7.6.). dla is clearly missing in Supella and Euphyllodromia. dla
is also missing in Nauphoeta, but in this species the whole right phallomere has been
strongly reduced. Second, the presence of a tve-tendon (6.2.4.). tve is absent in Supella
and Euphyllodromia and present in all members of this subgroup except Blaberus
(Byrsotria not investigated). The absence in Blaberus is regarded as a secondary loss. The
holophyly of this subgroup is strongly supported by a clearly derived feature of the
females: They perform a rotation of the ootheca within the vestibulum (into a horizontal
orientation; McKittrick 1964; termed advanced rotation by Roth 1967). In Supella and
Euphyllodromia the ootheca retains a vertical orientation till it is dropped (Roth 1967).

Subgroup 2.2.3.2.2.2.2.1.: Parcoblatta

Features of Parcoblatta which are derived as compared with all other investigated species
are the vge-groove on the via-process (fig.273) and the rotation of the via-process and of
some adjacent elements, which includes a ventrad shift of the genital opening (6.2.4.).
Some other features of Parcoblatta are derived at least as compared with Blaberus
(characters mostly not investigated in the other species of subgroup 2.2.3.2.2.): Sclerite
L4U has been lost (6.3.4.). Sclerite R2 has become distinctly curved (6.7.4.). The muscles
139 and 140 have developed (fig.277, 278). Muscle s14 has divided into two bundles s14a
and s14b (6.9.). The ventral insertion of muscle l6a and the dorsal and ventral insertions
of muscles s5a and s6a have shifted very far posteriad (6.2.4., 6.9.). Further derived
features are the fusion of the L2-sclerotisations of the via-process and of the Ive-pouch
and the loss of muscle 110 (6.2.4.; both like in Euphyllodromia and Blaberus: compare
(Q), (R) in 7.5.).

Subgroup 2.2.3.2.2.2.2.2.: Nyctibora + (Blaberus + Nauphoeta + Blaptica + Byrsotria)

This subgroup has some probably derived features on the right phallomere: There are two
new sclerites, R5 in the ventral part of the right phallomere (6.7.4.) and R4 in the dorsal
wall of the dla-lobe (6.7.6.). A new muscle rll runs from the dorsal dla-wall to the ventral
fda-wall (6.7.6.).

Subgroup 2.2.3.2.2.2.2.2.1.: Nyctibora

No certain autapomorphies have so far been found in the phallomere complex.

Subgroup 2.2.3.2.2.2.2.2.2.: Blaberus + Nauphoeta + Blaptica + Byrsotria

In all listed members of this subgroup, which corresponds to the Blaberidae sensu
McKittrick (1964), the Rit’-region (sclerite R1S’) and the rest of R1’ (sclerite RIP’) have

292

fused secondarily to form sclerite RIT’. Only in Blaberus, Blaptica, and Byrsotria the
ate-tendon has shortened and broadened (6.3.4.; ate has remained long and thin in
Nauphoeta), the L10’-sclerotisation has evolved (one sclerite in Blaberus, fig.299; many
small sclerites in Blaptica, fig.291, and Byrsotria), and the via-process and some adjacent
elements have rotated (6.2.4.; this rotation, which includes a dorsad shift of the genital
opening, is most advanced in Blaberus).

At least Blaberus has developed the muscles 141 to 146 and r12 to r18 (not investigated
in Nauphoeta, Blaptica, and Byrsotria) and lost muscle s7 (6.9.; s7 is present at least in
Nauphoeta; not investigated in Byrsotria and Blaptica).

7.4. Survey of phylogeny and aut/synapomorphies

In this section, the character states assumed to be autapomorphies of subgroups are listed,
and a phylogenetic tree is given (diagram 1). The autapomorphies are termed by bold
printed numbers put in brackets. Some symbols give additional information: !: The same
apomorphic character state has evolved in at least one other subgroup, too, and homology
is not contradicted by morphological data or functional arguments but only by the
distribution of the apomorphic states of other characters (i.e. by parsimony). The
apomorphic state has the same number in all subgroups concerned. ?: The position of the
autapomorphy in the tree is questionable, due either to lack of investigations or to not
definitely interpretable morphology. (? element): The homology of the named element and
hence that of the respective apomorphic state in the various species included in the
subgroup is questionable. // separates different conceivable morphological interpretations
of character states. In the tree bold print, the brackets, and the symbols except for ? are
omitted.

The plesiomorphic character states are given in brackets, and for each state the taxon or
subgroup is named within the range of which it is plesiomorphic: (1) In most cases one
character state is listed which is plesiomorphic within Blattaria and Mantodea as a whole
(i.e. which is present in the common ground-plan of Blattaria and Mantodea). Such a
character state is preceded by “BM:” = Blattaria + Mantodea. (2) In some cases one
character state is listed whose categorisation as plesiomorphic is related to the range of
Blattaria, of Mantodea, or of a subgroup of Blattaria or Mantodea which is superordinate
to the subgroup under consideration and includes it. Such a character state is preceded by
the name of the respective superordinate subgroup or taxon, e.g. “SG2.2.:” = subgroup
2.2., “SG1.:” = Mantodea, “SG2.:” = Blattaria. This is practised if the character concerns
a property of an element whose presence in the common ground-plan of Blattaria and
Mantodea is uncertain, if the character is for any reason not assessable in the species
outside the named subgroup or taxon, or if an exact description of the character is only
possible within the named subgroup (e.g. if a sclerotisation has divided and changed its
shape previously, and the character concerns a further derivation of such a sclerotisation).
(3) If several states of a character form a transformation series, all states which are more
-plesiomorphic than the named apomorphic state are listed, and for each of them the
respective subgroup is given as in (1) and (2).

293

Subgroup 1.2.: (1) Region L4b between L4-regions L4l, L4v, and L4n present (BM: L4b
absent) (2) Apodeme swe vestigial or absent (BM: swe present along most of L4l) (3)
Region L4l in dorsal wall extending far to the right (BM: L4l restricted to left margin of
dorsal wall) (4) Region L4d not prominent from outline of region L41 // L4d absent (BM:
L4d prominent from outline of L4l // L4d present) (5) Membranous part of pne-wall on
right side (BM: Membranous part of pne-wall dorsal or right-dorsal) (6) Distal part of
region Lim curving into dorsal wall of pouch lve (BM: Distal part of Lim not curving
into dorsal wall of Ive) (7) Pouch lve narrow (BM: Ive moderately broad) (8) Sclerite L2
ribbon- or plate-like (BM: L2 arch-shaped) (9) Process paa long (BM: paa very short)
(10) Sclerotisations of processes pda and paa = regions L4l and L2d separated (BM: L4l
and L2d connected) (11) Region Ric divided by stripe 4: sclerites RIA, RIB, or RIA,
RIC (BM: Rlc undivided, stripe 4 absent).

Subgroup 1.2.2.: (12) Region L4l divided by articulation Al: sclerites L4A, L4B (BM:
L4l undivided, Al absent) (13) Region L4b as heavily sclerotised as L4-regions LAl, L4v,
and L4n (SG1.2.: L4b weaker than L4-regions L4l, L4v, and L4n; BM: L4b absent) (14)
Region LIm rather broad (BM: Lim very narrow) (15)(? afa) Region Lim occupying
process afa (SGl.: afa membranous) (16) Region Lim divided in its posterior part by
membranous stripe 2 (BM: Lim completely undivided, stripe 2 absent) (17) Left part of
apodeme age abruptly deepened (BM: Left part of age not or gradually deepened).

Subgroup 2.1.: (18) Pouch lve: anteroventral part deeply invaginated to the left,
posteroventral part receded to the right (BM: anterior or anteroventral part of Ive not
deeper invaginated to the left than posterior part, posterior or posteroventral part not
receded to the right) (19) Process paa far on the right and far to the right of process pda
(BM: paa on the left and immediately to the right of pda) (20) Sclerotisations of processes
pda and paa = regions L4l and L2d separated (BM: L4l and L2d connected) (21) Sclerite
L4F present (BM: L4F absent) (22) Ventral insertion of muscle 15 far posteriorly and on
sclerite L4F // Muscle 15 present, not homologous with 15 of other subgroups (BM: Ventral
insertion of 15 far anteriorly // This 15 absent) (23) Left insertion of muscle b4b on top of
pouch pne (BM: Left insertion of muscle b4b in dorsal wall outside pouch pne) (24)!
Region Lla level (BM: Lla hood-shaped) (25)! Muscle 11 absent (BM: 11 present) (26)
Region Rit rather large, occupying entire dorsal and anterior walls of invagination cbe
(BM: RIt narrow, occupying only part of dorsal wall of cbe).

Subgroup 2.2.: (27) Region L41 divided by articulation A5 or by a farther separation
homologous with A5: sclerites L4K, L4N (BM: L4l undivided, A5 absent) (28) Apodeme
swe absent (BM: swe present along most of L4l) (29)? Muscle 110 present // Muscle 14
divided into anterior and posterior (= 110) bundle and posterior bundle shifted posteriad
to common sclerotisation of pda and paa (BM: 110 absent // 14 undivided and inserted on
L4l left-anterior to sclerotisation of paa and pda) (30) Region L4d directed anteriad (BM:
L4d directed to the right or right-anteriad) (31) Tooth pia absent (BM: pia present) (32)
Regions Rld and Rlv broadly connected posterior to membrane 17: sclerite R1J (BM:
Rld and Rlv not or narrowly connected posterior to membrane 17) (33)? Muscle s2
extremely reduced (SG2.: s2 moderately reduced; BM: s2 not reduced and as strong as
sl) (34) Posterior part of sclerite L1 forming a ring (BM: L1 not forming a ring).

294

? Subgroups 2.2.1. + 2.2.3.: (35) Membranous basal part 30 of hook hla extensive, hla
rather deeply retractable (SG2.: 30 very narrow, hla hardly retractable) (36) Base of hla
in middle part of left ventral wall of left complex (SG2.: Base of hla in anterior left ventral
? Subgroups 2.2.2. + 2.2.3.: (37) Anterior and posterior parts of region L4l far separated
(SG2.2.: Anterior and posterior parts of L4Al separated but articulated in A5; BM: L4l
undivided) (38) Utmost right-anterior part of region L41 absent (BM: Utmost right-anterior
part of L41 present) (39) Region L4d directed to the left (SG2.2.: L4d directed anteriad;
BM: L4d directed to the right or right-anteriad) (29)? Muscle 110 present // Muscle 14
divided into anterior and posterior (= 110) bundle and posterior bundle shifted posteriad
to common sclerotisation of pda and paa (BM: 110 absent // 14 undivided and inserted on
LAI left-anterior to sclerotisation of paa and pda) (33)? Muscle s2 extremely reduced
(SG2.: s2 moderately reduced; BM: s2 not reduced and as strong as sl).

Subgroup 2.2.1.: (40) Membranous part of pne-wall on left side (BM: Membranous part
of pne-wall dorsal or right-dorsal) (41) Region LIm plate-like, articulation A2 very broad
(BM: LIm ribbon-like, A2 narrow) (24)! Region Lla level (BM: Lla hood-shaped).

Subgroup 2.2.2.: (42) Right-ventral part of sclerite L4K missing (SG2.2.: Right-ventral
part of L4K present) (43) Process nla absent (SG2.: nla present) (44) Right insertion of
muscle 12 on top of pouch pne (BM: Right insertion of 12 in left wall of pne) (45) Left
insertion of muscle 12 on anterior left edge of left complex, on sclerite L4K and region
L41 (BM: Left insertion of 12 in posterior two thirds of left edge of left complex, on region
L4l) (46) Anterior face of pouch pne and sclerite L1 plateau-like (BM: Anterior face of
pouch pne and of sclerite LI pointed or ridge-like) (47)?! Muscle s2 absent (SG2.2.: s2
extremely reduced; SG2.: s2 moderately reduced; BM: s2 not reduced and as strong as
sl).

Subgroup 2.2.2.2.: (48) Sclerite L8 present (BM: L8 absent) (49) Sclerite L7 present
(BM: L7 absent) (50) Muscle 111 present (BM: 111 absent) (51) Muscle 112 present (BM:
112 absent) (52) Muscle s12 present (BM: s12 absent) (53) Region L4a bearing dorsal s3-
insertion present (BM: L4a absent, dorsal s3-insertion on membrane) (54) Left insertion
of muscle 12 on anterior left edge of left complex, anterior to sclerite L4K and region L41
(SG2.2.2.: Left insertion of 12 on anterior left edge of left complex, on L4K and LAl; BM:
Left insertion of 12 in posterior two thirds of left edge of left complex, on L4l) (55) Pouch
lve almost reaching left edge of left complex (BM: Ive by far not reaching left edge of
left complex) (56) Muscle s7 absent (SG2.: s7 present) (25)! Muscle 11 absent (BM: 11
present) (47)?! Muscle s2 absent (SG2.2.: s2 extremely reduced; SG2.: s2 moderately
reduced; BM: s2 not reduced and as strong as sl).

Subgroup 2.2.2.2.2.: (57) (Dorsal part of) Sclerite L4K in posteroventral part of hla-base
(SG2.2.2.: L4K in dorsal part of hla-base; SG2.2.: Dorsal part of L4K left-dorsal to hla-
base) (58) Muscle 114 absent (SG2.: 114 present) (59) Region L4x bearing left I2-insertion
present (BM: L4x absent) (60) Left insertion of muscle 12 in left anterior ventral wall of
left complex, on region L4x (SG2.2.2.2.: Left insertion of 12 on anterior left edge of left
complex, anterior to L4K and L41; SG2.2.2.: Left insertion of 12 on anterior left edge of
left complex, on LAK and LAl; BM: Left insertion of 12 in posterior two thirds of left
edge of left complex, on L41) (61) Lobe Iba present (BM: Iba absent) (62) Sclerites R2

295

and R3 fused, articulation A7 absent (SG2.: R2 and R3 separated, articulated in A7) (63)
Regions Rit and Ric reaching far posteriad and probably fused with parts of the broadly
interconnected regions Rld and Rlv: articulations A8 and A9 and membranous area 17
absent; sclerite RIM (SG2.2.: Rlt and Rlc restricted to a more anterior area and separated
from the broadly interconnected Rld and Rlv: A8, A9, and 17 present; SG2.: Rit and
Rlc restricted to a more anterior area and separated from the narrowly interconnected Rid
and Rlv: A8, A9, and 17 present; BM: RIt restricted to a more anterior area and not
connected with Rld or Rlv; Ric restricted to a more anterior area ventrally but possibly
extending more posteriad dorsally, connected with both Rld and Rlv which are separated
from each other: A8 and A9 absent but 17 present) (64)! Muscle r3 absent (BM: r3 present)
(65) Ridge pva longitudinally orientated (BM: pva transversely or obliquely orientated)
(66) Sclerite R2 occupying left-ventral and anterior walls of invagination cbe, broadly
connected with Rit (SG2.: R2 restricted to left-ventral base of cbe, separated from RIt)
(67) Articulation A6 absent (SG2.: A6 present).

Subgroup 2.2.3.: (68) Membranous basal part 30 of hook hla very extensive, hla very
deeply retractable (SGs2.2.1.+2.2.3.: 30 extensive, hla rather deeply retractable; SG2.: 30
very narrow, hla hardly retractable) (69) Base of hook hla at left posterior edge of left
complex (SGs2.2.1.+2.2.3.: Base of hla in middle part of left ventral wall; SG2.: Base of
hla in anterior left ventral wall) (70) Infolding fpe present (BM: fpe absent) (71) Ive-
apodeme present (BM: Ive-apodeme absent) (72) Common sclerotisation of processes paa
and pda with stout basal ring (BM: Common sclerotisation of paa and pda not with stout
basal ring) (73)! Tendon tre and muscles s8 and b4a,b absent (SG2.: tre, s8, and b4a,b
present; BM: presence of tre and s8 unclear, b4a,b present) (74) Region Ric fused with
the broadly interconnected regions Rld and R1v: articulations A8 and A9 and membranous
area 17 absent; sclerite RIN (SG2.2.: Ric separated from the broadly interconnected R1d
and Rlv: A8, A9, and 17 present; SG2.: Rlc separated from the narrowly interconnected
Rid and Rlv: A8, A9, and 17 present; BM: Ric connected with both Rld and Rlv which
are separated from each other: A8 and A9 absent but 17 present) (64)! Muscle r3 absent
(BM: r3 present) (75) Groove rge absent (SG2.: rge present) (76) Hook-curvature on
median end of region Rit present (BM: Hook-curvature on RIt absent) (77) Muscle sl
absent (BM: s1 present) (78) Ventral insertion of muscle 15 on region L4n near process
nla // Muscle 15 present, not homologous with 15 of other subgroups (SG2: Ventral insertion
of 15 not on L4n // This 15 absent).

Subgroup 2.2.3.2.: (79) Anterior insertion of muscle 114 on region L2a (SG2.: Anterior
insertion of 114 on L4n) (80) Lobe vla small (BM: vla large) (81) Process via present
(BM: via absent) (82) Right posterior dorsal part of left complex small (SG2.: Right
posterior dorsal part of left complex large) (83) Sclerite L2 divided by articulation A10:
sclerites L2D, L2E (BM: L2 undivided, A10 absent) (84) Sclerite L4K divided: sclerites
LAU, LAV (SG2.2.: L4K undivided) (85) Region L4v absent (BM: L4v present) (86) Right
insertion of muscle 12 on basal part 30 of hook hla (BM: Right insertion of 12 in left wall
of pouch pne on L1) (87) Muscle 130 present (BM: 130 absent) (47)! Muscle s2 absent
(SG2.2.: s2 extremely reduced; SG2.: s2 moderately reduced; BM: s2 not reduced and as
strong as sl) (88) Muscles s5 and s6 divided into s5a,b and s6a,b (BM: s5 and s6

296

undivided) (89) Insertions of muscles p4 median to lateral margin of subgenital plate (BM:
Insertions of p4 on lateral margin of subgenital plate).

Subgroup 2.2.3.2.2.: (90) Groove hge and notch 45 on hook hla present (SG2.: hge and
45 absent) (91) Tendon ate very long and narrow, including region L4n or the membranous
area corresponding to L4n (SG2.2.3.: tendon ate short and broad, not including L4n) (92)
Region L4n = sclerite L4V very small or absent, not forming a process nla // Region L4n
and process nla absent, new sclerite L4V present (SG2.: L4n large, forming a process nla
// LAn and nla present, sclerite L4V absent) (93) Process via unbranched, subordinate
processes paa and pda not distinct (SG2.2.3.2.: via branching into distinct paa and pda;
BM: via absent, paa and pda distinct) (94) Right posterior branch of sclerite L2 or L2D
= right arm of L2-arch absent (BM: Right posterior branch of L2 = right arm of L2-arch
present) (95)! Region L4d absent (BM: L4d present) (96) Extension 28 on sclerite L2
absent (SG2.2.3.: 28 present; BM: 28 or sclerite L5 possibly present).

Subgroup 2.2.3.2.2., or 2.2.3.2.2.2., or 2.2.3.2.2.2.2.: (97) That part of apodeme age along
posterior right margin of sclerite R3 absent, age by far not reaching articulation A3 (BM:
That part of age along posterior right margin of R3 present, age reaching A3) (98) Muscle
114 divided into two bundles 114a and b (SG2.: 114 undivided) (99) Muscle 13 absent (BM:
13 present) (100) Muscle 136 present (BM: 136 absent) (101) Muscle 137 present (BM: 137
absent) (102) Muscle 138 present (BM: 138 absent) (103) Muscle 15 absent // Muscle 15
integrated into muscle 16b (BM or SG2.2.3.: 15 present // 15 not integrated into 16b) (104)
Muscle 16a very large (SG2.: 16a of moderate size) (105) Muscle s3 divided into two
bundles s3a and b (BM: s3 undivided) (106) Muscle s14 present (BM: s14 absent) (107)
Dorsal insertion of muscle s6b on entire right margin of sclerite R3 (SG2.2.3.2.: Dorsal
insertion of s6b restricted to anterior right margin of R3; SG2.2.: Dorsal insertion of s6
restricted to anterior right margin of R3; BM: Dorsal insertion of s6 in ventral wall of
genital pouch, possibly extending to anterior margin of R3) (108) Ventral insertions of
muscles s5a and, less distinctly, s6a more posteriorly (SG2.2.3.2.: Ventral insertions of s5a
and s6a more anteriorly; BM: Ventral insertions of s5 and s6 more anteriorly) (109)!
Sclerite L1 absent (BM: LI present) (110)! Process(es) dca absent (SG2.: dca present)
(25)! Muscle 11 absent (BM: 11 present) (111) Pouch pne absent (BM: pne present).

Subgroup 2.2.3.2.2.2.: (112) Median ends of region RIt and sclerite R2 connected:
articulation A6 absent (SG2.: Median ends of R1t and R2 separated and articulated in A6)
(113) Thickening cwe present (BM: cwe absent) (114) Region Rit separated from region
Ric: sclerites RIP, RIS (BM: RIt connected with Ric).

Subgroup 2.2.3.2.2.2.2.: (115) Lobe dla present (BM: dla absent) (116) Tendon tve present
(BM: tve absent) (117) Females: Advanced rotation of ootheca present (BM: Advanced
rotation absent).

Subgroup 2.2.3.2.2.2.2.2.: (118) Sclerite R5 present (BM: R5 absent) (119) Sclerite R4
present (BM: R4 absent) (120) Muscle rll present (BM: rll absent).

Subgroup 2.2.3.2.2.2.2.2.2.: (121) Region Rlt connected with region Ric: sclerite RIT
(SG2.2.3.2.2.2.: Rit separated from Ric; BM: Rit connected with Ric).

12, 13, 14,
Dr

48, 49,50, 51,
52053654855:
56, 25, 47?

Dy

ily ksh, CE 7
100, 101, 102,
103, 104, 105,
106, 107, 108,
10921107111

115, 116, 117
118, 119, 120

297

Sphodromantis

Metallyticus

Chaeteessa

Mantoida
Archiblatta
Blatta
Deropeltis
Periplaneta

Eurycotis

Tryonicus

Cryptocercus

Lamproblatta

Polyphaga

Ergaula

Anaplecta

Nahublattella

Supella

Euphyllodromia

Parcoblatta
Nyctibora

Nauphoeta
Blaberus
Byrsotria
Blaptica

Diagram 1: Phylogenetic tree of the investigated representatives of Blattaria and Mantodea, with the

assumed aut/synapomorphies

298

7.5. Remarks on the polarity and evolution of some characters

For some characters the polarity assumptions in 7.1.-7.4. are not yet sufficiently
substantiated. In some cases the polarity question can be settled by a detailed discussion
of morphology, homology, or functional intercorrelations. The respective discussions will
be largely independent of the phylogenetic hypothesis in 7.4. In other cases a solution of
the polarity question can only be approached by a reciprocal consideration of the various
arguments or possibilities in terms of parsimony, i.e. a weighing of the various possible
polarity assumptions against the assumed autapomorphies of the subgroups defined in 7.2.
and 7.3. and against the outgroup comparison between Blattaria and Mantodea. The
respective discussions will be done in interdependence with the phylogenetic hypothesis
in 7.4.

The following discussions under (A)-(C) are concerned with the polarity of three
characters for which the outgroup comparison between Blattaria and Mantodea is
somewhat conflicting since the same two character states are present in Blattaria as well
as in Mantodea. The question arises whether that character state represents the ground-
plan condition which in 7.1. has been assumed to do this. (A similar conflict is also present
in (G), which will be discussed below.) These discussions will be independent of the
phylogenetic hypothesis in 7.4.

(A) The connection or separation of the sclerotisations L2d and L4l of the processes paa
and pda and the length of paa and pda

The area bearing the paa- and pda-processes is very similar in Mantoida (fig.44) and
Tryonicus (fig.91): The sclerotisations of paa (L2d-region) and pda (L4l-region) are
connected; paa and pda are completely sclerotised, are bulge-like and short, paa being
somewhat upcurved, and are close to each other. This has been regarded as the condition
of the common ground-plan of Blattaria and Mantodea (6.2.1., 7.1.).

In other Blattaria (6.2.1., 6.2.4.) as well as in other Mantodea (6.2.3.) paa and pda can
be longer, and their sclerotisations can be separated from each other. Alternatively, these
two states could be regarded as the ground-plan states of the respective characters, but
there are some arguments against this view: Within Blattaria the sclerotisations are
separated in Eurycotis (fig.66), Archiblatta (fig.53), and Cryptocercus (fig.150). This
separation is accompanied (1) by a reduction of the sclerotisation of one of the processes
in Cryptocercus (pda) and in Archiblatta (paa), and (2) by a far separation of the processes
themselves in Eurycotis and Archiblatta, which feature is correlated with the recess of the
posterior part of the Ive-pouch to the right (6.2.1.). These accompanying conditions in
Eurycotis, Archiblatta, and Cryptocercus are clearly derived features, and the separation
of the paa- and pda-sclerotisations can easily be explained as correlated with these and
as being derived, too. In Chaeteessa (fig.34), Metallyticus (fig.25), and Sphodromantis
(fig.11) the sclerotisations of paa and pda are separated from each other without being
reduced (with the exception that in Chaeteessa the whole pda has been lost), and the close
vicinity of paa and pda has been retained. According to these very different concomitant
circumstances in Cryptocercus, in Archiblatta + Eurycotis, and in the respective Mantodea,

299

the separation of the sclerotisations of paa and pda is clearly suggested to have been
achieved three times independently. As regards the length of paa and pda, there are no
peculiar similarities in the shape of paa and pda in the Blattaria and Mantodea with these
processes being long. On the other hand, both paa and pda of Tryonicus and Mantoida
are rather similar, and in my view it is the most parsimonious solution to regard this
similarity as homologous and as representing the ground-plan condition.

(B) Sclerite L2 arch-shaped or plate-like

In Mantoida, Archiblatta, and Polyphaga L2 extends like an arch along the margins of
the Ive-pouch, and this has been regarded as the condition of the common ground-plan of
Blattaria and Mantodea (6.2.1., 7.1.; fig.324d,f,k). A plate-like L2, with the arms of the
arch (regions L2p and L2m) probably fused to each other, is present in Chaeteessa,
Metallyticus, and Sphodromantis (6.2.3.), but also in Eurycotis (6.2.1.) and, more or less,
in Cryptocercus (6.2.4.) (fig.324a,b,c,e,h). However, in Cryptocercus the indistinctness of
the arch-shape of L2 is due to a reduction of the right part of the L2-arch (L2m-region,
area of articulation A2) and to a broadening of the L2a-region; vestiges of the arch-shape
are still recognisable in this L2. Eurycotis has the posterior left-ventral part of L2
extremely reduced. In the respective Mantodea L2 is only narrowed, with none of its parts
reduced. Thus, the morphology of L2 and the concomitant circumstances of its plate-like
condition are very different in Eurycotis, in Cryptocercus, and in the respective Mantodea,
and this supports the assumption that the plate-like shape of L2 has arisen several times
by parallel evolution. Additionally, the plate-like condition is in my opinion more liable
to homoplasy than the more complicated arch-shape.

(C) Sclerite L5 present or absent and in various positions

As discussed in 6.5., the presence of L5 — somewhere in the dorsal wall of the vla-lobe
— might be the plesiomorphic state for Blattaria or even Blattaria + Mantodea, but a definite
decision is not possible. The extension 28 (a ventral part of sclerite L2) of Anaplecta and
Nahublattella is possibly homologous with L5 (6.5.). The position of L5 or 28 is very
similar in Polyphaga, Anaplecta, and Nahublattella, but since it is not clear which position
of L5 or 28 has to be regarded as primitive, this similarity could also represent the
plesiomorphic state of Blattaria or Blattaria + Mantodea. At the present state of knowledge,
the characters of these sclerotisations are not yet suitable for a phylogenetic analysis since
the polarity of their states remain uncertain.

The following discussions will be concerned with the polarity of some characters, termed
(D)-(K) and (L), for which the reciprocal outgroup comparison between Blattaria and
Mantodea suggests another polarity as it has been assumed in 7.1.-7.4. (or, in the case of
(G), this outgroup comparison is conflicting in the same way as in the characters discussed
under (A)-(C)). The question is whether some features present in some Blattaria but not
in the other Blattaria and in Mantodea are either ground-plan features of Blattaria as stated
in 7.3. or autapomorphies of Blattarian subgroups as it is suggested by the outgroup
comparison. If the polarity assumptions are based on the latter alternative, then (1) the
distribution of the states of the characters (D)-(K) would suggest a phylogenetic hypothesis

300

which is roughly the reverse of the hypothesis in 7.4., and (2) the distribution of the states
of character (L) would suggest a grouping which is inconsistent with the phylogenetic
hypothesis in 7.4. as well as with the hypothesis discussed in the context of the characters
(D)-(K). These possibilities have to be tested.

(D) The presence or absence of the curvature of the right parts of sclerite L2 and of the
Ive-pouch

That the right parts of L2 and of lve curve dorsad and back to the left (6.2.1., 6.2.4.) has
been assumed to be a ground-plan feature of Blattaria. This curvature is present in
Archiblatta, Eurycotis, Tryonicus, Lamproblatta, Polyphaga, Ergaula, and Anaplecta but
absent in the remaining Blattaria (Cryptocercus and subgroup 2.2.3.2.) and in Mantodea.

(E) The presence or absence of the nla-process

The nla-process (6.3.1., 6.3.4.) has been assumed to be a ground-plan element of Blattaria.
nla is present in Eurycotis, Archiblatta, Tryonicus, Anaplecta, and probably Nahublattella
but absent in the remaining Blattaria (subgroups 2.2.2. and 2.2.3.2.2.) and in Mantodea.

(F) The presence or absence of the dcea-processes

The dca-processes (6.1.1., 6.1.4.) have been assumed to be ground-plan elements of
Blattaria. dea are present in Eurycotis, Archiblatta, Tryonicus, Cryptocercus, Polyphaga,
Ergaula, Nahublattella, and possibly Lamproblatta but absent in the remaining Blattaria
(Anaplecta and subgroup 2.2.3.2.2.) and in Mantodea.

(G) The connection or separation of the regions Rlt and Ric

In most Blattaria and in the Mantodea Chaeteessa and Sphodromantis the regions R1t and
Ric are firmly connected, and several Blattaria and Chaeteessa have a distinct edge 16,
which has been defined as the border between Rit and Ric. This has been regarded as
the condition of the common ground-plan of Blattaria and Mantodea (6.7.1., 7.1.). In most
Blattaria of subgroup 2.2.3.2.2.2. (6.7.6.) as well as in the Mantodea Mantoida and
Metallyticus (6.7.3.) Rit and Ric are separated from each other.

(H) The connection or separation of the regions Ric, Rid, and Riv (= the absence or
presence of the articulations A8 and AQ)

That Rid as well as Rlv are separated from Ric by the articulations A8 and A9 (6.7.1.,
6.7.6.), respectively, has been regarded as a ground-plan feature of Blattaria (7.3.). This
separation is present in Eurycotis, Archiblatta, Tryonicus, Lamproblatta, and Cryptocercus,
but in the Blattarian subgroups 2.2.2.2.2. and 2.2.3. (6.7.6.) and in the ground-plan of
Mantodea (6.7.1.) the regions Ric, Rid, and R1v are contained in one sclerite.

(I) The presence or absence of the tre-tendon and of muscle s8

The tre-tendon (6.7.1., 6.7.5.) and the s8-muscle (6.9.) have been assumed to be ground-
plan elements of Blattaria. tre and s8 are present in Eurycotis, Archiblatta, Tryonicus,
Cryptocercus, Polyphaga, and Ergaula (s8 not studied in Archiblatta and Tryonicus) but
absent in the remaining Blattaria (Lamproblatta and subgroup 2.2.3.) and in Mantodea.

301

(J) The presence or absence of the rge-groove

The rge-groove (6.7.1., 6.7.6.) has been assumed to be a ground-plan element of Blattaria.
rge is present in Archiblatta, Eurycotis, Tryonicus, Cryptocercus, Lamproblatta,
Polyphaga, and Ergaula but absent in the remaining Blattaria (subgroup 2.2.3.) and in
Mantodea.

(K) The presence or absence of muscle r6

The r6-muscle (6.7.1., 6.7.6.) has been assumed to be a ground-plan element of Blattaria.
r6 is present in Eurycotis, Lamproblatta, Polyphaga, and Ergaula but absent in the
remaining Blattaria (Cryptocercus and subgroup 2.2.3.) and in Mantodea. (Archiblatta and
Tryonicus not investigated.)

A comparison between Blattaria and their outgroup Mantodea could hence lead to the view

that in the ground-plan of Blattaria nla, dca, rge, tre, s8, r6, and the curvature of L2 and

Ive are missing; Rit is separated from Rle, and Ric is connected with both Riv and Rid

(no articulations A8 and AQ). The counterparts of these character states would then have

developed within Blattaria and would have to be regarded as possible autapomorphies of

Blattarian subgroups. From this view the following phylogenetic hypothesis could arise:

1. Part of subgroup 2.2.3.2.2.2. (Euphyllodromia + Parcoblatta + Nyctibora) is the sister-
group of the other Blattaria. All remaining Blattaria have connected Rit and Ric.

2. The next offshoots are Supella and Blaberus + Nauphoeta + Blaptica + Byrsotria.
All remaining Blattaria have developed the nla-process (lost again in Lamproblatta,
Ergaula, Polyphaga, and Cryptocercus) and the dca-processes (lost again in
Anaplecta), and they have also developed, for example, sclerite L1 and region L4d
(both lost again in Anaplecta), the posterior branching of sclerite L2 (the first hint of
the arch-shape), and the division of the via-process into paa and pda.

3. The next offshoot is Nahublattella. All remaining Blattaria have developed the
curvature of L2 and Ive (lost again in Cryptocercus), and they have also translocated
the anterior insertion of 114 from the Ive-apodeme to nla (L4n-region).

4. The next offshoot is Anaplecta. All the remaining Blattaria have developed the rge-
groove, and they have also lost the lve-apodeme and developed the contact between
L1 and L2 (articulation A2).

5. The next dichotomy would be questionable: Polyphaga + Ergaula could be the next
offshoot, with the remaining species having separated Ric+Rit and Rid+R1v (new
articulations A8 and AQ). Alternatively, Lamproblatta could be the next offshoot, with
the remaining species having developed the tre-tendon and the s8-muscle.
Alternatively, Cryptocercus could be the next offshoot, with the remaining species
having developed the r6-muscle.

6. Eurycotis + Archiblatta and Tryonicus would represent a holophyletic group which
has, for example, developed a close contact between the anterior and posterior parts
of the L4l-region and rotated the L4d-region anteriad (from a formerly leftward
orientation).

7. In Eurycotis + Archiblatta the anterior and the posterior parts of the L4l-region have

302

fused, L4l has developed an swe-apodeme, and the right phallomere has evolved the

pia-tooth.
This phylogenetic hypothesis as a whole is for several reasons extremely improbable: (1)
If accepting it, one would have to suppose that many of the features assumed for the
common ground-plan of Blattaria and Mantodea (listed in 7.1.) are not ground-plan
features but similarities born by parallel evolution. Some of the most important have been
mentioned in the steps 1.-7. Concerned are, for example: the detailed similarities of the
L4l- and L4d-regions and of the swe-apodeme in Mantoida and Archiblatta; the arch-
shape of L2 and articulation A2 in e.g. Mantoida, Archiblatta, and Polyphaga; the similar
morphology of paa and pda in Mantoida and Tryonicus; the pia-teeth of Mantodea and
e.g. Archiblatta. (2) In addition, some elements present in subgroup 2.2.3. or its subordi-
nate subgroups but not in the other Blattaria and in Mantodea would have to be regarded
as ground-plan elements of Blattaria, e.g. the hook-like curvature of the median end of
the Rit-region, inclusive of its ewe-thickening, and the Ive-apodeme. Features of the
females could be added to this “ground-plan” list, e.g. the advanced rotation of the ootheca.
Hence, as regards (1) and (2), this hypothesis would be extremely conflicting with the
outgroup comparison of Blattaria with Mantodea — much more than the hypothesis
proposed in 7.4. (3) This hypothesis is rather inconsistent in itself: In several cases
secondary reductions (e.g. nla-process in 2.) or parallel evolution (situation in 5.) have to
be assumed. If the developments contained in the steps 1.-5. are arranged in another way,
some of these assumptions could be avoided, but they would only unavoidably be replaced
by other assumptions of secondary loss or parallel evolution.

If only some or even only one of the polarity statements of this alternative hypothesis are
accepted, this would still cause extensive inconsistencies either with the ground-plan
hypothesis given in 7.1. or with the clusters of assumed autapomorphies given in 7.4. If
it is, for example, supposed that rge (J) is an autapomorphy of a Blattarian subgroup
comprising Archiblatta, Eurycotis (and the other species assigned to subgroup 2.1.),
Tryonicus, Lamproblatta, Cryptocercus, Polyphaga, and Ergaula, and that the lack of rge
in subgroup 2.2.3. is the plesiomorphic state, one would have to accept many cases of
parallel evolution. Which features would have to be regarded as having evolved several
times depends on which type of phallomere complex is regarded as plesiomorphic for this
hypothetical grouping: (1) If the basic phallomere morphology is supposed to resemble
Archiblatta, all the assumed autapomorphies of subgroup 2.2. must have developed two
times independently. (2) If the basic phallomere morphology is supposed to resemble either
Tryonicus, Lamproblatta, Cryptocercus, Polyphaga, or Ergaula, most of the similarities
in the morphology of the L4l- and L4d-regions (including the similar insertions of 12 and
14; 6.3.1.) and in the posterior part of the right phallomere (with the fda-lobe and the pia-
tooth; 6.7.1.), which have revealed in the comparison between Mantoida, Chaeteessa,
Archiblatta, and Eurycotis, would have to be regarded as cases of parallel evolution. (3)
If any combination or mixture of these types is supposed to represent the basic phallomere
morphology, the extent of parallel evolution having to be accepted could not be decreased,
but only the assumptions of parallel evolution necessary for (1) and (2) would mingle.

If it is assumed that either tre and s8, or r6, or the L2-curvature, or nla, or dea, or the
separation of Rld and Rlv from Ric, or the connection of Ric and RIt is an

303

autapomorphy of the respective Blattarian grouping, one would likewise have to accept
extensive parallel evolution — in an analogous way as described for rge. It is thus certainly
by far most parsimonious to regard all these elements or properties as ground-plan features
of Blattaria (like in 7.3.) and to assume secondary loss or change when any of these
elements or properties is missing in any of the Blattarian species investigated in this paper.

As regards (E) and (I), this view is additionally supported by arguments concerning the
functional intercorrelation of phallomere elements. (E): compare discussion in 7.5. (M),
(N). (I): The b4-muscles, which in Blattaria insert with s8 on tre, are probably elements
of the common ground-plan of Blattaria and Mantodea (6.7.1., 7.1.). When present all
together, tre, b4, and s8 are certainly functionally intercorrelated elements (and in this
case the function of the b4-muscles is certainly different from that of the b4-muscles of
Mantoida). If reduction occurs in such a situation, all three elements can be expected to
be concerned. Hence, the lack of b4 in Lamproblatta and subgroup 2.2.3. (there are no
muscles in a similar position as b4a and b4b are in Mantoida) might indicate that tre and
s8 were present in former times.

(L) The presence or absence of muscle s7

s7 (6.9.) has been assumed to be a ground-plan element of Blattaria (7.3.). s7 is present
in the subgroups 2.1. and 2.2.3., and vestiges are probably present in subgroup 2.2.2.1.
(Cryptocercus). In subgroup 2.2.2.2. and in Mantodea s7 is absent. Subgroup 2.2.1.
(Tryonicus) has not been investigated. Hence, s7 could be a synapomorphy of the
subgroups 2.1. and 2.2.3. and possibly Cryptocercus. However, since this assumption
would be inconsistent with the assumed autapomorphies of the subgroups 2.2. and 2.2.2.,
it is assumed that s7 has been lost secondarily in Lamproblatta, Polyphaga, and Ergaula.
The lack of s7 in Blaberus (6.9.) is certainly a secondary loss since s7 is present in all
other investigated species of subgroup 2.2.3. (inclusive of Nauphoeta).

The following discussions under (M) + (N) and (O) will be concerned with the polarity
of some characters of Blattaria for which an outgroup comparison with Mantodea is not
possible since the respective elements (hla-hook or dea-processes) are present in all
Blattaria (hla) or at least in the Blattarian ground-plan (dea, compare (F)) but not in
Mantodea. A result can be achieved in interdependence with the phylogenetic hypothesis
in 7.4., but, mainly in the case of (M) and (N), also independently of this hypothesis, if
correlations with other elements for which an outgroup comparison with Mantodea is
possible are considered.

(M) The position of the hla-base
(N) The extension of the membranous basal part 30 of hla

The hla-hook and the L3-sclerite are present in all Blattaria (6.4.3.). In 7.3. it has been
Stated that in the ground-plan of Blattaria the hla-base takes a position in the left anterior
ventral wall of the left complex, and that the introversible membranous basal part 30 of
hla is narrow (and hla is therefore — almost — non-retractable). These statements have to
be substantiated.

304

In dependence on the phylogenetic hypothesis in 7.4. the following view results: A far
anterior position of the hla-base and a narrow membrane 30 (and a non-retractable hla)
are present in the subgroups 2.1. and 2.2.2.2. Thus, the more posterior position of the hla-
base and the expansion of membrane 30 (and the retractility of hla) can be regarded as
apomorphic states developed in the Blattarian subgroups 2.2.1. and 2.2.3. (These are
possibly synapomorphies of these two subgroups: (35) and (36) in 7.4.) The more posterior
position of the hla-base in subgroup 2.2.2.1. (Cryptocercus, fig.151), which is, however,
not accompanied by an expansion of membrane 30 and by a retractility of hla, is also
derived.

There are some functional intercorrelations between the various characters of hla and
between these and some other elements of the left complex. Consequently, the species
with a completely retractable hla (subgroup 2.2.3.) show some concomitant morphological
and functional differences to the species with a non-retractable hla. Since for some of
these correlated elements or properties an outgroup comparison with Mantodea and hence
a polarisation of the respective character states is possible, these intercorrelations permit
assessing the polarity of the characters of hla independently of the phylogenetic hypothesis
in 7.4. The following intercorrelations and evolutionary changes are assumed:

— The extension of membrane 30, the retractility of hla, and the position of the hla-base
are intercorrelated: If membrane 30 is more extensive, hla can be retracted more deeply,
and then its base can be more posteriorly without having hla exceeding the subgenital
plate or protruding from the genital pouch during its inactive state. From a functional
point of view, the more posterior position of the hla-base and the retractility of hla
might have the advantage that in its active state (during copulation) hla protrudes farther
from the genital pouch and is more flexible. In Cryptocercus the more posterior position
of the hla-base is not accompanied by an extension of membrane 30 but by a shortening
of hla (fig.151).

— The retractility of hla, the length of 114, and the positions of the 114-insertions are
intercorrelated: In the species with a non-retractable hla, muscle 114 has only the
function to move hla; to accomplish this function 114 has to be contracted for a short
distance only and can be rather short (fig.72). In the species retracting hla (subgroup
2.2.3.) the contraction distance of 114 has to be much longer, and thus 114 itself has to
be longer. This requirement is in part fulfilled by the posteriad shift of the posterior
114-insertion together with the hla-base. Additionally, the anterior insertion of 114 has
shifted anteriad. At a first stage, this latter shift has evidently been achieved by an
anteriad shift of the L4n-region and of the nla-process, the primary anterior insertion
area of 114; this stage is represented by Anaplecta (fig.222; compare Eurycotis, fig.68,
72, where nla and the 114-insertion are by far more posteriorly). At a later stage, in
subgroup 2.2.3.2., the anterior 114-insertion has been translocated to the top of the Ive-
pouch (lve-apodeme), and the 114-insertion is shifted further anteriad by a lengthening
of this Ive-apodeme (compare Anaplecta, fig.222, and Nahublattella and Parcoblatta,
fig.249, 276).

— The position of the anterior 114-insertion and the condition of the swe- and Ive-apodemes
and of the nla-process are intercorrelated: A lot of force seems to be needed in operating

305

the hla-hook, and there are various supporting mechanisms present to give the
contraction of 114 the proper effect, which is the movement or retraction of hla, and to
prevent a distortion of the areas (1) between the anterior 114-insertion and the hla-base
and (2) immediately around the anterior 114-insertion. (1) In Archiblatta and Eurycotis
the swe-apodeme might fulfil the former function (fig.53, 65), and since the morphology
of the L4l-region and swe is the same in Mantoida (6.3.1.), this mechanism is certainly
the most primitive within Blattaria. In subgroup 2.2.3. swe is missing, and the anterior
part of the Ive-pouch has been elaborated as a long and stiff Ive-apodeme. In Anaplecta
(subgroup 2.2.3.1.) the anterior 114-insertion on nla is stabilised by nla being firmly
rested upon the top of the Ive-apodeme (fig.209, 210). The stabilisation of nla might
be the primary function of the Ive-apodeme. In the further evolution (in subgroup
2.2.3.2.) the anterior 114-insertion has shifted even further anteriad, and it has apparently
been preferred to make more directly use of the supporting function of the Ive-apodeme
and to translocate the 114-insertion to the top of Ive. (2) The nla-process itself assumedly
has the function to stiffen the immediate vicinity of the anterior 114-insertion — as long
as this insertion is on the L4n-region — and is therefore bulge-like (Archiblatta,
Eurycotis, Anaplecta). In species having this insertion translocated to the lve-apodeme,
nla is either modified in its shape (Nahublattella: whip-like) or has been lost
(Parcoblatta, Blaberus).

— The position of the anterior 114-insertion and the condition of the phallomero-sternal
muscles sl and s7 are intercorrelated: Additional stabilisation preventing a distortion of
the area around the anterior 114-insertion is probably achieved by muscles conducting
much of the force which 114 exerts to this area to the left apophysis of the subgenital
plate. These phallomero-sternal muscles insert immediately anterior to the anterior I14-
insertion. In Eurycotis this function is accomplished by muscle sl, which inserts
between nla and the anterior end of swe (fig.70). In the species of subgroup 2.2.3.,
where the supporting function has been transferred from swe to Ive and where, except
for Anaplecta, the muscle-insertion for which the support is needed (114) has been
transferred from nla to Ive, the function of the “conductor’-muscle has consequently
been transferred from sl to s7, which inserts anteriorly on the Ive-apodeme. Muscle sl
has been lost in all species using a lve-apodeme for support (subgroup 2.2.3.; 6.9.).

The reciprocal outgroup comparison between Blattaria and Mantodea clearly suggests that
swe (6.3.1.) and s1 (6.9.) are ground-plan elements, that the primitive position of the L4n-
region is like in Eurycotis (6.3.1.; compare fig.325c and e), and that there is no tube-
shaped Ive-apodeme. The lack of swe and sl, the extremely far anterior position of the
L4n-region (or its lack), and the Ive-apodeme — the features of subgroup 2.2.3. — are
certainly derivations within Blattaria. Especially the assumedly primitive position of the
L4n-region (like in Eurycotis) is clearly correlated with a short 114 inserted on L4n, with
the non-retractility of hla, and with a narrow basal membrane 30, and a non-retractable
hla can be expected to have its base far anteriorly. Thus, the non-retractable hla with its
base in the anterior ventral wall can be assumed to represent the plesiomorphic condition
within Blattaria. A bulge-shaped nla-process can also be regarded as a ground-plan element
of Blattaria, since it is present in all species showing the primitive condition that the
anterior 114-insertion is on a well-developed L4n-region.

306

In subgroup 2.2.2. the anterior 114-insertion (114 present in Lamproblatta and Cryptocercus
only) is not stabilised by cuticular elements (no swe- or Ive-apodemes, L4n-region highly
reduced, no nla-process), and the force acting on the anterior 114-insertion seems to be
completely conducted to the subgenital plate by the s1-muscles, which insert immediately
anterior to 114 (fig.157, 158, 184, 185). Also in subgroup 2.2.1. (Tryonicus) neither an
swe- nor an lve-apodeme are present, and the stabilisation by cuticular elements does not
seem to be very effective. The mechanism cannot be assessed here since the muscles are
not known.

(O) The shape and sclerotisation of the dca-processes

dca-processes are restricted to Blattaria (6.1.1.) and are probably ground-plan elements of
this taxon (7.5. (F)). In 7.3. it has been stated that in the Blattarian ground-plan the dea
are two cushion-like processes posterior to L1; this statement has to be substantiated. Two
membranous cushion-like dea are present in representatives of both of the basic Blattarian
subgroups 2.1. (Archiblatta, fig.54) and 2.2. (Tryonicus angustus, fig.107, Cryptocercus,
fig.153, Polyphaga, fig.120). More or less completely sclerotised (by L1) dea are also
present in both the subgroups 2.1. (Eurycotis, fig.67) and 2.2. (Tryonicus parvus, fig.94,
Nahublattella, fig.243, 244), but since the L1m-region is a ribbon-like extension in the
common ground-plan of Blattaria and Mantodea, and since L1m is expanded posteriad in
the species with sclerotised dea-processes (to contribute to this sclerotisation), this is
assumed to be a derived state.

In those species of subgroup 2.2. having two membranous dca, these dca are very similar
(fig.107, 120, 153; these are the members of the subsroups 2-22.13 and > 7 excep
Ergaula, fig.105, and Lamproblatta, fig.177, but no member of subgroup 2.2.3. is
concerned). This peculiar shape of the dca is assumed to represent the plesiomorphic state
within subgroup 2.2. It is not regarded as a synapomorphy of the respective species since
this assumption would be inconsistent with the many assumed autapomorphies of the
subgroups 2.2.2. and 2.2.2.2. It also cannot be regarded as a synapomorphy of the
subgroups 2.2.1. and 2.2.2. (assuming a secondary change in Lamproblatta and, less
drastic, in Ergaula) since the shape of the dea in their primitive membranous condition
is not assessable in subgroup 2.2.3.: Here the dea are either completely sclerotised
(Nahublattella) or missing (remaining species), situations which are both derived.

A sclerotised peak 18 in between the dca-processes is present only in Tryonicus angustus
and Cryptocercus (fig.107, 153). To regard this as a synapomorphy (assuming secondary
loss in Tryonicus parvus) would be inconsistent with the autapomorphies of subgroup
2.2.2. only, but in my view the latter are much more conclusive. The peaks 18 are possibly
non-homologous in the two species, or they are again an element of the ground-plan of
subgroup 2.2.

The following discussions under (P) and (Q) + (R) will be concerned with the evolution
of three characters for which the polarity is essentially clear, but for which the distribution
of the character states within Blattaria is somewhat in conflict with the phylogenetic
hypothesis in 7.4.; reversals of apomorphic character states seem to have taken place.

307

These discussions will be dependent on the phylogenetic hypothesis in 7.4., and the
soundness of the results depends on the soundness of this hypothesis.

(P) The presence or absence of a sclerite-ring formed by the regions Lim, L1l, and Lir

A distinct extension Lim and possibly also a less distinct extension L1l are elements of
the common ground-plan of Blattaria and Mantodea (6.1.1.). However, only in species of
the Blattarian subgroup 2.2. L1l and Lim curve ventrad (L1m does this in a different
way as in the Mantodean subgroup 1.2.) and approach each other again (6.1.4.); this is
certainly a derived feature. The sclerite ring is either complete (Ergaula, Cryptocercus),
and in this case it sometimes additionally expands onto the dca-processes (Tryonicus
parvus, Nahublattella), or the ring has a short gap (ventrally in Polyphaga, dorsally in
Tryonicus angustus). In all other species of subgroup 2.2. the feature is not assessable
since L1 has been completely lost. The only exception is Lamproblatta, which shows no
trace of a ring though L1 is present; this is assumed to be a secondary loss because of
the many assumed autapomorphies of the subgroups 2.2.2. and 2.2.2.2. A complete sclerite-
ring 1s assumed to be a ground-plan feature of subgroup 2.2.

(Q) The connection or separation of the sclerotisations of the Ive-pouch (L2 or L2D) and
of the paa- and pda- (or via-) processes (L2+L4N or L2ZE+L4N)
(R) The absence or presence of muscle 110

In the common ground-plan of Blattaria and Mantodea the sclerotisation of the lve-pouch
is connected with the common sclerotisation of the paa- and pda-processes (6.2.1., 7.1.),
and the same is true of the ground- plan of the Blattarian subgroup 2.2. Muscle 110 con-
necting these two sclerotisations is certainly also a ground-plan element of subgroup 2.2.
(matmleastaol) 222, + 2.2.3. since Tryonieus = 2.2.1. has not been investigated; 7.3.).
Within subgroup 2.2.3. (fig.328) the sclerotisations concerned can be connected with
(Anaplecta, Euphyllodromia, Parcoblatta, Blaberus) or separated from each other
(Nahublattella, Supella, Nyctibora, Nauphoeta), and muscle 110 is present in the species
showing a separation (and in Anaplecta) but absent in the species showing a connection
(except for Anaplecta).

Anaplecta is the first offshoot within subgroup 2.2.3. The outgroup comparison with the
subgroups 2.2.1. and 2.2.2. clearly suggests that the connection of the sclerotisations in
Anaplecta corresponds to the plesiomorphic state. Nahublattella is the next offshoot, and
Supella follows. Both species show an (probably homologous) apomorphic separation of
these sclerotisations. In the remaining species, which together form subgroup 2.2.3.2.2.2.,
both conditions — connection and separation — are represented. The outgroup comparison
with the subgroups 2.2.3.2.1. (Nahublattella) and 2.2.3.2.2.1. (Supella) clearly suggests
that within subgroup 2.2.3.2.2.2. the connected condition is an apomorphic state. This view
is supported by the additional lack of 110, which is certainly a derived feature, and which
is probably correlated with this secondary fusion of the sclerotisations.

The fusion of the sclerotisations and the concomitant loss of 110 have certainly happened
several times within subgroup 2.2.3.2.2.2. since the distribution of these two derived
features is completely inconsistent with some other well-founded groupings: (1) Blaberus

308

and Nauphoeta are (together with Blaptica and Byrsotria) members of the certainly
holophyletic taxon Blaberidae (McKittrick 1964). (2) Parcoblatta, Nyctibora, Blaberus,
and Nauphoeta are (together with Blaptica and Byrsotria) members of a probably
holophyletic taxon characterised by the autapomorphies (115)-(117) in 7.4. Hence, the
fusion and the loss of 110 can be assumed to have been achieved independently in
Euphyllodromia, Parcoblatta, and Blaberus.

7.6. Conflicts in the distribution of character states

In some characters whose polarity is rather clear, the distribution of the apomorphic state(s)
over the taxa is inconsistent with the phylogenetic hypothesis in 7.4.; these inconsistencies
can only be resolved by the assumption of either parallel evolution in the one or reversal
in the other species. In some other characters, mainly in those discussed in 7.5., the polarity
might be supposed to be the reverse of that assumed in 7.4. or resulting from the
discussions in 7.5. The discussions have shown that this is highly improbable, but some
doubt may remain. If the reverse polarity is supposed, the distribution of the surmisedly
apomorphic character states would also be inconsistent with the hypothesis in 7.4.

The (clearly or possibly) apomorphic character states concerned and the groupings they
might support will be listed here, and the respective plesiomorphic states are given in
brackets. This will be done in the same manner as in 7.4. The conclusiveness of many of
these “apomorphic” states is in terms of their value as possible autapomorphies decreased
by uncertain homology relations, uncertain or even improbable polarity assumptions, or
other circumstances. Some of these states, however, could really be autapomorphies and
could hazard the phylogenetic hypothesis in 7.4., but only in very few points.

In the species provided with “?” some or all of the respective characters have not been
investigated. For muscles the two insertion areas are given.

Arguments for alternative groupings within Mantodea
Grouping A: Metallyticus + (Mantoida + Sphodromantis)

(122) Groove on region Ric behind articulation A3 and anterodorsal to right insertion of
muscle r3 present (BM: Groove on Rlc behind A3 and anterodorsal to left r3-insertion
absent).

This groove on the Rlc-region (fig.6, 20, 41), not homologous with the rge-groove of
Blattaria (6.7.1.), 1s very distinct in Mantoida and Sphodromantis, hardly recognisable in
Metallyticus, and missing in Chaeteessa (6.7.3.). To interpret the groove as an
autapomorphy of this grouping would be in conflict with the many assumed
autapomorphies of the subgroups 1.2. and, since the groove is more distinct in Mantoida
and Sphodromantis, 1.2.2. It is assumed that this groove is a ground-plan element of
Mantodea and has been reduced in Metallyticus and lost in Chaeteessa. The following fact
supports this interpretation: The age-apodeme on R3 probably reaches articulation A3 in
the common ground-plan of Blattaria and Mantodea (6.7.1., 7.1.), but in Chaeteessa age
does not reach A3, and the right posterior part of age has evidently been reduced. A
concomitant reduction of the groove posterior to A3 would be plausible.

309

Grouping B: Mantoida + Metallyticus

(123) Region Rit separated from region Ric: sclerites RID, RIE or RID, RIC (BM: RIt
connected with Rlc).

To regard this certainly apomorphic state (6.7.1., 6.7.3.; compare (G) in 7.5.) as a
synapomorphy of Metallyticus and Mantoida would be in conflict with the assumed
autapomorphies of the Mantodean subgroups 1.2. and 1.2.2. Parallelism is thus highly
suggested. Support for this assumption comes from the fact that some other taxa have also
achieved this state independently: According to LaGreca (1955), at least Amorphoscelis
abyssinica (Amorphoscelididae), Tarachodes insidiator (Mantidae), and Polyspilota sp.
(Mantidae) (fig.13, 15, 16 in LaGreca) show the separation of R1t from Rlc, and this is
certainly not a synapomorphy of these distantly related species. The Blattarian subgroup
2.2.3.2.2.2. also shows the separation of R1t and Ric.

Arguments for alternative groupings within the Blattarian subgroup 2.2.

Grouping C: Tryonicus + Cryptocercus + (Anaplecta + (Nahublattella + (Supella +
(Euphyllodromia + (Parcoblatta + (Nyctibora + (Blaberus + Nauphoeta + Blaptica +

Byrsotria)))))))

(124) Left edge 61 of lobe vla extending far anteriad (BM: 61 ending far posteriorly).
This certainly apomorphic state (6.2.1., 6.2.4.) is distinct in Tryonicus, Cryptocercus,
Anaplecta, and Nahublattella, and the character is hardly assessable in the other species
listed. Hence, it might be an autapomorphy of this grouping. This character is inconsistent
with the assumed autapomorphies of subgroup 2.2.2.

Grouping D: Cryptocercus + (Anaplecta + (Nahublattella + (Supella? +
(Euphyllodromia? + (Parcoblatta + (Nyctibora? + (Blaberus + Nauphoeta? + Blaptica?
+ Byrsotria?)))))))

(125) Muscle s10 present: from subgenital plate to ejaculatory duct (BM: s10 absent) (126)
Muscle r6 absent: from region Ric to region Rld (SG2.: r6 present) (127) Sclerites L1
and L2 far separated: articulation A2 absent (BM: L1 and L2 articulated in A2).

(127) is certainly the apomorphic state (6.2.4.), and the same is probably true of (125)
(6.9.) and (126) (6.7.6., (K) in 7.5.). These characters are inconsistent with the assumed
autapomorphies of subgroup 2.2.2. but consistent with (124) of grouping C. (s10 and r6
not investigated in Supella, Euphyllodromia, Nyctibora, Nauphoeta,- Blaptica, and
Byrsotria.) As regards (127), however, non-homology is suggested for Cryptocercus,
Anaplecta, and the other species: Cryptocercus and Nahublattella have lost the right part
of L2, which curves upwards and bears articulation A2 distally in the Blattarian ground-
plan, but have retained L1. Anaplecta has retained the upcurved right part of L2 but has
lost L1. Hence, A2 has possibly been lost in different ways. If this is true, A2 has been
lost three times since Anaplecta and Nahublattella are clearly more closely related
(autapomorphies of subgroup 2.2.3. in 7.4.).

310

Grouping E: Lamproblatta + (Anaplecta + (Nahublattella + (Supella + (Euphyllodromia
+ (Parcoblatta + (Nyctibora + (Blaberus + Nauphoeta + Blaptica + Byrsotria)))))))

(73) Tendon tre and muscles s8 and b4a,b absent (SG2.: tre, s8, and b4a,b present; BM:
Presence of tre and s8 unclear, b4a,b present).

This is probably the apomorphic state (compare (I) in 7.5.; 6.7.5., 6.9.). To regard it as
an autapomorphy of this grouping would be inconsistent with the many assumed
autapomorphies of the subgroups 2.2.2. and 2.2.2.2.

Grouping F: (Polyphaga + Ergaula) + (Anaplecta + (Nahublattella + (Supella +
(Euphyllodromia + (Parcoblatta + (Nyctibora + (Blaberus + Nauphoeta + Blaptica +
Byrsotria)))))))

(128) Region Ric probably fused with at least part of the broadly interconnected regions
Rid and Rlv: articulations A8 and A9 and membranous area 17 absent; sclerites RIM or
RIN (SG2.2.: Ric separated from the broadly interconnected Rld and Rlv: A8, A9, and
17 present; SG.2.: Rlc separated from the narrowly interconnected Rld and Riv: A8, A9,
and 17 present; BM: Rlc connected with both Rld and Rlv which are separated from
each other: A8 and A9 absent but 17 present) (64) Muscle r3 absent: from region Ric to
region Rlv (BM: r3 present).

(128) is a less specific formulation of (63) as well as (74) (compare list in 7.4.) and is
true of both the subgroups 2.2.2.2.2. and 2.2.3. (128) and (64) are certainly apomorphic
— and the characters are probably intercorrelated (6.7.6.; compare (H) in 7.5.). The special
kind of fusion, however, is rather different in Polyphaga + Ergaula (subgroup 2.2.2.2.2.)
and in the other species (subgroup 2.2.3.); consequently, it is highly questionable whether
the fusion, the resulting sclerites RIM and RIN, and the loss of r3 are each homologous.
If these states are still regarded as autapomorphies of this grouping, these characters would
be inconsistent with the many assumed autapomorphies of the subgroups 2.2.2. and 2.2.2.2.

Grouping G: Tryonicus + Lamproblatta

(129) Apodeme age absent (BM: age present) (130) Extension R2m present (SG2.: R2m
absent).

Both states are certainly apomorphic (6.7.4.). To regard them as autapomorphies of this
grouping would be inconsistent with the many assumed autapomorphies of the subgroups
222, andy 222.2,

Grouping H: Cryptocercus + (Polyphaga + Ergaula)

(131) Muscle s5 absent: from subgenital plate to left wall of genital pouch (BM: s5
present).

This state is certainly apomorphic (6.9.). To regard it as an autapomorphy of this grouping
would be inconsistent with the assumed autapomorphies of subgroup 2.2.2.2.

Grouping J: Cryptocercus + Polyphaga
(132) Sclerite RIK present (BM: RIK absent).

The presence of a separate RIK is certainly apomorphic. However, the homology of the

311

RIK of these two species is uncertain since it is only indicated by a roughly similar
position (6.7.6.). If this state is still regarded as an autapomorphy of this grouping, the
character would be inconsistent with the many assumed autapomorphies of the Blattarian
subgroups 2.2.2.2. and, if RIK is not assumed to have been lost secondarily in Ergaula,
Dede:

Grouping K: Cryptocercus + (Lamproblatta + Anaplecta)

(133) Muscle 14 reduced or absent: from sclerite L2 to region L41 (BM: 14 present and
stout).

14 is certainly stout in the common ground-plan of Blattaria and Mantodea (6.2.1., 6.3.1.).
Its reduction in Cryptocercus and its loss in Lamproblatta and Anaplecta are certainly
apomorphic states (6.3.4.). To regard them as autapomorphies of the respective groupings
would be inconsistent with the many assumed autapomorphies of the subgroups 2.2.2.,
2.2.2.2., and 2.2.3. To regard the reduction of 14 as a synapomorphy of Lamproblatta and
Cryptocercus (assuming parallel loss in Anaplecta) would be inconsistent with the assumed
autapomorphies of subgroup 2.2.2.2. However, the possibility must be considered that 14
was strongly reduced in the ground-plan of subgroup 2.2.2. and has enlarged secondarily
at the base of subgroup 2.2.2.2.2. (Polyphaga + Ergaula) where the hla-muscle 114 has
been lost and 14 has acquired a new function in moving the hla-hook (6.3.4.). In this case
the reduction in Cryptocercus and Lamproblatta would at least be homologous.

Grouping L: Anaplecta + (Supella? + (Euphyllodromia? + (Parcoblatta + (Nyctibora? +
(Blaberus + Nauphoeta? + Blaptica? + Byrsotria?)))))

(109) Sclerite L1 absent (BM: L1 present) (110) Process(es) dca absent (SG2.: dca present)
(134) Pouch pne indistinct (BM: pne distinct) (25) Muscle 11 absent: from pouch pne to
region L4d (BM: 11 present) (95) Region L4d absent (BM: L4d present).

These states are certainly all apomorphic (6.1.4., 6.3.4.), and they are probably all
intercorrelated since (109), (110), (134), and (25) relate to reductions in the same area,
and (95) relates to the opposite insertion area of muscle 11. L1, dea, a distinct pne, 11,
and L4d are all clearly present in Nahublattella. In Supella and Euphyllodromia (109),
(110), (134), and (25) have not been investigated. In Nyctibora, Nauphoeta, Blaptica, and
Byrsotria (25) has not been investigated.

If these character states are regarded as autapomorphies of this grouping, the characters
would be inconsistent with the assumed autapomorphies of subgroup 2.2.3.2. Since all
states relate to the reduction of elements and are intercorrelated, the assumed
autapomorphies of subgroup 2.2.3.2. are regarded as much more convincing: These include
more complicated features, e.g. the 114-translocation (79), the L4K-division (84), and the
12-shift (86). Moreover, the presence of L1 in another species of Anaplecta (McKittrick
1964) might suggest that (109) is not an autapomorphy of this grouping; however, since
the phylogenetic position of this species within subgroup 2.2.3. is not known (the genus
Anaplecta is not necessarily holophyletic), this argument is not of high value.

32

Arguments for other alternative groupings within Blattaria

Grouping M: (Archiblatta? + Periplaneta + Blatta + Deropeltis + Eurycotis) + Tryonicus?
+ (Anaplecta + (Nahublattella + (Supella? + (Euphyllodromia? + (Parcoblatta +
(Nyctibora + (Blaberus + Nauphoeta + Blaptica? + Byrsotria?)))))))

(135) Muscle s7 present: from subgenital plate to L2 (SG2. and BM: s7 absent) (24)
Region Lla level (BM: Lla hood-shaped).

s7 (6.9.) has been regarded as a ground-plan muscle of Blattaria (7.3.) which has been
lost in subgroup 2.2.2.2. (compare (56) in 7.4., with the reverse polarity assumption, and
(L) in 7.5.). The anterior part of L1 (L1a) is hood-shaped in the common ground-plan of
Blattaria and Mantodea (6.1.1.). In Archiblatta, Periplaneta, Blatta, Deropeltis, and
Eurycotis, in both species of Tryonicus, and in Nahublattella Lla has become level
(6.1.4.). In the other species listed this character has not been investigated (Supella,
Euphyllodromia) or ıs not assessable for the complete loss of L1 (remaining species).
Lamproblatta does not reveal (24) since it has a distinct vestige of the hood- or even of
the plateau-shape (L1 bends into the dorsal pne-wall; 6.1.4.).

To regard (24) and (135) as autapomorphies of this grouping would be inconsistent with
the assumed autapomorphies of subgroup 2.2. If Cryptocercus really has vestiges of $7,
(135) would moreover be inconsistent with the assumed autapomorphies of subgroup 2.2.2.

Grouping N: (Archiblatta? + Periplaneta + Blatta + Deropeltis + Eurycotis) +
(Tryonicus? + (Cryptocercus + (Polyphaga + Ergaula)))

(136) Tendon tre and muscle s8 present (SG2. and BM: tre and s8 absent).

The presence of tre (6.7.1., 6.7.5.) and s8 (6.9.) has been regarded as a ground-plan feature
of Blattaria (7.3.; compare (73) in 7.4., with the reverse polarity assumption, and (I) in
7.5.). (s8 not investigated in Archiblatta and Tryonicus). If this state is regarded as an
autapomorphy of this grouping, the character would be inconsistent with the assumed
autapomorphies of the subgroups 2.2. and, if tre and s8 are not assumed to have been lost
secondarily in Lamproblatta, 2.2.2. and 2.2.2.2.

Grouping O: (Archiblatta? + Periplaneta + Blatta + Deropeltis + Eurycotis) +
(Tryonicus? + (Lamproblatta + (Polyphaga + Ergaula)))

(137) Muscle r6 present: from region Ric to region Rld (SG2. and BM: r6 absent).
The presence of r6 (6.7.6.) has been regarded as a ground-plan feature of Blattaria (7.3.;
compare (126) of grouping D, with the reverse polarity assumption, and (K) in 7.5.). If
this state is regarded as an autapomorphy of this grouping, the character would be
inconsistent with the assumed autapomorphies of the subgroups 2.2. and 2.2.2.

Grouping P: (Archiblatta + Periplaneta + Blatta + Deropeltis + Eurycotis) + (Tryonicus
+ (Cryptocercus + (Lamproblatta + (Polyphaga + Ergaula))))
(138) Groove rge present (SG2. and BM: rge absent).

The presence of rge (6.7.1., 6.7.6.) has been regarded as a ground-plan feature of Blattaria
(7.3.; compare (75) in 7.4., with the reverse polarity assumption, and (J) in 7.5.). If this

318

state is regarded as an autapomorphy of this grouping, the character would be inconsistent
with the assumed autapomorphies of subgroup 2.2.

Grouping Q: (Archiblatta + Periplaneta + Blatta + Deropeltis + Eurycotis) +
(Lamproblatta + Cryptocercus)

(139) Region Ric separated from regions Rld and Rlv: articulations A8 and A9 present
(SG2. and BM: Rlc connected with Rld and Rlv: A8 and A9 absent).

The presence of this separation and of A8 and A9 (6.7.1., 6.7.6.) has been regarded as a
ground-plan feature of Blattaria (7.3.; compare (128) of grouping F, with the reverse
polarity assumption, and (H) in 7.5.). If this state is regarded as an autapomorphy of this
grouping, the character would be inconsistent with the many assumed autapomorphies of
the subgroups 2.2., 2.2.2., and 2.2.2.2.

Grouping R: (Archiblatta? + Periplaneta + Blatta + Deropeltis + Eurycotis) +
(Tryonicus? + (Lamproblatta + (Polyphaga + Ergaula)) + (Anaplecta + (Supella? +
(Euphyllodromia? + (Parcoblatta + (Nyctibora? + (Blaberus + Nauphoeta? + Blaptica?
+ Byrsotria?)))))))

(25) Muscle II absent: from pouch pne to region L4d (BM: 11 present).

Muscle Il has been found only in Mantoida, Sphodromantis, Cryptocercus, and
Nahublattella, and since it is in the same relative position in all these species it has been
regarded as homologous and as a muscle of the common ground-plan of Blattaria and
Mantodea (6.1.1., 6.1.3., 6.1.4.). To regard the loss of Il as an autapomorphy of this
grouping would be inconsistent with the many assumed autapomorphies of the subgroups
2.2., 2.2.2. (since Cryptocercus is excluded), 2.2.3., and 2.2.3.2. (since Nahublattella is
excluded). It is thus clearly suggested that 11 has been lost several times (or that the 11
of Mantodea, Cryptocercus, and Nahublattella are not homologous despite their similar
positions).

7.7. Conclusions in terms of phylogeny

The phylogenetic ideas presented in 7.2.-7.4. are highly supported by many
autapomorphies for the various subgroups. The inconsistent characters supporting the
groupings listed in 7.6. can in most cases not compete with the clusters of autapomorphies
given in 7.4., and this is due to various reasons:

— The (certainly) apomorphic character state relates to the loss or reduction of an element
(such derivations are not as convincing in their role as possible autapomorphies as those
relating to the presence of new elements): (126), (127), (73), (64), (129), (131), (133),
(109), (110), (134), (25), (95).

— The (certainly or surmisedly) apomorphic character state is, if related as an
autapomorphy to one of the groupings in 7.6., the only one suggesting the respective
grouping, not supported by the distribution pattern of any other character and
inconsistent with the distribution pattern of many other characters: (25) as related to
grouping R, (122), (123), (124), (73), (131), (132), (133), (136), (137), (138), (139).

314

— For the (certainly) apomorphic character state the homology in the species concerned
is questionable since the possibility of parallel evolution is revealed by other species
having achieved the same apomorphic character state independently: (123).

— For the (certainly) apomorphic character state the homology in the species concerned
is questionable since except for a formal correspondence the morphology of the
respective elements is rather different: (127), (128), (132).

— The polarity of the character is unresolved or even suggested to be the reverse: (122),
(135), (136), (137), (138), (139).

— The polarity of the character is suggested to be the reverse in a certain part of the
phylogenetic tree, (i.e. the apomorphic character state has been secondarily reduced in
the crucial species excluded): (133).

In my view, the only conceivable alternative resulting from the list in 7.6. is that supported
by (124), (125), (126), and possibly (127): Cryptocercus might be the sister-group of
subgroup 2.2.3. (Blattellidae and Blaberidae) and not of subgroup 2.2.2.2. (Polyphaga,
Ergaula, Lamproblatta). And Tryonicus could well be the sister-group of Cryptocercus +
subgroup 2.2.3.: The possibility of a close relation between Tryonicus and subgroup 2.2.3.
has already been considered in 7.3., based on the similar morphology of the hla-hook
((35) and (36) in 7.4.). In Cryptocercus the hla-base has also shifted posteriad (fig.151),
and the retractility and the large extension of the membranous base 30 of hla present in
Tryonicus and subgroup 2.2.3. could well have been reduced in this species — in correlation
with the shortening of hla (compare (M), (N) in 7.5.). However, in my view, the very
similar reduction of sclerite L4K and of the nla-process and the shift of 12 in correlation
with the plateau-like shape of the anterior face of the pne-pouch, the arguments suggesting
that Cryptocercus belongs to subgroup 2.2.2., are somewhat more convincing ((42)-(46)
in 7243):

Another problematical issue is the assumed phylogeny of subgroup 2.2.3.2.2. Apart from
the fact that more species will have to be investigated in detail to get a really reliable
result, some character states of members of the genus Blattella are somewhat in conflict
with the hypothesis in 7.4. According to Mizukubo & Hirashima (1987), fig.41, there are
a dla-lobe and a R4-sclerite (= RD1d) and possibly also a R5-sclerite (= RD2v) present
in Blattella karnyi. (A muscle corresponding to r11 has not been found by these writers.)
In Blattella germanica (Linné, 1767) I could also find sclerites which are certainly R4
and R5. According to McKittrick (1964), the females of Blattella germanica rotate their
oothecae. These features suggest that Blattella belongs to subgroup 2.2.3.2.2.2.2.2. or, at
least (if R4 and RS5 are assumed to be secondarily reduced in Parcoblatta), to subgroup
2.2.3.2.2.2.2. (compare (115) and (117)-(120) in 7.4.). On the other hand, Blattella karnyi
(not B. germanica) resembles Nahublattella in that the posterior part of sclerite L2 is
branched, and each branch occupies a process. (The two branches of B. karnyi are LD2d
and LD2v in Mizukubo & Hirashima, fig.41; those of Nahublattella are the sclerotisations
of via and psa in fig.244, 245.) The morphology of this area would in B. karnyi be more
primitive than in all species included in subgroup 2.2.3.2.2. (compare (94) in 7.4.). This
might indicate that some of the apomorphic character states regarded as autapomorphies
of the subgroups 2.2.3.2.2., 2.2.3.2.2.2., and 2.2.3.2.2.2.2. are cases of parallel evolution,
or that R4, R5, and the rotation of the ootheca have developed earlier and have been

315

reduced again in various taxa belonging to subgroup 2.2.3.2.2. However, details of
morphology of the Blattella-species are not yet investigated. Thus, there is compelling
need for further investigations on the phallomeres of the various subgroups of Blattellidae
to resolve these problems in terms of the evolution and polarity of characters.

The most parsimonious phylogenetic hypothesis resulting from the discussions in chapter
7 ıs shown in diagram | in 7.4. If the species investigated in this paper are true
representatives of the Mantodean and Blattarian families and subfamilies they are usually
assigned to (compare the systems of McKittrick 1964 and Beier 1968 given in chapter 2),
the overall phylogeny of Mantodea and Blattaria is as follows:

In Mantodea, the basal dichotomy is between Mantoididae and all other families. The
second dichotomy is between Chaeteessidae and the remaining families. In Blattaria, the
basal dichotomy is between Blattinae + Polyzosteriinae and all other Blattaria. These
remaining Blattaria form three groups: The first consists only of the rather isolated
Tryonicinae. The second contains Cryptocercidae, Lamproblattinae, and Polyphaginae, the
two latter taxa being especially closely related. The third group corresponds to Blattellidae
+ Blaberidae. Blattellidae are clearly paraphyletic, with Blaberidae being a rather
subordinate subgroup. The earliest offshoot within Blattellidae (+ Blaberidae) are the
Anaplectinae; the three subsequent offshoots are various taxa previously comprised in
Plectopterinae. Blaberidae, Nyctiborinae, Blattellinae, and Ectobiinae form together a
holophyletic group. Nyctiborinae and Blaberidae are possibly sister-groups.

As regards Blattaria, this phylogenetic hypothesis is in several repects very different from
the system of McKittrick (1964):

— Tryonicinae are not related to Blattinae + Polyzosteriinae.

Lamproblattinae are also not related to Blattinae + Polyzosteriinae but to Polyphaginae.

Cryptocercidae are not the sister-group of Blattidae but probably of Polyphaginae +

Lamproblattinae (or possibly of Blattellidae + Blaberidae).

— Blattellidae are paraphyletic since Blaberidae are one of their subgroups. (McKittrick
has also expressed this idea in her phylogenetic trees — text figure 3 — but not in her
system).

— Plectopterinae are paraphyletic.

This hypothesis is based almost exclusively on male postabdominal and genital
morphology. Of course, there are still other character complexes which have proved to be
useful in analysing Dictyopteran phylogeny, e.g. the morphology of the female genitalia,
of the proventriculus (McKittrick 1964), or of the wings. The present knowledge on these
character complexes has been revised in a phylogenetic approach in Klass (1995), and a
study on the evolution of the ovipositor containing many new results has been completed
more recently (Klass, in press). The many characters which are now reliably interpretable
are consistent with the phylogenetic hypothesis presented here. Some characters, however,
are still problematic, due to insufficent (in quantity and quality) data. To improve the data
base for these character complexes, and also for the male genitalia, by detailed
morphological investigations should be the major task of future work on Dictyopteran

phylogeny.

316
7.8. Conclusions in terms of the side-reversal of the phallomere complex

Of the species discussed in this paper, Nahublattella, Supella, Euphyllodromia
(Plectopterinae), Blaberus, Byrsotria, Blaptica, and Nauphoeta (Blaberidae) have side-
reversed phallomeres, and this is certainly an apomorphic feature. According to Bohn
(1987), side-reversal also occurs in some species of Ectobius. All these species belong to
subgroup 2.2.3.2. If projected on the phylogenetic tree in 7.4., the distribution pattern of
this feature is as follows:

The three basal offshoots within subgroup 2.2.3.2. give rise to species with side-reversed
phallomeres (Nahublattella, Supella, Euphyllodromia). Subgroup 2.2.3.2.2.2.2., the sister-
group of Euphyllodromia, contains both normally orientated (Parcoblatta, Nyctibora, part
of Ectobius) and side-reversed (Blaberus, Byrsotria, Blaptica, Nauphoeta, part of Ectobius)
species. (Ectobius can be assigned to this subgroup since the females show the advanced
rotation of the oothecae, (117) in 7.4.).

This distribution can be interpreted in two ways: (1) Side-reversal is a ground-plan feature
of subgroup 2.2.3.2. Nahublattella, Supella, Euphyllodromia, the Blaberidae, and the
respective species of Ectobius have retained this orientation. Parcoblatta, Nyctibora, and
the other species of Ectobius have achieved their normal orientation by a second side-
reversal (independently in the various taxa concerned). (2) In the ground-plan of subgroup
2.2.3.2. the phallomere complex is still normally orientated. The basal offshoots
Nahublattella, Supella, and Euphyllodromia, and also the Blaberidae and the respective
species of Ectobius have independently reversed the phallomere complex.

Alternative (1) is highly supported by the fact that the three basal offshoots of subgroup
2.2.3.2. are side-reversed. However, a definitive decision, whether (1) or (2) or any
combination of these possibilities is true, 1s not possible at the present state of knowledge,
and more species will have to be investigated. At least, it is strongly suggested that the
orientation of the phallomere complex, side-reversed or normal, is not a very good criterion
for phylogenetic conclusions.

7.9. Remarks on the procedure in the phylogenetic analysis and on character lists
and character state matrices

Character lists, describing the characters, their states, and the assumed polarities, and
character state matrices, describing the distribution of the character states over the taxa,
have the function to present all the character states used and their distribution
independently of any previous assumptions on phylogeny — as an objective basis for the
phylogenetic analysis or as a starting-point for a computer-based cladistic analysis. The
applicability of this method of presentation in the frame of an analysis concerned with a
very complex type of character evolution, as it has been found in the male genitalia of
Blattaria and Mantodea, is discussed here.

The procedure in the present phylogenetic analysis is mainly hierarchical: Mantodea and
Blattaria have been, in the frame of the species investigated, regarded as sister-groups
(Isoptera disregarded); this basic assumption is well-founded (Hennig 1969, Klass 1995).

37

Blattaria and Mantodea have then been reciprocally used as outgroups, and many features
of the common ground-plan of Blattaria and Mantodea could be reconstructed. Then, in
the discussions of phallomere evolution in 7.2. and 7.3., holophyletic subgroups have been
established according to their hierarchy. It was begun with the search for apomorphic
character states common to several species, permitting the delimitation of higher-ranked
subgroups (subgroups 1.2., 2.1., and 2.2.). On this level, “apomorphic” relates to a
comparison with features well-ascertained for the common ground-plan of Blattaria and
Mantodea. These higher-ranked subgroups, if their holophyly could be well ascertained,
were then split into more subordinate subgroups, again by searching apomorphic character
states common to part of the species. At this level, “apomorphic” relates, if e.g. a Blattarian
subgroup is under consideration, to a comparison either with the common ground-plan of
Blattaria and Mantodea, or with the ground-plan of Blattaria, or with the ground-plan of
any Blattarian subgroup superordinate to and including the subgroup under consideration.
At last, in 7.6., the distribution of the states of the characters inconsistent with the majority
has been discussed in terms of parsimony.

This hierarchical analysis has to be continuously accomplished with a procedure of
reciprocal illumination: There has to be a mutual feedback between the characters used,
also concerning their evidence in terms of phylogeny. This includes a continuous feedback
to the delimitation of superordinate subgroups when working on subordinate subgroups,
since an autapomorphy of a superordinate subgroup might be absent within a subgroup
suggested to be subordinate to it, and whether a reversal has ocurred or whether the range
of the superordinate subgroup has to be modified by removing the subordinate subgroup
from it has to be discussed in terms of parsimony.

The feedback between characters and also the resulting preliminary assumptions on
phylogenetic relationships can be necessary at various levels of the phylogenetic analysis:
for the interpretation of morphology in terms of homology relations, for the assignment
of a certain morphological condition present in certain species to a certain character state,
as well as for recognising the polarity of character states within a certain subgroup (and,
consequently, for the definition and formulation of characters and character states, too).
Hence, in the present analysis, the assumptions and conclusions related to these issues and
concerning certain subgroups are in many characters dependent on the distribution of
apomorphic states of other characters regarded as autapomorphies of a subgroup
superordinate to that under consideration. A character list and a matrix independent of
previous reciprocal illumination and preliminary assumptions on phylogeny do not include
this kind of feedback between characters (and their evidence) and are consequently
incomplete or even highly misleading in some characters. The following examples shall
illustrate this topic.

(1) Concern: Interpretation of morphology in terms of homology relations.

As discussed in 6.3.4., the fused sclerites L3 and L4K and the muscle 14 of Ergaula
capucina resemble L3 and 114 of Blattellidae and Blaberidae, Anaplecta excluded. These
14 and 114 have been regarded as non-homologous, and the similar position of the anterior
insertion of the muscle moving hla — 14 or 114 — is not a synapomorphy of these taxa.
This hypothesis is only in part based on a homology analysis — using the criteria of relative

318

position and special structure — since it is not possible to identify the muscle of Ergaula
reliably as the 14 by a morphological comparison alone. This identification also depends
on a preliminary assumption of phylogenetic relationships between Polyphaga and Ergaula
on the one hand and Anaplecta and the remainder of Blattellidae and Blaberidae on the
other, and this assumption results from the distribution of the apomorphic states of other
characters (autapomorphies of the subgroups 2.2.2.2.2. and 2.2.3. in 7.4.). Hence, many
characters referring to the properties of 14 and 114 (e.g. (58) and (79) in 7.4.) would have
to be regarded as not (reliably) assessable in Ergaula without preliminary assumptions on
phylogeny.

(2) Concern: Assignment of a certain morphology to a certain character state.

The description of autapomorphy (27) of subgroup 2.2. (division of region LAl, see in
7.4.) is not valid for Parcoblatta since this species has the anterior part of region L4l
completely lost (fig.268; compare sclerite L4U’ of Blaberus, fig.299), and it is not a priori
decidable if this loss was preceded by a division of L4l or not. That the condition in
Parcoblatta has to be assigned to character state (27), or is derived from it, can only be
recognised by regarding the evidence from the distribution of apomorphic states of other
characters revealing the close relationship between Parcoblatta and Anaplecta and
especially Blaberus (e.g. most autapomorphies of the superordinate subgroups 2.2.3. or
2.2.3.2. in 7.4.) — i.e. by practising reciprocal illumination between characters and after
having made preliminary assumptions on phylogenetic relationships.

(3) Combination of concerns: Assignment of a certain morphology to a certain character
state and recognition of the polarity.

In the common ground-plan of Blattaria and Mantodea the L2-sclerotisation within the
Ive-pouch (regions L2m, L2a, L2p), the paa-sclerotisation (region L2d), the pda-
sclerotisation (posterior part of region L4l), and the region L4d are all firmly connected
within one sclerite (e.g. Mantoida, fig.44-47). The apomorphic division of the left part of
L2 and the named parts of L4 (= sclerite L4N in the ground-plan of subgroup 2.2.) is
clearly different, and non-homologous, in Lamproblatta (resulting sclerites L2A+L4S and
L2C+L4T, fig.178-180) and in Nahublattella (resulting sclerites L2D and L2E+L4N,
fig.242-245; discussions in 6.2.4. and 6.3.4.). Non-homology can be recognised only by
the different position of the L4d-region: In Lamproblatta L4d is connected with the L2-
sclerotisation within Ive (fig.178, 186); in Nahublattella L4d is connected with the
sclerotisation of the insertion area of muscle 110, fig.244, 250); this is also the only
property that can serve for a description of the difference in the formulation of the
respective characters:

Character 1: Division in the left posterior part of main sclerite L2 and the associated
parts of L4 which separates region L4d from the L2-sclerotisation within Ive but does
not separate L4d from the sclerotisation of the posterior insertion area of muscle 110
(division = articulation A10). Character states: (0) absent; (1) present (in Nahublattella);
Polarity: 0>1.

Character 2: Division in the left posterior part of main sclerite L2 and the associated
parts of L4 which separates region L4d from the sclerotisation of the posterior insertion

319

area of muscle 110 but does not separate L4d from the L2-sclerotisation within Ive.
Character states: (0) absent; (1) present (in Lamproblatta); Polarity: O>1.

Nahublattella and the members of subgroup 2.2.3.2.2. (e.g. Blaberus, Parcoblatta) reveal
many apomorphic features in common, and together they form the certainly holophyletic
subgroup 2.2.3.2.: (79)-(89) in 7.4. The morphology near the left posterior end of L2 is
in some species of subgroup 2.2.3.2.2. (e.g. Nyctibora) rather similar to Nahublattella and
can easily be derived from it, and the division of L2 (articulation A10) is certainly
homologous (fig.328b,h). However, in all members of subgroup 2.2.3.2.2. the L4d-region,
and thus the only element yielding a criterion by which this special division of L2 can be
recognised or charcterised, is absent: (95) in 7.4. (In the L2-division as present in subgroup
2.2.3.2., L4d has primitively also retained the connection with the common sclerotisation
of paa and pda, and this would be a second criterion for recognising the difference to
Lamproblatta, whose L2- and L4-division separates L4d from the sclerotisation of paa
and pda; fig.329f,g, 6.3.4. However, this criterion can be applied neither to Nahublattella
— this species shows a certainly autapomorphic division of sclerite L2E+L4N into a basal
and a distal sclerite by the membranous ring 39, fig.244, which separates L4d from the
sclerotisation of paa and pda — nor to subgroup 2.2.3.2.2. for the loss of L4d).

Hence, there is a first problem in the L2-division of subgroup 2.2.3.2.2. concerning the
assignment of a certain morphology to a certain apomorphic character state: Without
having used reciprocal illumination between characters previously, i.e. without having the
preliminary assumption of the holophyly of the superordinate subgroup 2.2.3.2. resulting
from the consideration of other characters, the characters | and 2 would have to be
regarded as not assessable in those members of the subordinate subgroup 2.2.3.2.2. which
show a L2-division (Table la). The assessment and the matrix entries of the characters |
and 2 can only be proper if these two characters are considered in interdependence with
other characters having apomorphic states common to Nahublattella and subgroup
2.2.3.2.2., 1.e. if the probable holophyly of the superordinate subgroup 2.2.3.2. has been
recognised previously (Table 1b).

Moreover, some members of subgroup 2.2.3.2.2. reveal a secondary fusion of L2D and
L2E+LAN and a secondary loss of muscle 110. That these are reversals results clearly
from the hierarchical analysis. Concerning the sclerotisations, the highly apomorphic
character state achieved by this reversal conforms exactly with the state present in e.g.
Polyphaga (all parts of L2 and L4N form together one sclerite), and it also conforms with
the most plesiomorphic state within Blattaria and Mantodea (all parts of L2 and the
posterior part of region L4l — LAN not yet differentiated as a separate sclerite — are
contained in one sclerite). As regards 110, its absence does likewise correspond with the
most plesiomorphic state within Blattaria and Mantodea.

Hence, there is a second problem in the L2-division of subgroup 2.2.3.2.2. concerning the
recognition of the polarity and the definition of character states: In elaborating a character
state matrix without having used reciprocal illumination between characters previously,
the morphology of the sclerotisation and of 110 would have to be regarded as representing
rather or most plesiomorphic states of the respective characters (Table la). Only the dis-
tribution of the states of other characters and their evidence in terms of phylogeny reveals

320

Table la,b: Character lists and character state matrices of the characters 1, 2, and 3; a) without
regarding evidence of other characters in terms of phylogeny; b) under consideration of evidence of
other characters in terms of phylogeny.

Matrices: 0 = most plesiomorphic state; 1 = apomorphic state derived from 0; 2 = apomorphic state
derived from 1; / = character not investigated; ? = character not assessable; C = Character; Sp =
Sphodromantis,; Me = Metallyticus; Ch = Chaeteessa; Ma = Mantoida; Ar = Archiblatta, Eu =
Eurycotis, Tp = Tryonicus parvus; Po = Polyphaga; Er = Ergaula capucina; Cr = Cryptocercus; La
= Lamproblatta, An = Anaplecta; Na = Nahublattella; Su = Supella; Ep = Euphyllodromia; Pa =
Parcoblatta, Ny = Nyctibora; Np = Nauphoeta; Bb = Blaberus; Bp = Blaptica; By = Byrsotria.

To the characters | and 2 the criterion of the connection of the paa-sclerotisation (region L2d) has
been added in order to distinguish the described divisions from the division between the paa- and
pda-sclerotisations as present in some Blattaria and Mantodea (compare (A) in 7.5.).

a)

Character 1: Division in the left posterior part of main sclerite L2 and the associated parts of L4
which separates region L4d as well as the sclerotisation of paa from the L2-sclerotisation within lve
but does not separate L4d from the sclerotisation of the posterior insertion area of muscle 110 (=
articulation A10). Character states: (0) absent; (1) present; Polarity: 0>1.

Character 2: Division in the left posterior part of main sclerite L2 and the associated parts of L4
which separates region L4d from the sclerotisation of paa and from sclerotisation of posterior
insertion area of muscle 110 but does not separate L4d from the L2-sclerotisation within lve. Character
states: (0) absent; (1) present; Polarity: 0>1.

Character 3: Presence of muscle 110. Character states: (0) absent; (1) present; Polarity: O>1.

C Sp Me Ch Ma Ar Eu Tp Po Er Cr La An Na Su Ep Pa Ny Np Bbe Bp by

On nen mie sO Or 2 2 O @Q
OR Mn OO OO Orn vem Oo? 7 O QO Q
eo oe re Ei nmel, il 0 0 ©
SOO wi 2o2.352.,2
b)

Character 1: Division in the left posterior part of main sclerite L2 and the associated parts of L4
which separates region L4d as well as the sclerotisation of paa from the L2-sclerotisation within Ive
but does not separate L4d from the sclerotisation of the posterior insertion area of muscle 110 (=
articulation A10). Character states: (0) absent; (1) present; (2) secondarily absent; Polarity: 0>1>2.
Character 2: Division in the left posterior part of main sclerite L2 and the associated parts of L4
which separates region L4d from the sclerotisation of paa and from the sclerotisation of the posterior
insertion area of muscle 110 but does not separate L4d from the L2-sclerotisation within lve. Character
states: (0) absent; (1) present; Polarity: O>1.

Character 3: Presence of muscle 110. Character states: (0) absent; (1) present; (2) secondarily absent;
Polarity: 0>1>2.

C Sp Me Ch Ma Ar Eu Tp Po Er Cr Ea An Na Su’ Ep Pa NyeNp Bb spmey,

r 0 0.0 0-0-0. 00 9 0.070 177712 E27 7 es tee ee
0.070.000 80270, nee Oe 0 O @ OY Y

i i a ao Ue I il 2 0202
Subgroup, DD EDER

321

that the absence of the division described in character 1 and the absence of 110 are ple-
siomorphic for Blattaria and Mantodea as a whole but apomorphic within subgroup 2.2.3.2.
(discussion in 7.5. (Q), (R)). The character states achieved by these reversals can then be
properly defined as highly apomorphic states (Table 1b).

As a consequence of the combined presence of the first and the second problem, in the
character state matrix in Table la the items relating to the L2-divisions would suggest that
these L2-divisions have originated independently in Nahublattella and in subgroup
2.2.3.2.2. and are non-homologous. In a computer-based cladistic analysis this would cause
a misleading trend away from a holophyly of Nahublattella + subgroup 2.2.3.2.2. In the
matrix in Table 1b this misleading impression is eliminated.

As a result, the assessment of homology relations, the definition of character states, the
assignment of morphological conditions to certain character states, and the polarity
assumptions, and hence also the respective entries of items into the matrix, can in some
cases only be proper in dependence on a previous hierarchical analysis with reciprocal
illumination and on the resulting preliminary assumptions in terms of phylogeny. It is, at
least in the frame of the analysis presented here, not suitable to give a character list and
a character state matrix with the characters considered independently of each other and of
preliminary assumptions on phylogeny.

8. HOMOLOGY RELATIONS ACCORDING TO MIZUKUBO & HIRASHIMA (1987)
AND GENERAL REMARKS ON THE ANALYSIS OF HOMOLOGY RELATIONS

The assumptions and procedures of Mizukubo & Hirashima

Mizukubo & Hirashima (1987) investigate the sclerites and the muscles of the phallomeres
of Periplaneta fuliginosa (Blattidae / Blattinae), of 3 species of Blattella (Blattellidae /
Blattellinae), and of Opisthoplatia orientalis (Blaberidae). Additionally, they use data from
other writers concerning various species of Blattinae. The phallomeres of Blattinae are
regarded as the most primitive. The results of the authors comprise: (1) Homologies of
the phallomere elements of the different species. (2) Side-homologies of the elements of
the left and the right halves of the phallomere complex. (3) A ground-plan for the
sclerotisations of the phallomere complex of Blattaria, which is mainly based on the
morphology of Blattinae.

As regards (1), the supposed homology relations are fundamentally different from those

I assume for the respective close relatives Eurycotis (Blattidae), Parcoblatta (Blattellinae),

and Blaberus (Blaberidae). For example, Mizukubo & Hirashima suppose that the hooks

designated here as hla have developed from completely different elements in the three

groups. (In my view these hla are strictly homologous.) Their opinions concerning the

ground-plan of the Blattarian phallomeres are also completely contradictory of my results.

The paper of Mizukubo & Hirashima must therefore be discussed in detail.

Mizukubo & Hirashima procede as follows:

— They divide both the left and the right side of the phallomere complex into 11
“subregions”. The definition of “subregion” is: “The smallest and indivisible unit which

322

is a part of the bordered region in a plane and, in this entire region, possesses its own
relative position determined by its relations with other surrounding subregions.” (p.251).
The relationships between sclerites and subregions are characterised on p.251: (A)
Principally a single sclerite (often inclusive of the surrounding membrane) corresponds
with a single subregion. (B) A single sclerite can spread over two or more subregions,
or two or more subregions can participate in a single sclerite. (C) A subregion can be
completely membranous.

— They do not explicitly say whether (B) and (C) are exclusively regarded as derived
states, or if they can be already realised in the hypothetical ground-plan of Blattaria.
However, the definition of the subregions makes sense only if the demarcation of
“smallest indivisible units” obeys a uniform principle — and this can only be the
possession of an own sclerite according to (A). Hence, I interpret Mizukubo &
Hirashima in the following way: A set of 11 subregions on each side, each subregion
with one sclerite of its own, is regarded as the ground-plan pattern of Blattaria.
(According to (A); (B) and (C) realised in derived states only.)

— They deduce the basic pattern of 11 subregions per side from the morphology of the
various Blattinae: By considering several Blattinae and by combining their features, the
dividing of the phallomere complex into subregions can be accomplished in a way that
the relative positions of the subregions of the left side are a mirror-image of the relative
positions of the subregions of the right side.

— In both Blattella and Opisthoplatia the dividing into subregions can be accomplished
in a way that the subregions of the left side as well as those of the right side have the
same relative positions as they have in Blattinae and in the hypothetical basic pattern.
In their dividing procedure the authors assume losses or fusions for some sclerites
(according to (B) and (C)).

— From these equal relative positions of the subregions they deduce homology relations
between the subregions of the left and of the right side of the phallomere complex as
well as between the subregions of the phallomeres of different species. Side-homologous
subregions get the same names — except for L (left) or R (right) in the first position.

— Concerning the closeness of the relations between neighboring subregions, they
distinguish four categories which describe the closeness or intensity of the mutual
relations between the respective sclerites: weak adjacency — adjacency — articulation —
fusion. All the relations between all neighboring subregions together are the association
pattern of the phallomere complex.

— They investigate the muscle insertions on the various subregions.

— In their homology analysis the authors largely neglect the musculature. If the course of
a muscle is consistent with the homology assumptions deduced from the relative
positions of the subregions, this is regarded as a confirmation. If there is inconsistency,
the authors do not regard this as a matter of conflict.

The statements of Mizukubo & Hirashima include, or result in, the following assumptions

regarding the ground-plan of Blattaria:

— The left as well as the right side of the phallomere complex are provided with (exactly)
ll separate sclerites. (This results from the definition and characterisation of
“subregion”, compare above).

323

— The whole phallomere complex is bilaterally symmetrical (p.256). Consequently, all
side-homologous subregions would have to be regarded as completely symmetrical, too.

The procedure and the argumentation of Mizukubo & Hirashima have some weak points,
and many of their statements are in contradiction to my results. My critique concerns the
topics discussed subsequently.

The negligence of the musculature as a reference frame for the homology analysis

Mizukubo & Hirashima base their homology analysis on the relative positions of the
cuticular subregions to each other but largely neglect the musculature. Regarding their
results, the courses of most muscles are inconsistent with the homology assumptions
deduced from cuticular morphology. Referring to Matsuda (1976), they assume that, since
the muscles develop independently of the exoskeleton, the insertions of muscles generally
have a too large shifting potential in their evolution to be reliable landmarks in the analysis
of homology relations. Matsuda (1976) regards the musculature as a valuable
supplementary criterion only “when the structures within a relatively narrow range of
species — within a family, or perhaps an order — are under study” (p.36). The question
arises, therefore, how valuable the muscles are in homologising the phallomere elements
of Blattaria and Mantodea and to what extent they deserve to be considered.

In Eurycotis and Mantoida, which are only distantly related, the cuticular elements of the
phallomeres are quite similar in their principal arrangement. The arrangement of the
musculature is to a large extent consistent with the homology relations resulting from the
relative positions and special features of the cuticular elements: The main muscles of the
phallomere complex have the same course (12, 13, 14, 16, rl, r2, r3). In many cases the
musculature can help in confirming homology assumptions. This is not inconsistent with
the views of Mizukubo & Hirashima, but in my opinion it also suggests that one should
not a priori regard the musculature as highly variable as these authors do.

In my view, (1) to regard the insertions a priori as conservative and (2) to assume shifts
of insertions only if inconsistencies arise is the better approach. (1) In the comparison
between Mantoida and Sphodromantis, the evolution of some sclerotisations could be
reconstructed in detail, because the musculature was taken as an integral instrument of the
homology analysis (compare L4 in 6.3.3.). In the comparison of distantly related species
in which homologous cuticular elements show, apart from a somewhat similar position,
hardly any similarities, the insertions of muscles can in my opinion be extremely valuable
landmarks. For example, the insertions of 14, 12, and 114 suggest that sclerite L4K of
Cryptocercus is homologous with a part of sclerite L4H of Eurycotis (L4n-region and
anterior L4l-region). The insertion of 14 suggests (partial) homology for L4K of Ergaula
and LAK of Cryptocercus, though L4K of Ergaula has shifted to the ventral hla-base and
fused to sclerite L3 (compare in 6.3.4.). (2) On the other hand, of course, the muscle
insertions have a certain shifting potential. By studying enough species, however, these
shifts can often be “observed” step by step, and in many cases it becomes obvious whether
the insertions of the muscles or the similarities in cuticular morphology are the better basis
for homology assumptions. Drastic shifts of muscle insertions are e.g. that of 114 from
the L4n-region (Eurycotis) to the L2a-region (Nahublattella) and that of 12 from sclerite

324

L1 (Mantoida) to the base ot the hla-hook (Nahublattella). In these two cases, the
morphology of Anaplecta reveals how these shifts have taken place (and that with high
probability the insertions have shifted).

In my view the arrangement of the musculature is a very important element of the
homology analysis. A simultaneous consideration and a mutual weighing of similarities in
the cuticular elements and in the musculature — combined with the investigation of a larger
sample of species — has proved most useful in this work. Moreover, in this kind of
proceding, the consideration of the musculature has the advantage that the information
about the shifts of insertions, the losses, divisions, fusions, or de-novo-formations of
muscles can provide many autapomorphies — in addition to those gathered from cuticular
morphology. In my view, a homology hypothesis on the Blattarian phallomere elements
which accepts extensive inconsistencies in the arrangement of the musculature is not very
convincing.

The division into 11 subregions per side

Mizukubo & Hirashima deduce the presence of a natural division into 11 subregions on
each side of the phallomere complex from the morphology of various Blattinae. However,
in my view their special kind of procedure is debatable. (The subsequently used terms of
Mizukubo & Hirashima can be distinguished from mine by D or V in the second position.)
Mizukubo & Hirashima assume that the sclerotisations comprised in R1 in my terminology
represent 7 subregions (compare Eurycotis, fig.74-78, 330g, 33le, 332e, and Archiblatta,
fig.330f):

1. RDid essentially region Rid (sclerite RIH)

2. RD11 sclerotisation of process pra, part of region Rld
3. RDim sclerotisation of spine sra, part of region Rld
4. RD1v essentially region R1v (sclerite R1G)

5. RDivm a ribbon-like sclerotisation connecting RIH and R1G; missing in Eurycotis
but present in Archiblatta in the ventral wall of lobe fda (compare fig.330f

and g)
6. RD21 essentially region Ric
7. RD2d essentially region R1t (with ridge pva)
The remaining subregions of the right phallomere are:
8. RD2v sclerite R2
9. RD3 sclerite R3
10. RVv sclerite L4G (region L4v on lobe vla)
11. RVd right part of sclerite L5 of Periplaneta (within ejaculatory duct, compare

in 6.5.); RVd is supposed to have fused with its left counterpart LVd =
left part of sclerite LS.
In my view, R1 is in the common ground-plan of Blattaria and Mantodea either one
undivided sclerite (more probable) or composed of three sclerites (RIF, R1G, R1H;
separated by the articulations A8 and A9; 6.7.1.). For the ground-plan of Blattaria I
assumed the latter condition, which is still present in Eurycotis. For the subregions RD1I,
RDim, RDivm, RD2d, and RD2I there is no indication that they have been separate

325

sclerites in the ground-plan of Blattaria. (The separation of RD2d and RD21 = Rit and
Ric is realised as an apomorphic feature in some Mantodea and Blattellidae only, compare
in 6.7.3., 6.7.4., 7.5.(G). RD1l and RD1m are separate sclerites in some Blattinae, but the
outgroup comparison with Mantodea suggests that this is not a ground-plan feature of
Blattaria.) As regards the ventral sclerotisation of the vla-lobe (RVv = L4G), it is not
impossible that this is an element of the right half of the phallomere complex (according
to Quadri 1940), but in my view this is not very probable (discussion in 3.1.).

The left complex is divided into the following subregions (compare Eurycotis, fig.65-69,
323e, 324e, 325e, and Archiblatta, fig.53-57, 323f, 324f, 325f):

1. LD1d left part of sclerite L1 (part of region Lla)

2. LD1l right part of sclerite L1 (parts of regions Lla and L1m)

3. LD1m rightmost part of sclerite L1, near articulation A2 (part of region Lim)

4. LDiv sclerite L4F (posterior part of region L4c) + sclerotisation of paa (region
L2d)

5. LDivm essentially sclerite L2 (except region L2d)

6. LD21 posterior part of sclerite L4C of Archiblatta (posterior part of region L4l)

7. LD2d sclerite L3 (on hook hla)

8. LD2v sclerite L4D of Archiblatta (region L4n)

9. LD3 anterior part of sclerite L4C of Archiblatta (anterior part of region L4l)

10. LVv sclerite L4E of Archiblatta (anterior part of region L4c)

11. LVd left part of sclerite L5 of Periplaneta

In 6.3.1. it has been shown that the sclerotisation of the L4l-region (LD21 and LD3) is
undivided in the common ground-plan of Blattaria and Mantodea as well as in the ground-
plan of Blattaria, and that this situation is retained in Archiblatta (sclerite L4C) and
Eurycotis (sclerite L4H). And there is no indication that L4l was present as two separate
sclerites in still earlier times.

In some cases the division into subregions is based on apomorphic features of Blattinae
(and Polyzosteriinae): L4F is a sclerite peculiar to these groups and certainly not a ground-
plan element of Blattaria. The L4n-region is only in Blattinae an isolated sclerite (L4D),
not in Eurycotis, Tryonicus, or Anaplecta; the ground-plan situation of Blattaria, however,
is unclear in this respect. The branching of the posterior part of L1 into several lobe-like
extensions (LD1d, LD1m, LD1I) is a consequence of the posteriad expansion of L1 onto
the dca-processes, and in this distinct form it is certainly an apomorphic state; that L1 is
a fusion product of several previously isolated sclerites cannot be deduced fom this
situation.

As a result, there are two principal reasons to refute (1) the division into 11 subregions

in Blattinae as well as (2) the ascription of this division to the ground-plan of Blattaria:

— Neither the left nor the right side of the phallomere complex of Periplaneta (or other
Blattinae or Polyzosteriinae) shows a priori a composition of exactly 11 subregions, nor
does the Blattarian ground-plan show such a pattern. Most of the dividing into
subregions is based either on apomorphic situations in a subgroup of Blattaria (Blattinae)
or on arbitrary — and in my view wrong — assumptions on which sclerotisations were
isolated from each other in the Blattarian ground-plan.

326

— Though the demarcation or identification of a ground-plan subregion has a clear
theoretical background (indivisible unit = 1 sclerite), no uniform principle can be
recognised in the practical application to extant species (analysıs of Blattinae), let alone
the attempt to come close to the definition or to explain discrepancies. It is not
comprehensible why Mizukubo & Hirashima assume for some sclerites of Blattinae a
contribution of several subregions and why for other sclerites they do not. The dividing
procedure seems to aim to have subregions with corresponding relative positions on the
left and on the right side. So the division of the L4l-region (into LD21 and LD3) results
in having — like on the right side (RD21 and RD3) — one subregion for the sclerotisation
in the anterior ventral wall (LD3 and RD3) and one for the sclerotisation in the lateral
edge of the phallomere (LD21 and RD2I). (According to my results, only the division
on the right side is a ground-plan feature: articulation A3.)

The argumentation concerning homology assumptions

Mizukubo & Hirashima mainly make use of the first criterion of homology (relative
positions). However, the specific procedure of the dividing into subregions described above
makes the homology assumptions questionable: From the fact that the left as well as the
right side can be (in a largely arbitrary way) divided into 11 areas having the same relative
positions cannot be decuced that these areas are side-homologous because of their equal
relative positions (circular argumentation). Assumptions of homology would only be
justified, if (1) these areas have specific features in common (i.e. if there are similar
structures in the same relative positions, e.g. similar sclerites, muscle insertions,
processes, apodemes, etc.), or if (2) an equal arrangement on both sides results from a
uniform principle of dividing. The same critique applies to the homology assumptions that
concern the comparison of different species: Again, the surface of the phallomeres is
divided largely arbitrarily into subregions with equal relative positions, and the subregions
are then supposed to be homologous because their relative positions are equal.

Moreover, the reliability of the homology hypothesis becomes further diminished by the

fact that neither the side-homologies nor the homologies between different species are

supported by similarities in the arrangement of the musculature or in the intensities of the
mutual relations between the subregions / sclerites:

— Side-homologies: Related to the side-homologies assumed by Mizukubo & Hirashima,
the musculature of Periplaneta is completely different in the left and in the right half
of the phallomere complex (of 20 phallomere muscles only two are a pair, Mizukubo &
Hirashima, fig.6). As regards the principal relative positions of the subregions, the
association patterns of the left and of the right side of Periplaneta are very similar. This
simply results from the fact that the two halves of the phallomere complex have been
arbitrarily divided into subregions which are in the same relative positions. However,
the subregions of the left and of the right side supposed to be homologous hardly have
any intensity of the mutual relations in common (Mizukubo & Hirashima, fig.2).

— Homologies between species: In Periplaneta and Blattella germanica, of 14 or 7,
respectively, intrinsic muscles of the left half of the phallomere complex only 2 have

827

the same course (Mizukubo & Hirashima, fig.6, 8). The intensities of the mutual
relations between the subregions are almost never the same in the two species.

The special morphology of the supposedly homologous subregions (e.g. position in a
pouch, formation of a process) is not considered at all.

The symmetry of the phallomere complex in the Blattarian ground-plan

The investigations and conclusions of Mizukubo & Hirashima are restricted to Blattaria;
Mantodea are not mentioned at all. With their statement “We cannot detect proto-types of
the genitalia .... indirectly on evidence obtained from other insect groups.” (p.250) the
authors deprive the phallomeres of Mantodea of any value to contribute to the
reconstruction of the ground-plan of the Blattarian phallomeres. However, in the
reconstruction of the ground-plan of any group an outgroup comparison can be very useful.
In the case of the Blattarian phallomeres the consideration of the Mantodean phallomeres
was of great value for the determination of the polarities of characters within Blattaria
(and within Mantodea). Since Mizukubo & Hirashima neglect Mantodea, the statement
“We believe that, at the period of the formation of the order, the early Blattaria had
symmetrical genitalia” has no foundation at all. According to this statement, the asymmetry
of the Blattarian and the Mantodean phallomeres is a case of parallel evolution. However,
my results clearly suggest that the very special kind of asymmetry present in Blattaria and
Mantodea is homologous and a feature of their common ground-plan.

Mizukubo & Hirashima recognise the side-reversed similarities in the phallomeres of

Blattellidae (Blattella) and Blaberidae (Opisthoplatia). However, they do not assume

homology for these similarities but parallel evolution due to similar selective pressure.

Hence, they assume completely symmetrical phallomeres even for the last common

ancestor of Blattellidae and Blaberidae. These opinions are refuted:

— Since extreme asymmetry had already been established in the common ground-plan of
Blattaria and Mantodea, it must have been present in the common ancestors of
Blattellidae and Blaberidae, too.

— The similarities of the left complexes of Blaberidae and the more derived Blattellidae
(Nyctibora, Parcoblatta) are so detailed and peculiar that the probability for parallel
evolution is in my view infinitely small; side-reversal is substantially ascertained by my
results (compare in 6.13.).

— That a reversal of the left-right asymmetry must be considered as a possible evolutionary
pathway is clearly demonstrated by those species of Ectobius (Ectobiinae) having side-
reversed phallomeres (compare in 6.13.).

I: HOMOLOGY RELATIONS ACCORDING TO GRANDCOLAS (1994) AND THE
PHYLOGENETIC POSITION OF CRYPTOCERCUS

Apart from other morphological studies, Grandcolas & Deleporte (1992) and Grandcolas
(1994) investigate the phallomere sclerites of some Blattaria. The latter paper contains
nearly all the information given in the former, and also some additional data, and will be
referred to in the following discussions.

328

Grandcolas (1994) investigated the phallomeres of Periplaneta americana (Blattinae),
Cryptocercus punctulatus, and several Polyphaginae and proposes a homology hypothesis
for the phallomere sclerites. He finds many synapomorphies suggesting Cryptocercus to
be a subordinate taxon of Polyphaginae. However, his homology hypothesis is very
different from my homology assumptions for Archiblatta (Blattinae), Eurycotis
(Polyzosteriinae), Polyphaga, Ergaula (Polyphaginae), and Cryptocercus, and these
discrepancies and the resulting assignment of Cryptocercus have to be discussed.

9.1. Discussion of the homology relations assumed by Grandcolas

The data base of Grandcolas

Grandcolas gives data on phallomere morphology in figures showing the cuticular
phallomere elements of Periplaneta americana (fig.1), Heterogamodes ursina (fig.3),
Therea petiveriana (fig.5), and Cryptocercus punctulatus (fig.6), and in sketches showing
the principal sclerite pattern in Blattinae (fig.2) and Polyphaginae (fig.4). He terms the
sclerites in the same manner as McKittrick (1964), but due to different homology
assumptions the names of the sclerites are in many cases different, too. Some differences
result from Grandcolas’ assumptions on side-homologies, which are expressed by giving
side-homologous sclerites the same names (except for L or R in the first position to name
the side). The sclerite terminology of Grandcolas is rather different from mine, and table
2 gives the synonymy and the homology assumptions. To distinguish them from mine the
terms of Grandcolas will be provided throughout with *.

Table 2: Synonymy of the sclerite terms of Grandcolas (1994) and those used in this paper.

Terms of Grandcolas are provided with *. Somewhat questionable synonymies are provided with ?.
L1 of Periplaneta (2nd column): In his fig.2 Grandcolas 1994 assigns L1 to L2*, but it is not clear
whether he assigns it to L2d* or to L2v*.

Periplaneta
Archiblatta Cryptocercus Therea Heterogamodes
ALE JS) Bl 11 eit
20s LAC and L1? 13 L4N? LAN?
ILE e2randele ty? UL 12 162
L3d* 13 part of LAN LAK? 13
Eav: L4D part of LAN 132 L4K or L4M?
vp* L4G L4G erie —-
N* — R2 L8 L8
RIZE R2 and RIF RIF R2 R2
R3d* RIG and RIH R1J RIM RIM
R3v* R3 R3 R3 R3

In Grandcolas’ fig.3 and 5 showing Heterogamodes and Therea at least some of the
apomorphies listed in 7.4. can be recognised, and these permit the integration of these
species into my phylogenetic hypothesis. Therea, fig.5, shows five of these apomorphies:

329

(48) L8 = N* is present. (55) L2 = L2v* and the Ive-pouch almost reach the left edge of
the left complex. (62) R2 = R2* and R3 = R3v* are fused. (63) RIM = R3d* is present.
R2 = R2* is so broad that R3 = R3v* is for most of its breadth confluent with it (compare
in 7.3., subgroup 2.2.2.2.2.2., and fig.330m). Hence, Therea can be assigned to subgroup
2.2.2.2. (Polyphaga + Ergaula + Lamproblatta) by (48) and (55), to subgroup 2.2.2.2.2.
(Polyphaga + Ergaula) by (62) and (63), and to subgroup 2.2.2.2.2.2. (Ergaula) by the
breadth of R2, and Grandcolas is probably right in assuming a close relation between
Ergaula and Therea. Heterogamodes, fig.3, shows at least (48) L8 = N* and (63) RIM
= R3d* and can be assigned to subgroup 2.2.2.2.2. (Polyphaga + Ergaula). These
assignments permit treating Therea and Heterogamodes — independently of the
assumptions of Grandcolas — as true representatives of subgroup 2.2.2.2.2.
(“Polyphaginae”) in the following discussions and to assume that at least all
autapomorphies of the subgroups 2.2., 2.2.2., and 2.2.2.2. are also present in these species
(if there have not occurred secondary changes). Also, the morphology of the phallomere
sclerites can be expected to be at least similar to Polyphaga and Ergaula.

The way Grandcolas uses his terminology in Polyphaginae (excluding Cryptocercus) and
his homology assumptions between Polyphaginae and Cryptocercus or Periplaneta can
only be inferred from the figures showing Therea and Heterogamodes. The terminology
applied to these two species can largely be transferred to the Polyphaginae I have studied,
Polyphaga and Ergaula, since for most phallomere elements the homology relations
between Therea / Heterogamodes and Polyphaga / Ergaula are quite evident; in some
other cases, however, problems arise. In combination with Grandcolas’ figures on
Periplaneta and Cryptocercus, this transfer allows the comparison and discussion of the
homology relations which are assumed for Polyphaginae (in general), Blattinae, and
Cryptocercus by Grandcolas and by me. Polyphaga and Ergaula (alone), Periplaneta, and
Cryptocercus can be compared independently of this transfer, since the homology relations
between these species have been discussed in chapter 6. In the following discussions (A)-
(F) the phallomere morphology of Therea, Heterogamodes, Polyphaga, and Ergaula
(designated as Polyphaginae) will be compared with that of Cryptocercus and Blattinae
(and, in part, Polyzosteriinae, which are closely related to Blattinae).

Sclerite L1* sensu Grandcolas and the genital opening (A)

Grandcolas names the sclerite next to the genital opening L1*. In Cryptocercus and Therea
and probably also in Heterogamodes L1* is sclerite L1 (compare fig.3, 5, 6 of Grandcolas
and fig.151). As regards the homology of L1 = L1* of these three species, I agree with
Grandcolas. However, if L1* of Heterogamodes really is the homologue of the L1 = L1*
of Cryptocercus, Polyphaga, and Ergaula, the opening concerned would not be the genital
but the phallomere-gland opening (compare black arrow in fig.3 of Grandcolas and P in
fig.106, 121, 153). In Cryptocercus, Polyphaga, and Ergaula the genital opening is much
more ventrally: the ejaculatory duct (D in fig.122, 151) opens into the Ive-pouch, next to
sclerite L2 = L2v*. This relation resulted clearly from own investigations of the internal
anatomy.

In Periplaneta L1* is sclerite L5, which is situated inside the true ejaculatory duct
(compare in 6.5.). Since the assumption of Grandcolas that this L5 is homologous with

330

L1 of Polyphaginae and Cryptocercus is only based on the similar position next to the
genital opening and since the genital opening has been misidentified in Polyphaginae and
Cryptocercus, the homology of these sclerites is no longer supported. McKittrick (1964)
assumes homology for the L1 of Polyphaginae, Cryptocercus, and Blattinae (as I do), and
this assumption is confirmed by the similar morphology of the sclerites, by a similar
position relative to other sclerites, by similar muscle insertions, and by a position next to
the phallomere-gland opening (discussion in 6.1.).

Sclerites L2d*, L3d*, and L3v* sensu Grandcolas (B)

As regards L2d*, L3d*, and L3v* of Therea and Heterogamodes, neither the muscles nor
the exact morphology and relative position of the sclerites are shown in fig.3 and 5 of
Grandcolas, and an exact homologisation with the sclerites of Polyphaga, Ergaula,
Cryptocercus, and Periplaneta is, therefore, not possible.

L3v* of Periplaneta is sclerite L4D (L4n-region, fig.325f). L3v* of Heterogamodes might
correspond to either L4M or L4K of Polyphaga and Ergaula (fig.325k). However, neither
L4M nor L4K nor any other sclerite of Polyphaga and Ergaula is strictly homologous
with L4D (discussion in 6.3.4.). Hence, the sclerites L3v* of Heterogamodes and
Periplaneta are certainly not homologous. L3v* of Therea is possibly homologous with
L3 (on the hla-hook) of Ergaula, Polyphaga, and Periplaneta.

L2d* of Periplaneta is sclerite L4C (L4- and L4d-regions, fig.325f). L2d* of Therea
and Heterogamodes probably correspond to LAN (on the pda-process) of Polyphaga and
Ergaula (fig.325k). L4N, however, is not strictly homologous with L4C but only with the
posterior part of L4C (discussion in 6.3.4.).

L3d* of Periplaneta and Heterogamodes — I agree with this homology assumption —
correspond to sclerite L3 of Blattinae, Polyphaga, and Ergaula (on the hla-hook in fig.53,
117). L3d* of Therea occupies a shallow bulge (not a long hook as hla is), and hla is
hence supposed to be reduced; however, the long and somewhat hook-like process, whose
sclerotisation is designated L3v*, is in my view more likely to be hla. I suppose that
L3d* of Therea is sclerite L4K, which is on a shallow bulge like in Ergaula (fig.326d).

In Cryptocercus (fig.6 of Grandcolas, fig.150, 151) the sclerites are designated as follows:
L3* (Grandcolas probably assumes a fusion of L3v* and L3d*) is L4N. Hence, L4N and
pda of Cryptocercus are regarded as the homologues of L3 and hla of Polyphaga and
Ergaula (fig.117) and Blattinae (fig.53). The hla-hook of the other species 1s thus supposed
to be quite reduced in Cryptocercus (as Grandcolas also supposes for Therea, which
assumption, however, is probably not true).

L2d* is L3. Hence, L3 and hla of Cryptocercus are regarded as the homologues of L4N
and pda of Polyphaga and Ergaula (fig.117) and of L4C and pda of Blattinae (fig.53).

In my view, hla and L3 of Cryptocercus are homologous with hla and L3 of the other
species (discussion in 6.4.3.), and pda and LAN of Cryptocercus are homologous with
pda and LAN (or the posterior part of L4C, respectively) of the other species (discussion
in 6.3.4.). These relations are clearly demonstrated by the muscles inserting on these
elements (compare e.g. 114 in Eurycotis and Cryptocercus, fig.72, 157) and by the relative
positions of the respective elements (compare the dorsoventral arrangement of the posterior

331

part of L4C, pda, L3, and hla in Archiblatta and of L4N, pda, L3, and hla in
Cryptocercus, fig.65, 150). As a result, Grandcolas has certainly misidentified L2d* and
L3* in Cryptocercus (as compared with Blattinae, Polyphaga, Ergaula, and probably
Heterogamodes).

As regards Cryptocercus and Therea, the sclerites L3* (or L3v* and L3d*) as well as the
L2d* are probably also not homologous (compare the interpretation of the Therea-sclerites
above), since in Therea L3v* is ventral to L2d*, whereas in Cryptocercus L3* is dorsal
to L2d*. No argument is given to explain this difference.

Sclerite R2* sensu Grandcolas (C)

As regards Periplaneta, Grandcolas follows McKittrick (1964) in the definition and
demarcation of R2*. This R2* includes sclerite R2 but additionally the regions R1t and
Ric (= sclerite RIF). The two sclerites of this R2*, R2 and RIF, are dorsoventrally
articulated with each other in A6 (fig.75, 76). The homology relations which Grandcolas
assumes concerning the R2*-sclerotisations are in some respects not completely clear. This
is in part due to incomplete information about which sclerotisations are assigned to R2*
and to the incompleteness of the figure showing the phallomeres of Periplaneta.

In fig.1 showing Periplaneta, Grandcolas labels that sclerite R2* which I designate R2
in Blattinae / Polyzosteriinae (fig.75-77). However, RIF is not contained in this figure.
Since in McKittrick RIF is assigned to R2* (McKittrick 1964, fig.108: the central and
the right parts of the slerite termed R2), RIF is probably part of R2* in the terminology
of Grandcolas.

For Polyphaginae and Cryptocercus Grandcolas evidently assumes a fusion of the two
sclerites of R2* (R2, RIF) and a concomitant loss of articulation A6. This is indicated
by one of the supposed synapomorphies of Cryptocercus and Polyphaginae: “Sclerite R2*
with two tubercles, which are not articulated dorso-ventrally” (p.151). (The only
articulation within McKittrick’s R2 is A6 between my sclerites R2 and RIF, fig.75, 76,
which is a ground-plan feature of Blattaria). In fig.3 and 5 showing Heterogamodes and
Therea, Grandcolas labels that sclerite R2* which I designate R2 in Polyphaga and
Ergaula (fig.135-137); the entire sclerite RIM is designated R3d*, without any
contribution of R2*. In fig.6 showing Cryptocercus, Grandcolas labels that sclerite R2*
which I designate RIF (fig.163); sclerite R2 is designated N* (see below in (D)). Hence,

RIF = R2* of Cryptocercus and R2 = R2* of Polyphaginae are regarded as the results

of this fusion and as strictly homologous. I cannot agree with these homology assumptions:

— That RIF of Cryptocercus has developed by a fusion of RIF and R2 of Periplaneta is
certainly wrong: In Cryptocercus R1F and R2 take the same relative positions as RIF
and R2 do in Periplaneta (and in Eurycotis, compare fig.75 and 161 and in 6.7.4.,
6.7.6.), and these sclerites are certainly strictly homologous. Articulation A6 is in
Cryptocercus as well-developed as in Eurycotis.

— That R2 of Polyphaginae has developed by a fusion of RIF and R2 of Periplaneta is,
in my view, also wrong: In Polyphaga and Ergaula the sclerotisation homologous with
RIF of Periplaneta (regions Ric and RIt) is contained in the anterior part of sclerite
RIM (compare fig.332e and i), and the R2-sclerites of these species are strictly

332

homologous (fig.332e,1). Since the morphology of the respective part of the right
phallomere of Therea and Heterogamodes is similar to Polyphaga and Ergaula, the
same relations are assumed for these species.

— The resulting assumption that RIF of Cryptocercus is homologous with R2 of
Polyphaginae is also refuted.

Sclerite N* sensu Grandcolas (D)

Sclerite N* of both Therea and Heterogamodes certainly corresponds to L8 of Polyphaga
and Ergaula (fig.117), which is situated on the rightmost part of the left complex, close
to the right phallomere. According to Grandcolas, a sclerite N* is also present in
Cryptocercus. From its relative position shown in Grandcolas’ fig.6 results that this N* is
R2 (fig.161-163): it articulates with both RIF = R2* and the left posterior end of R3 =
R3v* (articulations A6 and A7), and its general shape and position also fit. However, in
my opinion (compare in (C)) this R2 = N* of Cryptocercus is not homologous with L8
of Polyphaga and Ergaula but with R2 of Polyphaga, Ergaula, Periplaneta, and Eurycotis
(fig.75, 135, 161; compare in 6.7.4.), and L8 is missing in Cryptocercus.

The vp*-lobe (ventral phallomere) sensu Grandcolas and its sclerotisation (E)

In Periplaneta, Therea, and Cryptocercus Grandcolas (fig.1, 5, 6) designates a sclerotised
lobe in the median ventral wall of the phallomere complex as the ventral phallomere vp*
(vla-lobe in my terminology). As regards Cryptocercus and Periplaneta I agree with him:
The lobe is the true vla with sclerite L4G in its ventral wall (fig.63, 148). In Therea,
however, since this species is closely related to Ergaula, the lobe concerned can be
regarded as homologous with the Iba-lobe of Polyphaga and Ergaula, which bears sclerite
L7 in its ventral wall (fig.115).

The figures showing the general phallomere structure of Blattinae and Polyphaginae
(Grandcolas’ fig.2, 4) furthermore reveal that Grandcolas assumes homology for L7 of
Polyphaginae (including Polyphaga, Ergaula, and Therea) and L4G of Cryptocercus and
Blattinae / Polyzosteriinae (the white posteromedian sclerites in these figures). In my
opinion this assumption is wrong: In 6.2. and 6.3. the area belonging to the ventral
phallomere or vla-lobe of Polyphaga has been identified, and the sclerotisation
homologous with L4G of the other species has proved to be contained in L4M
(fig.325e,f,h,k); the position of the ventral insertion of muscle 16a, the position of the vla-
lobe relative to the Ive-pouch, and the position of the genital opening are the main
arguments. In 6.5. it has been shown that the Iba-lobe corresponds only to the rightmost
part of the vla-lobe of the other species, and that L7 is a new sclerite of Lamproblatta +
Polyphaga + Ergaula (and probably of at least some other Polyphaginae).

Side-homologies according to Grandcolas (F)

In Cryptocercus, Polyphaginae, and Blattinae Grandcolas assumes side-homologies for the
sclerites of the left half and of the right half of the phallomere complex. The only argument
is that an identical number of sclerites with similar form and position were recognisable
on each side of the ejaculatory duct opening (p.146).

333

I cannot agree with these assumptions: The number of sclerites is not identical on both
sides (compare in Grandcolas’ fig.2, 4), and the shapes and relative positions of the
sclerites supposed to be side-homologous are far from being similar on both sides (compare
e.g. L2v*/L2d* and R2* in fig.1, 3, 5, 6 of Grandcolas). Furthermore, the position of the
genital opening, defined as the center of symmetry of these supposed side-homologies,
has been identified incorrectly in Polyphaginae and Cryptocercus (compare (A)).
Generally, a superficial correspondence in the number, arrangement, and shapes of the
sclerites on the left and on the right side could well indicate side-homologies, but the
muscles should be investigated in terms of confirmation or contradiction. The musculature,
however, does not at all support the side-homologies assumed by Grandcolas.

9.2. The phylogenetic position of Cryptocercus

Cryptocercus punctulatus, Polyphaga aegyptiaca, and Ergaula capensis are, apart from
Blattinae, the only species investigated in both Grandcolas (1994) and this paper.
According to my results Ergaula and Polyphaga are more closely related, according to
Grandcolas Ergaula and Cryptocercus are more closely related.

Grandcolas lists many autapomorphies suggesting the holophyly of various groupings of
Polyphagidae. The autapomorphies of all groupings containing Cryptocercus will be
discussed subsequently according to their hierarchy, focused on the question whether the
features listed provide arguments to include Cryptocercus in the respective grouping. The
first three groupings include Cryptocercus, Ergaula, and Polyphaga and are not
contradictory of my results. The fourth grouping includes Cryptocercus and Ergaula but
not Polyphaga and is directly in conflict with my results. If this latter grouping — with or
without Crytocercus — proves to be holophyletic, the last two groupings subordinate to it
are also in conflict with my results.

Many of these autapomorphies relate to those phallomere sclerites for which Grandcolas’
homology assumptions for Cryptocercus and Polyphaginae have been refuted in 9.1. (A)-
(E), and they are in my opinion not valid; they will be commented with “misidentifica-
tion”, and the letter of the respective discussion in 9.1. will be added for reference. The
autapomorphies are numbered like in Grandcolas (no numbers used in the first two
groupings). From the quotations references like “in male genitalia” will be omitted. The
autapomorphies concerned with tibial, head, “paraproct” (= subanal lobe), or female genital
morphology have been reinvestigated. The autapomorphies 10, 16, and 17 of Grandcolas
have been omitted since they refer to characters of the wings, which are completely absent
in Cryptocercus.

Polyphagidae (including Cryptocercus)

— ”Sclerite L2v* with the form of an arch invaginated in ventro-posterior direction.” An
arch-shaped L2 = L2v* extending along an invagination (Ive-pouch) is a feature of the
common ground-plan of Blattaria and Mantodea (6.2.1., 7.1.).

— ”Sclerite L1* with thick-lipped edges.” This probably refers to the plateau-like anterior
face of the pne-pouch and of L1, and this is probably a synapomorphy of the respective

334

species (compare in 7.3., subgroup 2.2.2.; secondary reduction has been assumed for
Lamproblatta).

— ”Sclerite L1* turned on the ejaculatory duct and its opening.”

— ”Apical apodema of sclerite L1* curved around the ejaculatory duct opening.”

These two features probably refer to the hood-like shape of sclerite L1 and to its close
relation to the opening of the phallomere-gland (= “ejaculatory duct’: compare in (A)).
Both features, however, are present in the common ground-plan of Blattaria and
Mantodea (6.1.1., 7.1.).

— *Sclerite R2* with two tubercles, which are not articulated dorso-ventrally.”
Misidentification (C).

— "Female paraprocts with a membranous area in their sub-basal and internal parts.”
According to fig.13 of Grandcolas (= fig.334a in this paper) this refers to a membranous
area at the median base of the subanal lobes (similar to Y in fig.321c). However, a
membranous area taking the same position is also present in the females of e.g.
Periplaneta (fig.334e), Deropeltis (fig.334d), and Lamproblatta (fig.334c) and is
certainly not an autapomorphy of Polyphagidae or Polyphagidae + Cryptocercus.

— "Straight, long and narrow paratergites.” According to my own investigations (Klass,
in press: te and tg in fig.15, 16), the fused paratergites of the abdominal segments 8
and 9 of the females are, as compared with Periplaneta, somewhat lengthened and
narrowed in Cryptocercus but not in Polyphaga. Apart from this, a slight change of the
proportions of sclerite elements is in my view not very convincing as an autapomorphy.

— ”L1* pourvu d’une dilatation basale” (Grandcolas & Deleporte 1992). This feature
probably relates to the transverse expansion of L1 at its posterior margin, which
continues towards both sides into the extensions Lil and Lim (fig.120, 153, 3231,]).
However, a similar expansion, with at least one extension Lim, is also present in e.g.
Mantoida (fig.49, 323d), and this is certainly a feature of the common ground-plan of
Blattaria and Mantodea (6.1.1.). The extensions or regions Lil and Lir are also not
restricted to Cryptocercus and Polyphagidae (compare e.g. Tryonicus angustus, fig.107,
323h, and Nahublattella, fig.243, 244, 323n).

a) Polyphaga b) Cryptocercus c) Lamproblatta d) Deropeltis e) Periplaneta
aegyptiaca punctulatus albipalpus sp. americana

334

Fig.334: Paraprocts of female Blattaria. — Ventral view of left subanal lobe; posterior, anterior\,
median—. Sclerotised areas (paraproct) are stippled, membranous areas are white. Fig.334a according
to Grandcolas (1994).

335

Polyphaginae (including Cryptocercus)

— "Hook sclerite L3d* directed internally and posteriorly.” Misidentification (B).

— "Tubercles of the sclerite R2* fused together.” Misidentification (C).

— *Sclerite L3v* plate-like.” Misidentification (B).

— ”Spermatheque des femelles nettement bifide” (Grandcolas & Deleporte 1992). The
polarity of this character (spermatheca bifid or unbranched) is unclear, but the bifid
condition is certainly not a synapomorphy of Cryptocercus und Polyphaginae since it
is also present in e.g. Blattinae, Polyzosteriinae, Lamproblatta, and Mastotermes
(McKittrick 1964).

Cryptocercus + Therea + Eucorydia + Ergaula + Polyphaga + Eupolyphaga +

Anisogamia

1 ”Sclerite R2* with the fore tubercle showing a sharp outer apophysis.” Misidentifica-
tion (C).

2 ~Neoformation N*, right to L1*, presenting a ventral loop.” Misidentification (D).

Cryptocercus + Therea + Eucorydia + Ergaula

9 "Inner apophysis of sclerite L2d* less sharp.” Misidentification (B).

11 ”Presence of an expanded and warty area on the inner basal part of the anterior arch.”
In the female genitalia, the left and right second valvifers are narrowly connected with
each other at their anterior margins by a median transverse bridge (anterior arch of
McKittrick 1964). Grandcolas probably refers to a posteriad expansion of the
sclerotisation of the second valvifers towards the bases of the second and third valves,
which is lateral to this transverse bridge (compare Klass, in press: fig.2, 3). “warty”
might refer to the small and thick setae in this area. The expansions as well as the
bristles are present in Cryptocercus, but also in e.g. Sphodromantis, Lamproblatta, and
Eurycotis (own investigations). If I have understood this autapomorphy correctly, it
has to be refuted.

12 ”Apical spur lacking on the outer caudal margin of the fore tibiae.” I have investigated
the spurs of the fore tibiae in Polyphaga, Ergaula capucina, Cryptocercus,
Lamproblatta, and Deropeltis (fig.335a-e). All these species have 5 apical spurs, whose
bases are either outside or inside the sclerotisation of the tibia. These apical spurs can
be homologised one by one, if the slightly curved row of spurs Z, y, x .... is taken as
a landmark. The apical spur at the distal end of this row, which is always outside the
tibial sclerotisation, has been arbitrarily termed 1. Ergaula, Polyphaga, Lamproblatta,
and Deropeltis correspond in their sets of apical spurs: Two adjacent spurs at the inner
caudal margin of the fore tibiae are outside the tibial sclerotisation (1,5), three other
spurs at the outer caudal margin are inside the tibial sclerotisation (2,3,4). Only in
Cryptocercus spur 5 is inside the tibial sclerotisation. Hence, the apical spurs of
Cryptocercus and Ergaula do not show any special situation in common differing from
the other species. The autapomorphy is refuted.

13 ”Neoformation N* adjacent to L1 horizontal.” Misidentification (D).

14 ”Spermatheca sclerite vertical.” Grandcolas probably refers to the orientation of the
spermathecal plate of the female genitalia (McKittrick 1964). This sclerite, which is

336

a) Polyphaga b) Ergaula ©) Cryptocercus d) Lamproblatta e) Deropeltis
aegyptiaca capucina punctulatus albipalpus sp.

335

Fig.335: Spurs on fore-tibiae of Blattaria. — The sclerotisation and the spurs of the left fore-tibia are
shown; basalT, distall. The area bordered by straight or undulate lines is the sclerotisation of the
tibia. This sclerotisation is cut lengthwise along the ventral = inner edge of the tibia (undulate lines)
and unfolded. Black arrows mark the dorsal = outer edge of the tibia. Black dots represent the bases
of spurs. Most spurs are labelled with numbers (apical spurs) or small letters — according to the
homology relations assumed. Some apical spurs have their base outside the tibial sclerotisation.

vestigial in Cryptocercus, has a vertical orientation also in Lamproblatta and
Sphodromantis, and the posterior main part of the sclerite of Blattinae and
Polyzosteriinae is vertical, too. (These sclerotisations lie within the posterior wall of
the bulge containing the spermathecal opening (compare Klass, in press: fig.2, 3).
Thus, the vertical orientation is certainly not an autapomorphy of this grouping.

15 ”Sclerite R2* with a hind tubercle large and rounded.” Misidentification (C).

Cryptocercus + Therea + Eucorydia

25 ”Fore tubercle of R2* very small.” Misidentification (C).

26 ”Hind tubercle of R2* fused with R3v*.” Misidentification (C). Moreover, RIF (=
R2*) of Cryptocercus is in no place fused with R3 (= R3v*). (According to fig.6 of
Grandcolas the articulation A3, fig.163, is probably regarded as the point of “fusion”).

27 ”Frontal maculae of circular outline.” These frontal maculae are more or less clearly
demarcated cuticular areas median to the antennal bases. I have investigated them in
the following species (from externally only): In Ergaula capucina they are clearly
demarcated and — like in Grandcolas, fig.16 — drop-shaped. In Deropeltis and
Polyphaga they are nearly circular. In Cryptocercus and Lamproblatta no maculae
could be found. According to this distribution of the character states the autapomorphy
is refuted.

28 ”Postclypeus little or even not rounded.” It is not clear whether this feature refers to
(1) the bulging of the postclypeus or to (2) the arch-like course of its anterior margin
(= sutura epistomalis). According to (1), this feature would be like in e.g. Periplaneta

337

or Sphodromantis, whose clypei are hardly bulged. According to (2), I could not find
a sutura epistomalis in Cryptocercus. In both cases the autapomorphy has to be refuted.

29 ”Arch of L2v* horizontal.” “Arch of L2v*” is probably the part of L2 within the Ive-
pouch. However, the orientation of L2 in Cryptocercus is not or hardly different from
that in Polyphaga or Mantoida. If, however, the lack of an upcurving of the right parts
of L2 and Ive is referred to (6.2.1., 6.2.4.), this feature, if really present in the three
species, would be derived. However, according to fig.5 of Grandcolas, in Therea L1
and the right end of L2 are still in contact (articulation A2), whereas in Cryptocercus
the loss of this contact A2 and the loss of the right part of L2 (upcurved in other
Blattaria) are probably intercorrelated. Hence, the levelness of the right part of L2
would probably not be homologous in Therea and Cryptocercus.

30 “Basis of inner apophysis of L2d* widened.” Misidentification (B).

31 ”Neoformation N* protruding.” Misidentification (D).

32 ”L3v* as a narrow plate in dorso-caudal position.” Misidentification (B).

Cryptocercus + Therea

34 ”L3d* very shortened.” Misidentification (B).
35 ”Neoformation N* as a rod.” Misidentification (D).

Conclusions

All surmised synapomorphies suggesting that Ergaula, Eucorydia, or Therea are more
closely related to Cryptocercus than to Polyphaga are not valid or at least (only 29)
questionable. On the other hand, in 7.3. many apomorphies have been listed which clearly
suggest that at least Ergaula (and Lamproblatta) is more closely related to Polyphaga than
to Cryptocercus (autapomorphies of the subgroups 2.2.2.2. and 2.2.2.2.2. ın 7.4.). That
Therea and Eucorydia are true members of Polyphaginae and that they are closely related
to Ergaula is not questioned or even confirmed in the case of Therea, which shares at
least one synapomorphy with Ergaula and several synapomorphies with Ergaula and
Polyphaga (compare in 9.1.).

The synapomorphies of Grandcolas suggesting Cryptocercus to be a member of
Polyphaginae or Polyphagidae are all not valid either. The only exception is the plateau-
like anterior face of sclerite L1. However, Cryptocercus is probably closely related to
Polyphaginae (autapomorphies of subgroup 2.2.2. in 7.4., but compare in 7.7.), and, if
Lamproblatta is also included, Cryptocercus might well be assigned to the Polyphaginae
sensu Grandcolas (representing the basalmost offshoot).

As regards the various other groups usually assigned to Polyphagidae (Holocompsinae,
Euthyrrhaphinae, Latindiinae, and Tiviinae in Grandcolas, and some others), hardly
anything is known about the morphology of their male and female genitalia, and their
phylogenetic relationships are still open to question.

ERRERISEUREZEINED

Beier, M. (1968): 12. Ordnung Mantodea (Fangheuschrecken). In: Helmcke, J.-G., D.
Starck & H. Wermuth (ed.): Handb. Zool. 4 (2) 2/12, pp. 1-47 — de Gruyter, Berlin.

— (1970): Dictyoptera. In: Tuxen, S.L. (ed.): Taxonomist’s glossary of genitalia in insects,
2nd edition, pp. 31-34 — Munksgaard, Copenhagen.

Bohn, H. (1987): Reversal of the right-left asymmetry in male genitalia of some
Ectobiinae (Blattaria: Blattellidae) and its implications on sclerite homologization and
classification. — Ent. scand. 18: 293-303.

Cholodkowsky, N. (1891): Die Embryonalentwicklung von Phyllodromia (Blatta)
germanica. — Mém. Acad. Sci., St. Pétersbourg, Sér.7, 38 (5): 1-120.

Chopard, L. (1917): Note préliminaire sur la conformation de l’extremité abdominale
des Orthopteres. — Archs. Zool. exp. gén. 56, Notes et Revue 5: 105-112.

Ford, N. (1923): A comparative study of the abdominal musculature of orthopteroid
insects. — Trans. R. Can. Inst. 14: 207-319.

Gorg, I. (1959): Untersuchungen am Keim von Hierodula (Rhombodera) crassa Giglio-
Tos, ein Beitrag zur Embryologie der Mantiden (Mantodea). — Dt. ent. Z., N.F. 6 (5):
389-450.

Graber, V. (1890): Vergleichende Studien am Keimstreifen der Insekten. — Denkschr.
Akad. Wiss. Wien, math.-nat. Kl. 57: 1-114.

Grandcolas, P. (1994): Phylogenetic systematics of the subfamily Polyphaginae, with
the assignment of Cryptocercus Scudder, 1862 to this taxon (Blattaria, Blaberoidea,
Polyphagidae). — Syst. Entomol. 19: 145-158.

Grandcolas, P., & P. Deleporte (1992): La position systématique de Cryptocercus
Scudder au sein des Blattes et ses implications évolutives. — C.R . Acad. Sci. Paris 315
2317322

Gupta, P.D. (1947): On copulation and insemination in the cockroach Periplaneta
americana (Linn.). — Proc. Natn. Inst. Sci. India 13: 65-71.

Hagan, H.R. (1917): Observation on the embryonic development of the mantid
Paratenodera sinensis. — J. Morph. 30: 223-243.

Hennig, W. (1969): Die Stammesgeschichte der Insekten. 436pp. — Senckenbergbuch
49, Frankfurt / Main.

Heymons, R. (1895): Die Segmentierung des Insektenkörpers. — Anhang zu Phys. math.

Abh. K. Akad. Wiss. Berlin 1895: 1-39.

Klass, K.-D. (1995): Die Phylogenie der Dictyoptera. 400pp. — München. Univ., Diss.;
Cuvillier, Göttingen

— (in press): The ovipositor of Dictyoptera (Insecta): Homology and ground-plan of the
main elements. — Zool. Anz.

Kristensen, N.P. (1991): Phylogeny of extant hexapods. In: CSIRO (ed.): The Insects
of Australia, 2nd edn, pp. 125-140 — Melbourne University Press, Carlton, Victoria.

— (1995): Forty years’ insect phylogenetic systematics. — Zool. Beitr., N.F. 36 (1): 83-124.

Kumar, R. (1973): The biology of some Ghanaian mantids. — Bull. Inst. fond. Afr. noire,
SER A\ a) (G)E DES 75,

339

LaGreca, M. (1954): Sulla struttura morfologica dell’apparato copulatore dei Mantodei.
— Ann. Ist. sup. Sci. Lett. S. Chiara (Napoli) 1953/54: 1-28.

Ba@reea MI & A. Rainome (1949): I dermascheletro 'e 1a” muscolatura
dell’addome di Mantis religiosa. — Annuar. Ist. Mus. Zool. Univ. Napoli I (5): 1-43.
Levereault, P. (1936): The morphology of the Carolina Mantis (Stagmomantis

carolina) I: Skeleton — Univ. Kans. Sci. Bull. 24: 205-259.

— (1938): The morphology of the Carolina Mantis (Stagmomantis carolina) I: Musculature
— Univ. Kans. Sci. Bull. 25: 577-633.

Matsuda, R. (1976): Morphology and evolution of the insect abdomen, 50Ipp. —
Pergamon Press, Oxford etc.

McKittrick, F.A. (1964): Evolutionary studies of cockroaches. - Mem. Cornell Univ.
achiGmE xp. Sin. 389: 1-197.

McKittrick, F.A., & M.J. Mackerras (1965): Phyletic relationships within the
Blattidae. — Ann. ent. Soc. Am. 58 (2): 224-230.

Mizukubo, T., & Y. Hirashima (1987): Homology of male genital sclerites in
cockroaches (Blattaria) by means of analysis of their association patterns. — J. Fac. Agr.
Kyushu Univ. 31 (3): 247-277.

Pipa, R.L. (1988): Muscles and nerves of the posterior abdomen and genitalia of male
Periplaneta americana (L.) (Dictyoptera: Blattidae). — Int. J. Insect Morphol. Embryol.
17 (6): 455-471.

Quadri, M.A.H. (1940): On the development of the genitalia and their ducts of
orthopteroid insects. — Trans. R. ent. Soc. Lond. 90 (6): 121-175.

Remane, A. (1952): Die Grundlagen des natiirlichen Systems, der vergleichenden
Anatomie und der Phylogenetik; Theoretische Morphologie und Systematik I., 400pp.
— Akademische Verlagsgesellschaft Geest & Portig K.-G., Leipzig.

Roth, L.M. (1967): The evolutionary significance of rotation of the oötheca in the
Blattaria. — Psyche 74 (2): 85-103.

— (1974): A new cockroach genus (Gurneya) previously confused with Pinaconota
(Blaberidae: Epilamprinae). — Psyche 81 (2): 288-302.

— (1989): Sliferia, a new ovoviviparous cockroach genus (Blattellidae) and the evolution
of ovoviviparity in Blattaria (Dictyoptera). — Proc. ent. Soc. Wash. 91 (3): 441-451.
Scudder, G.G.E. (1971): Comparative morphology of insect genitalia. — Annu. Rev.

Ent. 16: 379-406.

Snodgrass, R.E. (1933): Morphology of the insect abdomen II: The genital ducts and
the ovipositor. — Smithson. misc. Collect. 89 (8): 1-148.

— (1935): The abdominal mechanisms of a grasshopper. — Smithson. misc. Collect. 94 (6):
1-89.

— (1936): Morphology of the insect abdomen III: The male genitalia (including arthropods
other than insects). — Smithson. misc. Collect. 95 (14): 1-96.

— (1937): The male genitalia of orthopteroid insects. — Smithson. misc. Collect. 96 (5): 1-
107.

— (1957): A revised interpretation of the external reproductive organs of male insects. —
Smithson. misc. Collect. 135 (6): 1-60.

340

Walker, E.M. (1919): The terminal abdominal structures of orthopteroid insects: a
phylogenetic study. Part I. Introduction / The terminal abdominal structure of the female.
— Ann. ent. Soc. Am. 12: 267-316.

— (1922): The terminal abdominal structures of orthopteroid insects: a phylogenetic study.
Part II. The terminal abdominal structures of the male. — Ann. ent. Soc. Am. 15: 1-76.

Weidner, H. (1970): 14. Ordnung Isoptera (Termiten). In: Helmcke, J.-G., D. Starck &
H. Wermuth (ed.): Handb. Zool. 4 (2) 2/14: 1-147 — de Gruyter, Berlin.

— (1982): 11. Morphologie, Anatomie und Histologie. In: Helmcke, J.-G., D. Starck & H.
Wermuth (ed.): Handb. Zool. 4 (2) 1/11: 1-531 — de Gruyter, Berlin.

Wheeler, W.M. (1889): The embryology of Blatta germanica and Doryphora
decemlineata. — J. Morph. 3: 291-372.

APPENDIX
Synonymy of the terminology of the phallomere elements

LaGreca: Mantodea

LaGreca (1954) has introduced special terms for the phallomere elements of Mantodea
(left column, taken directly from LaGreca 1954, p.27). Some of these are used for
formative elements as well as for sclerites situated inside or upon them. In the following

table the synonymy with the terms used in the present paper (right column) is given.

Fallomero dorsale di sinistra (fs, not fv)

Lamina dorsale (Id)
Lamina ventrale (lv)
Processo apicale (pa)
Apofisi falloide (af)
Processo anteriore (pn)
Lobo membranoso (lo)

Fallomero dorsale di destra (fd)
Corpo del fallomero (fd)

Braccio mediale del fallomero (bm)
Apodema anteriore (an)

Processo ventrale sclerificato (pv)
Piastra ventrale (pi)

Area sensoria (as)

Fallomero ventrale (fv)

Processo articolare (ar)
Lobo mediale (Im)
Processo distale (pd)
Pene (p)

Left complex minus vla-lobe and sclerite L4A
Sclerite L4B

Pouch Ive and sclerite L2 (mainly region L2a)
Process paa and region L2d

Process afa and sclerite L1B

Pouch pne and sclerite LI or LIA

Process loa

Lobe fda and region Rid

Lobe fda and region Rid

Leftmost part of lobe fda and region Rld
Sclerite R3 including apodeme age

Tooth / ridge pva and region RIt

Tooth / ridge pia and region Rlv

(not treated in the present paper)

Lobe vla and remaining ventral wall of left complex and
sclerite L4A

The part of sclerite L4A near articulation Al
Rightmost part of lobe vla and sclerite L4A
Process pda and pertaining parts of L4 or L4A
Lobe(s) goa next to genital opening

341

McKittrick: Blattaria

Synonymy is given for the terms used in the present paper and those used in McKittrick
1964, fig. 106-126. Since my results concerning the homologies of the sclerites are different
from those of McKittrick, the synonymy is different in the various subgroups, and some
representative groups are selected. This synonymy is also valid for many taxonomic papers
in which the terminology of McKittrick has been used (e.g. Roth 1974).

Blaberidae Blattellinae Anaplecta
Supella Nyctiborinae
McKittrick present paper McKittrick present paper McKittrick present paper
ul R3, R2, RIT, R4 L2vm L2 (inside lve: L2D) Li BIT)
L2vm L2 (inside Ive: L2D) L2d (virga) L2 (on via: L2E+LAN) L2v LAK
L2d (virga) L2 (on via: L2E+L4AN) L2 LAU L2vm L2 (inside lve)
R2 L3 (on hla), L4U 1e3 L3 L2d L2 (on paa), L4N
R2 R2, RIS, RIP 13 L3
R3 R3 R2 R2, RIN, L4G
R3 R3
Nahublattella Cryptocercus Blattinae,
Lophoblatta Polyzosteriinae
McKittrick present paper McKittrick present paper McKittrick present paper
Ll R2, RIN, R3 El El Ll Ll
L2vm L2 (inside Ive: L2D) L2v L2 (inside lve) L2v E2
L2d IEA L2d L2 (on paa) L2d L4C,D,E; L4H
Ri L2E+L4N L3 L3 L3 163)
R2 L3, L4U Rl R1J Rl RIG, RIH
R3 LAV R2 R2, RIF R2 R2, RIF
R3 R3 R3 R3

Author’s address:
Dr. Klaus-Dieter Klass, Zoologisches Institut der Ludwig-Maximilians-Universitat
München, Karlstr. 23, 80333 München, Germany

£ ; re ig Boah
| oe Oe, ba
i ” Q mit =
4 u hi
\ 1 r ve 3
er u
24 > {
x a. i
1
F r ;
va:
{
} at ve
Se
t .
LS i IR,
f ie Pi: , i Be
a
g STE i
}
= ‘ i

der Serie BONNER ZOOLOGISCHE MONOGRAPHIEN sind erschienen:

an: C.M.: Untersuchungen zur Systematik und Phylogenese der holark-
Sag tischen Sesiiden (Insecta, Lepidoptera), 1971, 190 S., DM 48,—

Soy Ziswiler, V., H.R. Güttinger & H. Bregulla: Monographie der Gattung
ü j Erythrura Swainson, 1837 (Aves, Passeres, Estrildidae). 1972, 158 S., 2 Tafeln,
DM 40,—

Eisentraut, M.: Die Wirbeltierfauna von Fernando Poo und Westkamerun.
"Unter besonderer Berücksichtigung der Bedeutung der pleistozänen Klimaschwan-
‚kungen für die heutige Faunenverteilung. 1973, 428 S., 5 Tafeln, DM 106,—
rrlinger, E.: Die Wiedereinbürgerung des Uhus Dube bubo in der Bundes-
ublik Deutschland. 1973, 151 S., DM 38,—

8) rich, H.: Das Hypopygium der Dolichopodiden (Diptera): Homologie und
ıdplanmerkmale. 1974, 60 S., DM 15,—

, ©.: Zur Ökologie der Wasseramsel (Cinclus cinclus) mit besonderer Berück-
gung ihrer Ernährung. 1975, 183 S., DM 46,—

‚P., & E. Kulzer: Physiology of hibernation and related lethargic states
amals and birds. 1976, 93 S., 1 Tafel, DM 23,—
‘a J.: Secondary contact zones of birds in northern Iran. 1977, 64 S., 1 Falt-

X a Les batraciens de Madagascar. 1978, 144 S., 82 Tafeln, DM 36,—
ler, E.: Das Aktionssystem von Winter- und Sommergoldhähnchen (Regulus
us, R. ignicapillus) und deren ethologische Differenzierung. 1979, 151 S.,

no of lan and in species (Teleostei: Percoidei: Cichlidae). "1980,
IM 38,—

inger, W.: Zur Ethologie der Fortpflanzung und Jugendentwicklung des
chtskauzes (Strix uralensis) mit Vergleichen zum Waldkauz (Strix aluco). 1980,

ie G. (arse Die Wirbeltiersammlungen des Museums Alexander
. 1984, 239 S., DM 60,—

, G. & C. Andrén: The Mountain Vipers of the Middle East — the
nthina complex (Reptilia, Viperidae). 1986, 90 S., DM 23,—

oeve, H.: Bibliographie der Säugetiere und Vögel der Türkei. 1986,

GELGELENKS

’ atari “Herausgeber:

6) | FORSCHUNGSINSTITUT
JSEUM ALEXANDER KOENIG
BONN

Ly
Ye

UND EVOLUTIO

MORPHOLOGIE UND EVOLUTION DES FLÜGELGELENKS
DER COLEOPTERA UND NEUROPTERIDA

von

THOMAS HORNSCHEMEYER

BONNER ZOOLOGISCHE MONOGRAPHIEN, Nr. 43
1998

Herausgeber:
ZOOLOGISCHES FORSCHUNGSINSTITUT
UND MUSEUM ALEXANDER KOENIG
BONN

Dissertation
zur Erlangung des Doktorgrades
der Mathematisch-Naturwissenschaftlichen Fakultäten

der Georg-August-Universität zu Göttingen

Referent: Prof. Dr. Rainer Willmann
Koreferent: Prof. Dr. Ulrich Ehlers

Tag der mündlichen Prüfung: 4.11.1997

Die Deutsche Bibliothek — CIP-Einheitsaufnahme

Hörnschemeyer, T.:

Morphologie und Evolution der Flügelgelenks der Coleoptera und Neuropterida / von
Thomas Hörnschemeyer. Hrsg.: Zoologisches Forschungsinstitut und Museum

Alexander Koenig, Bonn. — Bonn: Zoologisches Forschungsinst. und Museum
Alexander Koenig, 1998

(Bonner zoologische Monographien ; Nr. 43)
Zugl.: Göttingen, Univ., Diss., 1997
ISBN 3-925382-47-X

DBN: 95.368915.8
SG: 32

INHALT

Seite

EVD ENING. aus ner il ed NR TER LE NR EN LE Baer Lone Rey 5
DDIM SETTING, el ee RES RR EN er 6
NiarenralsundaN\lethoden as u, 2 Rn WM LAeetar rane ort Pat Tene enc mage 6
Daparallaon rer ele Bar N AML RN cae Foes a BAER 6
ialistollagiiselne Selina rer sete yi ache a Re U
Methodexders\/erwandtschaftsanalyser 2.2.2... vorm. u el ola ein. Ü
lintersuchteglaxare en re a TE OD ON ec 7
Alblairzıungen so wer aa Bro le Le ee alle Wkdrunanaer ease 9
Diesskeleitelemente,derneopterennHlügelbasis .. 2.2.22. 2 ee. 10
STS NOUS. 0. Eee ee NS er a A N BEL RE ig
Coleopiere s 3" ale ans Be EN eI ORL uals EURE st a cotietimelnn se cet 11
ANCE DOS USDA Go Star oe. AIMS Nn EM ORTE TO 12
COUDSCIGAS. 5.320 3 loot ne tin er ee SRR oe ae 12
ICTR UG AVS a ges, ch ee en eee Roane ar: und ner een 14
Adkeplage ne Seal Be Rais ivy Me celiyeh ital wend austere: TEE EHE NE 14
Dis cl ac mem rum ee a ne arena 14
Garabidaea@ieindelinaes nn a ey. 2 el IA ell ie erbelu: 16
Carabidae: Zabrinae, Harpalinae, Pterostichinae. ..2.... .2u22m. 022.2. 17
MERODITASEL 0 aa a Rp N ee ee Sa IE Ca 2, 18
khydioscaphidacsn. eyes Van er I N ei, 18
Däterosponidao Ma WO ee RL. 19
Kokupkbasasee a N Akt I. NER aa . 19
FiydiopmilidaezHelophorinaer 2 een mi an. 19
HiydreophilidaeHydtophilinae? 22a... 2 zen lols Ss: eos at Sts 20

Sill plaid Ae Me terrae ciety tga ee er DE. 22
Sraploylinidaoh m tent teen BAR ee 23
lBncanidace mar ne ERNST Da AA RL. 25
Scatabaeidaen a an 2 ewe aka gem ete ae Ne pa 26
[BNVRETIGIES: se ee DEN A el 27
Bupresuldacse u u u en a ec NOTE A Tam 28
IARI AC eR ww rasen ae RR ar. 29
lampymdacs er 0 ea a edel ae. 30
Santhanidace Pet ans ekNehsti lin 32
Dermesudaeren m er. ES A a ok commer ean ahaa mops ay a 33
Tllermide ner SER le SE RR SEEN ee EST 34
DIEhyuldaC HP an a EL A N ET Rn ey 34

JL ymiexslon dach NEN u u ee ehe ee EE 36
SOccinelidac N a euere de Some wiacle ve SRR 37

INICIO Cae eerie el, ik Od U INH BERN NN deere hte 38

Tenebrionidae a... rm a pm u SR ee OU a a 39

Cerambyeidäen.: 2.0 Dt Eee 40
Chrysomelidae2Criocerinae, ChrysomelinaesGalerucinae mE 41
Chrysomelidae: Hispinaev.e 3. za 2 el ae ee 42
Cureulionidae, N. Saas SOI ae Ree I U SS N 43
Neuröpteridan art ee Re Re UN A NE 44
Plänipennia 2... 2. zn a Er VO 44
Megaloptera; ... „+... ocr aly ae are AE ies 45
Raphidiöpiera: 4... 2 Ae ar ee ee re ae 47
Diskussion... 2.2... 2er ae ee Ne 48
Das Grundmuster der Neoptera. . ... 21.22 2... N Era a Er 48
Das ‘Grundmuster/der Holometabolay 2.22 SZ
Neuroptertda . a... 22 2 ee ee Ghee ele ee eet fet 54
Das: Grundmuster‘, 2.00. 08.0 2 2.20 an RE DS N 54
Autapomorphien der Neuropterida 22 2.2. 54.4.0 Si

Die Verhältnisse innerhalb der Neuropterida 72 2. 5.25.5.) i,
Megaloptera + Raphidioptera 2.2... ..... 22.0022. eee a
Planipennia vun... #2. en een er ee Re a 58
Coleoptera: 2. en 3-4 eens ee ee ee RA 58
Das; Grundmuster. „u... adhe a au hie obs eee oe Glee N ER 58
Autapomorphien der Coleoptera... 55. s 242. = 3 ee 60

Die Teilgruppen der Coleoptera)... ... 2... 2... 2. Ree ee 61
Archostemata ...... + 2: 22S SP SER oe 61
Adephaga. . oes gees ann re cee eee 62
Myxophaga ... 20. ee cede ee a ee 2 SE 62
Polyphaga :. ... 2.2.2. un sauren rn A 62
Myxophaga’+-Polyphaga N... 2... eee en ann oe 62
Adephaga + (Myxophaga + Polyphaga) <2) 23> 323-5 > eee 63
Archostemata + (Adephaga + (Myxophaga + Polyphaga) .............. 63
Vergleich mit anderen Verwandtschaftsanalysen 7. = 25542550.) eee 64
Computeranalyse: 2c. 2... au. ea ee er 65
Die Merkmale und ihre Zustände 2... 22. 2... 2 a 2 22 ER 65
Coleoptera: ses 8... 00 Be Ate nen ar a rr zul
Neuropterida. . u... 2... en ee ae 1
Strepsiptera: gs Ge «un. ar et ee een ee SE 75
Neuropterida + Coleoptera. . 22.0). cmc u ee 76
Holometabola. os. ace ale 2... iets oa ellie a Rei eee 76
Einfluß der Körpergröße auf die Rlügelgelenkstrukturen 22.2.2. ease ee 76
Ursprung: der Axillarsklerite . ....... na nn onen 76
Zusammenfassung. en ee A V7
Abstraet: ese 2. un. ie tee sed aie wore ee Bae ee el 78
Biteratüur u. een we rr 719
Tabellen 2. ee es er ee er go 0 5 0 uo 6 0 0 84

EINLEITUNG

Die Basis des Insektenflügels hat eine Vielzahl von Funktionen zu erfüllen. In ihr liegt
einerseits das Gelenk für den Flügelschlag, gleichzeitig aber finden sich hier auch Ele-
mente, über die eine Feinjustierung der Flügel, wie z.B. die Veränderung des Anstellwin-
kels, erfolgen kann. Zusätzlich zu diesen direkt den Flug betreffenden Funktionen haben
die Neoptera die Fähigkeit, die Flügel über dem Abdomen zusammenzulegen. Auch dafür
hat die Flügelbasis spezielle Strukturen, die sowohl die notwendige Beweglichkeit schaffen,
als auch einen Arretierungsmechanismus bereitstellen, der in der Ruhe- und in der Flug-
stellung für die erforderliche Fixierung des Flügels sorgt. Aus diesen komplexen und teil-
weise widersprüchlichen Funktionen ergibt sich zwangsläufig eine komplizierte Struktur
der beteiligten Skelettelemente.

In Anbetracht der Bedeutung der Flügelbasis für die Funktionsfähigkeit der Flügel und da-
mit für die Evolution des Insektenfluges ist es verwunderlich, daß sich nur relativ wenige
Arbeiten intensiver mit diesen Strukturen beschäftigen. Die Aderung der Flügel (z.B. Roger
1875, Adolph 1879, Redtenbacher 1886, Comstock & Needham 1898, 1899, Orchymont
1920, 1921, Forbes 1922, 1943, Hamilton 1972a, 1972b, Wallace 1971, Wallace & Fox
1975, Wootton 1979, Kukalova-Peck & Lawrence 1993) und die Muskulatur und das Ske-
lett des Thorax (z.B. Snodgrass 1908, 1927, Weber 1927a, 1927b, Maki 1938, Larsen 1966,
Matsuda 1970, Pringle 1976) sind seit jeher Gegenstand umfassender Untersuchungen. Die
zwischen Flügel und Thorax vermittelnden Elemente werden dabei aber oft ausgeklammert
oder nur beiläufig erwähnt. Untersuchungen, die sich gezielt mit der Flügelbasis ausein-
andersetzen, wurden z.B. von Snodgrass (1908, 1909), Stellwaag (1914), Crampton (1918),
Tannert (1958), Onesto (1959a, 1959b, 1960, 1961, 1963, 1965), Pfau (1977, 1986, 1991)
und Schneider (1978, 1987) vorgelegt. Besonders umfassende Arbeiten zu Struktur, Funk-
tion und Evolution des Flugapparates wurden von Brodskiy (1979a, 1979b, 1986, 1987,
1988, 1992, 1994) verfaßt. Untersuchungen der Flügelbasis, die eine Analyse der Phylo-
genie der für die vorliegende Darstellung bearbeiteten Taxa zum Ziel haben (Kukalova-
Peck & Lawrence 1993, Brown 1991, Brown, Scholtz & Kukalova-Peck 1993, Brown &
Scholtz 1994, 1995, 1996), stammen erst aus jüngster Zeit und beschränken sich auf die
Coleoptera bzw. auf Teilgruppen der Coleoptera.

Ziel der vorliegenden Untersuchung ist es, über eine Rekonstruktion der Grundmuster des
Flügelgelenks der Neoptera und der Holometabola Aufschlüsse hinsichtlich der Evolution
der Gelenkstrukturen innerhalb der Holometabola zu gewinnen. Im Zentrum stehen dabei
die Coleoptera und Neuropterida. Sie gelten als sehr alte Teilgruppen der Holometabola,
und zum Teil werden bei ihnen relativ ursprüngliche Verhältnisse in der Flügelbasis er-
wartet. Nach der Ansicht mancher Autoren sind die Coleoptera und Neuropterida Schwe-
stergruppen, obwohl man zugeben muß, daß die phylogenetischen Beziehungen zwischen
den Großgruppen der Holometabola noch wenig verstanden sind. Alles in allem aber sind
die Coleoptera und Neuropterida Taxa, aus deren Untersuchung entscheidende Hinweise
auf das Grundmuster der Flügelgelenkung der Holometabola zu erwarten sind. Daher wurde
im Zuge dieser Arbeit die Morphologie der Flügelgelenksklerite und die relative Lage der
einzelnen Skelettelemente zueinander untersucht.

DANKSAGUNG

Mein herzlicher Dank gilt Prof. Dr. Rainer Willmann für seine umfassende Unterstützung
meiner Arbeit in der Arbeitsgruppe Morphologie und Systematik des II. Zoologischen
Instituts.

Für anregende Diskussionen bedanke ich mich bei allen Mitgliedern der Arbeitsgruppe,
besonders bei Dr. Gert Tröster und Dr. Jes Rust.

Den technischen Assistentinnen des Instituts danke ich für die Unterstützung des prak-
tischen Teils der Arbeit. Besonders Frau R. Grahneis stand mir jederzeit mit Rat und Tat
zur Seite.

Prof. Dr. U. Ehlers danke ich für seine Bereitschaft, sich als Zweitgutachter zur Verfügung
zu stellen. Prof. Dr. M. Schaefer half mit der Erstellung eines Gutachtens bei der Erlangung
eines DAAD-Stipendiums.

Dr. habil. R.G. Beutel (FSU Jena) und Dr. M. Schmitt (Museum Alexander Koenig, Bonn)
gilt mein Dank für die kritische Durchsicht des Manuskripts.

Bei der Beschaffung des Untersuchungsmaterials waren Dr. habil. R.G. Beutel (FSU Jena),
Prof. Dr. J.F. Lawrence (CSIRO, Canberra, Australien), sowie Dr. G.N. House und Dr.
W.E. Steiner (Smithsonian Institution, Washington, USA) besonders hilfreich, da sie mir
seltenes Material der Archostemata und Myxophaga zur Untersuchung überließen. Dafür
gebührt ihnen mein besonderer Dank. Außerdem danke ich für die Überlassung von Unter-
suchungsmaterial Dr. H. Pohl (TU Darmstadt), Dr. M. Schmitt und Dr. T. Wagner (Mu-
seum Alexander Koenig, Bonn), Dr. F.-T. Krell (Universität Tübingen), Dr. J. Hevers
(Staatliches Naturhistorisches Museum, Braunschweig), Prof. Dr. R. Willmann, Dr. Michael
Ohl (Humboldt Universität Berlin) und Dipl.-Biol. V. Mauss (Universität Bonn).

Für die freundliche Aufnahme und die vielseitige Unterstützung während meines Aufent-
haltes an der Montana State University in Bozeman, USA, bedanke ich mit herzlich bei
Prof. Dr. Michael A. Ivie und seiner Familie sowie bei Prof. Dr. Richard Hurley. Die
Sammelreise in die USA wurde mir durch ein Kurzstipendium des DAAD (Kennziffer
D/96/05967) ermöglicht.

Mein herzlicher Dank gilt Sonja Wedmann für ihre Geduld, Kritik und tatkräftige Hilfe,
wann immer dies nötig war.

Abschließend möchte ich mich bei meinen Eltern bedanken, ohne deren Unterstützung die-
se Arbeit kaum hätte entstehen können.

MATERIAL UND METHODEN
Präparation

Für die Untersuchungen wurden ausschließlich in 80%-igem Alkohol oder in Bouin’scher
Lösung (Romeis 1968) fixierte Tiere verwendet. Versuche mit getrockneten Insekten, die
unter hoher Luftfeuchtigkeit wieder aufgeweicht wurden, ergaben keine verwertbaren Er-
gebnisse, da die feinen Strukturen der Flügelbasis beim Aufspannen der Flügel in der Regel
stark beschädigt wurden.

Die Untersuchung der Flügelbasis erfolgte hauptsächlich in freier Präparation direkt am
Objekt. Dazu wurden die Tiere mit aufgespannten Flügeln für ein bis drei Tage in Bouin
fixiert. Das Aufspannen der Flügel erfolgte sofort nach dem Abtöten der Tiere in Bouin,
da dann die Muskulatur noch geschmeidig war. Wenn nur die Skelettelemente untersucht

werden sollten, konnten auch Tiere genutzt werden, deren Flügel erst einige Zeit nach dem
Abtöten und Fixieren aufgespannt worden waren. Eine gute Fixierung der zu untersuchen-
den Tiere ist unerläßlich, da bei schlecht fixiertem Gewebe die Sklerite während der Prä-
paration leicht ihre natürliche Lage verlieren.

Für die Präparation stand ein Zeiss Stereomikroskop mit Zeichenspiegel und Fotoeinrich-
tung zur Verfügung. Für die Untersuchung im Rasterelektronenmikroskop (REM) wurden
die Objekte über die Alkoholreihe entwässert, am Kritischen Punkt getrocknet und mit
Gold besputtert.

Histologische Schnitte

Für die Anfertigung histologischer Schnitte wurden auschließlich Bouin-fixierte Tiere
verwendet. Die Kutikula aller für die Herstellung von Schnittserien bestimmten Objekte
wurde in Diaphanol (Romeis 1968) erweicht.

Besonders kleine Objekte wurden in Kunstharz (Araldit) eingebettet. Die Polymerisation
erfolgte in einem evakuierbaren Heizschrank. Mit Glasmessern wurden Schnitte von Sum
Dicke angefertigt. Die fertigen Schnitte wurden mit Toluidinblau-Lösung (0,1% Tolui-
dinblau in 2,5% Natriumcarbonat) angefärbt.

Größere Objekte wurden in Paraplast eingebettet. In reinem Paraffin und in reinem Para-
plast wurden die Objekte jeweils für eine Stunde bei 60°C ins Vakuum gestellt, um sämt-
liche Luft aus den Präparaten zu entfernen. Geschnitten wurden die in Paraplast einge-
betteten Objekte mit C- und D-Stahlmessern. Die Schnittdicke betrug zwischen 5 um und
10um. Die Schnitte wurden mit der Azanfärbung nach Romeis (1968) angefärbt.

Für das Betrachten, Zeichnen und Fotografieren der Schnitte standen ein Zeiss Axioskop
mit Zeichenspiegel und ein Zeiss Axiophot zur Verfügung. Die Zeichnungen wurden mit
den Rechenanlagen der Gesellschaft für wissenschaftliche Datenverarbeitung Göttingen
(GWDG) nachbearbeitet.

Die Abbildungen mit den morphologischen Details der untersuchten Arten (Abb.8 bis 85)
sind am Ende des Buches zur besseren Vergleichbarkeit zusammengefaßt.

Methode der Verwandtschaftsanalyse

Die Analyse der Verwandtschaftsverhältnisse erfolgte nach der Methode der strikt phy-
logenetischen Systematik nach Hennig (1950, 1966, 1969, 1982). Die Leserichtung der
Mermale wurde über den Außengruppenvergleich (Watrous & Wheeler 1981, Farris '9%°
Nixon & Carpenter 1993) bestimmt. Für die Computeranalyse wurde das Programm PAUP
3.1 (Swofford 1993) benutzt.

Untersuchte Taxa
Die Teilgruppen der Coleoptera sind entsprechend dem System von Lawrence & Newton
(1995) angeordnet.
Aus den folgenden Taxa wurden, soweit verfügbar, jeweils mehrere Individuen untersucht:

Heteroptera
Coreidae Leach, 1815: Coreus marginatus (L., 1758)

Plecoptera
Perlodidae Klap., 1909: Gen. sp.

Coleoptera

Archostemata

Micromalthidae Barber, 1913: Micromalthus debilis LeConte, 1878.

Cupedidae Lap., 1836: Cupes capitatus F., 1801, Distocupes varians (Lea, 1902), Priacma serrata
(LeConte, 1861), Tenomerga concolor (Westw., 1830).

Adephaga

Dytiscidae Leach, 1815: Dytiscus marginalis L., 1758.

Carabidae Latr., 1802: Amara sp., Harpalus cordatus (Duft., 1812), Harpalus sp., Poecilus versicolor
(Sturm, 1824), Pterostichus metallicus (F., 1792), Cicindela lunulata F., 1781.

Myxophaga

Hydroscaphidae LeConte, 1874: Hydroscapha sp..

Microsporidae Crotch, 1873: Microsporus sp..

Polyphaga

Hydrophilidae Latr., 1802: Anacaena limbata (F., 1792), Helophorus sp., Hydrophilus piceus (L.,
158)

Silphidae Latr., 1807: Blitophaga opaca (L., 1758), Nicrophorus investigator (Zett., 1824),
Nicrophorus vespilloides (Hbst., 1783), Oeceoptoma thoracica (L., 1758).

Staphylinidae Latr., 1802: Ontholestes murinus (L., 1758), Quedius sp..

Lucanidae Latr., 1804: Sinodendron cylindricum (L., 1758).

Scarabaeidae Latr., 1802: Aphodius sp., Cetonia cf. aurata (L., 1761), Phyllopertha horticula (L.,
1758). .

Buprestidae Leach, 1815: Anthaxia sp., Chalcophora mariana (L., 1758).

Byrrhidae Latr., 1804: Byrrhus sp..

Elateridae Leach, 1815: Agrypnus murinus (L., 1758), Argiotes pilosellus (Schönh., 1817), Denticollis
linearis (L., 1758), Elater cf. ferrugineus L., 1758, Hemicrepidius niger (F., 1792), Hypnoidus sp..
Lampyridae Latr., 1817: Lamprohiza splendidula (L., 1767).

Cantharidae Imhoff, 1856 (1815): Cantharis nigricans (Müll., 1776), Cantharis pellucida F., 1792.
Dermestidae Latr., 1804: Dermestes lardarius L., 1758.

Lymexylonidae Fleming, 1821: Hylecoetus dermestoides (L., 1761).

Cleridae Latr., 1802: Thanasimus formicarius (L., 1758), Trichodes sp..

Melyridae Leach, 1815: Dasytes plumbeus (Müll., 1776), Malachius bipustulatus (L., 1758), Malachius
SP..

Coccinellidae Latr., 1807: Calvia quatuordecimguttata (L., 1758), Coccinella septempunctata L., 1758.
Tenebrionidae Latr., 1802: Tenebrio molitor L., 1758.

Meloidae Gyllenhal, 1810: Lytta vesicatoria (L., 1758).

Pyrochroidae Latr., 1807: Schizotus sp..

Cerambycidae Latr., 1802: Rhagium mordax (Geer, 1775), Clytus arietis (L., 1758), Agapanthia villo-
soviridescens (Geer, 1775), Dinoptera collaris (L., 1758), Strangalia melanura (L., 1758), Gaurotes
virginea (L., 1758).

Chrysomelidae Latr., 1802: Agelastica alni (L., 1758), Cassida sp., Crioceris asparagi (L., 1758),
Leptinotarsa decimlineata (Say, 1824), Chrysomela populi (L., 1758).

Curculionidae Latr., 1802: Furcipus rectirostris (L., 1758), Otiorhynchus sp., Phyllobius sp.1,
Phyllobius sp.2, Chlorophanus sp..

Neuropterida

Megaloptera

Corydalidae Leach, 1815: Chauliodes rastricornis Rambur, 1842, Corydalus cornutus (L., 1758).
Sialidae Leach, 1815: Sialis lutaria (L., 1758).

Raphidioptera

Raphidiidae Latr., 1810: Raphidia ophiopsis L., 1758, Agulla adnixa Hagen, 1861.

Planipennia

Chrysopidae Hagen, 1866: Chrysopa perla (L., 1758), Chrysotropia ciliata (Wesm., 1841).

Myrmeleonidae Latr., 1804: Cueta beieri Asp. & Asp., 1964.

Osmylidae Leach, 1815: Osmylus fluvicephalus (Scop., 1763).

Mecoptera

Panorpidae Steph., 1836: Panorpa communis L., 1758, Panorpa germanica L., 1758.
Trichoptera

Philopotamidae Steph., 1829: Wormaldia copiosa McLachl., 1868.

Strepsiptera
Elenchidae Perkins, 1905: Elenchus sp.

Abkürzungen

(a) = Achse durch die Spitze des ANP und die Basis des PNP
(b) = Achse durch die Spitzen von ANP und PNP

A = Analader

1Ax = erstes Axillare

2Ax = zweites Axillare

3Ax = drittes Axillare

4Ax = viertes Axillare

AMD = Muskelplatte, axillary muscle disc

ANP = vorderer Flügelgelenkfortsatz des Notum, anterior notal wing process
AxC = axillary cord

Ba = Basalare

BaRK = Rastknopf des Basalare

BC = Basis der Costa

BR = Basiradiale

BSc = Basis der Subcosta

cF2Ax = caudaler Fortsatz des 2Ax

C = Costa

Cu = Cubitus

DMP = distale Medianplatte, distal median plate

Epm = Epimeron

Eps = Episternum

ie = Fulcrum = Gelenkkopf des PWP

H = Humeralplatte

M = Media

MNP = mittlerer Flügelgelenkfortsatz des Notum, median notal wing process
N = Notum

PIS = Pleuralnaht, pleural suture

PMP = proximale Medianplatte, proximal median plate

PN = Postnotum

PNP = hinterer Flügelgelenkfortsatz des Notum, posterior notal wing process
Poab = Postalarbriicke, Postalararm, postalar bridge

PRA = Praealarsklerit

Prab = Praealarbriicke, Praealararm, prealar bridge

PWP = pleuraler Flügelgelenkfortsatz, pleural wing process

R = Radius

Sc = Subcosta

Sb = Subalare

Tg = Tegula

Bei der Benennung der Muskeln folge ich der Nomenklatur von Matsuda (1970).

Die von Brown & Scholtz (1994) eingeführten Bezeichnungen der Sklerite des Flügelge-
lenks finden hier keine Anwendung. Die von ihnen benutzte Nomenklatur basiert auf der

10

Interpretation der Flugelgelenksklerite durch Kukalova-Peck (1983, 1991), deren Ergebnisse
auf Untersuchungen fossil überlieferter Flügelgelenke beruhen. Da diese Interpretationen
umstritten sind und durch Beobachtungen an rezenten Taxa nicht bestätigt werden, benutze
ich neutrale Beschreibungen bzw. die Nomenklatur nach Snodgrass (1935).

DIE SKELETTELEMENTE DER NEOPTEREN FLÜGELBASIS

Das Flügelgelenk der Neoptera besteht aus folgenden Elementen, die in allen Taxa wieder-
zufinden sind (Snodgrass 1909, 1927):

Das Notum bildet in der Regel drei Fortsätze aus, über die die Verbindung zu den Sklerit-
elementen der Flügelmembran hergestellt wird (Abb.1). Der vordere und der mittlere Ge-
lenkfortsatz des Notum (ANP bzw. MNP) artikulieren mit dem ersten Axillare (1Ax). Der
hintere notale Gelenkfortsatz (PNP) steht mit dem dritten Axillare (3Ax) in gelenkiger
Verbindung. Hinter dem PNP setzt der verstärkte Flügelhinterrand als axillary cord (AxC)
am Notum an.

Der Vorderrand des Notum (Praescutum) kann lateral einen Fortsatz tragen, der dann nach
ventral umgebogen ist und mit dem Episternum (Eps) in Verbindung steht. Dieser Fortsatz
wird als Praealararm oder Praealarbrücke (Prab) bezeichnet (Abb.1, 2). Das Postnotum bil-
det in der Regel ebenfalls einen seitlichen Fortsatz aus, der oft direkt in das Epimeron
übergeht. Dieser Fortsatz wird als Postalarbrücke (Poab) bezeichnet (Abb.1, 2).

Der Vorderrand des Notum (Praescutum) kann lateral einen Fortsatz tragen, der dann nach
ventral umgebogen ist und mit dem Episternum (Eps) in Verbindung steht. Dieser Fortsatz
wird als Praealararm oder Praealarbrücke (Prab) bezeichnet (Abb.1, 2). Das Postnotum bil-
det in der Regel ebenfalls einen seitlichen Fortsatz aus, der oft direkt in das Epimeron
übergeht. Dieser Fortsatz wird als Postalarbrücke (Poab) bezeichnet (Abb.1, 2).

Distal an die vorderen beiden Fortsätze des Notum schließt das 1Ax an (Abb.1). Von
dessen vorderem distalen Rand geht die Basis der Subcosta ab, an den mittleren und

H Prab

1

2
Abb. 1,2: Skelettelemente der neopteren Flügelbasis. 1: Ansicht von dorsal. 2: Ansicht von lateral ohne
Flügel und Axillarsklerite

11

Kopf
ANP =
Hals FE
Körper
2Ax
1Ax
Caudalfortsatz
Sehne
Winkel o DE ~~
AMD
Distalarm
Achse (a) Caudalarm
Achse (b) 3Ax
3 Winkel ß

Abb.3: Schema zur Erläuterung der an der Flügelbasis gemessenen Winkel und Strecken sowie der
Benennung der Teilbereiche der Axillarsklerite

hinteren distalen Rand ist das zweite Axillare (2Ax) dicht angelagert. Vom vorderen
Bereich des 2Ax entspringt die Basis des Radius.

Dem hinteren Gelenkfortsatz des Notum (PNP) ist das 3Ax angeschlossen, von dem die
Analadern ausgehen. In einigen Taxa findet sich zwischen PNP und 3Ax ein weiteres
Sklerit, welches dann als viertes Axillare (4Ax) bezeichnet wird. Zwischen 3Ax und 2Ax
vermittelt die proximale Medianplatte (PMP), an welche die distale Medianplatte (DMP)
anschließt, von der Media und Cubitus abgehen.

Am Übergang des Flügelvorderrandes zum Notum finden sich an der Basis der Costa die
Humeralplatte (H) und zwischen dieser und dem Notum die in der Regel schwach skle-
rotisierte und mit feinen Borsten besetzte Tegula (Tg) (Abb.1).

Zur Flügelgelenkung gehörende Elemente der Körperseitenwand (Abb.2) sind das Basalare
(Ba), das unter dem ANP liegt, das Subalare (Sb), das unter dem PNP liegt, und der pleu-
rale Flügelgelenkfortsatz (PWP), der als dorsale Verlängerung der Pleuralleiste ausgebildet
ist. Der PWP liegt zwischen Ba und Sb und bildet den Gelenkkopf (Fulcrum, F) aus, auf
dem in der Regel das 2Ax ruht.

ERGEBNISSE

Coleoptera

Aus den Coleoptera wurden fünf Arten der Archostemata, sieben Arten der Adephaga, zwei
Arten der Myxophaga und 53 Arten der Polyphaga untersucht.

Das Flügelgelenk der Coleoptera weist alle zu erwartenden Elemente auf (s.o.). Die Axil-
larsklerite sind sehr kräftig und im Verhältnis zum Notum besonders groß ausgebildet. Das
1Ax ist deutlich in einen verbreiterten Kopf, einen schmalen Hals und einen großflächigen
Körper gegliedert. Der basale Bereich des Basalare ist relativ schmal und gegen den erwei-
terten, komplex gestalteten Kopf deutlich abgesetzt. Mit Ausnahme der Archostemata liegt
das Fulcrum bei den Coleoptera unter dem Kopf-Hals-Bereich des ersten Axillare.

12

Archostemata
Cupedidae

Material: Cupes capitatus, Tenomerga concolor, Distocupes varians, Priacma serrata

Notum (Abb.8, 11, 13, 16A)

Der vordere Gelenkfortsatz des Notum (ANP) ist sehr lang und spitz dreieckig ausgezogen.
Die Spitze des ANP liegt deutlich vor dem cranialen Rand des Notum. Der mittlere Ge-
lenkfortsatz (MNP) ist klein und nur durch eine leichte Einbuchtung des Notumseitenrandes
hervorgehoben. Der hintere Gelenkfortsatz (PNP) ist lang ausgezogen und distal zumindest
leicht verbreitert. Das Ende des PNP liegt auf gleicher Höhe oder etwas craniad des MNP.
Es ist durch einen Streifen kräftig ausgebildeter Membran mit dem Subalare verbunden.
Ein weiteres Band stellt die Verbindung zum caudalen Fortsatz des zweiten Axillare (2Ax)
her. Bei Cupes, Tenomerga und Distocupes ist der PNP am Ende geringer sklerotisiert als
an der Basis. Die Achsen (a) und (b) durch das Ende und die Basis des PNP (Abb.3,
Tab.1) schließen einen Winkel B von ca. 40° ein. Dabei liegt das Ende des ANP weiter
distal als das Ende des PNP. In beiden Fällen bildet Priacma eine Ausnahme: hier beträgt
der Winkel ß nur ca. 30° und die Spitze des PNP liegt weiter distal als die Spitze des
ANP.

Axillar-Region (Abb.8, 11, 13, 16A)

Das erste Axillare (1Ax) ist dem ANP angelagert; es ist eine Sklerotisierung der dorsalen
Flügelmembran. Kopf und Hals des 1Ax liegen dorsad des ANP. Die Spitze des ANP trifft
etwa in der Mitte des Kopfes auf das 1Ax. Der proximale Rand des Körpers des 1Ax liegt
unterhalb des Notumrandes. Das 1Ax hat keinen Kontakt zum MNP. Der Kopf des 1Ax
ist relativ breit. Sein Vorderrand ist schmal nach ventral umgeschlagen und weist in der
Regel eine flache Einkerbung auf, die bei geöffnetem Flügel einen entsprechenden Fortsatz
der Basis der Subcosta aufnimmt. Bei Priacma ist der Kopfvorderrand nahezu gerade. Im
hinteren Bereich ist der Kopf des 1Ax von Priacma deutlich breiter als bei den anderen
Archostemata. Ventral hat die distale Kante des Kopfes eine kleine Aussparung, die dem
Gelenkkopf des pleuralen Flügelgelenkfortsatzes (F) aufliegt. Bei allen untersuchten
Archostemata ist der Hals des 1Ax im Verhältnis zu Kopf und Körper ausgesprochen
schmal. Er ist von distad her stark eingezogen. In die so geformte Bucht des 1Ax ist das
2Ax eingepaßt. Der Körper des 1Ax ist asymmetrisch ausgebildet. An der dem Notum an-
liegenden Seite trägt er einen nach caudal gerichteten fingerförmigen Fortsatz, der ca. drei
Achtel der Gesamtlänge des 1Ax ausmacht. Die dem 2Ax zugewandte Kante ist leicht kon-
vex. Der Winkel « zwischen der Gelenkachse von 1Ax und Notum und der disto-cranialen
Kante des Körpers des 1Ax ist größer als 50° (Abb.3, Tab.1).

Das zweite Axillare (2Ax) ist annähernd dreieckig geformt. Es ist so in die distale Bucht
zwischen Kopf, Hals und Körper des 1Ax eingepaßt, daß eine Spitze zum Notum weist.
Die Verbindung zwischen erstem und zweitem Axillare ist ausgesprochen fest und erlaubt
nur sehr geringe Bewegungen der beiden Sklerite gegeneinander. Die proximale Spitze des
2Ax liegt auf dem Gelenkkopf des pleuralen Flügelgelenkfortsatzes (F). Das 2Ax ist ein
von der dorsalen zur ventralen Flügelmembran durchgängig sklerotisiertes Element. Der
dorsale Bereich des 2Ax besteht aus einer Hauptregion, die V-förmig den proximo-crani-
alen und den proximo-caudalen Rand beinhaltet. Dieser Bereich ist sehr stark sklerotisiert.
Die von der Hauptregion eingeschlossene Fläche kann schwächer sklerotisiert sein und ist
gegen die Ränder leicht abgesenkt. Der proximo-caudale Rand ist caudad verlängert und

13

durch ein Band derber Membran mit dem PNP verbunden. Ventral hat das 2Ax einen vom
hinteren Rand ausgehenden kurzen, kräftigen Fortsatz, von dem ein Band zum Hals des
pleuralen Flügelgelenkfortsatzes zieht.

Dorsal an der proximalen Spitze des 2Ax inseriert das Basiradiale (BR). Es ist bei allen
untersuchten Cupedidae lang und schmal und gut sklerotisiert. Von der Spitze des 2Ax
zieht das BR nach disto-cranial in Richtung der Basis der Subcosta. Subcosta und Radius
liegen hier dicht nebeneinander, sind aber nicht verschmolzen. Das Basiradiale ist eine
Sklerotisierung der dorsalen Flügelmembran.

Das dritte Axillare (3Ax) ist eine Sklerotisierung der ventralen und dorsalen Flügelmem-
bran. Es besteht aus einem bei geöffnetem Flügel quer zur Körperlängsachse liegenden Be-
reich, dessen distales Ende zugespitzt und leicht nach caudal umgebogen ist. Über diesen
Bereich läuft eine nach proximal höher werdende Aufwölbung, an der ein Band ansetzt,
das zu einer kleinen Muskelplatte (AMD) in der Membran zwischen 1Ax und 3Ax zieht.
Das 3Ax setzt sich in einen nach proximo-caudal weisenden Arm fort, der membranös mit
dem PNP verbunden ist. Die Außenkante dieses Armes ist einfach, die Innenkante S-förmig
geschwungen.

Die Medianplatten sind innerhalb der Cupedidae unterschiedlich ausgebildet. Bei Priacma
sind zwei deutlich differenzierte Medianplatten vorhanden. Die proximale (PMP) liegt zwi-
schen 2Ax und 3Ax, direkt neben der distalen Spitze des Körpers des 1Ax. Craniad der
PMP schließt sich die distale Medianplatte (DMP) an. Von dieser gehen zwei Adern ab
(Media (M) und Cubitus (Cu)). Beide Medianplatten sind annähernd gleich groß und ähn-
lich geformt. Bei Tenomerga sind die Medianplatten proximal verschmolzen. Die PMP ist
deutlich schmaler als die DMP. Nur die Basis des Cubitus hat direkten Kontakt zur DMP.
Die Basis der Media ist als kurzer Stumpf neben dem Kreuzungspunkt von Radius, Media
und Cubitus erkennbar. Bei Distocupes ist der proximale Bereich der DMP reduziert.
Weder Media noch Cubitus haben direkten Kontakt zur DMP. Die Media ist soweit ver-
kürzt, daß sie den Kreuzungspunkt von Radius und Cubitus nicht erreicht. Bei Cupes sind
die Medianplatten, ähnlich wie bei Tenomerga, proximal verschmolzen. Die PMP ist sehr
schmal. Die Basen von Media und Cubitus sind schwach sklerotisiert. Die Media erreicht
den Kreuzungspunkt von Radius und Cubitus nicht.

Pleural-Region (Abb.9, 10, 12, 14, 15, 16B)

Der pleurale Flügelgelenkfortsatz (PWP) wird überwiegend vom Epimeron gebildet. Der
Anteil des Episternum beschränkt sich auf einen schmalen Streifen an der Vorderkante des
PWP. Das Fulcrum ist bei Cupes, Tenomerga und Distocupes kurz und proximal abge-
schrägt. Dadurch ist die Auflagefläche sehr klein. Dem Gelenkkopf liegt die proximale
Spitze des 2Ax auf. Bei Priacma ist das Fulcrum verlängert, so daß es nicht allein vom
2Ax überdeckt werden kann. Seinem vorderen Drittel liegt der Kopf des 1Ax auf. Bei allen
untersuchten Cupedidae befindet sich etwas unterhalb des Gelenkkopfes eine flache Ein-
buchtung, an der ein Band ansetzt, das den PWP mit dem 2Ax verbindet.

Das Basalare (Ba) liegt vor dem PWP und ist durch einen schmalen Membranstreifen von
ihm getrennt. Das ventrale Ende des Basalare ist mit dem Episternum verschmolzen. Kurz
unterhalb des Fulcrum bildet das Basalare einen komplex strukturierten Kopf aus, der aus
einer nach außen gerichteten großen, blasigen Erweiterung und einem nach vorn oben ge-
richteten löffelförmigen Fortsatz besteht. Die Spitze dieses Fortsatzes liegt etwa auf
gleicher Höhe wie der Gelenkkopf des PWP. Der craniale Fortsatz ist durch einen derben
Membranstreifen fest mit der Vorderkante des Flügels verbunden (Basis der Costa / Hume-

14

ralplatte / Basis der Subcosta). Die laterale Vorwölbung dient zusammen mit entspre-
chenden Strukturen der ventralen Sc-Basis der Fixierung des Flügels in geöffnetem und in
angelegtem Zustand. Von dieser breiten Kopfkonstruktion ausgehend verschmälert sich das
Basalare schnell nach ventral.

In der Membran hinter dem PWP unterhalb des hinteren Gelenkfortsatzes des Notum (PNP)
liegt das relativ große scheibenförmige Subalare (Sb). Es ist über Bänder (verstärkte
Membranstreifen) mit dem PNP verbunden.

Micromalthidae
Material: Micromalthus debilis

Notum (Abb.17)

Die Ausbildung des Notum entspricht weitgehend den Verhältnissen bei den Cupedidae.
Der MNP ist sehr klein und nur als kurzer nach vorn gerichteter Haken direkt hinter der
caudalen Spitze des 1Ax vorhanden. Der PNP ist, ähnlich wie bei Cupes, Tenomerga und
Distocupes, zum Apex hin geringer sklerotisiert. Im Unterschied zu den anderen Ar-
chostemata ist eine distale Verbreiterung des PNP nicht erkennbar.

Axillar-Region (Abb.17)

Auch die Elemente der Axillar-Region sind bei Micromalthidae und Cupedidae nahezu
gleich gestaltet. Die proximale Spitze des 2Ax liegt auf dem Gelenkkopf des PWP auf. Bei
Micromalthus ist das Basiradiale allerdings bis auf den Ursprung am 2Ax nicht sklero-
tisiert. Die distale Fläche des 2Ax ist ebenso wie der caudale Arm des 3Ax nur schwach
sklerotisiert. Die Medianplatten sind nicht identifizierbar.

Pleural-Region

Das Fulcrum ist wie bei Cupes, Distocupes und Tenomerga sehr kurz. Es liegt unter der
proximalen Spitze des 2Ax.

Adephaga
Dytiscidae

Material: Dytiscus marginalis

Notum (Abb.18A)

Der vordere Gelenkfortsatz (ANP) ist relativ groß, flach ausgezogen und kräftig sklero-
tisiert. Das Ende des ANP ist schmal gerundet und reicht nur wenig weiter nach cranial als
der Notumvorderrand. Craniad des mittleren Gelenkfortsatzes (MNP) ist der Notumseiten-
rand tief eingekerbt, so daß der MNP deutlich abgesetzt erscheint. Er überragt nicht den
Seitenrand des Notum und ist stumpf zweispitzig ausgebildet. Der hintere Gelenkfortsatz
(PNP) ist lang ausgezogen, proximal schmal und distal deutlich verbreitert. Der Winkel
ß zwischen der Spitze und der Basis des PNP mit Bezug zur Spitze des ANP beträgt ca.
25° (Abb.3, Tab.1).

Axillar-Region (Abb.18A)

Das erste Axillare (1Ax) hat einen sehr breiten, massigen Kopf, der mit etwa einem Drittel
seines proximalen Randes dem ANP aufliegt. Der Vorderrand ist nach ventral umgeschla-

15

gen und proximal in einen langen Zahn ausgezogen, der über Membranen gelenkig mit der
Subcosta-Basis (BSc) und dem Basalare (Ba) verbunden ist. Der Vorderrand weist außer-
dem ein bis zwei senkrecht verlaufende Rippen auf, die sich bei geöffnetem Flügel mit der
BSc verhaken. Der Hals des 1Ax ist im Verhältnis zu Kopf und Körper sehr kurz und
schmal. Er liegt dem Gelenkkopf des pleuralen Flügelgelenkfortsatzes (F) auf. Durch den
großen Unterschied in der Breite von Kopf, Hals und Körper ergibt sich eine schmale
Bucht, in der das 2Ax liegt. Der Körper des 1Ax ist asymmetrisch, die proximale Hälfte
ist wesentlich länger als die distale. Etwa zwei Drittel des proximalen Randes liegen unter
dem Seitenrand des Notum. Der asymmetrische Teil des Körpers des 1Ax macht ca. zwei
Fünftel der Gesamtlänge des 1Ax aus (Tab.1). Die disto-craniale Kante des Körpers des
1Ax ist leicht konvex. Der Winkel & zwischen der Achse durch den vorderen und den hin-
teren Kontaktpunkt von 1Ax und Notum und dem disto-cranialen Rand beträgt ca. 75°. Das
Notum ist ca. 1,5 mal so lang wie das gesamte 1Ax.

Das 2Ax ist in die Bucht zwischen Kopf, Hals und Körper des 1Ax eingepaßt und durch
schmale Membranstreifen fest mit dem 1Ax verbunden. Von der schmal gerundeten proxi-
malen Spitze des 2Ax geht das lange, schmale und durchgehend sklerotisierte Basiradiale
(BR) ab. Distal ist das 2Ax mit der proximalen Medianplatte (PMP) verschmolzen. Ventral
hat das 2Ax einen lang ausgezogenen caudalen Fortsatz, an dem ein Band ansetzt, das die
Verbindung zum PNP herstellt. Ein weiterer ventraler Fortsatz, der am proximo-caudalen
Rand des 2Ax entspringt, ist zweihöckrig ausgebildet. Er liegt dicht am Halsbereich des
pleuralen Flügelgelenkfortsatzes (PWP) und ist durch ein kurzes, breites Band mit dem
PWP verbunden.

Das 3Ax ist kompakt und unregelmäßig geformt. Es hat einen kurzen caudalen Arm, an
dessen distalem Ende sich eine kurze Gabelung befindet. Von der Mitte des proximalen
Randes geht ein deutlich abgesetzter Fortsatz aus, an dem eine kurze Sehne ansetzt, die zu
einer kleinen Muskelplatte (AMD) in der Membran zwischen 1Ax und 3Ax zieht. Die
AMD ist der Insertionspunkt der Muskulatur des 3Ax.

Die proximale Medianplatte (PMP) ist gut abgegrenzt, ihr proximaler Rand ist mit dem
2Ax verschmolzen. Distal der PMP liegt eine etwa gleich große, sklerotisierte Fläche mit
einem kurzen, proximalen Fortsatz, die als distale Medianplatte (DMP) interpretiert werden
kann. Die Basis der Media ist als Stumpf an der Kreuzungsstelle von Radius, Media und
Cubitus erkennbar und erreicht die DMP nicht. Die Basis des Cubitus ist stark erweitert.

Pleural-Region (Abb.18B)

Der Gelenkkopf des pleuralen Flügelgelenkfortsatzes (F) ist leicht länglich-oval geformt.
Er erreicht nur ca. ein Zehntel der Länge des 1Ax und liegt unter dessen hinterem Hals-
bereich. Unterhalb des Fulcrum setzt ein Band an, das die Verbindung zum ventralen Fort-
satz des 2Ax herstellt.

Vor dem PWP liegt das Basalare (Ba). Es ist basal mit dem Episternum verschmolzen und
bildet dorsal kurz unterhalb des Fulcrum einen breiten, flachen Kopf aus, der dem PWP
dicht angelagert ist. Cranial hat es einen langen, nach dorsal gerichteten, flachen Anhang,
der mit dem basalen Vorderrand des Flügels (Costabasis/Humeralplatte) verbunden ist.

In der Membran hinter dem PWP und direkt unterhalb des PNP liegt das große plattenför-
mige Subalare (Sb). Es ist durch einen schmalen Membranstreifen mit dem PNP verbunden.

16

Carabidae: Cicindelinae
Material: Cicindela lunulata

Notum (Abb.19)

Der vordere Gelenkfortsatz (ANP) ist annähernd dreieckig, flach ausgezogen und gut skle-
rotisiert. Die Spitze liegt deutlich hinter dem Vorderrand des Notum. Der Berührungspunkt
zwischen ANP und 1Ax befindet sich nur knapp hinter der Vorderkante des 1Ax-Kopfes.
Der mittlere Gelenkfortsatz (MNP) wird durch eine vor ihm liegende tiefe Einkerbung des
Notum deutlich abgesetzt. Der Fortsatz selbst ist breit, mit je einer vorderen und einer
hinteren Spitze. Das caudale Ende des 1Ax ist durch einen schmalen Membranstreifen mit
dem MNP verbunden. Der hintere Gelenkfortsatz (PNP) ist basal sehr schlank und verbrei-
tert sich distal. Der distale Rand des PNP ist flach eingekerbt, so daß ein kleiner vorderer
und ein großer hinterer Zahn entstehen. Der PNP ist über eine derbe Membran mit dem
hinteren Fortsatz des 2Ax und dem caudalen Arm des 3Ax verbunden. Der Winkel B zwi-
schen den Achsen (a) und (b) (Abb.3, Tab.1) beträgt ca. 27°.

Axillar-Region (Abb.19, 21B)

Das 1Ax hat einen sehr breiten Kopf, dessen nach ventral umgeschlagener Vorderrand drei
Einkerbungen aufweist, die sich bei geöffnetem Flügel mit entsprechenden Strukturen der
Subcosta-Basis verhaken. Der proximale Bereich des Vorderrandes ist zusätzlich in einen
langen, ventral weisenden Zahn ausgezogen, der durch Membranen mit der Basis der Sub-
costa und dem Basalare verbunden ist. Der Kopf des 1Ax geht rasch in den verhältnismä-
Big schmalen Hals über, dessen hinterer Bereich auf dem Fulcrum (F) liegt. Gegen den
Körper ist der Hals durch eine kleine Einkerbung des distalen Randes deutlich abgegrenzt.
Kopf und Hals liegen etwas höher als der ANP. Der Kopf des 1Ax ist wenig über den
Rand des ANP geschoben. Der Körper des 1Ax ist asymmetrisch, er verschmälert sich aus-
gehend von seiner breitesten Stelle schräg nach proximo-caudal. Seine proximale Kante ist
fast über ihre gesamte Länge unter den Rand des Notum geschoben. Die disto-craniale
Kante des Skleritkörpers ist deutlich konvex, die distale Spitze ist nach hinten umgebogen.
Die disto-craniale Kante und die Achse durch den vorderen und den hinteren Kontaktpunkt
von 1Ax und Notum schließen einen Winkel von ca. 56° ein. Der abgeschrägte Teil des
1Ax-Korpers macht etwa zwei Fünftel der Gesamtlänge des 1Ax aus (Tab.1). Durch das
rasche Schmalerwerden des 1Ax-Kopfes entsteht zwischen Kopf, Hals und Körper eine
relativ große Bucht, in der das zweite Axillare liegt.

Das 2Ax ist proximal breit gerundet. Nach caudal läuft es in einen relativ langen spitzen
Fortsatz aus, der bis in die ventrale Flügelmembran sklerotisiert und durch ein Band mit
dem PNP verbunden ist. Cranial ist ein Rest des Basiradiale (BR) als kurzer Stumpf vor-
handen. Ventral hat das 2Ax einen kräftigen Fortsatz, dessen Ende zweihöckrig ausgebildet
ist. Dieser Fortsatz liegt dicht hinter dem PWP unterhalb des Fulcrum und ist mit diesem
durch ein kräftiges Band verbunden.

Das 3Ax ist stabförmig ausgebildet. Cranial endet es in einer kurzen Gabelung. Etwa in der
Mitte des proximalen Randes trägt es einen relativ großen Zahn, an dem eine Sehne an-
setzt, die zu der kleinen Muskelplatte (AMD) in der Membran zwischen 1Ax und 3Ax
zieht.

Die Medianplatten sind zu schmalen sklerotisierten Streifen umgebildet. Die proximale Me-
dianplatte (PMP) verläuft in weit geschwungenem Bogen vom proximalen Zahn des 3Ax
zur gemeinsamen Basis von Media und Cubitus. Die distale Medianplatte (DMP) liegt als

17

Y-förmiges Gebilde zwischen PMP, 2Ax und BSc. Sie hat keine sklerotisierte Verbindung
zu einem dieser Elemente oder einer Aderbasis.

Pleural-Region (Abb.20, 21A)

Das Fulcrum ist etwas weniger als ein Zehntel so lang wie das gesamte 1Ax. Es liegt unter
dessen hinterem Halsbereich. Von dorsal betrachtet ist der Gelenkkopf längs oval. Im un-
teren Bereich hat das Fulcrum eine starke Vorwölbung, so daß der gesamte Kopf von late-
ral sehr massig wirkt. Der Vorderrand des PWP weist einen Fortsatz auf, der unter den
Kopf des Basalare faßt. Am Hinterrand des PWP setzt kurz unterhalb des Kopfes ein Band
an, das zum ventralen Fortsatz des 2Ax zieht.

Das vor dem PWP liegende Basalare ist mit einem breiten flachen Kopf ausgestattet, des-
sen Oberkante direkt an die Vorwölbung unterhalb des Fulcrum anschließt. Der craniale
Rand des Ba ist in eine lange, flache, dorsal weisende Platte ausgezogen, die durch Mem-
branen mit dem basalen Flügelvorderrand verbunden ist. Die Basis des Ba ist mit dem Epi-
sternum verschmolzen.

Das Subalare ist als große Skleritplatte ausgebildet und liegt direkt unter dem PNP, mit
dem sie durch einen derben Membranstreifen fest verbunden ist.

Carabidae: Zabrinae, Harpalinae, Pterostichinae

Material: Amara sp., Harpalus sp., Harpalus cordatus, Poecilus versicolor, Pterostichus
metallicus

Notum (Abb.22, 24)

Der ANP ist flach ausgezogen, bei Harpalus am Ende breit gerundet, bei Amara zugespitzt.
Er bleibt nur wenig hinter dem Notumvorderrand zurück. Der Berührungspunkt zwischen
ANP und 1Ax liegt direkt hinter der Vorderkante des 1Ax-Kopfes. Der MNP ist durch eine
tiefe Einkerbung des Notumseitenrandes deutlich hervorgehoben. Er ist kurz, breit und
zweispitzig. Bei Harpalus ist die cranial gelegene Spitze größer als die caudale und berührt
fast das Ende des 1Ax. Der PNP ist durch Bänder mit dem 2Ax und dem 3Ax verbunden.
Bei Amara ist er basal relativ schmal, distal verbreitert, mit einer Kerbe im distalen Rand,
so daß ein vorderer kleiner und ein hinterer großer Zahn entstehen. Bei Harpalus ist der
PNP stark verbreitert und fast kreisförmig. Die Achsen (a) und (b) durch das Ende und die
Basis des PNP (Abb.3) schließen einen Winkel zwischen 25° und 28° ein (Tab.1).

Axillar-Region (Abb.22, 24)

Der Kopf des 1Ax ist sehr breit, sein Vorderrand weist in der distalen Hälfte einen senk-
recht verlaufenden Grat auf. Der Vorderrand ist nach unten umgeschlagen und proximal
in einen langen, mit der Subcosta-Basis und dem Basalare verbundenen Zahn ausgezogen.
Der sehr schmale Hals des 1Ax liegt auf dem Gelenkkopf des PWP. Der Körper ist asym-
metrisch ausgebildet. Sein disto-cranialer Rand ist leicht konvex. Der schräg ausgezogene
Teil des 1Ax-Körpers ist etwa halb so lang wie das gesamte 1Ax. Kopf und Hals liegen
etwas höher als der ANP, der Kopf ist mit seinem proximalen Rand ein wenig über den
Rand des ANP geschoben. Der proximale Rand des 1Ax-Körpers liegt zu etwas mehr als
einem Drittel unter dem Rand des Notum. Der Winkel zwischen der disto-cranialen Kante
des 1Ax-Körpers und der Gelenkachse von 1Ax und Notum beträgt ca. 63°. Das caudale
Ende des 1Ax reicht sehr nah an den MNP heran.

18

Das zweite Axillare ist dem distalen Rand des 1Ax dicht angelagert. Sein proximales Ende
ist bei Harpalus breiter gerundet als bei Amara. Vom Basiradiale ist je ein Stumpf am 2Ax
und am Radius neben der Basis der Subcosta vorhanden. Der ventrale Fortsatz liegt dicht
neben dem PWP und ist durch ein Band mit diesem verbunden. Diese Verbindung ist zu-
sätzlich oberflächlich sklerotisiert. Ein weiteres Band zieht vom caudalen Fortsatz zum
PNP.

Das 3Ax hat bei Harpalus einen relativ großen, bei Amara einen kleineren proximalen
Fortsatz, von dem aus eine Sehne zu einer Skleritplatte (AMD) in der Membran zwischen
1Ax und 3Ax zieht. Der distale Bereich des 3Ax läuft spitz aus und ist leicht nach caudal
umgebogen. Die proximale Kante des caudalen Armes des 3Ax ist S-förmig geschwungen.
Er wird durch ein Band mit dem PNP verbunden.

Die Medianplatten sind zu einer großen, schwach sklerotisierten Platte verschmolzen und
laufen proximal in dünne Arme aus. Diese sind bei Harpalus ebenfalls verschmolzen und
erreichen gemeinsam das 3Ax. Bei Amara bleiben sie getrennt, und nur der Ausläufer der
PMP ist mit dem 3Ax verbunden. Die DMP hat einen in Richtung Subcosta-Basis weisen-
den Ausläufer.

Pleural-Region (Abb.23)

Der Gelenkkopf des PWP ist stempelartig verbreitert. Er ist etwas mehr als ein Zehntel so
lang wie das gesamte 1Ax (Tab.1). Unterhalb des Gelenkkopfes setzt ein Band an, das zum
ventralen Fortsatz des 2Ax zieht. Im Verlauf dieses Bandes ist die Membran oberflächlich
sklerotisiert. Der Hinterrand des PWP ist glatt, der Vorderrand hat einen Fortsatz, der unter
dem Kopf des Basalare liegt.

Der Kopf des Basalare ist breit und flach mit einem cranial gelegenen, nach dorsal weisen-
den Stab, der mit dem basalen Flügelvorderrand (Humeralplatte/Costabasis) verbunden ist.
Die Basis des Ba ist mit dem Episternum verschmolzen.

Das Subalare liegt als flache Skleritplatte unter dem PNP und ist mit diesem durch einen
schmalen Membranstreifen verbunden.

Myxophaga
Hydroscaphidae
Material: Hydroscapha sp.

Notum (Abb.25A)

Der ANP ist groß, annähernd dreieckig und überragt mit fast seiner ganzen Lange den Vor-
derrand des Notum. Sein Ende trifft knapp hinter dem Vorderrand des 1Ax-Kopfes auf des-
sen proximalen Rand. Ein MNP ist nicht erkennbar. Der PNP ist kurz und distal breit abge-
rundet. Der Winkel zwischen den Achsen (a) und (b) (Abb.3) beträgt ca. 13°.

Axillar-Region (Abb.25A)

Kopf und Hals des 1Ax sind sehr groß und massig. Der Vorderrand des 1Ax-Kopfes weist
eine Einbuchtung auf, in welche die relativ schmale Subcosta-Basis hineinfaßt. Der Kopf
verschmälert sich gleichmäßig zum Hals, der kurz und distal spitz eingekerbt ist. Der
Körper des 1Ax hat einen langen proximo-caudalen Fortsatz, dessen Ende in der von No-
tumseitenrand und PNP gebildeten Bucht liegt. Die distale Spitze des 1Ax-Körpers ist
leicht nach caudal umgebogen. Der Winkel « des 1Ax (Abb.3) beträgt ca. 32°, der caudale

19

Fortsatz des Körpers ist etwas mehr als ein Drittel so lang wie das gesamte 1Ax. Das
Notum ist nur ca. 1,5 mal so lang wie das 1Ax.

Das 2Ax ist relativ schmal. Seine distale Kante verläuft fast völlig gerade und parallel zur
Gelenkachse zwischen 1Ax und Notum. Die distale Fläche des 2Ax liegt tiefer als das di-
rekt an der proximalen Spitze des 2Ax entspringende Basiradiale. Dieses verläuft dicht
neben dem Kopf des 1Ax nach vorne.

Das 3Ax ist mehr als halb so lang wie das 1Ax. Es ist schlank stabförmig mit einer etwa
auf das Doppelte der sonstigen Breite erweiterten Basis. Die Muskulatur setzt direkt an der
vorderen Ecke der basalen Erweiterung an, eine AMD ist nicht vorhanden.

Beide Medianplatten sind als schwach sklerotisierte, ebene Strukturen erkennbar.

Microsporidae
Material: Microsporus sp.

Notum (Abb.25B)

Die Gelenkfortsätze des Notum von Microsporus entsprechen in ihrer Ausbildung weit-
gehend denen von Hydroscapha. Der ANP ist etwas kürzer und kompakter. Der Winkel
zwischen den Achsen (a) und (b) durch die Spitze und die Basis des PNP (Abb.3) beträgt
Cael]:

Axillar-Region (Abb.25B)

Der hauptsächliche Unterschied zu Hydroscapha besteht im 3Ax. Es entspricht in seiner
Form dem von Hydroscapha, ist aber wesentlich breiter und massiger.

Polyphaga
Hydrophilidae: Helophorinae
Material: Helophorus sp.

Notum (Abb.27A)

Der ANP ist im Verhältnis zu Notum und 1Ax relativ klein. Sein Ende bleibt hinter dem
Vorderrand des Notum zurück. Es trifft den proximalen Rand des 1Ax im hinteren Drittel
des 1Ax-Kopfes. Der MNP ist nur als sehr flache Einbuchtung des Notumrandes erkennbar.
Der PNP ist relativ kurz und hakenförmig ausgebildet. Der Winkel zwischen den Achsen
(a) und (b) (Abb.3) beträgt ca. 12°.

Axillar-Region (Abb.27A)

Der Kopf des ersten Axillare (1Ax) ist nur unwesentlich breiter als der Hals. Der Vorder-
rand ist nach ventral umgeschlagen, er weist eine zentral gelegene, flache Einbuchtung auf,
in die bei geöffnetem Flügel eine entsprechende Struktur der Subcosta-Basis einrastet. Der
umgeschlagene Rand ist proximal in einen relativ langen, ventral gerichteten Zahn ausge-
zogen. Der Hals des 1Ax ist nicht deutlich gegen Kopf und Körper abgegrenzt. Kopf und
Hals liegen mit ihrer proximalen Kante dem Rand des ANP auf. Unter dem hinteren Kopf-
bzw. dem vorderen Halsbereich liegt der Gelenkkopf des PWP. Der Körper des 1Ax ist
dreieckig ausgebildet, der caudale Rand ist flach konkav, der disto-craniale gerade bis
leicht konkav. Der proximale Rand liegt unter dem Rand des Notum. Der Winkel « beträgt
ca. 32° (Abb.3), das Notum ist etwa doppelt so lang wie das gesamte 1Ax.

20

Das 2Ax ist relativ lang und schmal. Es hat ventral einen großen, flachen Fortsatz, der etwa
bis zur Mitte unter den Körper des 1Ax ragt, und einen kleineren zentral gelegenen Fort-
satz, von dem aus ein Band zum PWP zieht. Das dünn ausgezogene caudale Ende ist über
ein schmales, leicht sklerotisiertes Band mit dem PNP verbunden. Das Basiradiale ent-
springt an der äußersten proximalen Spitze des 2Ax und verläuft dicht neben dem Kopf des
1Ax als sehr dünner, durchgehend sklerotisierter Streifen in Richtung Subcosta-Basis.

Das 3Ax ist im Verhältnis zu den anderen Elementen der Flügelbasis auffällig groß. Bei
geöffnetem Flügel liegt es leicht schräg zur Körperlängsachse. Die distale Kante ist flach
eingekerbt. Der Ansatzpunkt der AMD-Sehne ist durch eine caudal liegende Einbuchtung
gegen den caudalen Arm abgesetzt. Dieser weist eine lange, gerade proximale Kante auf,
die mit dem PNP verbunden ist.

Distale und proximale Medianplatte sind verschmolzen und nur leicht sklerotisiert.

Pleural-Region (Abb.27B)

Das Fulcrum verbreitert sich nach dorsal gleichmäßig, die Gelenkfläche ist nahezu eben.
Das 1Ax ist ca. 4 bis 4,5 mal so lang wie der PWP-Gelenkkopf. Am Hinterrand des PWP
entspringt kurz unterhalb des Kopfes ein Band, das zum ventralen Fortsatz des 2Ax zieht.

Das vor dem PWP gelegene Basalare (Ba) ist basal mit dem Episternum verschmolzen. Der
Stiel ist relativ breit, der Kopf hat eine nach distal gerichtete Erweiterung mit einer
abgeschrägten dorsalen Fläche. Diese Erweiterung dient als Gleitfläche und Arretierung für
die ventrale Subcosta-Basis. Vom cranialen Rand des Ba-Kopfes entspringt eine nach dor-
sal gerichtete schmale Platte, die mit dem basalen Vorderrand des Flügels verbunden ist.

Das flache Subalare (Sb) liegt unter dem PNP, mit dem es durch einen leicht sklerotisierten
Membranstreifen fest verbunden ist.

Hydrophilidae: Hydrophilinae

Material: Anacaena limbata, Hydrophilus piceus

Notum (Abb.26A)

Der vordere Gelenkfortsatz (ANP) ist flach, annähernd dreieckig und leicht zugespitzt. Die
Spitze überragt nicht den Vorderrand des Notum. Sie trifft am Ende des vorderen Drittels
des 1Ax-Kopfes auf dessen proximalen Rand. Der mittlere Gelenkfortsatz (MNP) ist kurz
dreieckig. Er wird durch zwei flache Einbuchtungen des Notumseitenrandes begrenzt. Der
Abstand zwischen MNP und 1Ax ist relativ groß. Der hintere Gelenkfortsatz (PNP) ist spitz
hakenförmig ausgebildet. Sein Ende liegt deutlich weiter distal als der ANP und reicht bis
in die Region des MNP nach vorne. Er ist durch leicht sklerotisierte Membranbereiche mit
dem 2Ax und dem 3Ax verbunden. Der Winkel zwischen den Achsen (a) und (b) (Abb.3)
beträgt ca. 17°.

Axillar-Region (Abb.26A,C)

Das erste Axillare (1Ax) ist sehr kompakt gebaut. Sein Kopf ist breit und fast halb so lang
wie das gesamte 1Ax. Der Vorderrand hat distal eine flache Einbuchtung, in welche die
Basis der Subcosta einrastet. Er ist nach ventral umgeschlagen und distal in einen langen
nahezu dreieckigen Fortsatz ausgezogen. Dieser ist mit der Subcosta-Basis und dem Basa-

21

lare verbunden. Im Halsbereich weist die distale Kante eine schmale relativ tiefe Einbuch-
tung auf. Der Hals ist etwa halb so breit wie der Kopf, aber ausgesprochen kurz. Kopf und
Hals des 1Ax liegen auf dem Rand des ANP. Der hintere Kopfbereich und der Hals liegen
auf dem Fulcrum. Der Körper des 1Ax ist relativ kurz und kompakt. Sein caudaler Rand
ist flach konkav mit einer leicht verlangerten distalen Ecke. Eine proximale Verlangerung
wie bei Archostemata und Adephaga ist nicht ausgebildet. Der proximale Rand liegt unter
dem Rand des Notum. Der disto-craniale Rand verlauft gerade bis leicht konkav. Der Win-
kel « zwischen der disto-cranialen Kante und der Achse durch den vorderen und den hin-
teren Kontaktpunkt von 1Ax und Notum beträgt ca. 36° (Abb.3, Tab.1).

Das 2Ax ist durch einen sehr schmalen, leicht sklerotisierten Membranstreifen fest mit der
disto-cranialen Kante des 1Ax-Körpers verbunden. Ventral hat es einen großen, flachen
Fortsatz, der weit unter den Körper des 1Ax ragt. Die proximale Ecke des 2Ax ist ver-
längert, so daß ein kurzer Fortsatz entsteht, der in die schmale Einbuchtung im proximalen
Rand des Halses des 1Ax hineinragt. Von einem schmalen caudalen Fortsatz zieht ein sta-
biles Band zum PNP. Von einem ventralen Fortsatz geht ein weiteres Band aus, das die
Verbindung zum pleuralen Flügelgelenkfortsatz herstellt. Das gesamte 2Ax ist gleichmäßig
stark sklerotisiert. Das Basiradiale ist ebenfalls vollständig sklerotisiert und distal leicht
verschmälert.

Das 3Ax ist distal breit abgeflacht. Nach proximal erhebt sich aus dieser Fläche ein - bei
geöffnetem Flügel - etwa senkrecht zur Körperlängsachse stehender Grat. An dessen proxi-
malem Rand setzt eine Sehne an, die zur AMD in der Membran zwischen 1Ax, Notum und
3Ax führt. Die craniale Kante der distalen Fläche des 3Ax ist tief eingekerbt, so daß ein
schmaler Zahn entsteht. Der caudale Arm des 3Ax ist schmal, spitz dreieckig, und seine
proximale Kante verläuft nahezu gerade. Er ist durch einen schmalen, leicht sklerotisierten
Membranstreifen fest mit dem PNP verbunden.

Die distale Medianplatte (DMP) ist nur leicht sklerotisiert. Sie hat einen proximo-cranialen
Fortsatz, der zur Basis des Radius zieht. Zwei schmale, schwach sklerotisierte Bänder ver-
binden sie mit der ebenfalls nur leicht sklerotisierten proximalen Medianplatte (PMP).

Pleural-Region (Abb.26B)

Das Fulcrum (F) ist pilzförmig erweitert. Von lateral gesehen hat die dorsale Fläche eine
im vorderen Drittel liegende Einbuchtung. Der vordere Teil des Fulcrum liegt unter dem
1Ax-Kopf, der hintere Teil unter dem Hals des 1Ax. Vom Hinterrand des PWP kurz unter-
halb des Kopfes zieht ein Band zum ventralen Fortsatz des 2Ax.

Das Basalare (Ba) ist dem Vorderrand des PWP dicht angelagert. Basal ist es mit dem Epi-
sternum verschmolzen. Der Stiel des Ba ist etwas um seine Längsachse gedreht, so daß die
Außenfläche fast nach vorne zeigt. Er erweitert sich dorsad zu einem Kopf, der aus einer
schräg nach außen weisenden Fläche und einem nach vorn-oben gerichteten Fortsatz be-
steht. Letzterer ist mit der basalen Flügelvorderkante und mit dem Vorderrand des 1Ax-
Kopfes verbunden. Die abgeschrägte Fläche dient als Gleitfläche und Arretierung für die
ventrale Basis der Subcosta.

Das relativ kleine Subalare (Sb) liegt unter dem hinteren Bereich des PNP, mit dem es
durch eine leicht sklerotisierte Membran fest verbunden ist.

22

Silphidae
Material: Nicrophorus vespilloides, Nicrophorus investigator, Oeceoptoma thoracica,
Blitophaga opaca

Notum (Abb.28, 30A, 31A)

Der vordere Gelenkfortsatz ist deutlich vom Notum abgesetzt. Er ist dreieckig bis lang
trapezförmig und im Verhältnis zum Notum relativ groß, er überragt aber in der Regel
nicht den Notumvorderrand. Die Spitze des ANP trifft direkt am Vorderrand des 1Ax oder
kurz dahinter auf dessen proximale Kante. Der MNP ist als kleiner Zahn am Notumrand
etwa in der Mitte zwischen ANP und PNP erkennbar. Der hintere Gelenkfortsatz ist kurz
hakenförmig. Hinter dem PNP ist der Notumrand eingekerbt. Hinter dieser Einkerbung ist
ein schmaler stabförmiger Fortsatz des Postnotum mit dem Notumrand verschmolzen. Das
Ende dieses Stabes steht mit dem Subalare in Verbindung. Die Achsen (a) und (b) durch
die Spitze und die Basis des PNP schließen einen Winkel von ca. 8° bis 10° ein (Abb.3,
Tab.1).

Axillar-Region (Abb.28, 29C, 30A,C,D, 31A,C,D)

Das 1Ax ist kompakt gebaut. Sein Kopf ist maximal doppelt so breit wie der Hals. Der
Kopf trägt distal eine Auslappung, die bei Nicrophorus zu einem zapfenformigen Fortsatz
verschmälert ist. Der Kopfvorderrand ist schräg nach disto-ventral ausgezogen. Dieser
Fortsatz hat ca. ein Drittel der Lange des gesamten 1Ax. Sein ventrales Ende ist löffel-
ähnlich nach vorne aufgebogen und relativ fest mit der ventralen Basis der Subcosta
verbunden. Die craniad gerichtete Fläche trägt eine Struktur aus Längsrippen, die bei
geöffnetem Flügel an der Subcosta-Basis einrastet. Der distale Rand des 1Ax ist nach
ventral umgeschlagen. Dieser umgeschlagene Bereich liegt bei Nicrophorus und Blitophaga
auf dem Gelenkkopf des PWP. Bei Oeceoptoma ragt die proximale Spitze des 2Ax unter
den Hals des 1Ax und liegt auf dem Gelenkkopf des PWP. Der caudale Rand des 1Ax-
Körpers ist zum Notum hin abgeschrägt, er verläuft gerade bis leicht konkav. Der disto-
craniale Rand ist schwach konvex. Die hintere proximale Ecke des Körpers ist verlängert,
und ihre Spitze liegt unter dem Notumrand. Der Winkel zwischen der Gelenkachse von
Notum und 1Ax und der disto-cranialen Kante des 1Ax-Körpers beträgt bei Nicrophorus
ca. 30°, bei Oeceoptoma und Blitophaga ca. 38° (Abb.3, Tab.1).

Das 2Ax ist innerhalb der Silphidae unterschiedlich ausgebildet. Bei Nicrophorus ist der
dorsal sichtbare Teil relativ klein, mit einem langen, schmalen Basiradiale, das an der
proximalen Spitze entspringt. Das caudale Ende erreicht das 3Ax nicht. Der proximale
ventrale Fortsatz des 2Ax ist groß und ragt weit unter den Körper des 1Ax. Bei Oeceop-
toma und Blitophaga ist der dorsale Bereich des 2Ax im Vergleich zum 1Ax größer als bei
Nicrophorus. Das caudale Ende reicht bis zum 3Ax. Das Basiradiale entspringt etwas distal
der proximalen Spitze. Bei Oeceoptoma hat die proximale Spitze des 2Ax eine Verlange-
rung, die unter den Hals des 1Ax reicht und dem Gelenkkopf des PWP aufliegt. Der ven-
trale Bereich ist ähnlich wie bei Nicrophorus ausgebildet.

Die distale Fläche des 3Ax ist bei Blitophaga relativ breit, bei Nicrophorus und Oece-
optoma etwas schmaler und länger. Durch eine Einfaltung der Membran disto-craniad des
3Ax liegt die distale Fläche teilweise über der proximalen Medianplatte. Eine tiefe
Einbuchtung des proximalen Randes des 3Ax trennt nach vorne einen langen, schmalen
Fortsatz ab, der als Ansatzpunkt für die AMD-Sehne dient. Die Muskelplatte (AMD) ist
relativ groß und kräftig sklerotisiert. Caudal der Einbuchtung liegt der lange, schlanke

23

caudale Arm des 3Ax. Dessen proximale Kante weist bei Nicrophorus auf Höhe der PNP-
Spitze einen scharfen Winkel auf, bei Oeceoptoma und Blitophaga verläuft sie sehr flach
konvex. Der caudale Arm liegt über eine relativ lange Strecke dem PNP an.

Die beiden Medianplatten sind kräftig sklerotisiert und durch zwei sklerotisierte Membran-
streifen miteinander verbunden. Der disto-caudale dieser Verbindungsstreifen liegt durch
eine Einfaltung der Membran teilweise unter der distalen Fäche des 3Ax. Die proximale
Medianplatte ist besonders lang und schmal, bei Nicrophorus erreicht sie den Ansatzpunkt
der AMD-Sehne am 3Ax.

Pleural-Region (Abb.29A,B,D, 30B, 31B)

Das Fulcrum ist kurz und kompakt; in der Aufsicht ist es unregelmäßig oval und auffallend
breit. Es erreicht nur ca. ein Sechstel der Länge des 1Ax. Bei Nicrophorus liegt das Ful-
crum unter dem hinteren Kopfbereich des 1Ax. Bei Oeceoptoma und Blitophaga liegt zu-
sätzlich die proximale Spitze des 2Ax auf dem Fulcrum. Die Ansatzstelle des zum ventra-
len Fortsatz des 2Ax ziehenden Bandes ist durch eine leichte Einbuchtung des Hinterrandes
des PWP direkt unterhalb des Gelenkkopfes markiert.

Das Basalare ist bis direkt unter seinen Kopf mit dem Episternum verschmolzen. Basal ist
es relativ schmal, nach dorsal erweitert es sich allmählich. Der Kopf des Basalare trägt eine
kegelige bis ellipsoide Vorwölbung, die nach distal weist. Sie dient als Arretierungspunkt
für eine entsprechende Struktur der ventralen Subcosta-Basis. An der Vorderkante befindet
sich ein nach dorsal gerichteter Fortsatz, der mit der Flügelvorderkante verbunden ist.
Das Subalare liegt unter dem PNP. Es ist relativ klein und leicht schräg angeordnet. Dorsal
trägt es einen nach caudal gerichteten Fortsatz. Dieser ist mit einem nach cranial gerich-
teten, unter dem PNP liegenden Fortsatz des Postnotum verbunden.

Das Epimeron weist eine große dorsale Auslappung auf, die bei Oeceoptoma und Nicro-
phorus bis direkt unter den PNP reicht.

Staphylinidae
Material: Quedius sp., Ontholestes murinus (Abb.32A, 33A)

Die Ausbildung der Gelenkfortsätze ähnelt stark den Verhältnissen bei den Silphidae. Der
vordere Gelenkfortsatz ist relativ groß, trapezförmig und deutlich vom Notum abgesetzt.
Sein Vorderrand liegt auf gleicher Höhe mit dem Vorderrand des Notum oder überragt die-
sen nach cranial. Auf das erste Axillare trifft er kurz hinter dessen Vorderkante. Der MNP
ist klein und zahnförmig. Er liegt bei geöffnetem Flügel der Ansatzstelle der AMD-Sehne
des 3Ax direkt gegenüber. Der PNP ist sehr kurz hakenförmig. Das Postnotum trägt einen
langen, schlanken craniad gerichteten Fortsatz, der direkt unter dem hinteren Bereich des
PNP liegt und mit diesem teilweise verschmolzen ist. Die Achsen (a) und (b) (Abb.3)
schließen einen Winkel von ca. 7° ein (Tab.1).

Axillar-Region (Abb.32A,C,D, 33A)

Kopf und Hals des 1Ax gehen ohne deutliche Abgrenzung ineinander tiber. Der Kopfvor-
derrand ist etwas verbreitert, nach ventral umgeschlagen und in einen langen, schräg nach
disto-ventral gerichteten Fortsatz ausgezogen. Dieser ist am Ende leicht löffelähnlich
aufgebogen und an seiner cranialen Fläche mit ein bis zwei dorso-ventral verlaufenden
Rippen versehen. Diese Strukturen dienen der Verbindung mit der Subcosta-Basis. Der

24

proximale Rand von Kopf und Hals ist wenig über den ANP geschoben. Der distale Rand
ist ventral umgeschlagen und bildet so eine Auflagefläche für den Gelenkkopf des PWP.
Der Körper des 1Ax ist vom Hals durch eine deutliche Einbuchtung des proximalen Randes
abgesetzt. Die hintere proximale Ecke des Körpers ist verlängert und am Ende breit
gerundet. Sie liegt unter dem Notumrand. Die Gelenkachse von 1Ax und Notum und die
disto-craniale Kante des 1Ax-Körpers schließen bei Ontholestes einen Winkel von ca. 25°,
bei Quedius von ca. 38° ein. Das Notum ist etwa 2,5 mal so lang wie das gesamte 1Ax.

Das 2Ax ist dicht an den Körper des 1Ax angelagert, sein proximaler ventraler Fortsatz
reicht wenigstens bis zur Hälfte unter den Körper des 1Ax. Das caudale Ende des 2Ax ist
durch ein Band mit dem PNP verbunden; von einem zentralen ventralen Fortsatz zieht ein
weiteres Band zum PWP. Die proximale Spitze des 2Ax ragt unter den distalen Rand des
1Ax und liegt auf dem distalen Bereich des PWP-Gelenkkopfes. Das Basiradiale entspringt
der proximalen Spitze des 2Ax, es ist durchgehend sklerotisiert und sehr schmal.

Das 3Ax ist dem der Silphidae sehr ähnlich. Sein proximaler Rand hat eine tiefe
Einbuchtung, die nach cranial einen sehr schmalen, langen Fortsatz abgliedert, von dem die
Sehne zur Muskelplatte (AMD) abgeht. Der hinter der Einbuchtung beginnende caudale
Arm ist auffallend lang und schlank ausgebildet. Er liegt über eine große Strecke dem PNP
dicht an. Die Spitze des caudalen Arms ist mit dem unter dem PNP liegenden Fortsatz des
Postnotum verbunden. Der distale Bereich des 3Ax ist flach und leicht verbreitert. Die
AMD in der Membran zwischen Notum, 1Ax und 3Ax ist verhältnismäßig groß und kräftig
sklerotisiert.

Die proximale Medianplatte verschmälert sich stark nach caudal. Mit der breiten, fast
rechteckigen distalen Medianplatte steht sie über eine breite Skleritbrücke in Verbindung.

Pleural-Region (Abb.32B, 33B)

Der Gelenkkopf des PWP ist bei Quedius in der Aufsicht sehr breit dreieckig, bei Ontho-
lestes im Verhältnis zum ANP deutlich schmaler und oval. In beiden Fällen hat die Gelenk-
fläche etwa ein Sechstel der Gesamtlänge des 1Ax. Auf dem Gelenkkopf liegt die gemein-
sam von der proximalen Spitze des 2Ax und dem umgeschlagenen distalen Rand des 1Ax
gebildete Axillargelenkfläche. Vom Hinterrand des PWP direkt unterhalb des Fulcrum ent-
springt ein zum 2Ax ziehendes Band. Unterhalb der Ansatzstelle dieses Bandes ist die
caudale Kante des PWP stark erweitert.

Das Basalare ist bis kurz unter seinen Kopf mit dem Episternum verschmolzen. Von der
relativ schmalen Basis her erweitert es sich allmählich zum Kopf hin. Dieser hat einen
schmalen zylinderförmigen Fortsatz, der fast waagerecht nach distal gerichtet ist. Er bildet
zusammen mit korrespondierenden Strukturen der Subcosta-Basis einen Rastmechanismus
für den angelegten Flügel. Cranial trägt der Ba-Kopf einen schräg nach dorsal gerichteten,
flachen Fortsatz, der die Verbindung zur basalen Flügelvorderkante herstellt.

Das Subalare liegt unter dem hinteren Bereich des PNP und ist mit diesem und dem mit
dem PNP verschmolzenen Ausläufer des Postnotum durch verstärkte Membranbereiche
verbunden. Die Muskelansatzfläche des Subalare ist etwa doppelt so groß wie die außen
in der Membran sichtbare Fläche.

25

Lucanidae

Material: Sinodendron cylindricum

Notum (Abb.34, 35)

Der ANP ist gleichseitig dreieckig, im Verhältnis zum Notum aber relativ klein. Er bleibt
deutlich hinter dem Notumvorderrand zurück. Die Spitze des ANP trifft am Ende des 1Ax-
Kopfes auf dessen proximalen Rand. Der MNP ist als annähernd senkrecht zur Körper-
längsachse stehender, gerader Dorn ausgebildet. Der PNP ist kurz hakenförmig. Sein
distaler Rand liegt auf einer Linie mit der distalen Ecke des 1Ax-Körpers. Der Winkel
zwischen Basis und Spitze des PNP mit Bezug zur Spitze des ANP (Abb.3) beträgt ca. 8°.

Axillar-Region (Abb.34, 35, 36B,C)

Das 1Ax hat einen durch eine scharfe U- bis V-förmige Kante gegen den Hals abgesetzten
Kopf. Dieser trägt am Vorderrand einen breiten distad gerichteten Fortsatz, der durch eine
flache Einkerbung gegen den proximalen Teil des Kopfes abgegrenzt ist. Dieser Fortsatz
geht frontal in einen zum Ende hin verschmälerten, schräg nach ventro-distal weisenden
Vorsprung über. Durch die Einkerbung zwischen dem proximalen Teil des Kopfes und dem
distalen Fortsatz entstehen zwei Vorwölbungen. Jeder dieser Vorwölbungen entspricht eine
gleichartige Struktur der Subcosta-Basis, so daß zwei Kontaktpunkte zwischen BSc und
1Ax-Kopf bestehen. Der Hals des 1Ax ist relativ kurz, breit und leicht gebogen. Er ver-
breitert sich kontinuierlich zum Körper hin. Der disto-ventrale Halsrand ist breit nach innen
umgeschlagen. Dadurch entsteht eine breite Auflagefläche für das Fulcrum. Die distale
Kante des 1Ax-Körpers verläuft schwach konvex, die caudale hingegen leicht konkav. Die
Gelenkachse von Notum und 1Ax und die disto-craniale Kante des 1Ax-Körpers schließen
einen Winkel von ca. 36° ein.

Die Dorsalseite des 2Ax ist langgestreckt dreieckig. Die kürzeste Kante bildet den Vor-
derrand. Die vordere, proximale Ecke trägt in Verlängerung der proximalen Kante einen
kurzen Fortsatz, an dessen disto-cranialer Kante das lange und sehr schmale Basiradiale
entspringt. Der ventro-laterale Vorsprung des 2Ax reicht bis etwa zur Mitte unter den
Körper des 1Ax. Der caudale Fortsatz, von dem ein Band zum PNP zieht, ist relativ kurz
und breit. Ein zweites Band verbindet das 2Ax mit dem PWP.

Das 3Ax ist insgesamt sehr langgestreckt. Der caudale Arm ist lang und hat eine nur leicht
geschwungene proximale Kante. Der distale Arm ist mehr craniad als distad gerichtet. Er
ist auf ganzer Länge gleich breit. Sein Ende ist breit abgerundet und nach distal gebogen.
Der distale Rand ist in seinem gesamten Verlauf mehr oder weniger stark ausgefranst. Die
Ansatzstelle der AMD-Sehne ist leicht erhöht und proximal vorgewölbt. Die spitz drei-
eckige AMD liegt auf Höhe des MNP sehr dicht beim 3Ax. Die schmalste Kante ist dem
1Ax zugewandt.

Bei der Struktur, die zwischen dem distalen Arm des 3Ax-und dem 2Ax liegt, handelt es
sich wahrscheinlich um einen Ausläufer der DMP. Als Rest der PMP ist nur noch ein kur-
zer dreieckiger Fortsatz am Hauptteil der DMP verblieben, der in den membranösen

Bereich zwischen dem Ende des distalen Arms des 3Ax und dem potentiellen Ausläufer
der DMP ragt.

Pleural-Region (Abb.35, 36A)

Der Gelenkkopf des PWP ist in der Dorsalansicht fast halbkreisförmig. In der Mitte des
Fulcrum läuft eine Vertiefung quer über die gesamte Breite, so daß zwei Gelenkpunkte, ein

26

vorderer und ein hinterer, entstehen. Auf diesen Gelenkpunkten liegt der umgeschlagene
Bereich des 1Ax-Halses (s.o.). Das Fulcrum ist nur leicht verlängert. Es ist nur ca. ein
Fünftel so lang wie das gesamte 1Ax. Das Band, das den PWP mit dem 2Ax verbindet,
inseriert unterhalb des Fulcrum an einer etwas vertieften Stelle des PWP-Hinterrandes.

Das Basalare ist nur im basalen Bereich mit dem Episternum verschmolzen. In den oberen
zwei Dritteln ist es durch einen Membranstreifen vom PWP getrennt. Der Ba-Kopf ist
gegenüber dem Stiel leicht erweitert, er wird praktisch vollständig von der großen lateralen
Erweiterung gebildet, die den Rastmechanismus mit Humerus und BSc bildet. Die dorsale
Fläche dieser Erweiterung ist schräg nach unten gerichtet. Der frontale Fortsatz des Ba ist
verhältnismäßig kurz und fast waagerecht nach vorne gerichtet.

Das Subalare ist mit etwas mehr als einem Siebtel der Notumlänge ausgesprochen kurz. Es
wird durch eine Längsnaht in eine dorsale und eine ventrale Hälfte geteilt. Sein hinterer
Dorsalrand ist über einen leicht sklerotisierten Membranstreifen mit einem langen, nach
vorne gerichteten Fortsatz des Postalararmes verbunden.

Scarabaeidae

Material: Phyllopertha horticola, Aphodius sp., Cetonia cf. aurata

Notum (Abb.37A)

Der ANP ist fast gleichseitig dreieckig und verhältnismäßig klein. Er bleibt deutlich hinter
dem Notumvorderrand zurück. Sein Ende trifft dicht hinter dem Kopf des 1Ax auf dessen
proximalen Rand. Direkt neben der hinteren proximalen Ecke des 1Ax-Körpers befindet
sich eine kleine Bucht im Notumseitenrand. Ihr Hinterrand verläuft gerade und bildet mit
dem Rand des Notum einen rechten Winkel. Der Vorderrand der Bucht ist stark konkav,
so daß sie distal von einem kurzen, nach hinten gerichteten Haken begrenzt wird. Diese
Bucht mit ihrem Vorder- und Hinterrand entspricht in Funktion und Lage dem MNP. Der
PNP ist als sehr kurzer, hakenförmiger Fortsatz ausgebildet, dessen Spitze und Basis mit
Bezug zur Spitze des ANP einen Winkel von ca. 5° einschließen (Abb.3).

Axillar-Region (Abb.37A,C,D)

Das 1Ax ist kompakt gebaut. Sein Kopf ist relativ kurz und breit. Der Kopfvorderrand ist
in einen nach unten deutlich schmaler werdenden, schräg nach disto-ventral gerichteten
Fortsatz ausgezogen. Der ebenfalls verhältnismäßig breite Hals ist nicht scharf von Kopf
und Körper abgesetzt. Der ventro-distale Halsrand ist so breit nach innen umgeschlagen,
daß er fast den proximalen Halsrand erreicht. Dieser umgeschlagene Rand liegt auf dem
Gelenkkopf des PWP. Die proximale Ecke des Körpers ist leicht verlängert, so daß sie ca.
ein Achtel der Gesamtlänge des 1Ax ausmacht. Ihr Ende liegt direkt vor dem caudalen
Rand der MNP-Bucht des Notum. Der Winkel zwischen der Gelenkachse von 1Ax und
Notum und der disto-cranialen Kante des 1Ax-Körpers beträgt ca. 44°.

Das 2Ax ist in der Ansicht von dorsal kurz dreieckig mit einer leicht konkaven Vorder-
kante, in deren proximalem Drittel das Basiradiale ansetzt. Dieses ist als kurzer, schmaler
Stumpf ausgebildet, der am Distalrand des 1Ax endet. Der latero-ventrale Fortsatz des 2Ax
ist breit dreieckig und ragt bis jenseits der Mitte unter den Körper des 1Ax.

Der distale Arm des 3Ax ist schräg nach disto-cranial gerichtet und stark keulenförmig
erweitert. Seine Vorderkante weist mittig eine sehr flache Kerbe auf. Der Insertionspunkt
der AMD-Sehne liegt als kurz dornförmiger Fortsatz direkt hinter der distalen Ecke des

Di

1Ax-Körpers. Der caudale Arm ist relativ kurz. Sein Ende ist abgerundet und liegt direkt
neben dem PNP.

Die DMP hat einen langen, schmalen Ausläufer, der dicht neben dem 2Ax zwischen die-
sem und dem distalen Arm des 3Ax liegt. Die PMP ist bis auf einen kurzen Fortsatz am
Hauptteil der DMP reduziert.

Pleural-Region (Abb.37B)

Das Fulcrum ist in der Dorsalansicht mit einem Viertel der Länge des 1Ax relativ lang. Es
ist gegenüber dem PWP nur wenig verbreitert. Vorne läuft es spitz aus, der Hinterrand ist
abgerundet. In der hinteren Hälfte hat es eine deutliche Aufwölbung. Der PWP ist gegen
die Senkrechte stark nach vorne gekippt.

Das Basalare ist nur basal mit dem Episternum verschmolzen. Über den größten Teil seiner
Höhe wird es durch einen schmalen Membranstreifen vom PWP getrennt. Wie der PWP
ist es deutlich nach vorne gekippt. Der Kopf ist gegenüber dem Stiel nur wenig erweitert.
Der Knopf, der mit der Humeralplatte und der BSc den Rastmechanismus bildet, ist relativ
klein, längsoval und liegt etwas unterhalb des Dorsalrandes des Ba. Der frontale Fortsatz
des Ba ist schräg nach oben gerichtet und so lang, daß sein Ende in einer Ebene mit dem
ANP und dem Fulcrum liegt.

Das Subalare ist sehr klein und durch einen leicht sklerotisierten Membranstreifen mit dem
PNP verbunden.

Byrrhidae
Material: Byrrhus sp.

Notum (Abb.38A)

Der ANP ist groß und länger als breit. Sein Vorderrand liegt, individuell etwas unter-
schiedlich, auf gleicher Hohe mit dem Vorderrand des Notum, oder er bleibt knapp hinter
diesem zurtick. Das Ende des ANP trifft im hinteren Kopfbereich auf den proximalen Rand
des 1Ax. Der MNP fehlt. Der PNP ist kurz und breit hakenförmig. Spitze und Basis des
PNP schließen mit Bezug auf die Spitze des ANP einen Winkel von ca. 8° ein (Abb.3).

Axillar-Region (Abb.38A,C,D)

Das 1Ax ist sehr schlank und langgestreckt. Sein Hals ist an der engsten Stelle etwa halb
so breit wie der Kopf, der wiederum nahezu so breit wie der Körper ist. Der Kopfvorder-
rand trägt neben dem senkrecht nach unten zeigenden, breiten Fortsatz einen zweiten, der
waagerecht nach distal gerichtet ist. Der distale Halsrand ist ventral schmal umgeschlagen
und bildet so eine Auflagefläche für das Fulcrum. Die proximale Ecke des 1Ax-Körpers
ist deutlich verlängert und erreicht etwa ein Fünftel der Gesamtlänge des 1Ax. Der Winkel
zwischen der disto-cranialen Kante des Körpers und der Gelenkachse zwischen Notum und
lAx beträgt ca. 30° (Abb.3, Tab.1).

Die proximale Ecke des 2Ax ist in Richtung 1Ax verlängert, so daß das BR nicht direkt
an der am weitesten proximal gelegenen Stelle entspringt. Der laterale Fortsatz des 2Ax
reicht bis jenseits der Mitte unter den Körper des 1Ax. Der distale Bereich des 2Ax ist
breit abgerundet.

Das 3Ax ist wegen der Sklerotisierung der Fläche zwischen dem caudalen und dem distalen
Arm stark vergrößert. Die proximale Kante des caudalen Arms ist im hinteren Bereich ge-

28

rade und liegt über eine längere Strecke dem PNP an. Der Insertionspunkt der AMD-Sehne
ist als kurzer, abgerundeter Fortsatz ausgebildet. Die AMD ist relativ klein und queroval.

Die Medianplatten sind breit bogenförmig verschmolzen und schieben sich als schwach
sklerotisierter Keil zwischen 3Ax und 2Ax.

Pleural-Region (Abb.38B)

Das Fulcrum ist sehr lang und schmal. Es ist ca. ein Viertel so lang wie das gesamte 1Ax
und liegt unter dessen proximalem Kopf/Halsbereich. Der Stiel des Basalare ist ausge-
sprochen schmal und nur basal mit dem Episternum verschmolzen. Der Kopf wird fast voll-
kommen von dem großen, cranio-ventrad gerichteten Rastknopf gebildet. Der frontale Fort-
satz ist relativ kurz und weist senkrecht nach dorsal. Das mittelgroße, leicht länglich ovale
Subalare liegt unter dem PNP, mit dem es durch einen breiten Membranstreifen verbunden
ist. Das Postnotum hat einen kompakten cranialen Fortsatz, der aber weder den PNP noch
das Sb erreicht.

Buprestidae

Material: Anthaxia sp., Chalcophora mariana

Notum (Abb.39, 41A)

Der vordere Gelenkfortsatz (ANP) ist auffallend klein. Sein Ende bleibt deutlich hinter dem
Vorderrand des Notum zurück. Die ANP-Spitze trifft etwa in der Mitte des Halses auf das
1Ax. Der MNP ist bei Chalcophora als kleine Einbuchtung des Notumseitenrandes direkt
hinter dem 1Ax erkennbar. Bei Anthaxia fehlt der MNP vollständig. Der PNP ist kurz und
breit, sein Ende ist schräg abgestutzt. Zwischen PNP und Notum ist eine Bucht ausgebildet,
die bei Chalcophora nicht ganz ein Drittel, bei Anthaxia fast die Hälfte der caudalen
Verlängerung des 1Ax umfaßt. Die Achsen (a) und (b) durch die Spitze und die Basis des
PNP schließen einen Winkel von ca. 16° ein (Abb.3, Tab.1).

Axillar-Region (Abb.39, 41A)

Das 1Ax zeichnet sich durch einige charakteristische Bildungen aus. Sein Kopf ist kurz und
nahezu rechteckig. Er ist deutlich gegen den langen, breiten Hals abgesetzt. Der nach ven-
tral umgeschlagene Vorderrand des 1Ax-Kopfes tragt in seiner distalen Halfte zwei kleine
Einkerbungen, die eine Vorwölbung einschließen. Diese Strukturen korrespondieren mit
entsprechenden Bildungen der Subcosta-Basis, die hier bei geöffnetem Flügel einrasten. Der
Hals des 1Ax liegt mit seiner proximalen Kante auf dem Rand des ANP. Zum 1Ax-Körper
hin verbreitert sich der Hals leicht; er ist vom Körper durch eine Einbuchtung des proxi-
malen Randes deutlich abgesetzt. Die vordere Hälfte des Halses liegt auf dem Gelenkkopf
des PWP. Der Hauptteil des Körpers des 1Ax ist relativ kurz. Proximal hat er eine caudale
Verlängerung, die ein Drittel bis fast die Hälfte der Gesamtlänge des 1Ax ausmacht
(Tab.1). Der Habitus des 1Ax ähnelt stark dem der Archostemata und der Adephaga. Der
Winkel zwischen der Gelenkachse von Notum und 1Ax und der disto-cranialen Kante des
1Ax-Körpers beträgt bei Chalcophora ca. 38° , bei Anthaxia ca. 40°. Das Notum ist ca. 1,8
mal so lang wie das 1Ax (Abb.3, Tab.1).

Das 2Ax ist bis auf den stark sklerotisierten proximalen Rand reduziert. Durch die nahtlose
Anlagerung des ebenfalls kräftig sklerotisierten Basiradiale (BR) entsteht eine bogen- bis
U-förmige Struktur, die in die weite Bucht zwischen Kopf und Körper der 1Ax eingepaßt

29

ist. Ventral hat das 2Ax zwei Fortsätze: einen dornförmigen, der etwa bis zu dessen Mitte
unter den Körper der 1Ax ragt, und einen zweiten zapfenförmigen, von dem aus ein Band
zum PWP zieht.

Das 3Ax ist kräftig ausgebildet. Der caudale Arm ist kurz, seine proximale Kante ist leicht
geschwungen, caudal läuft er spitz zu. Der distale Bereich des 3Ax ist relativ breit und
flach, sein distaler Rand ist bei Anthaxia gegabelt, bei Chalcophora schräg abgestutzt. Die
Ansatzstelle der AMD-Sehne an der proximo-cranialen Ecke des 3Ax ist etwas erhöht. Die
Muskelplatte in der Membran zwischen 1Ax und 3Ax (AMD) ist relativ groß.

Die proximale Medianplatte (PMD) bildet ein bogenförmiges Skleritelement. Sie liegt distal
des 2Ax und cranial des 3Ax und umschließt zusammen mit dem 2Ax einen Membranbe-
reich, durch den eine Falte läuft. Ein feiner Skleritfaden verbindet die proximale mit der
distalen Medianplatte. Die DMP ist als leicht erweiterte gemeinsame Basis von Media und
Cubitus identifizierbar.

Pleural-Region (Abb.40, 41B)

Das Fulcrum (F) ist ausgesprochen lang. Seine dorsale Fläche ist in der Mitte vertieft, so
daß ein vorderer und ein hinterer Auflagepunkt für den Hals des 1Ax entstehen. Durch die
Vertiefung in der Mitte des Gelenkkopfes läuft ein Tracheenast in den Flügel. Das 1Ax ist
bei Anthaxia fast neunmal, bei Chalcophora ca. 5,6 mal so lang wie das Fulcrum. Direkt
unterhalb des Gelenkkopfes setzt an der Hinterkante des PWP ein Band an, das zum ven-
tralen Fortsatz des 2Ax zieht.

Das vor dem PWP liegende Basalare ist bis dicht unter seinen Kopf mit dem Episternum
verschmolzen. Der Kopf des Ba weist eine große, halbovale Erweiterung auf, die distad
und schräg nach unten gerichtet ist. Sie bietet eine Gleitfläche und Arretierung für die
ventrale Subcosta-Basis. Frontal hat der Ba-Kopf eine kurze, annähernd senkrecht nach
dorsal weisende Platte, die mit dem Vorderrand des Flügels verbunden ist.

Das Subalare liegt unter dem PNP, ist leicht unregelmäßig, länglich geformt und steht über
eine sklerotisierte Brücke mit dem Postnotum in Verbindung.

Elateridae

Material: Hemicrepidius niger, Denticollis linearis, Argiotes pilosellus, Agrypnus murinus,
Elater cf. ferrugineus, Hypnoidus? sp.

Notum (Abb.42A, 43A, 44A)

Der vordere Gelenkfortsatz des Notum (ANP) ist im Verhältnis zum Notum relativ klein.
Er ist flach dreieckig ausgezogen, das Ende ist schmal gerundet. Der Vorderrand des ANP
liegt auf gleicher Höhe mit dem Vorderrand des Notum oder etwas dahinter. Die Spitze des
ANP trifft etwa auf halber Lange des 1Ax-Kopfes auf dessen proximalen Rand. Der MNP
ist bei Argiotes als minimale Vorwölbung des Notumseitenrandes erkennbar, bei Hemicre-
pidius und Denticollis ist er nicht identifizierbar. Der PNP ist kurz dreieckig, bei Argiotes
endet er zugespitzt, bei Hemicrepidius und Denticollis stumpf. Die Achsen (a) und (b)
(Abb.3) schließen einen Winkel von 10° bis 16° ein (Tab. 1).

Axillar-Region (Abb.42A, 43A, 44A,C)

Das erste Axillare (1Ax) hat einen breiten Kopf, dessen nach ventral umgeschlagener Vor-
derrand eine flache bis ausgepragte Einkerbung tragt. Die distale Ecke kann einfach spitz

30

ausgezogen sein, einen kurzen schmalen Fortsatz tragen, oder durch eine Einkerbung vom
Rest des Kopfes abgesetzt sein. Der Vorderrand ist nach ventral in einen breiten, trapez-
förmigen, relativ kurzen Fortsatz ausgezogen, der mit der Subcosta-Basis in Verbindung
steht. Der Sklerithals ist ein Drittel bis halb so breit wie die breiteste Stelle des Kopfes.
Der Übergangsbereich zwischen Kopf und Hals liegt auf dem Gelenkkopf des PWP. Die
Ausbildung des Körpers des 1Ax variiert zwischen den untersuchten Arten leicht. Generell
ist er relativ kurz dreieckig, die hintere proximale Ecke ist kaum bis stark verlängert. Die
caudale Kante ist deutlich konkav, die disto-craniale Kante fast gerade bis leicht konvex.
Der Winkel zwischen der disto-cranialen Kante des 1Ax-Körpers und der Achse durch den
ANP und den hinteren Anlagepunkt des 1Ax beträgt ca. 40°. Das gesamte 1Ax ist etwa
halb so lang wie das Notum.

Das 2Ax ist innerhalb des Taxon variabel ausgebildet. Bei Denticollis ist in der dorsalen
Membran nur ein schmaler Streifen direkt neben dem 1Ax stark sklerotisiert. Bei Denti-
collis und Argiotes liegt die proximale Spitze des 2Ax auf dem distalen Rand des Fulcrum.
Ventral hat das 2Ax einen flachen Fortsatz, der weit unter den Körper des 1Ax faßt. Dieser
Fortsatz kann als breite Platte oder schmaler Dorn ausgebildet sein. Von einem kleinen
zapfenförmigen Fortsatz an der Unterseite des 2Ax zieht ein Band zum PWP unterhalb des
Gelenkkopfes. Das caudale Ende des 2Ax ist durch einen kurzen, leicht sklerotisierten
Membranstreifen mit der proximo-cranialen Ecke des 3Ax verbunden. Das Basiradiale ist
unterschiedlich stark, aber durchgehend sklerotisiert.

Das 3Ax ist etwas länger als das 1Ax. Der distale Bereich ist breit und flach. Der Ansatz-
punkt für die AMD-Sehne ist leicht erhöht und in Richtung Notum vorgewölbt. Der cau-
dale Arm ist so lang wie oder etwas länger als der gesamte Rest des 3Ax. Die proximale
Kante verläuft gerade und ist durch einen Streifen verstärkter Membran mit dem PNP ver-
bunden.

Die distale und die proximale Medianplatte sind relativ schwach sklerotisiert und ver-
schmolzen oder durch zwei Skleritstreifen verbunden.

Pleural-Region (Abb.42B, 43B, 44B)

Das Fulcrum ist, in der Ansicht von dorsal, langgestreckt mit einem zugespitzten Vorder-
ende. Die dem ANP zugewandte Kante verläuft relativ gerade, die distale Kante ist konvex.
Bei Argiotes und Denticollis ist sie so weit vorgewölbt, daß sie unter der proximalen Spitze
des 2Ax liegt. Das 1Ax ist etwa 3,5 bis 4,5 mal so lang wie das Fulcrum. Vom PWP kurz
unterhalb des Gelenkkopfes zieht ein Band zum ventralen Fortsatz des 2Ax.

Der Stiel des vor dem PWP gelegenen Basalare ist weitgehend mit dem Episternum ver-
schmolzen. Der Kopf des Ba hat eine nach disto-cranial gerichtete, leicht abwärts geneigte
Erweiterung, die als Gleitfläche und Arretierungspunkt für die ventrale Subcosta-Basis
dient. Außerdem trägt er frontal einen nach dorsal gerichteten stabförmigen Fortsatz, der
mit dem basalen Flügelvorderrand verbunden ist.
Das flache, scheibenförmige Subalare liegt unter dem PNP und ist mit diesem und dem
Postnotum durch Bereiche leicht sklerotisierter Membran verbunden.

Lampyridae
Material: Lamprohiza splendidula

Notum (Abb.45)
Der vordere Gelenkfortsatz (ANP) ist dreieckig ausgebildet. Sein Ende überragt den
Vorderrand des Notum nur wenig. Der Berührungspunkt zwischen ANP und 1Ax liegt kurz

31

hinter dem Vorderrand des 1Ax-Kopfes. Ein mittlerer Gelenkfortsatz (MNP) ist nicht
vorhanden. Der PNP ist kurz hakenförmig. Die Achsen (a) und (b) durch die Spitze und
die Basis des PNP (Abb.3) schließen einen Winkel von ca. 12° ein (Tab.1).

Axillar-Region (Abb.45)

Das 1Ax ist fast halb so lang wie das Notum. Der Vorderrand des 1Ax-Kopfes ist nach
ventral umgeschlagen und trägt in der distalen Hälfte einen großen knopfartigen Fortsatz.
Der Kopf des 1Ax liegt auf dem Gelenkkopf des PWP. Der 1Ax-Körper ist schmal drei-
eckig. Seine hintere proximale Ecke ist lang ausgezogen und etwas verbreitert. Dieser
Anhang hat ca. ein Sechstel der Gesamtlänge des 1Ax. Die Gelenkachse zwischen Notum
und 1Ax (Abb.3) und die disto-craniale Kante des 1Ax-Körpers schließen einen Winkel
von ca. 29° ein. Bei den nicht flugfähigen Weibchen dieser Art ist das Hinterflügelgelenk
in unterschiedlich starkem Maße rückgebildet. Charakteristisch ist dabei, daß auch im am
weitesten fortgeschrittenen Reduktionsgrad, wenn 2Ax, 3Ax und die Medianplatten schon
völlig fehlen, das 1Ax noch in seiner typischen Form erkennbar ist. Allerdings ist es mit
dem Notum verschmolzen (Geisthardt 1974). Reduktionen in sehr ähnlicher Art und Weise
sind z.B. auch bei Curculioniden zu beobachten (s.u.).

Das 2Ax ist der disto-cranialen Kante des Körpers des 1Ax dicht angelagert. Ventral hat
es einen großen, näherungsweise dreieckigen proximalen Fortsatz, der unter den Körper des
1Ax ragt und fast dessen proximalen Rand erreicht. Von einem kleineren, zentral unter dem
2Ax gelegenen Fortsatz verläuft ein Band zum PWP. Die caudale Spitze des 2Ax ist durch
ein weiteres Band mit dem PNP verbunden. Der dorsal sichtbare Bereich des 2Ax ist spitz
dreieckig mit einem tief konkaven cranialen Rand. Das Basiradiale ist bis auf einen kurzen
Stumpf an der proximalen Spitze des 2Ax reduziert.

Das 3Ax ist kompakt und kräftig. Der Ansatzpunkt der AMD-Sehne ist als breit gerundeter
Fortsatz des proximalen Randes ausgebildet. Der caudale Arm ist relativ breit, sein
proximaler und distaler Rand verlaufen fast gerade, das Ende ist leicht zugespitzt. Der
distale Bereich des 3Ax ist abgeflacht und zum Ende hin sehr verbreitert. Die Muskelplatte
in der Membran zwischen Notum, 1Ax und 3Ax (AMD) ist relativ klein und dreieckig.

Die proximale Medianplatte ist schief dreieckig mit einem tief eingebuchteten cranialen
Rand. Sie steht mit der distalen Medianplatte über eine Skleritbrücke in Verbindung.

Pleural-Region (Abb.46)

Der Gelenkkopf des PWP ist mit einem Viertel bis zu einem Drittel der Länge des 1Ax
ausgesprochen lang, dabei aber sehr schmal. Er liegt unter der proximalen Hälfte von Kopf
und Hals des 1Ax. Die Gelenkfläche selbst ist leicht konkav. Vom Hinterrand des PWP
direkt unterhalb des Gelenkkopfes zieht ein Band zum 2Ax.

Das Basalare ist bis unter seinen Kopf mit dem Episternum verschmolzen. Der Kopf des
Basalare trägt distal eine große, schräg nach ventral gerichtete Erweiterung. Die Dorsal-
fläche dieser Erweiterung ist auffallend glatt und dient als Gleitfläche und Arretierungs-
punkt für die ventrale Subcosta-Basis. Am Vorderrand des Basalar-Kopfes befindet sich ein
kurzer, stumpfer Fortsatz, der mit der basalen Flügelvorderkante verbunden ist.

Das Subalare ist lang oval und recht groß. Es liegt unter dem PNP und ist mit diesem
durch einen verstärkten Membranbereich verbunden.

32

Cantharidae

Material: Cantharis nigricans, Cantharis pellucida

Notum (Abb.47)

Der ANP wird durch eine Einkerbung neben seiner proximalen Ecke deutlich vom Notum
abgesetzt. Der Vorderrand des ANP bleibt knapp hinter dem Vorderrand des Notum zurück.
Die Spitze des vorderen Gelenkfortsatzes trifft im mittleren Kopfbereich auf den pro-
ximalen Rand des 1Ax. Der mittlere Gelenkfortsatz ist vollständig reduziert. Der hintere
Gelenkfortsatz ist als flach gerundete Vorwölbung des Notumrandes ausgebildet. Da ein
langgezogener Fortsatz fehlt, ist die Basis des PNP schwer zu bestimmen. Hier wird der
Punkt, an dem die Aufwölbung des Notum auf den Notumseitenrand trifft, als PNP-Basis
festgelegt. Davon ausgehend kann der Winkel zwischen den Achsen (a) und (b) durch die
Spitze und die Basis des PNP mit ca. 7° bestimmt werden (Abb.3, Tab.1).

Axillar-Region (Abb.47)

Die vordere distale Ecke des Kopfes des 1Ax trägt eine pilzförmige Abschnürung. In die-
sem Bereich ist der Kopf ca. doppelt so breit wie der Hals des 1Ax. Der Vorderrand des
Kopfes ist nach ventral umgeschlagen und kurz und breit ausgezogen. Kopf und Hals lie-
gen mit ihrer proximalen Kante dem Rand des ANP auf. Kopf und Hals liegen auf dem
Gelenkkopf des PWP. Der Körper des 1Ax ist schlank dreieckig. Sein caudaler Rand ist
leicht konkav, der proximale Rand liegt unter dem Rand des Notum. Die Gelenkachse zwi-
schen 1Ax und Notum und der disto-craniale Rand des 1Ax-Körpers schließen einen Win-
kel von ca. 24° ein.

Das 2Ax ist schmal dreieckig, eine proximale ventrale Erweiterung ragt unter den Körper
des 1Ax. Von einem zentral gelegenen, zapfenförmigen ventralen Fortsatz zieht ein Band
zum PWP. Die caudale Spitze des 2Ax ist durch ein Band mit dem PNP verbunden. Das
kurze, breite Basiradiale entspringt etwa in der Mitte der cranialen Kante des 2Ax; es ist
vollständig sklerotisiert.

Das 3Ax hat einen stark verlängerten, sehr schmalen caudalen Arm. Die proximale Kante
des 3Ax trägt zwei Fortsätze. Der vordere ist leicht aufgewölbt und dient als Ansatzstelle
für die AMD-Sehne. Der zweite Fortsatz ist ein kleiner Zahn, der etwas caudad des ersten
Fortsatzes liegt. Der distale Bereich des 3Ax ist abgeflacht und leicht verbreitert, er läuft
in eine stumpfe Spitze aus.

Die Medianplatten sind schwächer sklerotisiert als die Axillaria, aber gut identifizierbar.
Die distale Medianplatte ist nur als leichte Erweiterung der Mediabasis ausgebildet; sie ist
mit der PMP verschmolzen.

Pleural-Region (Abb.48, 49)

Der Gelenkkopf des PWP ist fast halb so lang wie das 1Ax, unter dessen Kopf- und Hals-
bereich er liegt. An der caudalen Kante des PWP direkt unterhalb des Gelenkkopfes ent-
springt ein Band, das zum 2Ax zieht.

Der Stiel des vor dem PWP liegenden Basalare ist über den größten Teil seiner Länge mit
dem Episternum verschmolzen. Der Kopf des Ba hat disto-cranial eine relativ kleine, bla-
sige Erweiterung. Der Vorderrand trägt eine kurze, nach dorsal gerichtete Skleritplatte, die
mit der basalen Flügelvorderkante verbunden ist.

Das große, leicht schräg gestellte Subalare liegt dicht unterhalb des PNP.

33

Dermestidae

Material: Dermestes lardarius

Notum (Abb.50A)

Der vordere Gelenkfortsatz ist schief dreieckig und im Verhältnis zum Notum relativ klein.
Die Spitze des ANP überragt den Vorderrand des Notum nicht; sie trifft etwa auf halber
Länge des 1Ax-Kopfes auf dessen proximalen Rand, wobei der distale Rand des ANP ein
wenig unter den Kopf des 1Ax geschoben ist. Der mittlere Gelenkfortsatz ist nur als sehr
flache Welle im Notumrand ausgebildet. Der PNP ist hakenförmig, sein Ende ist relativ
kurz, schmal und spitz. Die Achsen (a) und (b) (Abb.3) schließen einen Winkel von ca. 8°
ein (Tab.1).

Axillar-Region (Abb.50A)

Kopf und Hals des 1Ax sind im Verhältnis zu seinem Körper relativ kurz. Das gesamte
1Ax ist ungefähr ein Drittel so lang wie das Notum. Der Vorderrand des 1Ax-Kopfes ist
nach ventral umgeschlagen und in der proximalen Hälfte in einen langen Vorsprung aus-
gezogen. Der hintere Kopfbereich liegt auf dem Gelenkkopf des PWP. Der Körper des 1Ax
ist unregelmäßig dreieckig. Sein caudaler Rand ist konkav, die proximo-caudale Ecke ist
etwas verlängert. Diese Verlängerung beträgt wenig mehr als ein Siebtel der Gesamtlänge
des 1Ax. Die disto-craniale Kante des 1Ax-Körpers und die Gelenkachse zwischen Notum
und 1Ax (Abb.3) schließen einen Winkel von ca. 28° ein.

Das zweite Axillare ist verglichen mit dem 1Ax relativ groß. Es hat einen proximalen,
ventralen Fortsatz, der weit unter den Körper des 1Ax reicht. Die caudale Spitze des 2Ax
ist über ein Band mit dem PNP verbunden. Mittig unter dem 2Ax sitzt ein weiterer Fort-
satz, von dem aus ein Band zum PWP verläuft. Das Basiradiale ist schmal, durchgehend
sklerotisiert und entspringt etwas distal der proximalen Spitze des 2Ax.

Das 3Ax ist kräftig sklerotisiert und leicht unregelmäßig geformt. Der caudale Arm ist
durch eine Einbuchtung des proximalen Randes vom Rest des 3Ax abgesetzt. Der Fortsatz,
an dem die AMD-Sehne ansetzt, ist kurz und flach. Die AMD in der Membran zwischen
Notum, 1Ax und 3Ax ist relativ klein.

Die proximale Medianplatte ist in drei kleinere sklerotisierte Bereiche unterteilt, die über
schmale Skleritbrücken miteinander verbunden sind. Das am weitesten distal gelegene Teil-
stück ist wiederum mit der distalen Medianplatte verbunden.

Pleural-Region (Abb.50B)

Der Gelenkkopf des PWP liegt unter dem hinteren Kopfbereich des 1Ax. Er ist in der Auf-
sicht dreieckig und relativ breit und kurz. Er ist etwas mehr als ein Siebtel so lang wie das
1Ax. Vom Hinterrand des PWP direkt unterhalb des Gelenkkopfes zieht ein Band zum
2AX.

Das Basalare ist bis knapp unter seinen Kopf mit dem Episternum verschmolzen. Der Kopf
hat eine große, blasige, distale Erweiterung, die in der Lateralansicht breit oval erscheint.
Sie dient als Gleitfläche und Arretierungspunkt der Subcosta-Basis. Der craniale Rand des
Ba-Kopfes trägt einen langen, stabförmigen Fortsatz, der schräg nach dorsal gerichtet ist
und mit dem basalen Flügelvorderrand in Verbindung steht.

Das sehr kleine Subalare liegt weit caudal unter dem PNP.

34

Cleridae

Material: Trichodes sp., Thanasimus formicarius

Notum (Abb.51A, 52A)

Der ANP ist bei beiden untersuchten Arten im Verhältnis zum Notum klein. Er überragt
den Vorderrand des Notum nicht und berührt das 1Ax in dessen hinterem Kopf- bzw.
vorderen Halsbereich. Der MNP fehlt bei Thanasimus vollständig. Bei Trichodes ist der
Notumseitenrand im Bereich des MNP craniad schmal eingebuchtet, so daß ein nach hinten
gerichteter, hakenförmiger Fortsatz entsteht. Der PNP ist als spitz auslaufender, schwach
sklerotisierter Haken ausgebildet. Basis und Spitze des PNP schließen mit Bezug auf die
Spitze des ANP einen Winkel von ca. 11° ein (Abb.3).

Axillar-Region (Abb.51A,C,D, 52A,C,D)

Das 1Ax hat einen kräftigen, breiten Hals. Der Kopf erreicht fast die Breite des Körpers
und trägt an der Vorderkante zusätzlich zu dem nach disto-ventral gerichteten Fortsatz
einen kurzen, waagerecht nach distal weisenden Vorsprung. Die proximale Ecke des Kör-
pers ist leicht verlängert, der Hinterrand deutlich konvex ausgebildet. Der Winkel zwischen

der disto-cranialen Kante des Körpers und der Gelenkachse zwischen Notum und 1Ax
beträgt 27° bzw. 31° (Abb.3, Tab.1).

Der vorderste Punkt des 2Ax ist der Kontaktpunkt der proximalen Ecke mit dem 1Ax. Die-
se Ecke liegt auf dem distalen Rand des Fulcrum. Das Basiradiale ist schmal und setzt an
der proximalen Spitze des 2Ax an. Bei Thanasimus ist die Ansatzstelle verbreitert. Der
latero-ventrale Fortsatz des 2Ax reicht bis zur Mitte unter den Körper des 1Ax.

Das 3Ax ist durch die Sklerotisierung der Fläche zwischen dem caudalen und dem distalen
Arm stark vergrößert. Der proximale Rand des caudalen Arms verläuft annähernd gerade
und liegt dem PNP über eine längere Strecke an. Die Verbindung zwischen 3Ax und PNP
wird über einen schmalen Membranstreifen hergestellt. Der Ansatzpunkt der AMD-Sehne
ist leicht erhaben und proximad vorgewölbt. Die AMD liegt relativ dicht neben dem 3Ax;
bei Trichodes ist sie auffallend groß.

Die Medianplatten sind nur leicht sklerotisiert. Der proximale Ausläufer schiebt sich als
breites Band zwischen 3Ax und 2Ax.

Pleural-Region (Abb.51B, 52B)

Das Fulcrum erreicht ca. ein Achtel bis ein Sechstel der Lange des 1Ax. Es ist deutlich
verbreitert und liegt unter dem distalen Kopf-Halsbereich des 1Ax. Der distale Rand ragt
allerdings unter dem Hals des 1Ax hervor und liegt unter der proximalen Spitze des 2Ax.
Das Basalare ist im dorsalen Drittel durch einen schmalen Membrankeil vom PWP ge-
trennt. Der Kopf des Ba wird vollständig vom Rastknopf eingenommen, der schräg nach
cranio-ventral gerichtet ist. Der frontale Fortsatz des Ba ist kurz und weist schräg nach
oben. Das Subalare ist bei beiden Arten ausgesprochen klein und liegt weit hinten unter
dem PNP. Es ist mit einem sehr langen, schmalen Ausläufer des Postnotum verbunden.

Melyridae
Material: Malachius bipustulatus, Malachius sp.

Notum (Abb.53A)

Der vordere Gelenkfortsatz (ANP) ist im Verhältnis zum Notum sehr klein. Sein Vorder-
rand bleibt deutlich hinter der Vorderkante des Notum zurück. Die Spitze des ANP trifft

35

im Halsbereich auf die proximale Kante des 1Ax. Der MNP ist nur durch eine leichte Ein-
buchtung des Notumseitenrandes direkt hinter dem 1Ax markiert. Der PNP ist hakenförmig
und wie der ANP sehr klein. An der hinteren Ecke der distalen Kante trägt er einen klei-
nen, distal weisenden Zahn. Die Achsen (a) und (b) durch die Spitze und die Basis des
PNP (Abb.3) schließen einen Winkel von ca. 8° ein.

Axillar-Region (Abb.53A)

Kopf und Halsbereich des 1Ax sind etwa gleich breit und nicht voneinander abgesetzt. Der
nach ventral umgeschlagene Vorderrand des 1Ax-Kopfes hat in der distalen Hälfte eine
Einkerbung, in die bei geöffnetem Flügel eine entsprechende Struktur der Subcosta-Basis
einrastet. Kopf und Hals des 1Ax liegen mit ihrem proximalen Rand auf dem ANP. Der
hintere Kopf- bzw. der vordere Halsbereich liegt auf dem Gelenkkopf des PWP. Der Kör-
per des 1Ax verbreitert sich stark nach caudal. Nur ein kurzes Stück des proximalen Ran-
des liegt unter dem Notumrand. Die caudale Kante ist deutlich konkav, die disto-craniale
Kante verläuft annähernd gerade. Das Notum ist ca. 2,3 mal so lang wie das 1Ax. Der
Winkel zwischen der Gelenkachse von 1Ax und Notum und der disto-cranialen Kante des
1Ax-Körpers beträgt ca. 34° (Abb.3).

Das 2Ax ist der disto-cranialen Kante des 1Ax-Körpers dicht angelagert. Es hat einen brei-
ten, flachen, ventralen Fortsatz, der unter den Körper des 1Ax ragt und dessen Spitze fast
den proximalen Rand des 1Ax erreicht. Vom Zentrum dieses Fortsatzes zieht ein Band zum
PWP. Die caudale Spitze des 2Ax liegt komplett unter der disto-caudalen Spitze des 1Ax.
Das gesamte 2Ax ist gleichmäßig sklerotisiert. Das von der proximalen Spitze des 2Ax ab-
gehende Basiradiale ist ebenfalls durchgehend sklerotisiert.

Das 3Ax ist mit ca. der eineinhalbfachen Länge des 1Ax auffällig groß. Distal hat es eine
etwas verbreiterte Fläche, die durch eine Einbuchtung des cranialen Randes etwas abgesetzt
ist. Der caudale Arm ist sehr lang und gerade. Die Ansatzstelle der AMD-Sehne ist nicht
deutlich hervorgehoben. Eine Muskelplatte zwischen Notum, 1Ax und 3Ax ist nicht er-
kennbar.

Die proximale und die distale Medianplatte sind durch einen schmalen, schwach sklero-
tisierten Membranstreifen getrennt. Beide Medianplatten sind nur leicht sklerotisiert.

Pleural-Region (Abb.53B)

Der Gelenkkopf des PWP ist nach dorsal deutlich erweitert. In der Aufsicht ist er schlank
längs-oval. Die Gelenkfläche ist etwas mehr als ein Drittel so lang wie das 1Ax. Am cau-
dalen Rand des PWP direkt unterhalb des Gelenkkopfes setzt ein Band an, das den PWP
mit dem ventralen Fortsatz des 2Ax verbindet.

Das Basalare (Ba) ist bis dicht unterhalb seines Kopfes mit dem Episternum verschmolzen.
Der Kopf des Basalare weist eine nach disto-cranial gerichtete Erweiterung auf, die als
Gleitfläche und Arretierung für die ventrale Subcosta-Basis dient. Cranial weist der Ba-
Kopf eine kurze, schwach sklerotisierte Platte auf, die annähernd senkrecht steht und mit
dem Flügelvorderrand verbunden ist.

Das relativ kleine Subalare liegt leicht schräg gestellt, caudal höher, unter dem PNP.

36

Lymexylonidae

Material: Hylecoetus dermestoides

Notum (Abb.54A)

Der vordere Gelenkfortsatz (ANP) ist im Verhältnis zu Notum und erstem Axillare aus-
gesprochen klein. Sein Ende bleibt deutlich hinter dem Vorderrand des Notum zurück. Der
Kontaktpunkt zwischen ANP und 1Ax liegt etwa in der Mitte des 1Ax-Kopfes. Der MNP
ist durch eine Einbuchtung des Notumseitenrandes direkt hinter dem 1Ax gekennzeichnet.
Ein eigentlicher Fortsatz ist nicht vorhanden. Der PNP ist hakenförmig und sehr kurz und
schmal. Die Achsen (a) und (b) durch die Spitze und die Basis des PNP (Abb.3) schließen
einen Winkel von ca. 5° ein.

Axillar-Region (Abb.54A)

Das erste Axillare ist an der Vorderkante ca. dreimal so breit wie im Halsbereich. Der nach
ventral umgeschlagene Vorderrand hat in der Mitte eine schmale Einkerbung; nach distal
ist er in eine kurze Spitze ausgezogen. Der hintere Teil des Kopfes liegt auf dem Gelenk-
kopf des PWP. Die Übergänge zwischen Hals und Kopf und Hals und Körper sind flie-
Bend, so daß der Hals nicht gut abgegrenzt ist. Hals und Körper liegen mit ihrem proxima-
len Rand dem Rand des ANP auf. Der Körper der 1Ax ist lang dreieckig geformt. Die
disto-craniale Kante ist leicht konvex, die caudale Kante ist deutlich konkav ausgebildet.
Der proximale Rand liegt unter dem Rand des Notum. Das Notum ist ca. 2,6 mal so lang
wie das 1Ax. Der Winkel zwischen der Achse des Gelenks zwischen Notum und 1Ax und
der disto-cranialen Kante des 1Ax-Körpers beträgt ca. 41° (Abb.3, Tab.1).

Der disto-cranialen Kante des 1Ax dicht angelagert ist das 2Ax. Der dorsal sichtbare Teil
des 2Ax ist spitz herzförmig ausgebildet. Ventral hat es einen flachen, breit dreieckigen
Fortsatz, der unter den Körper des 1Ax ragt. Ein kurzes Band verbindet einen zweiten,
zentral gelegenen ventralen Fortsatz des 2Ax mit dem PWP. Das Basiradiale ist relativ breit
und über seine gesamte Länge deutlich sklerotisiert.

Das 3Ax ist auffallend groß, allein der schlanke, gerade caudale Arm ist so lang wie das
gesamte 1Ax. Er ist über eine längere Strecke mittels derber Membran mit dem PNP ver-
bunden. Die Ansatzstelle der AMD-Sehne ist nur wenig hervorgehoben. Der distale Bereich
ist abgeflacht und erweitert sich distad deutlich. Die Muskelplatte in der Membran zwi-
schen 1Ax, Notum und 3Ax ist relativ groß. Ihre transversale Ausdehnung entspricht etwa
der Breite der Basis des caudalen Arms des 3Ax. Dementsprechend ist die Sehne zwischen
3Ax und AMD kurz und hat einen großen Durchmesser.

Distale und proximale Medianplatte sind verschmolzen und nur schwach sklerotisiert.

Pleural-Region (Abb.54B)

Eine leichte Einschnürung direkt unterhalb des Gelenkkopfes des PWP setzt diesen deutlich
gegen den Stiel des PWP ab. Die dorsale Fläche des Fulcrum ist leicht konkav. In der
Ansicht von dorsal ist erkennbar, daß es in einen massiven hinteren und einen schmalen
vorderen Bereich unterteilt ist. Die beiden Bereiche werden durch eine Einbuchtung des
proximalen Randes voneinander abgesetzt. Das Fulcrum liegt unter dem hinteren Kopf-
bereich des 1Ax. Das 1Ax ist ca. 4,5 mal so lang wie das Fulcrum.

Der größte Teil des Stiels des vor dem PWP gelegenen Basalare ist mit dem Episternum
verschmolzen. Der Kopf des Ba ist disto-cranial erweitert und bildet eine Gleitfläche und

37

Arretierung für die ventrale Subcosta-Basis. Cranial ist ein breiter, kurzer Fortsatz nach
dorsal gerichtet. Er ist mit dem basalen Flügelvorderrand verbunden.

Das Subalare ist relativ klein und stark längs oval. Es liegt unter dem PNP und ist von
diesem durch einen breiten Membranbereich getrennt.

Coccinellidae
Material: Calvia quatuordecimguttata, Coccinella septempunctata

Notum (Abb.55A, 56A)

Der ANP ist wenig langer als breit mit einer leicht konvexen cranialen Kante. Er reicht so
weit nach vorne, daß sein Ende auf gleicher Höhe mit dem Notumvorderrand liegt. Auf den
proximalen Rand des 1Ax trifft er kurz hinter dessen Vorderrand. Neben dem proximalen
Ende des 1Ax ist der Notumseitenrand tief eingebuchtet. Bei Calvia ist die Bucht so ge-
formt, daß ein deutlicher, nach caudal gerichteter, hakenförmiger Fortsatz entsteht. Bei
Coccinella reicht die Bucht nicht so weit nach vorne, so daß nur ein stumpfer Vorsprung
im Notumrand zustandekommt. Diese Struktur entspricht in Lage und Funktion dem MNP.
Der PNP ist nur als breiter, flacher Rand des Notum ausgebildet. Sein cranialer Rand ist
in zwei bis drei kurze Wellen gelegt. Der Winkel zwischen der Basis und der Spitze des
PNP in Bezug auf die Spitze des ANP beträgt ca. 9° (Abb.3, Tab.1).

Axillar-Region (Abb.55A,C,D, 56A,C,D)

Im Übergangsbereich vom Körper zum Hals des 1 Ax befindet sich im distalen und im pro-
ximalen Rand jeweils eine Einbuchtung. Die im distalen Rand liegt etwas mehr craniad als
die im proximalen Rand, so daß der basale Halsbereich gebogen erscheint. Der Hals erwei-
tert sich gleichmäßig zum Kopf hin. Dieser trägt distal einen Fortsatz, der durch eine flache
Einkerbung des Vorderrandes vom Rest des Kopfes abgesetzt ist. Dieser laterale Fortsatz
ist nicht mit dem ventralen Fortsatz des Kopfvorderrandes verbunden. Der caudale Rand
des 1Ax-Körpers ist deutlich konkav. Seine proximale Ecke ist so weit verlängert, daß sie
ein Sechstel bis ein Neuntel der Gesamtlänge des 1Ax ausmacht. Die disto-craniale Kante
des Körpers und die Gelenkachse von Notum und 1Ax schließen einen Winkel von ca. 30°
ein (Abb.3, Tab.1).

Das 2Ax ist relativ groß. Bei Coccinella ist die distale Kante deutlich konvex, bei Calvia
gerade. Das Basiradiale setzt mit einer sehr breiten Basis etwa mittig am Vorderrand des
2Ax an. Der latero-ventrale Fortsatz ragt bis jenseits der Mitte unter den Körper des 1Ax.
Der caudale Fortsatz des 2Ax ist bei beiden untersuchten Arten sehr kräftig ausgebildet.
Bei Coccinella ist er deutlich länger als bei Calvia.

Das 3Ax hat einen auffallend langen und schlanken caudalen Arm, der mit wenigstens
einem Drittel seiner Länge dem PNP anliegt. Sein caudales Ende ist bei Coccinella nahezu
rechtwinklig nach distal umgebogen. Der Ansatzpunkt der AMD-Sehne ist bei Calvia durch
eine caudal gelegene, tiefe Einbuchtung deutlich hervorgehoben. Die AMD ist verhältnis-
mäßig klein.

Im Bereich der Medianplatten befindet sich bei den untersuchten Arten nur eine einheit-
liche, leicht sklerotisierte Platte.

Pleural-Region (Abb.55B, 56B)

Das Fulcrum ist in der Ansicht von dorsal bei den untersuchten Arten annähernd so breit
wie lang. Bei Calvia ist es dreieckig, bei Coccinella viereckig geformt. Es ist ein Siebtel

38

bis ein Sechstel so lang wie das gesamte 1Ax und liegt unter dessen Hals. Da das Fulcrum
nach proximal schräg abfällt, ist die eigentliche Auflagefläche relativ schmal. Der leicht
schräg verlaufende PWP hat unterhalb des Fulcrum eine deutliche Biegung nach oben.

Das Basalare ist dem PWP dicht angelagert. Nur im oberen Drittel liegt ein schmaler Mem-
branstreifen zwischen PWP und Ba. Der gesamte Kopf des Ba ist stark erweitert. Seine
dorsale Fläche fällt leicht nach cranio-distal ab. Der frontale Fortsatz ist sehr kurz und
schräg nach vorne-oben gerichtet.

Das Subalare ist mit einem Sechstel der Notumlänge relativ kurz. Es hat eine annähernd
tropfenförmige Gestalt und ist ca. 1,5 mal höher als lang.

Das Postnotum hat einen kurzen cranialen Fortsatz, der bei keiner der untersuchten Arten
den PNP oder das Sb erreicht.

Meloidae
Material: Lytta vesicatoria

Notum (Abb.57A)

Der vordere Gelenkfortsatz ist unregelmäßig ausgebildet. An der proximalen Ecke der Vor-
derkante sitzt ein kleiner Zahn, der hintere Bereich der distalen Kante ist wellig. Im
Verhältnis zum Notum ist der ANP relativ klein, sein Vorderrand bleibt wenig hinter der
Vorderkante des Notum zurück. Die Spitze des ANP trifft im hinteren Kopfbereich auf den
proximalen Rand des 1Ax. Der MNP ist durch eine Einkerbung des Notumseitenrandes di-
rekt hinter dem 1Ax markiert. Der PNP ist hakenförmig und relativ lang. An der hinteren
Ecke der distalen Kante sitzt ein kleiner, distal weisender Zahn. Die Achsen (a) und (b)
durch die Spitze und die Basis des PNP schließen einen Winkel von ca. 13° ein (Abb.3).

Axillar-Region (Abb.57A)

Das erste Axillare ist groß und massig gebaut. Es ist fast halb so lang wie das Notum. Der
Kopf ist am vorderen Rand sehr breit. Die distale Ecke ist spitz ausgezogen. Die Vorder-
kante ist nach ventral umgeschlagen und weist in der proximalen Hälfte eine relativ breite,
flache Einbuchtung auf. Die schmalste Stelle des Halses ist etwas weniger als halb so breit
wie der Kopf. Der proximale Rand von Kopf und Hals liegt auf dem Rand des ANP. Das
Fulcrum befindet sich unter dem vorderen Halsbereich. Der Körper entspricht näherungs-
weise einem gleichschenkligen Dreieck; der caudale Rand ist deutlich konkav. Die Gelenk-
achse von Notum und 1Ax und die disto-craniale Kante des 1Ax-Körpers schließen einen
Winkel von ca. 30° ein.

Das zweite Axillare ist ebenfalls relativ groß. Es ist der disto-cranialen Kante des 1Ax
dicht angelagert. Ein großer, flacher, ventraler Fortsatz ragt weit unter den Körper des 1Ax.
Vom Zentrum dieses Fortsatzes zieht ein Band zum PWP. Die caudale Spitze des 2Ax ist
durch ein weiteres Band mit dem PNP verbunden. Das an der proximalen Spitze des 2Ax
entspringende Basiradiale ist basal relativ breit und verschmälert sich distal abrupt. Es ist
durchgehend sklerotisiert.

Das dritte Axillare ist wie 1Ax und 2Ax groß und kräftig ausgebildet. Die distale Kante
ist spitz nach caudal verlängert und leicht umgebogen. Die Ansatzstelle der AMD-Sehne
ist durch eine caudad gelegene, flache Einbuchtung hervorgehoben. Der caudale Arm ist
lang, seine proximale Kante ist gerade und durch einen schmalen Membranstreifen mit dem
PNP verbunden. Die in der Membran zwischen Notum, 1Ax und 3Ax gelegene Muskel-
platte (AMD) ist relativ groß und etwas unregelmäßig queroval ausgebildet.

39

Beide Medianplatten sind kräftig sklerotisiert und weisen eine deutliche Wellenstruktur aut.
Die zwischen 2Ax und 3Ax gelegene proximale Medianplatte (PMP) ist etwas kleiner als
die DMP und mit dieser durch eine kurze, schmale Skleritbrücke verbunden.

Pleural-Region (Abb.57B, 58A-C)

Das Fulcrum liegt unter dem vorderen Halsbereich des 1Ax, es ist durch eine Einschnürung
vom PWP abgesetzt. Die dorsale Fläche des Fulcrum ist leicht konkav, so daß ein vorderer
und ein hinterer Auflagepunkt für das 1Ax entstehen. Die Länge des Fulcrum entspricht
etwa einem Viertel der Gesamtlänge des 1Ax. An der Unterseite des Gelenkkopfes inseriert
caudal ein Band, das zum 2Ax zieht.

Der Stiel des Basalare ist basal über zwei Drittel seiner Höhe mit dem Episternum ver-
schmolzen. Der Kopf des Ba ist nach distal und frontal blasig erweitert. Distal bildet diese
Erweiterung eine schräge Fläche, die als Gleitfläche und Arretierung für die ventrale
Subcosta-Basis dient. Frontal ist von der Erweiterung eine kleine, nahezu senkrecht nach
dorsal gerichtete Platte abgeteilt, die mit der basalen Flügelvorderkante verbunden ist.

Das lang ovale, flache Subalare liegt direkt unter dem PNP, mit dem es durch einen relativ
breiten Membranstreifen verbunden ist.

Tenebrionidae

Material: Tenebrio molitor

Notum (Abb.59A)

Der ANP ist relativ klein und dreieckig. Die Spitze bleibt deutlich hinter dem Vorderrand
des Notum zurück, sie trifft im hinteren Bereich des 1Ax-Kopfes auf dessen proximalen
Rand. Im Bereich des MNP befindet sich eine kleine Einbuchtung im Notum-Seitenrand;
ein eigentlicher Fortsatz ist nicht vorhanden. Der PNP ist schmal hakenförmig, seine Spitze
reicht nicht bis in den Bereich der MNP-Einbuchtung nach vorne. Der Winkel zwischen
den Achsen (a) und (b) (Abb.3) beträgt ca. 13°.

Axillar-Region (Abb.59A)

Kopf und Hals des 1Ax sind nicht deutlich voneinander abgesetzt. Der Kopfvorderrand ist
etwa doppelt so breit wie die schmalste Stelle des Halses. Der Kopfvorderrand ist nach
ventral umgeschlagen und weist in der Mitte eine flache Einbuchtung auf, die eine ent-
sprechende Struktur der Subcosta-Basis aufnimmt. Kopf und Hals liegen mit ihrem proxi-
malen Rand auf dem distalen Rand des ANP. Der Körper des 1Ax ist relativ schmal drei-
eckig. Seine caudale Kante ist deutlich konkav, der disto-craniale Rand ist leicht konvex.
Der proximale Rand des Körpers liegt unter dem Rand des Notum. Das Notum ist etwa 2,5
mal so lang wie das 1Ax. Der Winkel zwischen der Achse durch vorderen und hinteren
Anlagepunkt des 1Ax und der disto-cranialen Kante des 1Ax-Körpers beträgt ca. 31°.

Das 2Ax ist in der vorderen Hälfte relativ breit und verschmälert sich nach caudal stark.
Es hat einen flachen, dornförmigen ventralen Fortsatz, der weit unter den Körper des 1Ax
ragt. Von einem zentral gelegenen ventralen Fortsatz des 2Ax zieht ein Band zum pleuralen
Flügelgelenkfortsatz. Das caudale Ende des 2Ax ist durch ein Band mit dem PNP verbun-
den. Das Basiradiale ist sehr schmal und durchgehend sklerotisiert.

40

Das 3Ax ist insgesamt deutlich länger als das 1Ax. Der disto-craniale und der caudale Arm
sind nahezu gleich lang. Die Ansatzstelle der AMD-Sehne ist leicht erhöht und proximad
vorgewölbt. Der caudale Arm ist keilförmig ausgebildet mit einer geraden proximalen Kan-
te, die durch einen schmalen verstärkten Membranstreifen mit dem PNP verbunden ist. In
der Membran zwischen 3Ax, 1Ax und Notum liegt ein kleines Sklerit (AMD), das durch
eine kurze Sehne mit dem 3Ax verbunden ist.

Die Medianplatten sind zu einem schwach sklerotisierten Streifen zwischen der Basis des
Radius, 2Ax und 3Ax verschmolzen.

Pleural-Region (Abb.59B)

In der Lateralansicht weist die Oberkante des Fulcrum eine deutliche vordere und hintere
Erhebung auf. In der Aufsicht ist der Gelenkkopf halbrund, die proximale Kante ist gerade,
die distale konvex. Das 1Ax ist etwa fünfmal so lang wie das Fulcrum. Direkt unter dem
Gelenkkopf entspringt ein Band, das den PWP mit dem ventralen Fortsatz des 2Ax ver-
bindet.

Der Stiel des Basalare ist über mehr als die Hälfte seiner Länge mit dem Episternum ver-
schmolzen. Der Kopf hat eine große distale Erweiterung, die leicht nach ventral gerichtet
ist. Sie bietet eine Gleitfläche und Arretierung für die ventrale Subcosta-Basis. An der
Vorderkante des Kopfes befindet sich ein nach dorsal gerichteter Stab, der mit der basalen
Flügelvorderkante verbunden ist.

Das Subalare ist als auffällig kleine Skleritplatte ausgebildet. Es liegt unter der caudalen
Hälfte des PNP, mit dem es durch derbe Membran verbunden ist.

Cerambycidae

Material: Clytus arietis, Agapanthia villosoviridescens, Rhagium mordax, Dinoptera
collaris, Strangalia melanura, Gaurotes virginea

Notum (Abb.60A, 61A)

Der ANP ist bei den untersuchten Arten im Verhaltnis zum Notum relativ klein. Bei Aga-
panthia ist er einfach dreieckig und etwa so lang wie breit. Der Vorderrand des ANP von
Clytus ist dagegen deutlich konvex, so daß der ANP insgesamt länger als breit ist. Bei
beiden Arten bleibt der Vorderrand des ANP aber deutlich hinter dem Notumvorderrand
zurück. Den proximalen Rand des 1Ax berührt der ANP jeweils im hinteren Kopfbereich.
Der MNP ist weitgehend reduziert und nur durch eine flache Einbuchtung des Notumsei-
tenrandes markiert. Der PNP ist bei beiden Arten kurz hakenförmig. Bei Agapanthia nimmt
der Grad der Sklerotisierung nach distal stark ab. Spitze und Basis des PNP schließen mit
Bezug auf die Spitze des ANP einen Winkel von 10° bzw. 13° ein (Abb.3).

Axillar-Region (Abb.60A,C,D, 61A,C,D)

Das 1Ax hat einen relativ kurzen, breiten Körper mit leicht konkavem Hinterrand und nur
wenig verlängerter proximaler Ecke. Die disto-craniale Kante des Körpers und die Gelenk-
achse zwischen Notum und 1Ax schließen einen Winkel von ca. 32° ein (Abb.3). Der Kopf
ist am cranialen Rand fast ebenso breit wie der Körper. Der Vorderrand ist nach ventral
umgeschlagen und in einen senkrecht nach unten weisenden Fortsatz ausgezogen.

Das 2Ax reicht nur soweit nach cranial wie der Berührungspunkt seiner proximalen Ecke

mit dem 1Ax. Diese Ecke ist deutlich proximo-craniad ausgezogen. Der laterale Fortsatz
des 2Ax reicht bis zur Mitte unter den Körper des: 1Ax. Das BR ist vollständig reduziert.

41

Der caudale Arm des 3Ax ist lang und schmal dreieckig. Sein proximaler Rand verläuft
gerade. Der distale Arm ist bei Agapanthia verbreitert und schräg nach vorne gerichtet. Bei
Clytus ist er deutlich schmaler und nach distal umgebogen. Die AMD ist sehr klein und
rundlich. Sie liegt proximo-craniad des Ansatzpunktes der zugehörigen Sehne am 3Ax.

Zwischen der proximalen und der distalen Medianplatte erstreckt sich ein schmales sklero-
tisiertes Band.

Pleural-Region (Abb.60B, 61B)

Das Fulcrum ist bei den untersuchten Arten stark verlängert und liegt dicht neben dem
ANP unter dem Kopf/Hals-Bereich des 1Ax. Der Kopf des Basalare wird fast vollständig
von dem Rastknopf eingenommen, der leicht nach cranio-ventral gerichtet ist. Der frontale
Fortsatz ist kurz und zeigt schräg nach oben. Das Subalare ist ausgesprochen klein. Bei
Clytus ist es direkt mit dem langen, cranialen Fortsatz der Postalarbrücke verbunden. Bei
Agapanthia ist dieser Fortsatz deutlich kürzer ausgebildet, und das Sb ist nicht direkt mit
ihm verbunden.

Chrysomelidae: Criocerinae, Chrysomelinae, Galerucinae
Material: Crioceris asparagi, Chrysomela populi, Leptinotarsa decimlineata, Agelastica alni

Notum (Abb.62A, 63A, 64A)

Der ANP ist schrag nach vorne gerichtet und uberragt den Vorderrand des Notum in der
Regel nicht. Er trifft im hinteren Kopfbereich auf den proximalen Rand des 1Ax, bei Chry-
somela ist er so weit verkürzt, daß er das 1Ax im Halsbereich berührt. Der MNP ist bei
den untersuchten Arten unterschiedlich ausgebildet. Bei Chrysomela und Crioceris ist der
Notumseitenrand im Bereich des MNP eingebuchtet. Bei Crioceris erstreckt sich die Ein-
buchtung nach vorne, so daß ein caudad gerichteter, hakenförmiger Fortsatz entsteht. Bei
Leptinotarsa ist der MNP völlig reduziert, der PNP ist ebenfalls stark rückgebildet. Am
Ende einer flachen Verbreiterung des Notum existiert jeweils nur ein kurzer, nach vorne
zeigender Fortsatz, der bei manchen Individuen von Crioceris auch ganz fehlt. Der Winkel
zwischen Basis und Außenkante des PNP beträgt zwischen 6° und 13° (Abb.3).

Axillar-Region (Abb.62A,C,D, 63A,C,D, 64A,C,D)

Das 1Ax hat einen kurzen, kompakten Körper mit einer deutlich verlängerten proximalen
Ecke. Der Kopf trägt an seinem distalen Vorderrand zwei Fortsätze, einen waagerecht nach
außen gerichteten und einen schräg nach disto-ventral weisenden. Insgesamt ist der Kopf
fast so breit wie der Körper. Der Winkel, den die Gelenkachse von 1Ax und Notum und
der disto-craniale Rand des 1Ax-Körpers einschließen, beträgt zwischen 29° und 40°.

Das 2Ax ist leicht verkürzt, so daß es nicht weiter nach vorne reicht als der Kontaktpunkt
seiner proximalen Ecke mit dem 1Ax. Der laterale Fortsatz reicht bis jenseits der Mitte
unter den Körper des 1Ax. Das Basiradiale setzt als relativ breiter Streifen etwas distad der
proximalen Ecke am 2Ax an.

Das 3Ax hat einen lang dreieckigen proximalen Arm mit gerader proximaler Kante. Der
distale Arm läuft schmal aus und ist nach caudal umgebogen. Bei Crioceris weicht das 3Ax
von dieser Form ab, da sein distaler Arm deutlich verbreitert ist und die proximale und die
distale Kante des caudalen Arms fast parallel verlaufen. Die AMD liegt jeweils als kleines,
rundliches Sklerit in der Mitte zwischen 3Ax und Notumseitenrand, bzw. dem distalen
Rand der verlängerten proximalen Kante des 1Ax-Körpers.

42

Die Medianplatten sind bei allen untersuchten Arten zu einer breiten, gebogenen Platte
verschmolzen, die sich zwischen 3Ax und 2Ax schiebt. Welche Anteile dieser Platte der
DMP und welche der PMP zuzurechnen sind, ist nicht erkennbar.

Pleural-Region (Abb.62B, 63B, 64B)

Das Fulcrum ist gegenüber dem PWP deutlich verlängert. Die dorsale Fläche ist in der Mit-
te leicht abgesenkt, so daß ein vorderer und ein hinterer Auflagepunkt für den Halsbereich
des 1Ax entstehen. Das Basalare hat einen gegenüber dem Stiel deutlich erweiterten Kopf,
der größtenteils von dem cranial gelegenen Rastknopf eingenommen wird. Der frontale
Fortsatz ist bei Leptinotarsa und Chrysomela relativ kurz, bei Crioceris lang und breit
ausgebildet. Das Subalare ist bei den untersuchten Arten durchweg sehr klein. Bei einigen
Individuen von Crioceris fehlt es völlig. Das Postnotum aller Arten hat einen langen crani-
alen Fortsatz, der dem PNP dicht anliegt.

Chrysomelidae: Hispinae

Material: Cassida sp.

Notum (Abb.65A)

Der ANP ist ca. 1,5 mal länger als breit. Sein Vorderrand bleibt hinter dem Vorderrand des
Notum zurück. Das Ende des ANP trifft in der Mitte der proximalen Kante auf den Kopf
des 1Ax. Craniad des MNP hat der Notumseitenrand eine schmale, lang nach vorne gezo-
gene Einbuchtung. Aus der Form und Lage der Einbuchtung resultiert ein neben dem Ende
des 1Ax liegender, nach hinten gerichteter hakenförmiger Fortsatz. Der PNP ist als breit
ausgezogene, flache Vorwölbung des Notumseitenrandes ausgebildet. Der distale Rand des
PNP und seine Basis, abgegrenzt durch den Übergang zum MNP, schließen mit Bezug auf
die Spitze des ANP einen Winkel von ca. 5° ein (Abb.3, Tab.1).

Axillar-Region (Abb.65A,C,D)

Der Hals des 1Ax ist stark eingeschnürt. Der Kopf ist am Vorderrand fast so breit wie der
Körper. Seine äußere Ecke trägt einen distad weisenden Fortsatz, der nicht mit dem schräg
nach unten gerichteten Fortsatz verbunden ist. Der Körper ist sehr schmal. Die Gelenkachse
zwischen Notum und 1Ax und die disto-craniale Kante des Körpers schließen einen Winkel
von ca. 33° ein (Abb.3). Die proximale Ecke des 1Ax-Körpers ist gegenüber der distalen
deutlich caudad verlängert.

Der laterale Fortsatz des 2Ax reicht bis deutlich jenseits der Mitte unter den Körper des
1Ax. Das BR ist auf ganzer Lange gleich breit und inseriert an der proximalen Spitze des
2Ax.

Der caudale Arm des 3Ax ist dreieckig und auffallend lang. Seine proximale Kante verläuft
gerade und liegt über eine kurze Strecke dem PNP an. Der distale Arm ist relativ schmal
und leicht gebogen. Die AMD liegt dicht neben dem Ansatzpunkt der zugehörigen Sehne
am 3Ax und ist sehr klein.

Die Medianplatten sind zu einer gebogenen und leicht sklerotisierten Platte verschmolzen,
die sich zwischen 3Ax und 2Ax schiebt. Welcher Bereich dieser Platte auf die DMP und
welcher auf die PMP zurückgeht, ist nicht erkennbar.

43

Pleural-Region (Abb.65B)

Das Fulcrum ist in der Aufsicht ausgesprochen lang und nur leicht verbreitert. Es liegt
unter dem Kopfbereich des 1Ax. Der PWP ist stark nach vorne gekippt und hat kurz unter-
halb des Fulcrum einen deutlichen Knick nach dorsal.

Der Stiel des BA ist sehr schlank und größtenteils mit dem Episternum verschmolzen. ‚Der
Rastknopf liegt als leichte Vorwölbung im vorderen Bereich des Kopfes. Der frontale Fort-
satz ist breit und kurz. Dabei weist er senkrecht nach oben.

Das Subalare ist sehr klein. Es ist etwa so lang wie breit und steht über einen schwach
sklerotisierten Membranstreifen mit dem Postnotum in Verbindung.

Curculionidae

Material: Phyllobius sp. 1, Phyllobius sp. 2, Chlorophanus sp., Furcipus rectirostris,
Otiorhynchus sp.

Notum (Abb.66A, 67, 68A)

Der ANP ist im Verhältnis zum Notum relativ klein und bleibt deutlich hinter dem Vor-
derrand des Notum zurück. Er trifft im vorderen Hals- bzw. im hinteren Kopfbereich auf
den proximalen Rand des 1Ax. Der MNP ist bei allen untersuchten Arten vollständig redu-
ziert. Der PNP ist kurz und spitz hakenförmig ausgebildet. Zum Ende des PNP hin nimmt
der Grad seiner Sklerotisierung deutlich ab. Basis und Spitze des PNP schließen mit Bezug
auf die Spitze des ANP einen Winkel von 8° bis 11° ein (Abb.3).

Axillar-Region (Abb.66A,C,D, 67, 68A,C,D)

Das 1Ax hat einen stark verbreiterten Kopf, der genauso breit ist wie der Körper des 1 Ax.
Sein Vorderrand weist neben dem schräg nach unten gerichteten Fortsatz einen weiteren,
waagerecht nach distal gerichteten Vorsprung auf. Die proximale Ecke des 1Ax-Körpers
ist gegenüber der distalen leicht verlängert. Bei den zur Untersuchung verfügbaren Indi-
viduen von Phyllobius sp. 2 sind die Alae stark rückgebildet. Obwohl die Flügel in ange-
legtem Zustand nur bis zum Ende des dritten Abdominaltergites reichen, sind alle Elemente
der Flügelbasis vorhanden. Nur das 1Ax ist unbeweglich mit dem Notum verschmolzen.
Dieser Zustand entspricht den Verhältnissen geringster Reduktion, wie Geisthardt (1974)
sie bei den flugunfähigen Weibchen von Lamprohiza (Lampyridae) vorfand.

Die proximale Ecke des 2Ax ist deutlich proximo-craniad ausgezogen. Das Basiradiale setzt
dementsprechend nicht exakt an der proximalen Kante des 2Ax an. Es ist auch bei Phyllo-
bius sp. 2 als schmales, durchgehend sklerotisiertes Band ausgebildet. Der ventro-laterale
Fortsatz des 2Ax ragt nicht ganz bis zur Mitte unter den Körper des 1Ax.

Das 3Ax hat einen langen distalen Arm, der bei Phyllobius sp. 1 und Chlorophanus deut-
lich caudad gebogen ist. Die proximale Kante des caudalen Arms ist annähernd gerade und
hat über eine längere Strecke Kontakt mit dem PNP.

Die Medianplatten sind schwach sklerotisiert, aber als breites, bogenförmiges Element zwi-
schen 3Ax und 2Ax identifizierbar. Es ist nicht erkennbar, welcher Bereich dieses Elements
auf die DMP und welcher auf die PMP zurückzuführen ist.

Pleural-Region (Abb.66B, 68B)

Das Fulcrum ist deutlich verbreitert und hat ein Fünftel bis ein Viertel der Lange des 1Ax,
unter dessen Kopfbereich es liegt. Das Basalare ist größtenteils nahtlos mit dem Episternum

44

verschmolzen. Nur von der dorsalen Kante her schiebt sich ein kurzer Membrankeil zwi-
schen Ba und PWP. Der Ba-Kopf trägt im vorderen dorsalen Bereich einen relativ kleinen
Rastknopf. Der kurze, breite, frontale Fortsatz weist schräg nach dorsal. Das Subalare ist
sehr klein. Bei einzelnen Individuen fehlt es ganz. Die Postalarbrücke hat unterhalb des
PNP einen kurzen cranialen Fortsatz, der frei endet, ohne direkten Kontakt zum Sb oder
zum PNP zu haben.

Neuropterida

Für die Neuropterida wurden drei Arten der Megaloptera, zwei Arten der Raphidioptera
und vier Arten der Planipennia untersucht. |

Bei allen untersuchten Vertretern der Neuropterida befindet sich zwischen dem hinteren
Bereich des Seitenrandes des Notum und dem 3Ax ein weiteres, als viertes Axillare be-
zeichnetes Sklerit. Das 1Ax zeichnet sich durch einen sehr kurzen, breiten Körper aus. Das
Fulcrum liegt bei allen untersuchten Arten zumindest teilweise unter dem Kopf-Hals-
Bereich des ersten Axillare. Zweites und drittes Axillare sind relativ schwach sklerotisiert.
Das 3Ax ist in mehrere Sklerite aufgelöst, von denen mit Ausnahme eines stabförmigen
Elements alle reduziert sein können.

Planipennia

Material: Osmylus fluvicephalus, Chrysotropia ciliata, Chrysopa perla, Cueta beieri

Notum (Abb.69A)

Der ANP ist teilweise als flach gegen das aufgewölbte Notum abgesetzter Fortsatz aus-
gebildet. Bei Cueta ist er völlig reduziert, seine Funktion als Anlagepunkt des 1Ax wird
direkt vom Notumrand übernommen. Der ANP überragt niemals den Vorderrand des
Notum. Ein MNP ist bei keinem der untersuchten Tiere erkennbar. Der PNP ist lang,
schlank und in der Regel leicht nach vorne gebogen. Er ist als 4Ax durch einen nicht
sklerotisierten Membranstreifen vom Notum getrennt. Der Winkel ß zwischen den Achsen
(a) und (b) beträgt bei Cueta ca. 32° (Abb.3).

Axillar-Region (Abb.69A,B)

Das 1Ax aller untersuchten Planipennia hat einen sehr kurzen und breiten Körper. Dessen
proximales Ende liegt unter dem Rand des Notum, mit dem es beweglich verbunden ist.
Der Hals kann relativ gerade und schlank ausgebildet sein, wie bei Chrysotropia und
Osmylus. Teilweise trägt er einen distalen Fortsatz, wie z.B. bei Cueta und Nemoptera. Ist
ein solcher Fortsatz vorhanden, so liegt dieser auf dem Fulcrum. Fehlt dieser Fortsatz, dann
befindet sich das Fulcrum unter dem schmalen (Chrysotropia) oder leicht verbreiterten
(Osmylus) Kopf des 1Ax. Der Hals setzt im proximalen Drittel des Körpers an, wobei er
sich leicht verbreitert. Der Kopf kann am Vorderrand einen nach distal und ventral gerich-
teten Fortsatz tragen (Cueta), oder er ist einfach abgerundet und nur leicht verbreitert
(Osmylus, Chrysotropia). Der Winkel & zwischen der disto-cranialen Kante des 1Ax-Kör-
pers und der Gelenkachse von Notum und 1Ax beträgt ca. 70° (Abb.3).

Das 2Ax ist bei den Vertretern der Planipennia durchweg verhältnismäßig klein und teil-
weise nur leicht sklerotisiert. Bei Cueta ist es plattenförmig und ragt in die Bucht zwischen
den distalen Fortsätzen von Kopf und Hals des 1Ax. Von der vorderen distalen Kante ent-
springt das schmale und kurze Basiradiale. Distal geht das 2Ax in leichte, großflächige

45

Sklerotisierungen der Flügelmembran über, die wahrscheinlich Reste der Medianplatten
sind. Die anderen untersuchten Planipennia besitzen ein leicht gebogenes, streifenförmiges
2Ax, das in der Bucht zwischen Kopf und Körper des 1Ax liegt. Das BR bildet hier wahr-
scheinlich den vorderen Teil der als 2Ax anzusprechenden Struktur, ist aber nicht deutlich
vom eigentlichen 2Ax abgesetzt. Gemeinsam ist allen Varianten des 2Ax ein sehr langer,
an der hinteren ventralen Kante ansetzender, stabförmiger Fortsatz, an dem ein Band inse-
riert, das zum Subalare führt. Der Fortsatz erstreckt sich über nahezu zwei Drittel der
Distanz zwischen 2Ax und Sb. Ein weiteres Band verbindet das 2Ax mit dem PWP kurz
unterhalb des Fulcrum.

Das 3Ax besteht bei allen untersuchten Arten aus einem relativ langen, schmalen Stab, der
caudal leicht gegabelt ist. In dieser Gabel liegt das disto-craniale Ende des 4Ax. Das
vordere Ende des 3Ax liegt hinter oder distal neben dem distalen Ende des 1Ax-Körpers.
Distal davon liegt in der Regel ein kleines, unregelmäßig geformtes Sklerit. Die Muskulatur
des 3Ax inseriert breit auf dem proximalen Rand.

Die Medianplatten sind bei den untersuchten Arten variabel ausgebildet. Allgemein sind
sie nur leicht sklerotisiert und unregelmäßig geformt. PMP und DMP sind in der Regel
nicht deutlich voneinander getrennt.

Pleural-Region (Abb.70, 71, 72)

Der Gelenkkopf des PWP (Fulcrum) liegt bei allen untersuchten Arten unter dem 1Ax (s.
o.). Das Fulcrum ist gegenüber dem PWP leicht erweitert, kurz oval oder rund, mit sehr
kleinem Durchmesser. Bei Chrysotropia ist der PWP unterhalb des Fulcrum stark nach
vorne gebogen, so daß der Gelenkkopf unter dem Kopf des 1Ax liegt. Vom Hinterrand des
PWP direkt unterhalb des Fulcrum zieht ein Band zum 2Ax.

Das Basalare ist dem PWP dicht angelagert und bis auf ein kurzes Stück des dorsalen Ran-
des mit diesem verschmolzen. Der hintere dorsale Rand des Ba trägt eine kleine knopf-
förmige Vorwölbung, hinter der bei angelegtem Flügel eine entsprechende Struktur der ven-
tralen Basis der Subcosta einrastet. Bei Cueta ist der basale Bereich des Ba nicht sklero-
tisiert, so daß eine spangenförmige Struktur entsteht.

Das Subalare ist durchweg sehr lang oval und mit ca. einem Drittel der Notumlänge auffal-
lend groß. Innen verläuft im oberen Drittel oder in der Mitte über die gesamte Länge des
Sb ein flacher Grat, der die Muskelansatzflächen voneinander trennt.

Megaloptera

Material: Sialis lutaria, Chauliodes rastricornis, Corydalus cornutus

Notum (Abb.73, 74, 76A, 77)

Bei allen untersuchten Arten ist der ANP als flacher Fortsatz gegen das aufgewölbte Notum
abgesetzt. Er trifft direkt hinter dem Vorderrand des 1Ax-Kopfes auf dessen proximale
Kante. Bei Sialis hat er etwa die Form eines gleichschenkligen Dreiecks, bei Corydalus und
Chauliodes ist er kurz und breit, so daß seine kürzeste Seite dem 1Ax zugewandt ist. Der
ANP überragt nie den Vorderrand des Notum. Bei Sialis und Chauliodes ist keine Struktur
erkennbar, die sich als MNP deuten läßt. Corydalus hat exakt an der Stelle, an der der
proximale Rand des Körpers des 1Ax unter dem Notum liegt, eine minimale Vorwölbung,
die möglicherweise den MNP repräsentiert. Bei den drei untersuchten Arten ist der PNP
als kurzes, näherungsweise dreieckiges oder quadratisches 4Ax vom Notum getrennt. Bei

46

Sialis und Chauliodes existiert am Notum jeweils noch ein breiter Fortsatz, an dessen Ende
das 4Ax liegt. Dieser Fortsatz ist die ursprüngliche Basis des PNP. Bei Corydalus fehlt ein
vergleichbarer Fortsatz und der Notumrand proximal des 4Ax ist annähernd gerade. Der
Winkel zwischen den Achsen (a) und (b) durch die Spitze und die Basis des 4Ax beträgt
bei Sialis und Chauliodes ca. 23°, bei Corydalus ca. 13° (Abb.3, Tab.1).

Axillar-Region (Abb.73, 74, 76A, 77)

Das 1Ax tritt bei den Megaloptera in zwei deutlich unterschiedlichen Formen auf. Bei
Corydalus und Chauliodes ist der Körper des 1Ax relativ großflächig sklerotisiert. Die
distale und die caudale Kante schließen einen Winkel von ca. 45° ein. Im proximo-cauda-
len Bereich befindet sich eine relativ große Aussparung, so daß die Strecke, mit der der
Körper des 1Ax dem Notum anliegt, sehr kurz ist. Der Winkel, den die Gelenkachse zwi-
schen 1Ax und Notum und der disto-craniale Rand des 1Ax-Körpers aufspannen (Abb.3),
ist bei allen untersuchten Arten größer als 50° (siehe Tab.1). Der Sklerithals entspringt
mittig (Corydalus) bis distal (Chauliodes) am Vorderrand des Körpers. Bei Chauliodes ist
er in eine enge Schleife gelegt. Er erweitert sich gleichmäßig nach vorne und geht fließend
in den Kopf über. Der Kopfvorderrand ist mehr oder weniger gerade abgestutzt. Bei Sialis
hingegen ist der Körper des 1Ax extrem kurz und breit. Seine proximo-caudale Ecke ist
in eine kurze, nach hinten weisende Spitze ausgezogen. Der Hals setzt sehr schmal im pro-
ximalen Drittel des Körpers an und erweitert sich gleichmäßig zum Kopf hin. Dessen Vor-
derrand ist breit abgerundet. Seitlich trägt er im ventralen Bereich eine leichte Erweiterung.

Das 2Ax ist bei Corydalus und Chauliodes wiederum sehr ähnlich ausgebildet. Es ist eine
große, einfach dreieckige Platte, die der disto-cranialen Kante des 1Ax-Körpers direkt
anliegt. Mittig (Corydalus) oder proximal (Chauliodes) an der Vorderkante des 2Ax setzt
das Basiradiale an, das fast die gesamte Breite des 2Ax einnimmt. Der Übergang vom 2Ax
zum BR ist fließend, es ist keinerlei Naht erkennbar. Cranial verschmilzt das BR mit der
Basis der Subcosta. Sialis hat ein ebenfalls sehr großflächiges 2Ax, dessen Form aber eher
viereckig mit sehr großzügig gerundeten Ecken ist. Auch das BR ist ausgesprochen breit
und setzt in der Mitte des Vorderrandes des 2Ax an. Cranial endet es breit gerundet, von
Membran umgeben. Es ist nicht mit der Basis der Sc verschmolzen. Allen drei Arten
gemeinsam ist ein ventro-caudaler Fortsatz des 2Ax, an dem ein Band ansetzt, das zur
vorderen oberen Kante des Subalare zieht. Der Fortsatz ist so lang, daß er fast ein Drittel
der Distanz zwischen 2Ax und Sb überbrückt. Bei Sialis ist die Basis des Fortsatzes noch
am dorsalen Hinterrand des 2Ax erkennbar.

Bei Corydalus liegen im Bereich des 3Ax drei relativ große und zwei sehr kleine Sklerite,
die offensichtlich Reste des 3Ax sind und zusammen seine Funktionen erfüllen. An einem
großen, schräg von disto-cranial nach proximo-caudal orientierten Sklerit und einem schma-
len, wesentlich kleineren Sklerit, das der proximo-cranialen Kante des großen Sklerits
angelagert ist, inserieren die Muskeln des 3Ax. Der Kontakt zwischen diesem muskel-
tragenden Sklerit und dem 4Ax wird von einem weiteren relativ großen, ovalen Element
vermittelt. Zwischen den Analadern und dem Muskelelement liegt das dritte große Sklerit,
das relativ gerade und stabförmig geformt ist. Das zweite kleine Element ist in den distalen
Spalt zwischen dem Muskelsklerit und dem Verbindungselement zu den Analadern
eingefügt. Das 3Ax von Chauliodes besteht aus einem grob halbmondförmigen Sklerit, das
an seiner langen pro-ximo-cranialen Kante eine kurzen Fortsatz trägt. Die Muskulatur
inseriert zwischen diesem Fortsatz und der hinteren, ebenfalls nach cranial gebogenen
Spitze. Sialis hat ein annähernd dreieckiges 3Ax, das im Verhältnis zu 1Ax und 2Ax etwas

47

kleiner ist als bei den beiden anderen Arten. Die distale Spitze des Dreiecks geht direkt in
eine Analader über. Die Muskeln setzen im vorderen Bereich der proximalen Kante des
3Ax an.

Der Bereich der Medianplatten ist bei allen untersuchten Arten schwach sklerotisiert. PMP
und DMP sind nicht identifizierbar. Bei Corydalus und Sialis liegen relativ großflächige,
bogenförmige Sklerotisierungen vor, die wahrscheinlich auf die Medianplatten zurückgehen.

Pleural-Region (Abb.74, 75, 76B, 78)

Der Gelenkkopf des PWP liegt bei allen untersuchten Arten jeweils zu ca. einem Drittel
unter dem Kopf/Hals-Bereich des 1Ax. Die übrige Gelenkfläche liegt unter dem angren-
zenden Bereich des 2Ax. Im Verhältnis zum PWP ist der Gelenkkopf relativ stark erwei-
tert: bei Sialis in einen fast kugelförmigen Kopf, der einen schmalen caudalen Anhang
trägt, bei Corydalus und Chauliodes in einen langgestreckten Kopf mit einer proximalen
Erweiterung. Vom Hinterrand des PWP knapp unterhalb des Fulcrum verläuft bei den un-
tersuchten Arten jeweils ein Band zum 2Ax. Bei Sialis ist dieses Band zusätzlich sklero-
tisiert.

Das Basalare ist bei allen drei Arten dorsal erweitert. Es ist fast vollständig mit dem
Episternum verschmolzen. Die Abgrenzung zum Episternum ist durch eine Naht äußerlich
erkennbar. Im hinteren Bereich des Dorsalrandes befindet sich ein ellipsoider Fortsatz, der
bei angelegtem Flügel in eine von der Subcosta-Basis und der Humeralplatte gebildete
Aussparung einrastet.

Das Subalare ist relativ einheitlich lang oval ausgebildet. Es hat bei allen untersuchten
Arten etwa ein Drittel der Notumlänge.

Raphidioptera
Material: Agulla adnixa, Raphidia ophiopsis

Notum (Abb.79)

Der ANP der untersuchten Raphidiopteren ist dem von Sialis und dem der Coleoptera sehr
ähnlich. Es handelt sich um einen flachen Fortsatz von näherungsweise gleichschenklig
dreieckiger Form, der den Vorderrand des Notum nicht überragt. Sein Ende trifft kurz hin-
ter dem Vorderrand des 1Ax-Kopfes auf dessen proximale Kante. Ein MNP ist nicht
erkennbar. Kurz vor dem 4Ax ist der Notumseitenrand leicht vorgewölbt. Der PNP ist als
kurzes, dreieckiges 4Ax vom Notum abgetrennt. Der Winkel zwischen den Achsen (a) und
(b) (Abb.3) beträgt ca. 11° (Tab.1).

Axillar-Region (Abb.79)

Der Körper des 1Ax ist kurz und breit mit einem nahezu gerade verlaufenden Hinterrand.
Sein proximaler Rand liegt unter dem Notumseitenrand. Der disto-craniale Rand des
Körpers und die Gelenkachse zwischen Notum und 1Ax (Abb.3) spannen bei beiden unter-
suchten Arten einen Winkel von mehr als 70° auf (Tab.1). Im proximalen Drittel des
Körpers setzt der verhältnismäßig schmale Hals an. Der Kopf ist nach distal ausgezogen
und ca. doppelt so breit wie der Hals. Die Subcosta-Basis hat einen kurzen Vorsprung, der
bei geöffnetem Flügel auf dem Vorderrand des 1Ax-Kopfes liegt.

Die Form des 2Ax ähnelt stark der Ausbildung der zweiten Axillaria, wie sie bei den Co-
leoptera vorkommen (vergl. z.B. Abb.8 etc.). Die Grundform ist annähernd dreieckig,

48

wobei eine Ecke dem 1Ax zugewandt ist. Die distale Kante ist konvex. Es ist ein langer,
ventral gelegener caudaler Fortsatz vorhanden, der über ein Band mit dem Subalare ver-
bunden ist. Die Basis dieses Fortsatzes ist von dorsal erkennbar. Ein weiteres Band ver-
bindet das 2Ax mit dem Hinterrand des PWP. Am Vorderrand der proximalen Spitze des
2Ax inseriert mit relativ breiter Basis das Basiradiale.

Im Bereich des 3Ax sind drei Sklerite erkennbar. Ein stabförmiges Element liegt dem 4Ax
an und verläuft schräg nach disto-cranial. Die beiden kurzen Seiten sind flach eingekerbt.
Die vordere lange Seite trägt mittig einen schmalen craniad gerichteten Fortsatz. Proximal
dieses Fortsatzes liegt das Ende des proximalen Astes des vorderen Sklerits. Dieses ist
umgekehrt V-förmig ausgebildet. Am proximalen, vorderen Bereich dieses Astes inseriert
die Muskulatur des 3Ax. Neben dem distalen Ast liegt das dritte Sklerit. Es ist schief
viereckig mit einem langen, schmalen Fortsatz an der vorderen proximalen Ecke. An der
gegenüberliegenden distalen Ecke schließen sich die Analadern an. Bei Agulla sind das
vordere und das distale der drei Sklerite deutlich schwächer sklerotisiert als das hintere,
dem 4Ax benachbarte Sklerit.

Bei beiden untersuchten Arten sind im Bereich der Medianplatten nur kleine, schwach skle-
rotisierte Flächen vorhanden.

Pleural-Region (Abb.80)

Das Fulcrum ist gegenüber dem PWP deutlich cranio-caudal verlängert und leicht verbrei-
tert. Es liegt zu etwa gleichen Teilen unter dem Kopf des 1Ax und unter der proximalen
Ecke des 2Ax. Vom ventralen Hinterrand des Fulcrum verläuft ein Band zum 2Ax.

Das Basalare ist großflächig mit dem Episternum verschmolzen, zum PWP ist es durch eine
Naht abgegrenzt. Diese Naht ist gleichzeitig die Achse, um die sich das Ba bei Kontraktion
seiner Muskulatur dreht. Der hintere Bereich des Dorsalrandes des Ba trägt einen etwa
halbkugeligen Fortsatz. Als Gegenstück zu diesem Fortsatz bilden die Humeralplatte und
die Subcosta-Basis eine Klammerstruktur, die bei angelegtem Flügel auf dem Knopf des
Ba einrastet und so den Flügel fixiert.

Das Subalare ist etwas mehr als ein Drittel so lang wie das Notum. Es wird durch eine
Längsnaht in eine obere und eine untere Hälfte geteilt.

DISKUSSION

Es wird zunächst das Grundmuster des Flügelgelenks der Neoptera rekonstruiert, damit es
für den Vergleich mit den bei den Holometabola vorgefundenen Strukturen zur Verfügung
steht. Würde man nur Vertreter rezenter Taxa als Außengruppe zum Vergleich heranziehen,
liefe man Gefahr, deren Autapomorphien in der Diskussion erhebliches Gewicht beizu-
messen.

Das Grundmuster der Neoptera

Vergleicht man die in den verschiedenen Gruppen der hemimetabolen Neoptera verwirk-
lichten Ausbildungen der Flügelbasis, so ergibt sich für das Notum die Existenz von drei
lateralen Fortsätzen, wie sie in Abb.1 dargestellt sind. Die beiden vorderen Fortsätze, der
ANP und der MNP, waren wahrscheinlich kurz, der hintere Fortsatz (PNP) dagegen ver-
hältnismäßig lang ausgezogen. Diese Ausbildung findet sich bei Vertretern der Plecoptera
(Snodgrass 1909, 1927, Onesto 1965, Brodskiy 1979a, 1979b), Mantodea (Onesto 1960),

49

Blattodea (Onesto 1959), Dermaptera (Snodgrass 1909, Onesto 1961, Kleinow 1966), Salta-
toria (Snodgrass 1909, Onesto 1963, Wootton 1979, Brodskiy 1987), Psocoptera (Brodskiy
1992), Heteroptera (Betts 1986) und Homoptera (Emeljanov 1977, Brodskiy 1992). Inner-
halb der einzelnen Taxa kommt es zu verschiedenen Umbildungen der Notalfortsätze. So
kann der PNP, oder ein Teil des PNP, vom Notum getrennt werden, so daß ein sogenanntes
viertes Axillare (4Ax) entsteht. Dies ist z.B. von Plecoptera (Brodskiy 1979b, 1994) und
Saltatoria (Snodgras 1909) bekannt. In der Regel tritt ein 4Ax nur am Metanotum auf,
während das Mesonotum den normalen PNP beibehält. Für das Grundmuster der Neoptera
ist ein 4Ax nicht anzunehmen. Eine weitere relativ häufige Umbildung ist die teilweise
oder vollständige Reduktion des MNP (Brodskiy 1994, 1992). In besonders ausgeprägter
Form ist dies bei den Hemiptera zu finden (Betts 1986, Brodskiy 1992).

Prae- und Postalarbrücken treten in allen Taxa der Pterygota auf (Snodgrass 1909, 1927,
Brodskiy 1988, 1992, 1994), so daß ihre Existenz auch für die Stammart der Neoptera
angenommen werden kann.

Das 1Ax des Grundmusters der Neoptera war wahrscheinlich dem der rezenten Plecoptera
sehr ähnlich (Onesto 1965, Brodskiy 1979b). Ursprünglich ist ein nur in der dorsalen
Membran sklerotisiertes Element anzunehmen, das einen langgestreckten, schmalen vor-
deren Bereich (Kopf/Hals) und einen verbreiterten hinteren Bereich (Körper) aufweist
(Abb.1). Das Vorderende ist beweglich mit der Basis der Subcosta verbunden. Es bildet
weder einen deutlichen Kopf aus, noch trägt sein Vorderrand ausgeprägte Fortsätze. Der
Skleritkörper ist asymmetrisch ausgebildet, mit einem dem Notum anliegenden, relativ
langen caudalen Fortsatz und einer bauchigen bis dreieckigen distalen Erweiterung. Der
Übergangsbereich zwischen Kopf/Hals und Körper des 1Ax grenzt an den ANP, der mitt-
lere Bereich des 1Ax-Körpers berührt den MNP. Die Verbindung zum Notum ist gelenkig.
Die distale Kante vom Vorderende bis zur Spitze der distalen Erweiterung des Körpers
bildet eine flache Bucht, in der das 2Ax liegt. Die Verbindung zum 2Ax ist membranös
und somit potentiell beweglich. Durch die gebogene Form des Kontaktbereichs ist die
Beweglichkeit aber stark eingeschränkt. Erste Axillaria in einer Ausbildung, die der für das
Grundmuster der Neoptera hypothetisierten Form sehr ähnlich ist, finden sich neben den
Plecoptera (Onesto 1965, Brodskiy 1979a, 1979b) bei Blattodea (Onesto 1959), Mantodea
(Onesto 1960), Dermaptera (Onesto 1961, Kleinow 1966) und Saltatoria (Snodgrass 1909,
Onesto 1963, Brodskiy 1987). Das 1Ax ist im Grundmuster nicht mit Muskulatur versehen.
Abwandlungen der Grundform finden sich z.B. bei den Caelifera (Saltatoria) (Snodgrass 1909), etwa bei
Locusta migratoria, bei der in Vorder- und Hinterflügel der proximale, hintere Fortsatz des 1Ax-Körpers bis
auf einen kleinen Vorsprung reduziert ist und in deren Hinterflügelgelenk der gesamte Kopf/Hals-Bereich
des 1Ax fehlt (Wootton 1979). Eine noch stärkere Reduktion des 1Ax ist bei den Hemiptera zu finden, in
deren Flügelgelenken das 1Ax nur als kleines, unregelmäßig geformtes Sklerit erhalten ist (Snodgrass 1909,
Emeljanov 1977, Betts 1986).

Das 2Ax der Stammart der Neoptera muß ein sowohl in der dorsalen als auch in der ven-
tralen Flügelmembran sklerotisiertes Element gewesen sein, das in der Ansicht von dorsal
halbkreisförmig ausgebildet ist und eine leicht vorgezogene caudale Ecke aufweist (Abb.1).
Die breit abgerundete Seite ist dem 1Ax zugekehrt. Nach vorne steht das 2Ax mit der Basis
des Radius (= Basiradiale (Onesto 1965)) in Kontakt. Die Verbindung wird über einen
schmalen Membranstreifen hergestellt und ist beweglich. An der distalen Kante schließt
sich die proximale Medianplatte an, die ebenfalls durch einen Membranstreifen beweglich
mit dem 2Ax verbunden ist. Der caudalen Spitze des 2Ax entspringt ein Band, das zum
dorso-cranialen Rand des Subalare zieht. Die ventrale Fläche des 2Ax liegt auf dem

50

Fulcrum. Ein Band verstärkter Membran spannt sich zwischen dem PWP und dem 2Ax und
schränkt dessen Beweglichkeit so weit ein, daß sein Abheben vom Fulcrum während des
Flügelschlages weitgehend verhindert wird. Muskulatur setzt am 2Ax nicht an. Zweite
Axillaria, die dieser Ausbildung relativ nahe kommen, finden sich bei Plecoptera (Snod-
grass 1909, Onesto 1965, Brodskiy 1979b), Blattodea (Onesto 1959), Mantodea (Onesto
1960), Dermaptera (Onesto 1961, Kleinow 1966) und Saltatoria (Onesto 1963, Wootton
1979):

Bei den Hemiptera ist das 2Ax im Zuge einer Umbildung des gesamten Flügelgelenkes stark vergrößert
(Snodgrass 1909, Emeljanov 1977, Betts 1986). Bei einigen Plecoptera, Blattodea, Saltatoria und Dermaptera
kommt es zu einer Verschmelzung von BR und 2Ax (Onesto 1961, 1963, Kleinow 1966, Wootton 1979,

Brodskiy 1979, 1987). Bei Vertretern der Blattodea wird dabei das Fulcrum vom eigentlichen 2Ax unter das
stark sklerotisierte BR (Brodskiy 1979b) verlagert.

Das 3Ax der Stammart der Neoptera weist - wie das 2Ax - Sklerotisierungen in der dor-
salen und in der ventralen Membran auf. Es liegt zwischen dem PNP und der caudalen
Ecke des 2Ax. Das dem PNP zugewandte Ende ist schmal, nicht breiter als das Ende des
PNP. Nach vorne verbreitert sich das 3Ax deutlich (Abb.1). Die proximale Kante trägt die
Ansatzstelle für wahrscheinlich einen Muskel, der vermutlich zum Episternum zieht. Aus
einem solchen Muskel sind wahrscheinlich durch Aufspaltung die bei den rezenten Neo-
ptera i.d.R. vorhandenen zwei Muskeln, von denen einer zur Pleuralleiste und einer zum
Episternum zieht, hervorgegangen. Die Kontraktion dieser Muskeln bewirkt das Einfalten
des Flügels. An die vordere Kante des 3Ax schließt die proximale Medianplatte an, an die
distale Kante die Basis der Analadern. Entlang der proximalen Kante des 3Ax und der
PMP verläuft von der Spitze des PNP zur cranialen Spitze der PMP eine Falte des Flügel-
gelenks. Um diese Achse dreht sich das 3Ax mit der anhängenden PMP beim Einfalten des
Flügels. Dabei wird das Analfeld unter das Remigium geschlagen.

Das 3Ax ist in allen Taxa starken Variationen unterworfen, so daß eine genaue Rekonstruk-
tion der ursprünglichen Ausbildung schwierig ist. Sehr häufig tritt jedoch eine zumindest
angedeutete Gabelung des disto-cranialen Randes auf, was auf die Existenz einer ähnlichen
Struktur im Grundmuster hindeuten könnte. Sie findet sich bei einigen Vertretern der Pleco-
ptera (Snodgrass 1909, Onesto 1965, Brodskiy 1979a, 1979b), Blattodea (Onesto 1959),
Saltatoria (Onesto 1963) und Hemiptera (Emeljanov 1977, Betts 1986).

Innerhalb der Saltatoria kommt es zu Umbildungen, bei denen teilweise der Ansatzpunkt der

Muskulatur vom 3Ax abgetrennt wird und als eigenständiges Sklerit in der Membran liegt (Brodskiy

1987). Ähnliche, wahrscheinlich konvergent entstandene Strukturen sind auch bei den Dermaptera zu
beobachten (Onesto 1961, Kleinow 1966).

Die Medianplatten sind Sklerotisierungen der dorsalen Flügelmembran. Die Aufteilung die-
ser ansonsten einheitlichen Sklerotisierung in die proximale und die distale Medianplatte
kommt durch eine zwischen der PMP und der DMP liegende Falte des Flügelgelenks
zustande (Abb.1). Die PMP liegt distal des 2Ax und cranial des 3Ax. Proximal und distal
wird sie von zwei Falten des Flügelgelenks begrenzt (s.o.). Ihre Verbindung zum 3Ax ist
relativ stabil, und beim Einfalten des Flügels macht sie die Drehung des 3Ax mit. Die
DMP liegt distal der PMP, sie geht wahrscheinlich auf die Basen von Media und Cubitus
zurück, die hier ihren Ursprung haben. Eine sehr ähnliche Ausbildung der Medianplatten
ist bei Blattodea und Mantodea zu finden. Bei einigen Vertretern dieser Taxa ist die DMP
allerdings mit der Basis des Radius verschmolzen (Onesto 1959, 1960).

Die Medianplatten werden in unterschiedlichster Weise abgewandelt. Bei den Plecoptera ist die PMP deutlich
ausgebildet, teilweise jedoch mit dem 3Ax verschmolzen. Die DMP hingegen ist weitgehend reduziert
(Onesto 1965, Brodskiy 1979a, 1979b). Eine starke Verkleinerung der PMP tritt bei Saltatoria, Hemiptera
und Dermaptera auf (Onesto 1961, Betts 1986, Brodskiy 1987).

51

Die Tegula (Tg) und die Humeralplatte (H) sind zwei einfache Sklerite an der Basis des
Flügelvorderrandes (Abb.1).

Die Humeralplatte entspricht der Basis der Costa. Bei Vertretern vieler rezenter hemi-
metaboler Neoptera ist sie durch einen membranösen Bereich von der Costa abgesetzt, z.B.
bei einigen Saltatoria (Wootton 1979), Blattodea (Onesto 1959), a (Onesto 1961)
und Heteroptera (Betts 1986).

Die Tegula ist eine relativ schwach sklerotisierte Platte im Vorderrand des Flügels direkt
am Übergang zum Notum. Sie ist dicht mit feinen Borsten besetzt, die wahrscheinlich sen-
sorische Funktion haben (Brodskiy 1994).

Die pleuralen Elemente des Flügelgelenks sind der pleurale Flügelgelenkfortsatz (PWP)
sowie die Basalar- und Subalarsklerite (Abb.2).

Der PWP ist ein dorsaler Fortsatz der Pleura in Verlängerung der Pleuralleiste. Das dorsale
Ende des PWP ist als Gelenkkopf (Fulcrum) ausgebildet, auf dem das 2Ax liegt. Der PWP
ist innerhalb der Neoptera ausgesprochen einheitlich ausgebildet. Nur der Winkel, den die
Pleuralleiste mit der Körperlängsachse bildet, ist stärkeren Variationen unterworfen. Für das
Grundmuster kann ein Winkel von annähernd 90° angenommen werden (Snodgrass 1909,
Crampton 1914, Matsuda 1970).

Das Basalare (Ba) ist aus einer Abspaltung des dorsalen Vorderrandes des Episternum her-
vorgegangen. Von ihm gingen im ursprünglichen Zustand wahrscheinlich fünf Muskeln aus,
von denen einer auf dem Trochanter (p-tr2), einer auf der Coxa (p-cx2) und einer auf dem
Sternum (p3) inseriert. Zwei weitere, kleinere Muskelzüge inserieren auf dem Praealararm
(t-p3) und am vorderen Scutumseitenrand (t-p7). Diese Muskeln wirken als Senker und
Extensoren des Flügels, da ihre Kontraktion durch einen Bereich verstärkter Membran auf
die Humeralplatte und somit auf den Flügelvorderrand übertragen wird (Snodgrass 1909).
Ob das Basalare im Grundmuster völlig vom Episternum getrennt ist, läßt sich aufgrund
der rezenten Verhältnisse nicht endgültig entscheiden. In vielen als relativ ursprünglich
geltenden Gruppen der Neoptera tritt ein gänzlich vom Episternum getrenntes Ba auf
(Snodgrass 1909, Matsuda 1970). Praktisch genauso häufig und gerade auch bei den Pleco-
ptera (Onetso 1965) treten aber Ba auf, die zumindest basal mit dem Episternum ver-
schmolzen sind. Da aufgrund der Ontogenese davon ausgegangen werden kann, daß das
Ba als Abspaltung des Episternum entstanden ist (Maki 1938, Matsuda 1970), liegt der
Schluß nahe, daß ein zumindest partiell mit dem Episternum verschmolzenes Ba dem
Grundmusterzustand entspricht. Mit guter Sicherheit kann davon ausgegangen werden, daß
das Ba als einfache Platte ausgebildet und nicht näher mit dem PWP bzw. dem Fulcrum
assoziiert ist.

Das Subalare trägt zwei Muskeln, von denen einer auf der Pleura (t-p16/19) und einer auf
der Coxa (t-cx8) entspringt. Über eine membranöse Verbindung zum PNP wirken diese
Muskeln indirekt auf das 2Ax und das 3Ax ein. Für das Subalare wird allgemein ebenfalls
angenommen (Weber 1924, Matsuda 1970), daß es ursprünglich der Pleura angehört. Maki
(1938) weist jedoch darauf hin, daß aufgrund der Ontogenie der Subalarmuskeln bei Blat-
todea und Caelifera (Saltatoria) ein Ursprung des Subalare als hinterer lateraler Bereich des
Notum bzw. Postnotum wahrscheinlich ist. Für das Grundmuster der Neoptera wird hier
die Existenz von nur jeweils einem Basalare und Subalare angenommen.

Zumindest bei Dermaptera (Onesto 1961), Saltatoria (Onesto 1963) und Mantodea (Onesto 1960) treten

rezent zwei Basalare auf (Snodgrass 1909). Hierbei handelt es sich wahrscheinlich um mehrfach unabhängig
entstandene Spezialisierungen, da z.T. innerhalb mancher Gruppen sowohl ein als auch zwei Basalare

32

auftreten (Onesto 1963). Bei einigen Vertretern der Perlidae (Plecoptera) tritt nach Snodgrass (1909) ein
zweites Subalare auf.

Direkt an den Elementen des Flügelgelenks inserieren bei der Stammart der Neoptera wahr-
scheinlich 12 Muskeln (Maki 1938, Keler 1963, Mickoleit 1969, Matsuda 1970). Neben
den sieben Muskeln, die Ba und Sb versorgen (s.o.), ziehen drei weitere Muskeln von der
Pleuralleiste zu den Gelenkfortsatzen des Notum (t-p10, t-p12, t-p15) und einer von der
Pleuralleiste zum Praealararm (t-p4). Ein Muskel ist zwischen der Pleura und dem 3Ax
aufgespannt (t-p13/14) und bewirkt das Einfalten des Flügels.

Das Grundmuster der Holometabola

Aus dem Vergleich der Morphologie der Flügelbasis der Teilgruppen der Holometabola
läßt sich schließen, daß der Bau des Flügelgelenks im Grundmuster der Holometabola im
wesentlichen mit den Verhältnissen bei der Stammart der Neoptera übereinstimmt. Unter-
schiede ergeben sich nur in der Ausbildung einzelner Elemente.

Das Notum und seine Fortsätze unterscheiden sich im Grundmuster der Holometabola nur
geringfügig von der Stammart der Neoptera. Der deutlichste Unterschied besteht in der
zumindest teilweisen Reduktion des Praealararmes. Dieser ist soweit verkürzt, daß er keinen
Kontakt zur Pleura hat. Als einziger weit lateral gelegener sklerotisierter Rest bleibt der
Insertionspunkt der beiden Muskeln t-p3 und t-p4 erhalten, der als Praealarsklerit bezeich-
net wird (Maki 1936, Czihak 1954, Kelsey 1957, Mickoleit 1968, 1971: Subtegula, Baehr
1975: Muskelscheibe). Der ANP ist wie im Grundmuster der Neoptera als flacher Fortsatz
gegen das aufgewölbte Notum abgesetzt.

Die drei notalen Gelenkfortsätze unterliegen vielfachen Variationen. So kommt es bei
Coleoptera und Neuropterida zur weitgehenden Reduktion des MNP (Abb.47, 69A) (Snod-
grass 1909, Ferris & Pennebaker 1939, Ferris 1940). Da aber in wahrscheinlich ursprüng-
lichen Gruppen wie z.B. den Archostemata ein MNP (Abb.8) vorhanden ist (Maki 1936,
Baehr 1975, Brodskiy 1988, 1994), kann davon ausgegangen werden, daß auch in der
Stammart der Holometabola ein relativ deutlich ausgeprägter MNP vorliegt.

Ein gut ausgebildeter ANP existiert bei allen rezenten geflügelten Holometabola (Snodgrass
1909). Der PNP kann, wie z.B. bei den Strepsiptera und manchen Diptera, vollständig redu-
ziert sein (Snodgrass 1909, Kinzelbach 1971). Eine Abtrennung des PNP als 4Ax ist weit
verbreitet und tritt im Mesothorax der Hymenoptera (Snodgrass 1909, 1911), im Meso- und
Metathorax der Mecoptera (Abb.82) (Mickoleit 1967, 1968, 1971), im Metathorax der Neu-
ropterida (Abb.69A, 76A) (Maki 1936, Ferris & Pennebaker 1939, Kelsey 1957) und im
Metathorax der Lepidoptera (Snodgrass 1909, Ivanov 1995) auf.

Die Axillarsklerite weisen praktisch keine Unterschiede zum Grundmuster der Neoptera
auf. Auch für die Stammart der Holometabola muß ein 1Ax angenommen werden, dessen
proximaler Rand caudal lang ausgezogen ist. Ein derartiger Zustand des 1Ax ist bei
Vertretern von Coleoptera (Abb.8, 11, 13, 16A, 18A etc.) (Snodgrass 1909, Kukalova-Peck
& Lawrence 1993, Baehr 1975), Hymenoptera (Snodgrass 1909, 1911, Weber 1926), Lepi-
doptera (Snodgrass 1909, Weber 1924, Onesto 1959, Sharplin 1963, Ivanov 1995), Tricho-
ptera (Crampton 1919), Diptera (Snodgrass 1909) und Strepsiptera (Kinzelbach 1971) zu
finden. Dem Grundmuster am nächsten stehen wahrscheinlich Vertreter der Symphyta
(Hymenoptera) sowie der Archostemata und Adephaga (Coleoptera).

53

Die genaue Ausbildung des 2Ax für das Grundmuster der Holometabola zu rekonstruieren
ist problematisch, da es innerhalb der Holometabola relativ starken Variationen unterliegt
(Snodgrass 1909, Matsuda 1970). Geht man davon aus, daß das Flügelgelenk der ursprüng-
lichen Coleoptera dem der Stammart der Holometabola recht nahe kommt, so dürfte das
2Ax im Grundmuster der Holometabola wenig von dem der Stammart der Neoptera abwei-
chen.

Die Verbindung zwischen dem Vorderrand des 2Ax und dem BR ist wahrscheinlich durch
eine durchgehende Sklerotisierung stabilisiert.

Auch das 3Ax der Stammart der Holometabola unterscheidet sich in Form und Lage kaum
von dem des Grundmusters der Neoptera. Eine wesentliche Veränderung betrifft jedoch die
Muskelversorgung. Der einzelne Muskel, der bei der Stammart der Neoptera das 3Ax mit
der Pleura verbindet, hat sich aufgespalten. Somit liegen nun zwei Muskeln vor, von denen
einer zur Pleuralleiste (t-p14) und einer zum Episternum (t-p13) zieht (Maki 1938, Matsuda
1970, Brodskiy 1988).

Im Bereich der Medianplatten kommt es innerhalb der Holometabola zu vielfältigen Umbil-
dungen, Verschmelzungen und Reduktionen. So ist bei den Neuropterida die PMP sehr
klein und dem 3Ax dicht angelagert. Die DMP hingegen ist durch großflächige leichte
Sklerotisierungen stark erweitert (Maki 1936, Ferris & Pennebaker 1939, Ferris 1940). Bei
den Lepidoptera ist die DMP häufig aufgelöst. Als ihre Reste können die verbreiterten
Basen von Media, Cubitus anterior und Cubitus posterior interpretiert werden. Die PMP
ist, soweit vorhanden, sehr klein und cranial des 3Ax gelegen. Oft fehlt sie vollständig
(Weber 1924, Onesto 1959, Sharplin 1963, Ivanov 1995). Vertreter der Hymenoptera besit-
zen in der Regel nur eine Medianplatte, die von ihrer Lage her als DMP identifiziert
werden kann (Snodgrass 1909, 1911, Weber 1926). Bei den Mecoptera sind beide Median-
platten in Form und Größe etwa gleich (Mickoleit 1967, 1968, 1971).

Daraus kann man schließen, daß im Grundmuster der Holometabola beide Medianplatten
vorhanden sind und daß sich ihre Ausgestaltung nicht wesentlich von der Stammart der
Neoptera unterscheidet.

Die Humeralplatte ist bei nahezu allen Holometabola zumindest als vergrößerte Basis der
Costa vorhanden (Snodgrass 1909). Ob sie im Grundmuster der Holometabola von der
Costa getrennt ist, kann aufgrund der rezenten Verhältnisse nicht eindeutig entschieden
werden. Der ventrale Bereich der Humeralplatte ist an einem Rastmechanismus mit dem
Ba beteiligt (siehe dort).

Die Tegula liegt bei der Stammart der Holometabola wahrscheinlich in ähnlicher Ausprä-
gung vor wie im Grundmuster der Neoptera. Bei einigen Vertretern der Holometabola
erfährt sie eine teilweise extreme Vergrößerung. Eine stark erweiterte Tegula tritt am
Vorderflügel der Hymenoptera, Lepidoptera, Trichoptera und Diptera auf (Snodgrass 1909,
Crampton 1919, Matsuda 1970, Ivanov 1995). Zumindest bei Hymenoptera und Lepido-
ptera ist eine Schutzfunktion für das Flügelgelenk sehr wahrscheinlich.

Der PWP der Stammart der Holometabola unterscheidet sich in seiner Ausbildung nicht
von dem der Stammart der Neoptera.

Basalare und Subalare liegen im Grundmuster der Holometabola mit großer Wahrschein-
lichkeit einzeln vor. Das Basalare ist dem PWP dicht angelagert. Basal ist es höchstwahr-

54

scheinlich mit dem Episternum verschmolzen. Es trägt an seinem dorsalen Rand Strukturen,
über die die ventrale Basis der Costa und/oder der Subcosta beim Ein- und Auffalten des
Flügels abrollen. Ein mit dem Episternum assoziiertes Ba ist innerhalb der hemimetabolen
Neoptera bei Psocopteren (Snodgrass 1909, 1927, Badonnel 1934) und Hemipteren (Taylor
1918) weit verbreitet. Es kann angenommen werden, daß es sich dabei um eine Synapo-
morphie von Holometabola und Acercaria - der potentiellen Schwestergruppe der Holome-
tabola - handelt (Kristensen 1981). Alternativ könnte diese Ausbildung des Ba konvergent
bei Acercaria und bei der Stammart der Holometabola entstanden sein. In diesem Fall ist
das Ba in seiner beschriebenen Form eine Autapomorphie der Holometabola.

Das Ba trägt dorsal im hinteren Bereich eine knopfartige Erweiterung, die mit den ven-
tralen Bereichen der Basis der Subcosta und der Humeralplatte einen Rastmechanismus bil-
det, der den Flügel in der Ruhelage fixiert. Dieser Rastmechanismus ist mit hoher Wahr-
scheinlichkeit eine Autapomorphie der Holometabola. Die veränderte Struktur des Basalare
und die zumindest leichte Vergrößerung seiner Muskulatur bewirken außerdem eine stär-
kere Beteiligung am Flügelabschlag. Eine derartige Ausprägung findet sich zum Beispiel
bei Mecoptera (Abb.81) (Mickoleit 1967, 1968, 1971), Neuropterida (Abb.73, 80) (Maki
1936, Ferris & Pennebaker 1939, Ferris 1940, Kelsey 1957), Trichoptera (Malicky 1973)
und Coleoptera (Abb.10) (Bauer 1910, Doyen 1965, Larsen 1966, Baehr 1975, Schneider
1987).

Wie im Grundmuster der Neoptera ist das Subalare im Grundmuster der Holometabola als
einfache Skleritplatte ausgebildet, die über einen kräftigen Membranstreifen mit dem PNP
verbunden ist. Möglicherweise ist es bei der letzten gemeinsamen Stammart der Holo-
metabola und ihrer potentiellen Schwestergruppe - den Acrecaria - zur Aufspaltung des zur
Pleura ziehenden Muskels in einen zum Epimeron führenden (t-p16) und einen zur Pleural-
leiste führenden Strang (t-p19) gekommen, so daß insgesamt drei Muskeln am Sb inserieren
(Maki 1938, Matsuda 1970). Hier besteht theoretisch wieder die Möglichkeit einer konver-
genten Entwicklung bei den Acercaria und den Holometabola. In diesem Fall ist die Ver-
sorgung des Sb mit drei Muskeln eine Autapomorphie der Holometabola.

Die sonstige Muskelausstattung der Flügelgelenkelemente ist im Vergleich zum Grund-
muster der Neoptera nicht verändert.

Neuropterida

Das Grundmuster

Der Vergleich von Vertretern der drei Teilgruppen der Neuropterida - Raphidioptera, Mega-
loptera und Planipennia - ergibt unter Berücksichtigung der Annahmen für das Grundmuster
der Holometabola folgende Merkmalszustände für die Stammart der Neuropterida:

Der vordere notale Gelenkfortsatz (ANP) ist deutlich ausgebildet und als flacher dreieckiger
Anhang vom aufgewölbten Notum abgesetzt. Er ist im Hinterflügel etwas größer als im
Vorderflügel. Der MNP ist praktisch vollständig reduziert. Der PNP des Mesothorax ist
schlank und relativ lang. Im Metathorax ist der PNP vom Notum getrennt und vermittelt
als frei in der Membran liegendes, langes und schmales 4Ax zwischen Notum und 3Ax.
Bei diesem Zustand des PNP handelt es sich wahrscheinlich um eine Autapomorphie der
Neuropterida.

Als Reste des reduzierten Praealararms liegen ein bis zwei Praealarsklerite als Muskelan-
satzstellen in der Membran cranial des lateralen Vorderrandes des Notum. Postnotum und

35

Postalararm sind gut ausgebildet. Der Postalararm steht mit dem Epimeron in Verbindung.
Rezent finden sich Nota mit sehr ähnlichen Ausbildungen bei Sialidae (Abb.73) (Czihak
1954), Corydalidae (Abb.76A, 77) (Snodgrass 1909, Maki 1936, Kelsey 1957), Raphidiop-
tera (Abb.79) (Ferris & Pennebaker 1939) und Planipennia (Ferris 1940).

Für das 1Ax ist ein Zustand anzunehmen, der den Verhältnissen bei den Sialidae (Abb.73),
Raphidioptera (Abb.79) (Ferris & Pennebaker 1939) und einigen Planipennia (Abb.69A,
69B) sehr nahekommt. Im Unterschied zum Grundmuster der Holometabola besitzt das 1Ax
der Neuropteriden-Stammart keine caudale Verlängerung des Proximalrandes. Der Hinter-
rand des 1Ax ist konkav, so daß proximal und distal zwei caudad weisende, etwa gleich
lange Spitzen entstehen. Der Körper des 1Ax ist im Verhältnis zu seiner Breite und zur
Gesamtlänge des 1Ax sehr kurz (Abb.69A, 73). Die Ansatzstelle des 1Ax-Halses teilt des-
sen Körper so, daß zwei Drittel seiner Breite distal und ein Drittel proximal der Ansatz-
stelle liegen. Der Hals des 1Ax ist schmal, der Übergang zum wenig verbreiterten Kopf
fließend.

Distalarm a
I =>
A p oe ra TER
Abb.4A-C,: Schema zur möglichen

Caudalarm Co . -
Evolution des 3Ax bei den Neurop-
fl? terida. A: 3Ax der Stammart der Neu-
ropterida. B: 3Ax der Raphidioptera.
4 C,: 3Ax der Planipennia. C, & C,: 3Ax
C3 der Megaloptera.

Für die Form des 2Ax im Grundmuster der Neuropterida kommen zwei Varianten in Frage:

1.) Ein im Verhältnis zum 1Ax großes 2Ax mit einem sehr breiten, angeschmolzenen BR.
Die Form des Komplexes 2Ax/BR kommt wahrscheinlich den Verhältnissen bei Sialis
(Abb.73) oder Chauliodes (Abb.76A) nahe.

2.) Ein 2Ax, das verglichen mit dem 1Ax relativ klein ist und an ein schmales BR grenzt,
wie bei Vertretern der Planipennia (Cueta beieri (Abb.69A) oder Osmylus fulvicephalus).

Im ersten Fall ist die Ausbildung von 2Ax und BR bei den Planipennia eine Autapomor-
phie dieses Taxon oder einer seiner Teilgruppen. Trifft die zweite Variante zu, so ist der
Zustand von 2Ax und BR bei den Megaloptera und den Raphidioptera als Synapomorphie
dieser beiden Taxa zu werten. Da nur sehr wenige Vertreter der Planipennia zur Unter-
suchung zur Verfügung standen, ist es hier nicht möglich, diese Frage zu entscheiden.

Sehr wahrscheinlich eine Autapomorphie der Neuropterida ist die auf mehr als zwei Drittel
ihrer Länge sklerotisierte Verbindung von der ventralen, hinteren Spitze des 2Ax zum Sb.
Im Grundmuster der Holometabola und wahrscheinlich auch der Neoptera ist sie als Band
angelegt. Diese Sklerotisierung findet sich in dieser Form bei allen untersuchten Vertretern
der Neuropterida.

Das 3Ax der Stammart der Neuropterida entspricht wahrscheinlich weitgehend dem der
Stammart der Holometabola. Das heißt, es handelt sich um ein relativ kompaktes Gebilde,

56

das zwischen dem 4Ax und dem distalen Ende des 1Ax liegt (Abb.4). Am proximo-crani-
alen Rand liegt der wahrscheinlich leicht erweiterte Ansatzpunkt der Muskulatur. Bei
Raphidioptera (Abb.79) und Megaloptera (Abb.77) finden sich 3Ax in ähnlicher Ausbil-
dung, mit dem Unterschied, daß keine komplette Sklerotisierung mehr vorliegt. Die stärkste
Abwandlung des Grundmusters findet sich bei Vertretern der Planipennia (Abb.69), bei
denen das 3Ax nur noch als schmaler Stab ausgebildet ist (Abb.4). Die Muskulatur des
3Ax entspricht mit zwei Muskeln (t-p13, t-p14) der des Grundmusters der Holometabola
(Mickoleit 1969).

Die Medianplatten sind nur leicht sklerotisiert und stark reduziert. Aufgrund der sehr
variablen Ausbildung der Medianplatten bei den rezenten Neuropterida ist die genaue Form
für das Grundmuster im Rahmen dieser Untersuchung nicht zu ermitteln. Es kann aber da-
von ausgegangen werden, daß die PMP deutlich kleiner ist als die DMP.

Die Humeralplatte ist mit der Costa verschmolzen und nur als leichte Verbreiterung und
stärkere Sklerotisierung der Basis der Costa erkennbar (Ferris & Pennebaker 1939, Ferris
1940).

Die Tegula ist in Vorder- und Hinterflügel als weichhäutige, relativ dicht mit Borsten
besetzte Aufwölbung zwischen vorderem Notumseitenrand und Basis der Costa (= Hume-
ralplatte) ausgebildet (Maki 1936, Ferris & Pennebaker 1939, Ferris 1940).

Der PWP der Stammart der Neuropterida unterscheidet sich in seiner Form nicht von dem
der Stammart der Holometabola, seine Lage ist allerdings leicht verändert: Bei den rezenten
Neuropterida liegt eine Verschiebung des Fulcrum, das im Grundmuster der Holometabola
und der Neoptera mit dem 2Ax artikuliert, in Richtung des 1Ax vor. Bei den Megaloptera
und Raphidioptera ist diese Verschiebung nicht vollständig. Das Fulcrum ist hier etwas
länger und breiter. Es liegt mit seinem vorderen, proximalen Drittel unter dem Kopf/Hals-
Bereich des 1Ax und mit den hinteren, distalen zwei Dritteln unter dem 2Ax (Abb.74). Bei
den Planipennia ist es komplett unter den Kopf/Hals-Bereich des 1Ax verlagert (Abb.69B).
Aus diesen Verhältnissen kann geschlossen werden, daß im Grundmuster der Neuropterida
das Fulcrum zumindest teilweise unter dem 1Ax liegt.

Das Ba ist bei der Stammart der Neuropterida nahezu vollständig mit dem Episternum ver-
schmolzen. Nur das dorsale Ende ist durch eine schmale Einkerbung vom PWP abgesetzt.
Es trägt an der Außenseite eine druckknopfartige Vorwölbung. Mit diesem Knopf korre-
spondiert eine von den ventralen Basen der Subcosta und der Costa (Humerus) gebildete
Bucht, die bei angelegtem Flügel auf dem Knopf des Ba einrastet und den Flügel in der
Ruhelage arretiert (Abb.75) (Ferris & Pennebaker 1939). Die Muskeln des Ba unterscheiden
sich nicht von denen im Grundmuster der Neoptera oder der Holometabola: Drei Haupt-
muskeln ziehen von der Sehnenkappe des Ba zum Sternum (p3), zur Coxa (p-cx2) und zum
Trochanter (p-tr2). Außerdem entspringen zwei kleinere Muskeln der dorsalen Fläche der
Sehnenkappe des BA und inserieren am Praealarsklerit (t-p3) und am ANP (t-p7) (Miller
1933, Maki 1936, Korn 1934, Czihak 1954, Kelsey 1957, Mickoleit 1969, Matsuda 1970).

Das Subalare ist eine schlanke, auffallend lang-ovale Platte - sie erreicht ca. ein Drittel der
Notumlänge - mit einer Längsnaht im dorsalen Drittel. Ihre Form und Lage ist innerhalb
der rezenten Neuropterida sehr stabil (Abb.71, 75, 76B, 78, 80) (Maki 1936, Ferris &
Pennebaker 1939, Ferris 1940, Kelsey 1957). Subalaria sehr ähnlicher Ausbildung finden
sich auch bei Mecopteren und Coleopteren, so daß davon ausgegangen werden kann, daß

317

es sich hier um einen plesiomorphen Zustand handelt. Die Muskelausstattung des Sb ent-
spricht mit drei Muskeln (t-p16, t-p19, t-cx8) der des Grundmusters der Holometabola.

Die Muskulatur der notalen Gelenkfortsätze und des Praealararmes entspricht den Ver-
hältnissen des Grundmusters der Holometabola (Mickoleit 1969).

Autapomorphien der Neuropterida

Aus der obigen Rekonstuktion des Grundmusters der Neuropterida können einige Merkmale
als potentielle Autapomorphien des Taxon hervorgehoben werden.

Der als 4Ax vom Metanotum abgetrennte, relativ lange, schlanke PNP ist mit hoher Wahr-
scheinlichkeit ein abgeleitetes Merkmal der Neuropterida. Im Grundmuster der Holometa-
bola ist ein stabförmiger, nicht vom Notum getrennter PNP anzunehmen, wie er z.B. bei
einigen Lepidoptera, Trichopera und Coleoptera vorliegt (Abb.8) (Snodgrass 1909,
Brodskiy 1994).

Ebenso mit großer Wahrscheinlichkeit autapomorph ist die partielle Verlagerung des Ful-
crum unter den Kopf/Hals-Bereich des 1Ax. Der ursprüngliche Zustand einer Gelenkung
zwischen 2Ax und Fulcrum, wie er für das Grundmuster der Holometabola anzunehmen
ist, ist bei den meisten Vertretern der Holometabola vorzufinden, wie z.B. bei Archo-
stemata (Coleoptera), Lepidoptera (Sharplin 1963, Ivanov 1995), Diptera (Ennos 1987) und
bei nicht holometabolen Neoptera (Brodskiy 1979b).

Die Gestalt des 1Ax, insbesondere die seines Körpers, ist ebenfalls als abgeleitetes Merk-
mal der Neuropterida anzusehen. Ein kurzer, breiter Körper ohne langen proximo-caudalen
Fortsatz ist in allen Teilgruppen der Neuropterida vertreten (Abb.69A, 73, 79). Ein dem
Grundmuster der Holometabola nahestehendes 1Ax mit einer langen proximalen Ecke und
konkavem Hinterrand konnte hingegen nicht nachgewiesen werden. Im Gegensatz dazu tre-
ten erste Axillaria in einer für die Holometabola wahrscheinlich ursprünglichen Form in
allen anderen Teilgruppen der Holometabola mit Ausnahme der Mecoptera (Mickoleit
1967, 1968, 1971) auf (Snodgrass 1909). Daraus ist zu folgern, daß die abgeleitete Form
des 1Ax mit hoher Wahrscheinlichkeit eine Synapomorphie der Raphidioptera, Planipennia
und Megaloptera ist.

Gleiches gilt fur das Band zwischen dem 2Ax und dem Subalare, das - ausgehend vom
2Ax - auf ca. zwei Drittel seiner Lange sklerotisiert ist.

Die Verhaltnisse innerhalb der Neuropterida

Die Strukturen des Flügelgelenks unterliegen innerhalb der Neuropterida einigen Verän-
derungen, die Rückschlüsse auf die Verwandtschaftsverhältnisse zwischen den Teilgruppen
des Taxon zulassen.

Megaloptera + Raphidioptera

Das ursprünglich schlank stabförmige 4Ax des Metathorax ist nur bei Megaloptera und
Raphidioptera kurz und kompakt geformt (Abb.73, 76A, 77, 79). Dies kann als Synapomor-
phie der beiden genannten Taxa gewertet werden.

Als Teiltaxon der Megaloptera sind die Corydalidae durch die abgeleitete Gestalt des 1Ax
des Hinterflügels charakterisiert. Der Körper des 1Ax ist sehr großflächig und schräg disto-
caudal leicht verlängert. Desweiteren besitzen sie einen sehr charakteristisch geformten
ANP, der fast viermal breiter als lang ist (Abb.76A, 77). Diese Form kann gegenüber dem

58

Sialidae
Corydalidae

Raphidioptera

Planipennia

6-7

Abb.5: Verwandtschaftsverhaltnisse der Teilgruppen der Neuropterida. Autapomorphien der jeweiligen
Taxa: 1-4: 4Ax vorhanden; Fulcrum liegt unter 1Ax und 2Ax; 1Ax-Körper kurz und breit; Sehne
zwischen 2Ax und Sb auf zwei Dritteln der Lange sklerotisiert. 5: 4Ax kurz und kompakt. 6-7:
Fulcrum liegt unter dem 1Ax; Rastmechanismus zwischen Ba und BSc / H reduziert. 8: Verbindung
zwischen 2Ax und PWP sklerotisiert. 9-10: 1Ax-Körper schräg, proximale Kante ausgestanzt; ANP
kurz und breit.

bei den Sialidae vorhandenen grundmusternahen ANP (Abb.73) als abgeleitet angesehen
werden.

Die Sialidae zeichnen sich durch die oberflächliche Sklerotisierung des Bandes zwischen
dem PWP und dem 2Ax (Abb.75) aus.

Planipennia

Bei allen untersuchten Vertretern der Planipennia liegt das Fulcrum vollständig unter dem
1Ax. Geht man davon aus, daß im Grundmuster der Neuropterida das 2Ax noch an der Ge-
lenkung beteiligt ist, so ist die komplette Verlagerung des Fulcrum unter das 1Ax eine
Autapomorphie der Planipennia. Ein weiteres möglicherweise apomorphes Merkmal ist die
Verkleinerung des Rastmechanismus zwischen Ba und ventralem Flügelvorderrand. Diese
Verkleinerung betrifft sowohl den Rastknopf des Ba als auch die Struktur in der ventralen
Flügelbasis, die den Rastknopf aufnimmt. Die Humeralplatte ist an der Bildung dieser
Struktur nicht mehr beteiligt.

Daraus ergeben sich für die Neuropterida folgende Verwandtschaftsverhältnisse (Abb.5):
Planipennia+(Raphidioptera+ Megaloptera). Durch potentielle Autapomorphien abgesichert
sind dabei die Neuropterida insgesamt, die Planipennia, das Taxon Raphidioptera + Mega-
loptera und innerhalb der Megaloptera die Corydalidae und die Sialidae.

Diese Interpretation der Verwandtschaftsverhälnisse wird unterstützt durch Untersuchungen
zur Morphologie der Larvalaugen der Neuropterida durch Paulus (1986), die ebenfalls ein
Schwestergruppenverhältniss zwischen Megaloptera und Raphidioptera wahrscheinlich
machen. Zu dem gleichen Ergebnis kommen Achtelig & Kristensen (1973) aufgrund der
Untersuchung von Larvalmerkmalen sowie Achtelig (1975) auf der Basis ethologischer
Merkmale und des Baues der Strukturen des Abdomen.

Coleoptera
Das Grundmuster
Aus der Untersuchung diverser Vertreter der Coleoptera läßt sich für das Grundmuster des
Hinterflügelgelenks folgende Ausprägung ableiten (die Situation des Gelenks der Vorder-
flügel ist wegen der Elytrenbildung grundsätzlich verschieden und wird hier nicht berück-
sichtigt):

39

Der ANP ist wie im Grundmuster der Holometabola als flacher, relativ lang ausgezogener
Fortsatz vom aufgewölbten Notum abgesetzt. Der MNP ist sehr klein und nur durch eine
flache Einbuchtung des Seitenrandes des Notum begrenzt. Er liegt hinter dem Ende des
1Ax und ist mit diesem durch ein Band verbunden. Der PNP ist lang und kräftig aus-
gebildet. Er reicht bis auf die Höhe des MNP nach vorne. Disto-cranial ist er deutlich
verbreitert. Mit dem caudalen Ende des 2Ax ist er durch ein Band verbunden. Ein Prae-
alarsklerit ist als Rudiment des Praealararmes vorhanden. Der Postalararm ist gut aus-
gebildet und hat Kontakt mit dem Epimeron (z.B. Abb.8, 19).

Das 1Ax hat im Grundmuster der Coleoptera einen ausgeprägten, breiten Kopf, dessen Vor-
derrand nach ventral umgebogen ist. Der Halsbereich ist durch eine starke Einschnürung
deutlich von Kopf und Körper des Sklerits abgesetzt. Der Körper des 1Ax ist breiter als
der Kopf. Der dem Notum anliegende Rand ist in einen langen caudalen Fortsatz ausge-
zogen. Dieser ist durch ein Band mit dem MNP verbunden (s.o.). Das 1Ax ist beweglich
mit dem Notum verbunden. Durch die Anlagepunkte des Kopfes am ANP und des Körpers
am Notumseitenrand läßt sich eine Achse legen, um die sich das 1Ax dreht. Diese Achse
spannt mit der disto-cranialen Kante des Körpers des 1Ax einen Winkel von mehr als 50°
auf (Abb.3). Dieser Zustand wurde wahrscheinlich aus dem Grundmuster der Holometabola
übernommen, da er auch bei Neuropterida und Mecoptera zu finden ist (Snodgrass 1909,
Ferris & Pennebaker 1939, Mickoleit 1967, 1968, 1971).

Das 2Ax ist der disto-cranialen Kante des Körpers des 1Ax dicht angelagert; es ist annä-
hernd dreieckig. Eine relativ breit gerundete Ecke liegt in der Bucht, die von Körper, Hals
und Kopf des 1Ax gebildet wird. Am stärksten sklerotisiert sind der craniale und der pro-
ximo-caudale Randbereich des 2Ax (Abb.3). Der proximo-caudale Rand ist in einen caudad
gerichteten Fortsatz ausgezogen, der unter der distalen Spitze des Körpers des 1Ax hin-
durchführt und durch ein Band mit dem PNP verbunden ist. Ein weiteres Band verbindet
das 2Ax mit der latero-caudalen Fläche des PWP kurz unterhalb des Fulcrum. Am Vorder-
rand der proximalen Spitze des 2Ax entspringt das BR. Es zieht als schmaler, sklerotisierter
Streifen schräg nach vorne.

Die Form des 3Ax ist sehr charakteristisch. Es besitzt einen caudalen Arm, dessen proxi-
male Kante S-förmig geschwungen ist. Im vorderen Bereich läuft ein Grat transversal über
das 3Ax, an dessen proximalem Rand eine Sehne ansetzt. Distal flacht der Grat allmählich
ab und läuft leicht zugespitzt aus. Die Sehne, die am proximalen Rand des Grates ansetzt,
zieht zu einem kleinen Sklerit (AMD) in der Membran zwischen 1Ax und 3Ax, das als In-
sertionspunkt für die Muskeln t-p13 und t-p14 fungiert.

Axillarsklerite und Nota in ähnlicher Ausprägung treten mit allen oder einem Teil dieser
Merkmale in allen Teilgruppen der Coleoptera auf (z.B. Abb.8, 19, 44A, 39). Dem Grund-
muster besonders nahe stehen wahrscheinlich die Archostemata (Abb.8 bis 17).

Die Medianplatten sind in ihrer Lage gegenüber dem Grundmuster der Holometabola leicht
verändert. Die DMP ist in Richtung 2Ax verlängert, und die von der Spitze des 1Ax zur
Spitze des 3Ax verlaufende Falte des Flügelgelenks trennt nicht die DMP von der PMP,
sondern läuft durch beide Medianplatten (z.B. Abb.8).

Die Humeralplatte ist als verbreiterte Basis der Costa ausgebildet. Ventral ist sie an dem
Rastmechanismus zwischen Ba und der Basis der Subcosta beteiligt. Eine Tegula fehlt.

Der PWP unterscheidet sich in seiner Ausbildung nicht von den Verhältnissen im
Grundmuster der Holometabola. Das Fulcrum liegt unter der proximalen Spitze des 2Ax.

60

Dieser Zustand ist rezent nur noch bei Vertretern der Archostemata zu finden (Abb.11, 13,
16A, 12, 14, 16B, 17).

Das Basalare ist in einen Stiel- und einen Kopfbereich unterteilbar. Der Stiel ist an seiner
Basis mit dem Episternum verschmolzen. Der Kopf ist stark erweitert. Er trägt lateral eine
Vorwölbung, die bei angelegtem Flügel in eine von der ventralen Sc-Basis und der Hume-
ralplatte gebildete Aussparung einrastet und so den Flügel in der Ruhelage fixiert. An der
Vorderkante des Kopfes befindet sich eine relativ lange, dorsad gerichtete Platte, die mit
der Basis von Costa und Subcosta verbunden ist (z.B. Abb.9, 10). An der Innenseite des
Ba setzt über eine sklerotisierte Sehne eine große Sehnenkappe an, die als Insertionspunkt
für die drei Hauptmuskeln (p3, p-cx2, p-tr2) des Ba dient.

Das Subalare liegt als relativ große, ovale Platte leicht schräg gestellt unter dem PNP, mit
dem es durch einen verstärkten Membranstreifen verbunden ist (z.B. Abb.9, 10).

Die Muskelausstattung der Gelenkelemente weicht im Grundmuster der Coleoptera nicht
vom Grundmuster der Holometabola ab (Bauer 1910, Stellwaag 1914, Larsen 1966, Micko-
leit 1969, Baehr 1975, Beutel 1986, Belkaceme 1991).

Das Grundmuster des Hinterflügelgelenks der Coleoptera, so wie es aus den Verhältnissen
bei den rezenten Coleoptera und den Annahmen für das Grundmuster der Holometabola
rekonstruiert werden kann, ist in fast allen Einzelheiten bei vielen Vertretern der heutigen
Archostemata verwirklicht (Abb.11, 13, 16A, 12, 14, 16B). Abweichungen treten nur bei
Priacma serrata auf, bei dem das Fulcrum zur Hälfte unter den Kopf des 1Ax verlagert ist
(Abb.8, 9, 10), und bei Micromalthus debilis, bei dem wohl, bedingt durch die geringe
Körpergröße, die Medianplatten und der PNP reduziert bzw. schwach sklerotisiert sind
(Abb.17).

Autapomorphien der Coleoptera

Obwohl das Hinterflügelgelenk der Stammart der Coleoptera in seiner grundsätzlichen Aus-
bildung dem Grundmuster der Holometabola sehr nahesteht, existieren im Detail doch eine
Reihe potentiell autapomorpher Bildungen.

Für das Grundmuster der Holometabola ist ein relativ langer, annähernd stabförmiger PNP
anzunehmen. Bei den Coleoptera ist der PNP distal verbreitert. Dieser Zustand ist bei
Archostemata und Adephaga zu beobachten (z.B. Abb.8, 18A) und als autapomorphes
Merkmal der Coleoptera zu werten.

Das 1Ax weist eine wahrscheinlich apomorphe Veränderung insofern auf, als der vordere
Abschnitt im Verhältnis zum Hals auf charakteristische Art stark verbreitert und somit erst
deutlich als Kopf ausgebildet ist.

Die Muskulatur des 3Ax inseriert bei den Coleoptera nicht direkt am 3Ax, sondern an
einem kleinen Sklerit (AMD) (= Praeaxillarsklerit von Belkaceme 1991) in der Membran
zwischen 1Ax bzw. Notum und 3Ax (z.B. Abb.8, 47). Dieses Sklerit kann entweder durch
Abspaltung vom 3Ax oder durch Neubildung in Form einer Sklerotisierung in der gemein-
samen Sehne der beiden Muskeln t-p13 und t-p14 entstanden sein. Bei anderen Holometa-
bola ist ein derartiges Sklerit nicht nachweisbar, so daß es wahrscheinlich ein abgeleitetes
Merkmal der Coleoptera darstellt.

Ein weiteres abgeleitetes Merkmal des 3Ax ist wahrscheinlich der S-förmig geschwungene
Verlauf des proximalen Randes des 3Ax, wie er sehr ausgeprägt z.B. bei Priacma serrata
und Distocupes varians zu beobachten ist (Abb.8, 12). Bei Vertretern der Adephaga und

61

DMP PMP

Abb.6: Verschiedene Aus-
bildungen der Medianplat-
ten. A: Grundmuster der
Neoptera. B: Priacma
serrata. C-I: Coleoptera,

6 _ eit) New s J,K: Neuropterida, L:
Strepsiptera.

der Polyphaga ist dieser Verlauf des Proximalrandes ebenfalls zu finden, wenn auch in
etwas abgeschwachter Form (Abb.19, 24, 28, 39).

Autapomorph fur die Coleoptera sind wahrscheinlich auch Gestalt und Lage der Median-
platten. Die DMP ist proximad verlangert und reicht fast bis an das 2Ax heran. Die PMP
besitzt einen distalen Auslaufer, der mit der DMP verschmolzen ist. Diese Form der
Medianplatten, wie sie bei Priacma verwirklicht ist, unterliegt in den Teilgruppen der Kafer
starken Abwandlungen (Abb.6). Dabei kann einerseits der proximale Teil der DMP redu-
ziert sein (Meloidae: Abb.57A), andererseits kann dieser Bereich auch eine Verlangerung
erfahren, so daß er sich zwischen das 2Ax und die weitgehend reduzierte PMP schiebt
(Silphidae: Abb.28).

Das Ba weist an seinem dorso-frontalen Rand einen flachen, nach dorsal gerichteten Fort-
satz auf (Abb.12), der ebenfalls als abgeleitet für die Coleoptera angesehen werden kann.

Die Teilgruppen der Coleoptera

Archostemata

Das Flugelgelenk der Archostemata kommt dem der Stammart der Coleoptera wahrschein-
lich sehr nahe. Als potentiell abgeleitetes Merkmal kann das im Gegensatz zum Grund-
muster der Holometabola und wahrscheinlich auch der Coleoptera in seinem Kopfbereich
stark erweiterte Ba angesehen werden. Im funktionellen Zusammenspiel mit dieser großen
lateralen Erweiterung des Ba ist auch die von der Basis der Subcosta und der Humeralplatte
gebildete Aufnahme für diese Struktur sehr weitlumig gestaltet (Abb.10).

Im Zusammenhang mit den Archostemata sollte erwähnt werden, daß den Strukturen der
Flügelbasis keine Hinweise dafür zu entnehmen sind, daß Micromalthus debilis (Abb.17),
wie z.B. von Kausnitzer (1975) angenommen, zu den Polyphaga zu stellen ist. Die Flü-
gelbasis von Micromalthus weist keines der für einen Polyphagen zu erwartenden Merk-
male auf. Insbesondere ist der PNP lang, und der Winkel « des 1Ax (Abb.3) beträgt ca.
57° (vergl. Polyphaga + Myxophaga). Ein weiterer Punkt, der gegen eine enge Verwandt-
schaft von Micromalthus mit den Polyphaga spricht, ist das Fehlen einer Kryptopleura im
Prothorax, die eine Autapomorphie der Polyphaga darstellt. Bei Micromalthus fehlt zwar

62

die Notopleuralnaht, dies rührt aber von einer einfachen Verschmelzung der Pleuralsklerite
her und ist nicht auf eine Kryptopleurie zurückzuführen (Lawrence & Newton 1982).
Außerdem sind die potentiellen Autapomorphien der Archostemata im Flügelgelenk von
Micromalthus erkennbar, und es existieren Gemeinsamkeiten im Bau des Aedeagus und in
einigen Larvalmerkmalen (Lawrence & Newton 1982). Somit ist eine Zugehörigkeit zu den
Archostemata sehr wahrscheinlich.

Adephaga

Ein relativ auffälliges, mit hoher Wahrscheinlichkeit abgeleitetes Merkmal der Adephaga
ist der gut ausgebildete mittlere Gelenkfortsatz des Notum (MNP). Bei allen untersuchten
Adephaga ist der MNP durch eine vor ihm liegende tiefe Einkerbung des Notumseitenran-
des deutlich abgegrenzt. Der so gebildete Gelenkfortsatz ist sehr breit und oft zweispitzig
(Abb.19, 22). Ein Band verbindet die weiter cranial gelegene Spitze mit dem Ende des
1Ax.

Wahrscheinlich ebenfalls abgeleitet ist der Zustand des Rastkopfes am Basalare. Dieser ist
sehr lang (er erstreckt sich fast über die gesamte dorso-laterale Kante des Ba) und sehr
flach ausgebildet (Abb. 18B, 21A, 23). Im Grundmuster der Holometabola und der Coleo-
ptera ist dieser Fortsatz etwas kürzer und deutlich höher gebaut (Abb.30B, 48, 75, 76B, 78,
80, 81). Mit der Form des Knopfes am Ba korrespondiert natürlich immer die Form der
von Subcosta und Humeralplatte gebildeten Öffnung an der ventralen Flügelbasis.

Myxophaga

Aus der Gruppe der Myxophaga konnten zwei Arten als rasterelektronenmikroskopische
Präparate untersucht werden (s. Ergebnisse) (Abb.25A, 25B). Beide Arten besitzen sehr
ähnlich ausgebildete Flügelgelenke, die sich besonders durch ein langes, stabförmiges 3Ax
auszeichnen. Aufgrund der gegenüber den anderen Teilgruppen der Coleoptera offensicht-
lich abgeleiteten Struktur und der Übereinstimmung bei beiden untersuchten Arten kann
man mit einiger Vorsicht die Ausbildung des 3Ax als autapomorphes Merkmal der Myxo-
phaga interpretieren.

Polyphaga
Ein wahrscheinlich abgeleitetes Merkmal der Polyphaga ist die Form des PNP: Im Gegen-
satz zu den Myxophaga wird der PNP der Polyphaga zum Apex hin schmaler (Abb.57A).

Das 3Ax der Polyphaga besitzt gegenüber dem 3Ax der Adephaga und Archostemata einen
deutlich verlängerten caudalen Arm (Abb.57A). Dieses Merkmal kann ebenfalls als Aut-
apomorphie gewertet werden und steht möglicherweise in funktionellem Zusammenhang
mit der veränderten Form des PNP.

Außer potentiell autapomorpher Merkmale der vier wahrscheinlich monophyletischen Groß-
gruppen der Coleoptera (Crowson 1960, 1967, 1975, Klausnitzer 1975, Lawrence & New-
ton 1982, 1995) lassen sich aus der Struktur der Flügelbasis auch Hinweise auf mögliche
Schwestergruppenverhältnisse dieser vier Gruppen zueinander ableiten.

Myxophaga + Polyphaga

Diese beiden Taxa weisen zwei Gemeinsamkeiten auf, die auf ein mögliches Schwester-
gruppenverhältnis hindeuten. Bei allen untersuchten Arten der Myxophaga und der Poly-
phaga ist der hintere Gelenkfortsatz des Notum (PNP) sehr kurz (Abb.25A, 25B, 39, 47,
57A). Im Gegensatz dazu besitzen alle untersuchten Vertreter von Archostemata und Ade-
phaga einen lang ausgezogenen PNP, wie er auch für das Grundmuster der Coleoptera an-

63

zunehmen ist (z.B. Abb.8, 11, 13, 19, 22). Aufgrund dieser Merkmalsverteilung ist anzu-
nehmen, daß die Verkürzung des PNP eine Synapomorphie von Polyphaga und Myxophaga
1st.

Eine weitere Gemeinsamkeit dieser beiden Taxa betrifft die Form des 1Ax und seine Lage
zum Notum. Wie oben bereits erwähnt, ist das 1Ax beweglich mit dem Notum verbunden.
Die Achse, um die sich das 1Ax bezogen auf das Notum drehen kann, läuft durch die Be-
rührungspunkte von 1Ax, ANP und Notumseitenrand (= MNP). Diese Drehachse spannt
mit dem disto-cranialen Rand des Körpers des 1Ax einen Winkel auf (Abb.3: a). Er ist bei
allen untersuchten Adephaga, Archostemata und Neuropterida größer oder gleich 50°, und
es ist anzunehmen, daß dies auch dem Zustand im Grundmuster der Holometabola und der
Coleoptera entspricht. Im Gegensatz dazu konnte bei keiner der untersuchten Arten der
Myxophaga und Polyphaga ein Winkel von mehr als 45° festgestellt werden (Tab.1).
Daraus kann geschlossen werden, daß ein Winkel a von weniger als 45° ein synapomor-
phes Merkmal der Myxophaga und Polyphaga ist.

Bei beiden Merkmalen besteht die Möglichkeit der Konvergenz. Dabei ist aber zu beden-
ken, daß Veränderungen der Lagebeziehungen der Gelenkelemente, die durch Änderung
der PNP-Länge und des Winkels « bewirkt werden, automatisch auch die Hebelverhältnisse
im Gelenk beeinflussen. Somit bewirkt eine oberflächlich recht einfach wirkende Ver-
änderung eine sehr komplexe Beeinflussung der Gelenkmechanik, was die mehrfach unab-
hängige Entstehung praktisch identischer Strukturen in diesem Bereich als recht unwahr-
scheinlich erscheinen läßt.

Adephaga + (Myxophaga + Polyphaga)

Das 1Ax und seine Beziehung zu benachbarten Elementen des Flügelgelenks liefert auch
einige Hinweise auf ein Schwestergruppenverhältnis zwischen den Adephaga und dem
potentiellen Monophylum Myxophaga + Polyphaga. Eine mögliche Synapomorphie dieser
Taxa ist die Verlagerung des Fulcrum von der proximalen Spitze des 2Ax unter das 1Ax.
Dabei ist davon auszugehen, daß das Fulcrum nach seiner Verlagerung unter dem Hals-
bereich des 1Ax liegt. Innerhalb der Polyphaga kommt es dann anscheinend mehrfach zu
einer Verschiebung in Richtung 1Ax-Kopf. Die dem Grundmuster entsprechende Position
des Fulcrum unter dem 2Ax findet sich bei den meisten Vertretern der Archostemata und
der restlichen Holometabola mit Ausnahme der Neuropterida.

Wahrscheinlich in funktionellem Zusammenhang mit der Lage des Fulcrum besitzt das 2Ax
der Adephaga und der Myxophaga + Polyphaga an seiner Ventralseite einen breiten, fla-
chen Fortsatz, der seitlich unter den Körper des 1Ax ragt. Bei den Archostemata findet sich
ein derartiger Fortsatz nicht, ebensowenig bei den Neuropterida. Dieser Fortsatz schränkt
die Beweglichkeit des 2Ax gegenüber dem 1Ax stark ein. Diese Bewegungseinschränkung
könnte mit der Verlagerung des Gelenks unter das 1Ax und den damit veränderten Hebel-
verhältnissen im Zusammenhang stehen. Ein Heranrücken des Gelenks an das Notum (=
Verlagerung vom 2Ax zum 1Ax) bewirkt ein stärkeres Anheben des Flügels bei gleicher
Verschiebung des Notum. Anders ausgedrückt: Der Flügel wird bei gleicher Kontraktions-
länge der Muskeln stärker angehoben.

Archostemata + (Adephaga + (Myxophaga + Polyphaga))

Aus der obigen Bewertung der Merkmalsausprägungen im Flügelgelenk ergeben sich fol-
gende Verwandschaftsverhältnisse der Teiltaxa der Coleoptera: Archostemata + (Adephaga
+ (Myxophaga + Polyphaga)). Dabei ist jedes Taxon durch potentiell autapomorphe Merk-

64

male begründbar. Die Interpretation dieser Merkmale als Autapomorphien beruht im
Wesentlichen auf den Annahmen für die Ausprägung des Flügelgelenks im Grundmuster
der Coleoptera, welches durch Vergleich der Merkmalszustände bei unterschiedlichen
Vertretern der Holometabola und der hemimetabolen Neoptera rekonstruiert wurde.

Vergleich mit anderen Verwandtschaftsanalysen

Insgesamt korreliert diese Interpretation der Verwandtschaftsverhältnisse gut mit
Hypothesen auf der Basis anderer Merkmalskomplexe (Crowson 1960, 1967, 1975, Klaus-
nitzer 1975, Lawrence & Newton 1982). Problematisch ist aber jeweils die Position der
Myxophaga. Insbesondere aufgrund des Vergleichs mit fossilen Coleoptera aus der Trias
und dem Jura (Schizophoridae, Catiniidae), sowie aufgrund einiger Merkmale des Flü-
gelgeäders kommen Ponomarenko (1969, 1971), Lawrence & Newton (1982) und Kuka-
lova-Peck & Lawrence (1993) zu dem Schluß, daß die Myxophaga den Adephaga näherste-
hen als den Polyphaga. Träfe dies zu, müßten die hier dargestellten potentiellen Syn-
apomorphien des Flügelgelenks sowie der Tibiotarsus der Larven bei Myxophaga und Poly-
phaga konvergent entstanden sein (Klausnitzer 1975).

Die von Kukalova-Peck & Lawrence (1993: 214) angeführten synapomorphen Merkmale
von Adephaga und Myxophaga sind ebenfalls problematisch. So wird das Fehlen der Quer-
ader r3 als Synapomorphie angegeben. In Abb.13 und 14 desselben Artikels ist aber im
Flügel von Macrogyrus sp. (Adephaga: Gyrinidae) eine Ader r3 eingezeichnet und als
solche beschriftet. Desweiteren kann eine proximade Verschiebung der Querader r4, wie
sie als gemeinsames abgeleitetes Merkmal von Myxophaga und Archostemata postuliert
wird, in den gegebenen Zeichnungen (Abb.13 bis 29) mit Bezug auf das Flügelstigma nicht
festgestellt werden. Vielmehr liegt bei Adephaga, Myxophaga und Polyphaga eine proxi-
made Verlagerung der gesamten Aderung bzw. eine Verlängerung des distalen aderarmen
Membranbereichs vor. Dies ist aber eher als Synapomorphie dieser drei Taxa zu werten
und nicht geeignet, ein Schwestergruppenverhältnis von Myxophaga und Adephaga zu
begründen.

Für die von denselben Autoren als weitere Synapomorphie von Myxophaga und Adephaga
angeführte unverzweigte AP,,, ist die Wahrscheinlichkeit der Konvergenz sehr groß. Dies
ist anzunehmen, da bei den kleinen Flügeln der Myxophaga der gesamte Analbereich stark
reduziert ist (Kukalova-Peck & Lawrence 1993: Abb.23-29) und auch bei vielen Polypha-
gen die AP,,, unverzweigt ist (Kukalova-Peck & Lawrence 1993: Abb.36-53, 56-59, 61-
68).

Kukalova-Peck & Lawrence (1993) nehmen außerdem ein Schwestergruppenverhältnis von
Myxophaga + Adephaga und Archostemata an. Dieses wird unter anderem damit be-
gründet, daß das 1Ax dieser drei Taxa einen langen caudalen Fortsatz trägt. Wie meine
Untersuchung zeigt, existiert ein solcher Fortsatz auch bei den Polyphaga (z.B. Abb.44A,
39, 41A, 45) und bei vielen weiteren Holometabola (siehe Grundmuster der Holometabola).
Er ist auch für die Stammart der Coleoptera anzunehmen. Damit ist er bei den Myxophaga,
Adephaga und Archostemata eine Symplesiomorphie und nicht geeignet, ein Schwester-
gruppenverhältnis von Myxophaga + Adephaga und Archostemata zu begründen.
Ebenfalls plesiomorph ist wahrscheinlich das Merkmal Nummer (2) (Numerierung entspre-
chend Kukalova-Peck & Lawrence 1993: 213): Die Existenz einer Gelenkstelle in der
medial bar (= MP, ,,nach der Nomenklatur von Kukalova-Peck & Lawrence (1993) ent-
spricht dem Cu der Nomenklatur von Forbes (1922, 1926) und Ponomarenko (1972)). Ein
entsprechender nicht oder nur leicht sklerotisierter Bereich findet sich an gleicher Stelle

65

auch bei einigen Polyphaga (Kukalova-Peck & Lawrence 1993: Abb.54-56) und war daher
wahrscheinlich schon bei der Stammart der Coleoptera vorhanden.

Ähnlich sind die Verhältnisse bei Merkmal Nr. (5): Existenz und V-Form des BAA (=
anterior anal basivenale). Nach Kukalova-Peck (1983, 1991) ist ein BAA für das Grund-
muster des Pterygotenflügels anzunehmen. Außerdem ist es in vielen Taxa der hemimetabo-
len Neoptera nachweisbar (Kukalova-Peck 1983), so daß es auch zum Grundmuster der
Neoptera gehört. Wenn es eine Synapomorphie von Myxophaga + Adephaga und Archoste-
mata wäre, müßte es spätestens in der Stammart der Coleoptera fehlen, um dann in der
gemeinsamen Stammart dieser drei Taxa wieder neu zu entstehen. Da aber auch bei Ver-
tretern der Polyphaga eine nach Kukalova-Peck & Lawrence (1993) als BAA zu interpre-
tierende Struktur vorhanden ist (Kukalova-Peck & Lawrence 1993: Abb.80, 81, mit BA
beschriftet), ist die Wahrscheinlichkeit, daß auch die Stammart der Coleoptera ein BAA
besaß, relativ groß. Somit ist die Existenz des BAA für Adephaga, Myxophaga und Archo-
stemata eine Plesiomorphie.

Computeranalyse

Um soweit wie möglich auszuschließen, daß wahrscheinlichere, weil in der Merkmalsent-
wicklung sparsamere, Hypothesen über Verwandtschaftsverhältnisse übersehen wurden,
wurden die Merkmale des Flügelgelenks einer Computeranalyse unterzogen. Die Gelenk-
strukturen wurden zu diesem Zweck in 47 Merkmale kodiert.

Die Merkmale und ihre Zustände

1) ANP: Form

0: Der flache Bereich des ANP ist etwa so lang wie breit (Abb.8, 73, 79).

1: ... ist deutlich breiter als lang (Abb.76A, 77).

2: ... ist deutlich länger als breit (Abb.57A, 55A).

3: Es ist kein gegenüber dem Notum deutlich abgeflachter Bereich vorhanden (Abb.69A).
4: Der ANP ist als caudad gerichteter Haken im Notumrand ausgebildet.

2) ANP: Abstand des Kontaktes mit dem 1Ax, gemessen vom Vorderrand des 1Ax, bezogen auf die Länge
des 1Ax

0: Der Quotient aus der Strecke vom Vorderrand des 1Ax zum Kontaktpunkt mit dem ANP und der Länge
des 1Ax ist kleiner als 0,07 (Abb.79).

1: Der Quotient liegt zwischen 0,07 und 0,114 (Abb.47).

2: ... zwischen 0,114 und 0,164 (Abb.57A).

3: ... zwischen 0,164 und 0,25 (Abb.66A).

4: ... ist größer als 0,25 (Abb.34).

3) Mittlerer Gelenkfortsatz (MNP)

0: Der MNP ist ein kurzer, schmaler Haken, die Spitze weist craniad (Abb.8).

l: ... ist kurz und breit. Sein Ende ist zweispitzig oder eingekerbt (Abb.18A, 19).

.. ist kurz und dornförmig (Abb.32A, 33A).

.. ist durch eine deutlich spitze Einkerbung des Notum markiert (Abb.57A).

.. ist durch eine deutlich gerundete Einbuchtung des Notum gekennzeichnet (Abb.56A).
... ist als caudad gerichteter Haken ausgebildet, der durch eine nach cranial gerichtete Einbuchtung des
Notumrandes entsteht (Abb.55A).

6: ... ist in Form einer langen, flachen Einbuchtung des Notumrandes ausgebildet (Abb.53A).
7: ... ist nicht als eigenständige Struktur erkennbar (Abb.73).

4) Hinterer Gelenkfortsatz (PNP)

0: Der PNP ist ein langer, schlanker Fortsatz.

1: ... ist lang und apikal deutlich verbreitert (Abb.18A, 19).

2: ... ist kurz hakenförmig. Das breite Ende ist craniad gerichtet (Abb.39).

Ma ID

66

3: ... Ist kurz hakenförmig und wird apikal deutlich schmaler (Abb.57A).

4: ... ist als relativ langes, stabförmiges 4Ax vom Notum getrennt (Abb.69A).
5: Das 4Ax ist kurz dreieckig (Abb.79).

6: Das 4Ax ist kurz und kompakt, annähernd viereckig (Abb.77).

7: Der PNP ist bis auf eine Vorwölbung des Notumrandes reduziert (Abb.49).
8: Der PNP fehlt (Abb.84).

5) PNP: Winkel ß zwischen den Achsen (a) und (b) (Abb.3)
0: Der Winkel B liegt zwischen 30° und 40° (Abb.8).

1: ... B liegt zwischen 25° und 30° (Abb.19).

2: ... B liegt zwischen 13° und 25° (Abb.57A).

3: ... B liegt zwischen 6° und 13° (Abb.47).

4: ... B ist kleiner als 6° (Abb.54A).

5: Der PNP fehlt.

6) Praealarbriicke (Prab)
0: Die Praealarbriicke ist vollstandig, sie erreicht das Episternum.
1: Die Praealarbrücke ist stark verkürzt und hat keinen Kontakt zum Episternum.

7) Postalarbrücke (Poab)

0: Die Postalarbrücke ist eine einfache, laterale Verlängerung des Postnotum, die das Epimeron erreicht
(Abb.57A).

1: ... besitzt unterhalb des PNP einen kurzen, cranialen Fortsatz (Abb.38B).

2: ... besitzt einen langen cranialen Fortsatz, der den PNP bzw. das Sb erreicht (Abb.36).

8) 1Ax: Kopfform

0: Der Kopf des 1Ax ist schmal, nicht deutlich vom Hals abgesetzt und vorne gerundet (Abb.73).

l: ... ist deutlich verbreitert und vorne abgerundet (Abb.18A).

2: ... ist breit und vorne abgestutzt (Abb.8).

3: ... ist breit und sein Vorderrand trägt in der distalen Hälfte eine knopfartige Abschnürung (Abb.47).
4: ... ist breit und hat einen distalen Fortsatz (Abb.34).

5: ... ist hammerartig abgeknickt, so daß die Vorderkante nach außen gedreht ist (Abb.39).

9) 1Ax: Kopfvorderrand (ordered)

0: Der Kopfvorderrand ist flach oder schmal nach ventral umgebogen

1: ... besitzt einen deutlichen, nahezu senkrecht nach unten weisenden Fortsatz (Abb.71).
2: Der Fortsatz ist schräg nach ventro-distal gerichtet (Abb.31B).

10) 1Ax: Halsbreite

0: Der Hals ist sehr schmal, mit einem annähernd gerade verlaufenden Distalrand (Abb.73).
1: ..., sein distaler Rand ist relativ breit konkav (Abb.34, 37A).

: ..., der distale Rand ist tief konkav bis spitz eingekerbt (Abb.11).

: Der Hals ist etwa so breit wie der Kopf (Abb.53A).

: Der Hals ist schlank und kräftig geschwungen (Abb.76A).

... sehr schlank, mit einem deutlichen Fortsatz am Distalrand (Abb.69A).

... schmal und abrupt rechtwinklig vom Kopf abgesetzt (Abb.84).

... extrem kurz oder fehlt ganz (Abb.77).

D

1) 1Ax: disto-ventraler Halsrand

: Der distale Rand des 1Ax-Halses ist ventral nicht nach innen umgeschlagen.
... Ist maximal bis zur Mitte des Halses untergeschlagen (Abb.30B)

... Ist über die Mitte des Halses hinaus untergeschlagen (Abb.37B)

Se re ee mo A

N

12) 1Ax: proximale Ecke des Körpers

0: Die proximo-caudale Ecke des 1Ax-Körpers ist deutlich länger als die disto-caudale Ecke; sie endet am
MNP (Abb.8, 63A).

1: ...; sie endet in der Bucht zwischen Notum und PNP (Abb.39).

2: Die proximo-caudale Ecke ist so lang wie die disto-caudale Ecke (Abb.34).

3: ... ist kürzer als die disto-caudale Ecke (Abb.77).

13) 1Ax: distale Ecke des Körpers
0: Die hintere distale Ecke des 1Ax-Körpers ist nach caudal umgebogen (Abb.45).

67

l: ... zeigt gerade nach distal (Abb.73).
2: ... ist schräg nach hinten gerichtet (Abb.77).

14) 1Ax: caudaler Körperrand

0: Der Caudalrand des 1Ax-Körpers ist schief konkav (Abb.18A).
1: ... ist gerade konkav (Abb.34).

.. ist einfach gerade (Abb.77).

.. ist konvex.

.. ist stark geschwungen (Abb.77).

. ist sehr tief konkav (Abb.84).

15) 1Ax: Winkel « (Abb.3)

0: Der Winkel « ist größer als 50° (Abb.8).

1: ... & liegt zwischen 25° und 50° (Abb.57A).
2: ... & ist kleiner als 25° (Abb.47).

16) 1Ax: Gesamtlänge des 1Ax in Bezug auf die Länge des Notum

0: Der Quotient aus der Notumlänge und der Lange des 1Ax ist größer als 3,8 (Abb.69A).
1: Der Quotient liegt zwischen 3 und 3,8 (Abb.47).

2: ... zwischen 2 und 3 (Abb.34).

3: ... zwischen 1,3 und 2 (Abb.39).

17) BR: Form

0: Das Basiradiale ist relativ breit und nicht mit dem 2Ax verschmolzen.
1: ... ist sehr breit und mit dem 2Ax verschmolzen (Abb.77).

.. ist mittelmäßig breit und mit dem 2Ax verschmolzen (Abb.8, 47).
.. ist bis auf einen Stumpf am 2Ax reduziert (Abb.37A).

.. Ist schmal und mit dem 2Ax verschmolzen (Abb.18A).

.. fehlt (Abb.60A).

18) BR: Ursprung

0: Das Basiradiale ist nicht mit dem 2Ax verschmolzen.

1: Das BR inseriert mit schmalem Ansatz in der Mitte der cranialen Kante des 2Ax (Abb.47).
2: ... mit breitem Ansatz in der Mitte der cranialen Kante des 2Ax (Abb.77).

3: ... an der distalen Kante des 2Ax (Abb.69A).

4: ... an der proximalen Spitze des 2Ax (Abb.57A).

5: ... fehlt (Abb.60A).

19) BSc: Kontakt zum 1Ax

0: Die Basis der Subcosta liegt frontal vor dem Kopf des 1Ax (Abb.19).

1: Die BSc faßt mit einem Fortsatz auf den Kopf des 1Ax (Abb.79).

20) BSc: Anzahl der Kontaktpunkte zum 1Ax

0: BSc und der Kopf des 1Ax berühren sich mit einem mehr oder weniger breiten Kontaktpunkt (Abb.47).
1: BSc und der Kopf des 1Ax berühren sich an mindestens zwei getrennten Punkten (Abb.19).

ed Bho oy

ad hin on or

21) BSc: Aussparung an der Ventralseite fiir den Ba-Knopf

0: Die ventralen Basen von Sc und C (Humerus) sind einfach ausgebildet.

1: Der Humerus und die BSc bilden eine Aufnahme für den Rastknopf des Ba (Abb.80).
2: Die Aufnahme fiir den Ba-Rastknopf wird nur von der BSc gebildet (Abb.70).

22) 2Ax: Form von dorsal

0: Das 2Ax ist dreieckig geformt. Eine Ecke ist dem 1Ax zugekehrt, die disto-craniale Ecke liegt weiter
vorne als der Berührungspunkt der proximalen Ecke mit dem 1Ax (Abb.8).

1: Das 2Ax ist annähernd halbkreisförmig, mit lang ausgezogenem caudalem Fortsatz (Abb.19).

2: Nur die proximalen Randbereiche des 2Ax sind sklerotisiert. Daraus ergibt sich eine V-förmige Gestalt,
deren Spitze dem 1Ax zugewandt ist (Abb.39).

3: Das 2Ax ist grob rundlich geformt (Abb.69A, 73).

4: Das 2Ax ist dreieckig. Die disto-craniale Ecke reicht nicht über den Berührungspunkt der proximalen Ecke
und des 1Ax hinaus nach vorne (Abb.34).

5: Das 2Ax ist kompakt dreieckig mit breiter Ansatzstelle des BR (Abb.77).

6: Das 2Ax ist sehr langgestreckt und stabförmig (Abb.84).

68

23) 2Ax: Fortsatz unter dem Körper des 1Ax (ordered)

0: Das 2Ax besitzt keinen lateralen Fortsatz, der unter den Körper des 1Ax ragt.

1: Der laterale Fortsatz des 2Ax reicht maximal bis zur Mitte unter den 1Ax-Körper (Abb.30B).
2: ... ragt bis jenseits der Mitte unter den Körper des 1Ax (Abb.56B).

24) 2Ax: caudaler Fortsatz

0: Der caudale Fortsatz des 2Ax ist relativ kurz. Er reicht nicht bis hinter das 1Ax.
1: ... ist so lang, daß er bis hinter das 1Ax reicht (Abb.15).

2: ... ist stark verlängert. Er überbrückt ca. ein Drittel der Distanz zum Sb.

25) 2Ax: proximale Ecke

0: Die proximale Ecke des 2Ax ist einfach gerundet (Abb.39).
1: ... ist in Richtung des 1Ax ausgezogen (Abb.60A).

2: ... ist sehr breit gerundet (Abb.19).

26) 3Ax: Länge des caudalen Arms

0: Der caudale Arm des 3Ax ist von mittlerer Länge (Abb.45).
1: ... ist deutlich verlängert (Abb.47).

2: ... ist verkürzt (Abb.39).

3: ... fehlt (Abb.73).

27) 3Ax: Form des proximalen Randes des Caudalarms

0: Der proximale Rand des Caudalarms verläuft nahezu gerade (Abb.28).
1: ... S-förmig geschwungen (Abb.8).

2: ... einfach geschwungen (Abb.39).

3: Der Caudalarm fehlt (Abb.73).

28) 3Ax: Kontakt zwischen dem caudalen Arm und dem PNP
0: Der Caudalarm und der PNP stehen fast nur über die Spitzen in Kontakt (Abb.37A).

1: ... berühren sich über eine längere Strecke (Abb.45).

2: Der Caudalarm fehlt (Abb.73).

29) 3Ax: Form des distalen Arms

0: Der distale Arm des 3Ax ist relativ schmal und am Ende abgerundet (Abb.34).
l: ... ist schmal und am Ende zugespitzt (Abb.8).

2: ... ist deutlich verbreitert (Abb.45).

3: ... ist als langer, gerader Stab ausgebildet (Abb.25A).

4: ... ist am Ende deutlich gegabelt (Abb. 19).

5: ... ist kurz abgestutzt (Abb.84).

6: ... fehlt (Abb.73).

30) 3Ax: Ausrichtung des distalen Arms

0: Der distale Arm des 3Ax ist mehr oder weniger direkt nach distal gerichtet (Abb.45).
l: ... weist deutlich nach schräg-vorne (Abb.32A).

2: ... weist fast direkt nach cranial (Abb.33A).

3: ... ist rückläufig orientiert (Abb.84).

4: ... fehlt (Abb.73).

31) 3Ax: Gesamtsklerotisierung

0: Das 3Ax besteht aus einem Element (Abb.8).

1: Das 3Ax ist in mehrere Elemente aufgelöst (Abb.79).
32) AMD

0: Eine AMD ist nicht vorhanden (Abb.79).

1: Die AMD ist rundlich geformt (Abb.45).

2: ... ist lang dreieckig (Abb.34).

3: ... kurz dreieckig (Abb.47).

4: ... schräg queroval (Abb.57A).

5: ... langlich oval oder leicht eckig (Abb.50A).

33) AMD: Abstand zum 3Ax
0: Eine AMD fehlt (Abb.79).

69

1: Die AMD liegt etwa auf halber Strecke zwischen dem 3Ax und dem Notum bzw. dem 1Ax (Abb.8).
2: Die AMD liegt sehr dicht beim 3Ax (Abb.47).

34) Medianplatten
Die Formvarianten der Medianplatten sind in den Schemata der Abbildung 6 dargestellt.
: siehe Abb.6A

: siehe Abb.6B

: siehe Abb.6C

: siehe Abb.6D

: siehe Abb.6E

: siehe Abb.6F

: siehe Abb.6G

: siehe Abb.6H

: siehe Abb.61

: siehe Abb.6J

: siehe Abb.6K
: siehe Abb.6L

35) Tegula
0: Eine Tegula ist vorhanden.
1: Die Tegula fehlt.

36) Ba: Gesamtform

0: Kopf und Stiel des Basalare sind nicht deutlich verschieden (Abb.83).

1: Es ist ein gegenüber dem Stiel deutlich erweiterter Kopf vorhanden (Abb.9).
2: Das Ba ist als Skleritspange ausgebildet (Abb.71).

37) Ba: Lage des frontalen Fortsatzes

0: Das Basalare trägt keinen frontalen Fortsatz

1: Der frontale Fortsatz des Ba ist nahezu senkrecht nach dorsal gerichtet (Abb.9).
2: ... ist schräg nach dorso-cranial gerichtet (Abb.33B).

3: ... zeigt fast waagerecht nach vorn (Abb.36).

D>vosaurwm Ho

38) Ba: Länge des frontalen Fortsatzes

0: Das Ba trägt keinen frontalen Fortsatz.
1: Der Frontalfortsatz ist kurz (Abb.59B).
2: ... lang (Abb.33B).

3: ... von mittlerer Lange (Abb.31B).

39) Ba: Lage des Rastknopfes am Ba-Kopf

0: Das Ba besitzt keinen Rastknopf (Abb.83).

1: Der Rastknopf des Ba liegt im hinteren Bereich des dorsalen Randes (Abb.75).
2: ... im mittleren Bereich des Dorsalrandes (Abb.40).

3: ... vorne am Dorsalrand (Abb.48).

4: ... nimmt fast die gesamte Lange des Ba-Kopfes ein (Abb.9).

40) Ba: Form des Rastknopfes

0: Das Ba besitzt keinen Rastknopf (Abb.83).

1: Der Rastknopf ist als einfache, mittelgroße Beule ausgebildet (Abb.48)

2: ... ist sehr klein (Abb.71).

3: ... ist lang und flach (Abb.20).

4: ... ist groß aufgeblaht (Abb.14).

5: ... ist als relativ große, schräg disto-ventrad gerichtete Zunge ausgebildet (Abb.40).
6: ... ist ähnlich geformt wie bei 5, aber schräg nach vorne-unten gerichtet (Abb.60B).
4
0
1
2

1) Fulcrum: Position

: Das Fulcrum liegt unter dem 2Ax.

: ... liegt zum Teil unter dem 1Ax, zum Teil unter dem 2Ax.
: ... liegt unter dem 1Ax (Abb.48).

42) Fulcrum: Lange
0: Der Quotient aus der Länge des 1Ax und der Länge des Fulcrum ist größer als 9 (Abb.19, 21A).

70

|: Der Quotient liegt zwischen 5,3 und 9 (Abb.32A, 32B).
2: ... zwischen 3,5 und 5,3 (Abb.45, 46).
3: ... ist kleiner als 3,5 (Abb.47, 48).

43) Fulcrum: Breite
0: Das Fulcrum ist deutlich verbreitert.
1: Das Fulcrum ist sehr schmal (Abb.45, 46).

44) Fulcrum: Form, von dorsal gesehen

0: Das Fulcrum ist etwa so breit wie lang (Abb.55A).

1: ... ist deutlich länger als breit (Abb.61A).

45) PWP

0: Der PWP ist leicht nach vorne geneigt und verläuft annähernd gerade (Abb.46).

1: ... ist unterhalb des Fulcrum deutlich geknickt (Abb.56B).

2: Der Hinterrand des PWP ist unterhalb des Fulcrum deutlich ausgelappt (Abb.33B).

46) Sb: Größe

0: Das Subalare besitzt ein Sechstel bis ein Drittel der Notumlänge (Abb.46).
1: ... ein Drittel der Notumlänge oder mehr (Abb.75).

2: ... weniger als ein Sechstel der Notumlänge (Abb.57B).

3: Es ist kein Subalare vorhanden.

47) Sb: Form

0: Das Subalare ist länger als hoch (Abb.75).
1: ... etwa so lang wie hoch (Abb.35).

2: ... höher als lang (Abb.56B).

3: Das Sb fehlt.

Die Analyse der Matrix (Tab.2) erfolgte mit dem Programm PAUP 3.1 (Swofford 1993).
Die Merkmale Nr.9 und 23 wurden als “ordered”, alle anderen als “unordered” kodiert. Die
Merkmale Nr. 6, 21 und 43 bilden nur Autapomorphien terminaler Taxa ab. Da sie somit
keine Informationen über die Verwandtschaftsverhältnisse liefern, wurden sie für die end-
gültige Berechnung ausgeschlossen, um die Baumlänge nicht künstlich zu vergrößern und
die Rechenzeit möglichst gering zu halten. Die Analyse wurde als heuristische Suche mit
dem Verfahren TBR (tree bisection-reconnection) durchgeführt, wobei die Ausgangsbäume
durch schrittweise Addition des jeweils nächst benachbarten Taxon ermittelt wurden.

Als Außengruppe wurden die Plecoptera vorgegeben. Die Merkmale für diese Gruppe wur-
den aus eigenen Beobachtungen an nicht näher bestimmten Individuen einer Art der Perlo-
didae und aus den Arbeiten von Onesto (1965) und Brodskiy (1979a, 1979b) gewonnen.
Die Merkmale der Mecoptera stammen aus eigenen Beobachtungen an Individuen von
Panorpa communis und Panorpa germanica sowie aus den Arbeiten von Mickoleit (1967,
1968, 1971). Die Daten für die Strepsiptera wurden aus der Revision von Kinzelbach
(1971) und eigenen Beobachtungen an Individuen von Elenchus sp. gewonnen. Ergebnis
der Suche waren 130 Bäume mit je 395 Schritten Länge und einem consistency index von
0,527. Aus diesen Bäumen wurde ein strict consensus tree berechnet, der in Abb.7
wiedergegeben ist (zur Angabe von Merkmalen in strict consensus trees siehe Nixon &
Carpenter 1996). Berechnungen mit anderen Methoden (SPR = subtree pruning-regrafting,
NNI = nearest neighbor interchange) und variierenden Startparametern ergaben weder
kürzere noch weitere gleich kurze Bäume.

Für die Entwicklung der Merkmale bietet das benutzte Programm zwei Optimierungs-
varianten: eine beschleunigte (ACCTRAN) und eine verzögerte (DELTRAN) Merkmals-
transformation. Je nach Optimierungsmethode kann sich die Anzahl der Autapomorphien,
die für einen Knoten ermittelt werden, ändern. Daraus resultieren die zum Teil nicht

71

eindeutigen Angaben für die Anzahl der Autapomorphien der einzelnen Taxa in der fol-
genden Analyse.

Coleoptera

Die obigen Annahmen zu den Verwandtschaftsverhältnissen der Teilgruppen der Coleoptera
werden von der Computeranalyse weitgehend bestätigt. Danach besitzen die Coleoptera
zwölf bzw. 13 Autapomorphien im Bereich der Flügelbasis:

- Der PNP ist lang und apikal verbreitert (Merkmal Nr. 4:1).

- Der Winkel ß zwischen den Achsen (a) und (b) des PNP (Abb.3) liegt zwischen 30° und
40°. Dies besagt, daß der PNP gegenüber dem Grundmuster der Holometabola (ß zwischen
6° und 13°) stark verlängert ist (Merkmal Nr. 5:0).

- Der Vorderrand des 1Ax-Kopfes ist umgeschlagen und zu einem nahezu senkrecht nach
unten gerichteten Fortsatz verlängert (Merkmal Nr. 9:1).

- Das Verhältnis zwischen Notumlänge und Länge des 1Ax beträgt zwischen 1,3 und 2.
Dies besagt, daß in Bezug auf das Grundmuster der Holometabola das 1Ax im Verhältnis
zum Notum an Größe zugenommen hat (Merkmal Nr. 16:3).

- Das Basiradiale ist vollständig sklerotisiert, schmal, und inseriert an der proximalen Ecke
des 2Ax (Merkmal Nr. 17:4).

- Der caudale Fortsatz des 2Ax ist verlängert. Diese Tatsache an sich ist wahrscheinlich
eine Synapomorphie der Coleoptera und Neuropterida (s. u.). Für die Coleoptera besteht
zum einen die Möglichkeit, daß sie den Zustand dieses Merkmals aus der gemeinsamen
Stammart übernommen haben und bei den Neuropterida eine weitere Verlängerung des
Fortsatzes erfolgte. Das heißt, der Merkmalszustand bei den Coleoptera wäre plesiomorph.
Zum anderen besteht die Möglichkeit, daß bei der gemeinsamen Stammart ein caudaler
Fortsatz des 2Ax in der Ausprägung der Neuropterida vorlag. Dieser wäre dann bei den
Coleoptera als Autapomorphie wieder verkürzt worden (Merkmal Nr: 24:1/2).

- Der proximale Rand des Caudalarms des 3Ax ist S-förmig geschwungen (Merkmal Nr.
DIT):

- In der Membran zwischen 3Ax und Notum bzw. 1Ax liegt eine Skleritplatte (AMD), an
der die Muskeln des 3Ax inserieren (Merkmal Nr. 32:1).

- Die Tegula fehlt (Merkmal Nr. 35:1).

- Das Basalare ist in einen schmalen Stiel und einen deutlich erweiterten Kopf untergliedert
(Merkmal Nr. 36:1).

- Das Basalare tragt einen frontalen Fortsatz, der mehr oder weniger senkrecht nach dorsal
gerichtet und relativ lang ist (Merkmale Nr. 37:1 und 38:2).

- Der Rastknopf des Basalare nimmt fast die gesamte Lange des Ba-Kopfes ein (Merkmal
Nr. 39:4).

Die Archostemata sind aufgrund von drei Autapomorphien als Monophylum ausgewiesen:

- Der Quotient aus der Lange des 1Ax und dem Abstand des Kontaktpunktes von ANP und
1Ax zum Vorderrand des 1Ax liegt zwischen 4 und 6,1. Dies besagt, daß der ANP relativ
weit hinten auf das 1Ax trifft. Urspriinglich liegt dieser Kontaktpunkt knapp hinter dem
Vorderrand des 1Ax (Merkmal Nr. 2:3).

- Die distale Ecke des 1Ax-Körpers ist gerade nach distal gerichtet und nicht, wie im
Grundmuster der Coleoptera und der Holometabola, caudad umgebogen (Merkmal Nr.13:1).

- Der Rastknopf des Basalare ist stark aufgetrieben und nimmt den gesamten Kopf des Ba
ein (Merkmal Nr.40:4).

72

Mecoptera
Tenebrionidae
Meloidae
Lymexylonidae
Staphylinidae
Silphidae
Lucanidae
Melyridae
Lampyridae
Hydrophilinae
Helophorinae
Elateridae
Dermestidae
Cantharidae
Scarabaeidae
1-4 Coccinellidae
Hispinae
Chrysomelidae part
19-23 Cerambycidae
Cleridae
Byrrhidae
Curculionidae
Buprestidae
Hydroscaphidae
Microsporidae
Micromalthidae
Cupedidae
Dytiscidae
Cicindelinae
5-9 A Carabidae part
Myrmeleonidae
Sialidae
Corydalidae
Raphidiidae
Strepsiptera
7 Plecoptera

10-18
24-26 29-32

Abb.7: Strict consensus tree von 130 gleich langen Bäumen aus der Analyse mit PAUP 3.1. Im
Folgenden ist jeweils die Merkmalsnummer und die Nummer der autapomorphen Auspragung ange-
geben (siehe Liste der Merkmale). Die in Klammern gesetzten Merkmale treten je nach Optimierungs-
methode der Merkmalstransformation (ACCTRAN oder DELTRAN) an unterschiedlicher Stelle als
Autapomorphie auf. 1-4: Holometabola, 6:1, 21:1, 39:1, 40:1. 5-9: Neuropterida + Coleoptera, 8:2,
16:1, (24:1/2), 41:1, 46:1. 10-18: Coleoptera, 4:1, 5:0, 9:1, 16:3, 17:4, (24:1/2), 27:1, 32:1, 35:1, 36:1,
37:1, 38:2, 39:4. 19-23: Myxophaga + Polyphaga, 2:2, (4:2), 5:2, 15:1, 27:2. 24-26: Archostemata, 2:3,
13:1, 40:4. 27-28: Adephaga, 3:1, 40:3. 29-32: Myxophaga, (4:2), 12:1, 26:3, 29:3. Zu Strepsiptera
siehe Diskussion.

Der strict consensus tree (Abb.7) enthält eine Trichotomie für die Taxa Archostemata,
Adephaga und Polyphaga + Myxophaga. In 100 der 130 Baume (77%) wird aber ein Taxon
aus Adephaga + Myxophaga + Polyphaga als Schwestergruppe der Archostemata durch vier
abgeleitete Merkmale unterstützt:

- Der Winkel ß zwischen den Achsen (a) und (b) des PNP (Abb.3) liegt zwischen 25° und
30°. Dies besagt, daß der PNP gegenüber dem Grundmuster der Coleoptera (ß zwischen
30° und 40°) leicht verkürzt ist (Merkmal Nr. 5:1).- Das 2Ax besitzt an seiner Ventralseite
einen lateralen Fortsatz, der unter den Körper des 1Ax ragt (Merkmal Nr. 23:1).

73

- Die AMD liegt dicht neben dem 3Ax (Merkmal Nr. 33:2).
- Das Fulcrum liegt unter dem 1Ax (Merkmal Nr. 41:2).

Die Adephaga besitzen zwei Autapomorphien:

- Der MNP ist kurz, breit und zweispitzig (Merkmal Nr. 3:1).

- Der Rastknopf des Basalare ist flach und erstreckt sich über die gesamte Länge des Ba-
Kopfes (Merkmal Nr. 40:3).

Das Taxon aus Myxophaga + Polyphaga ist durch vier bzw. fünf abgeleitete Merkmale aus-
gezeichnet:

- Der Quotient aus der Länge des 1Ax und dem Abstand des Kontaktpunktes von ANP und
1Ax zum Vorderrand des 1Ax liegt zwischen 6,1 und 8,8. Dies besagt, daß der Kon-
taktpunkt zwischen dem ANP und dem 1Ax gegnüber den Archostemata leicht nach vorne
verlagert wurde (Merkmal Nr. 2:2).

- Der PNP ist bei beiden Taxa kurz. Es besteht einerseits die Möglichkeit einer kon-
vergenten Verkürzung, was recht wahrscheinlich ist, da für die außerordentlich kleinen
Myxophaga eine derartige Reduktion aufgrund der Größe nicht ungewöhnlich wäre. Ande-
rerseits kann es sich um eine Synapomorphie beider Taxa handeln. Dann wäre für die Poly-
phaga (s.u.) nur das schmale Ende des PNP als Autapomorphie zu werten und nicht die
Verkürzung an sich. Die Stellung der Buprestidae als Schwestergruppe der Myxophaga im
consensus tree beruht ausschließlich auf der Form des PNP und dessen Verbindung zum
1Ax. Bei dieser Merkmalsausprägung handelt es sich aber mit sehr hoher Wahrscheinlich-
keit um eine konvergente Entwicklung bei Myxophaga einerseits und Buprestidae anderer-
seits (Merkmal Nr. 4:2).

- Der Winkel ß zwischen den Achsen (a) und (b) des PNP (Abb.3) liegt zwischen 13° und
25° (Merkmal Nr. 5:2).

- Der Winkel a zwischen der disto-cranialen Kante des 1Ax-Körpers und der Gelenkachse
von 1Ax und Notum ist kleiner als 50° (Merkmal Nr. 15:1).

- Der proximale Rand des Caudalarms des 3Ax verläuft leicht geschwungen (Merkmal Nr.
NE):

Die Myxophaga haben drei bzw. vier autapomorphe Merkmale:

- Der PNP ist kurz und apikal nicht verschmälert. Dies kann sowohl eine Autapomorphie
der Myxophaga sein, als auch eine Synapomorphie von Polyphaga und Myxophaga (s.0.).
Bei den Buprestidae tritt dieses Merkmal sehr wahrscheinlich konvergent auf (Merkmal Nr.
4:2).

- Die proximale Ecke des 1Ax reicht bis in die Bucht zwischen PNP und Notum. Auch in
diesem Merkmal liegt wahrscheinlich eine Konvergenz zu den Buprestidae vor (Merkmal
Ne2:1):

- Der caudale Arm des 3Ax fehlt (Merkmal Nr. 26:3).

- Der distale Arm des 3Ax ist als langer, gerader Stab ausgebildet (Merkmal Nr. 29:3).

Die Polyphaga besitzen sieben abgeleitete Merkmale:

- Der PNP ist kurz und schmal. Unabhängig davon, ob die Verkürzung des PNP eine Syn-
apomorphie von Myxophaga und Polyphaga ist, kann das schmale Ende des PNP als Aut-
apomorphie der Polyphaga gewertet werden (Merkmal Nr. 4:3).

- Der Winkel B zwischen den Achsen (a) und (b) des PNP (Abb.3) liegt zwischen 6° und
13° (Merkmal Nr. 5:3).

74

- Der Kopf des 1Ax besitzt an seinem Vorderrand einen mehr oder weniger waagerecht
distad weisenden Fortsatz (Merkmal Nr. 8:4).

- Der caudale Arm des dritten Axillare ist deutlich verlängert (Merkmal Nr. 26:1).

- Der Kontakt zwischen PNP und 3Ax wird über eine längere Strecke hergestellt (Merkmal
Nr. 28:1).

- Der Rastknopf des Ba sitzt als relativ große, schräg nach unten gerichtete Struktur vorne
am Kopf des Ba (Merkmale 39:3 und 40:6).

Für die Verwandtschaftsverhältnisse zwischen den Taxa der Polyphaga ergibt sich aufgrund
der hier verfügbaren Merkmale keine eindeutige Rekonstruktion. Die im strict consensus
tree beibehaltenen Schwestergruppenverhältnisse gründen sich haupsächlich auf Merkmals-
zustände, die wahrscheinlich konvergent entwickelt wurden, oder auf plesiomorphe Über-
einstimmungen. Die resultierenden Schwestergruppen sind bei Hinzuziehung weiterer Merk-
male (Crowson 1960, 1967, 1972, 1975, Beutel 1995, Browne & Scholtz 1995, Lawrence
& Newton 1995) größtenteils nicht aufrecht zu erhalten. Diese schlechte Auflösung inner-
halb der Polyphaga ist einerseits auf die relativ geringe Zahl an untersuchten Familien und
an untersuchten Vertretern der einzelnen Familien zurückzuführen. Daher kann nicht aus-
geschlossen werden, daß teilweise Merkmalsausprägungen in die Analyse eingegangen sind,
die charakteristisch für die Art oder die Gattung, nicht aber für das höherrangige Taxon
sind. Nach Arnett (1967) müßten für einen statistisch relevanten Überblick über die Coleo-
ptera, deren größten Teil die Polyphaga stellen, Arten aus ca. 500 Gattungen untersucht
werden, eine Zahl, die hier nicht einmal zu 20% erreicht wird. Andererseits müßten, um
abgeleitete Merkmale der recht nahe verwandten Polyphagentaxa zu finden, die Flügelge-
lenkstrukturen mit einer höheren Auflösung untersucht werden. Die Arbeiten von Browne
(1991), Browne, Scholtz & Kukalova-Peck (1993), Browne & Scholtz (1994, 1995, 1996),
Scholtz, Browne & Kukalova-Peck (1994) und Scholtz & Browne (1996) zur Systematik
der Scarabaeoidea zeigen, daß auch eine phylogenetische Analyse niederrangiger Taxa
aufgrund von Merkmalen der Flügelbasis möglich ist.

Für ein Taxon, die Buprestidae (Abb.39-41B), liegt wahrscheinlich trotz der gerade genann-
ten Einschränkungen eine autapomorphe Merkmalskombination vor. Die Form des 1Ax und
seine Lagebeziehung zum PNP und zum Notum sowie die Ausbildung des 2Ax und der
Medianplatten werden in dieser Form nur bei den Vertretern der Buprestidae gefunden.
Eine sehr ähnliche Ausprägung des 1Ax und seiner Lage zum PNP liegt auch bei den
untersuchten Vertretern der Myxophaga (Abb.25A, 25B) vor. Hier kann man aber, wie
oben erwähnt, von einer konvergenten Entwicklung ausgehen, die bei den Myxophaga mög-
licherweise mit der extremen Größenreduktion in Zusammenhang steht.

Ebenfalls relativ gut begründet erscheint das Schwestergruppenverhältnis zwischen Sta-
phylinidae (Abb.32A-33B) und Silphidae (Abb.28-31B), das auf die charakteristische Aus-
bildung des Basalarkopfes und des 3Ax zurückzuführen ist. Da aber keine weiteren Ver-
treter der Staphyliniformia untersucht wurden, sollte dies nur als Hinweis auf eine enge
Verwandtschaft der beiden genannten Taxa gewertet werden. Ein Schwestergruppenverhält-
nis zwischen Staphylinidae und Silphidae ist mit den hier zur Verfügung stehenden Merk-
malen nicht zu belegen, da es durchaus möglich ist, daß diese autapomorph für einen grö-
Beren Verwandtschaftskreis sind.

WS

Neuropterida

Die Taxa der Neuropterida werden im strict consensus tree als Polytomie mit den Coleo-
ptera und den Strepsiptera zusammengefaßt. In 112 der 130 gleichlangen Cladogramme
wird jedoch ein monophyletisches Taxon, welches die Neuropterida und die Strepsiptera
umfaßt, als Schwestergruppe der Coleoptera durch drei bzw. fünf Apomorphien unterstützt:
- Die proximale Ecke des 1Ax-Körpers ist verkürzt (Merkmal Nr. 12:2).

- Der Caudalrand des Körpers des 1Ax verläuft gerade (Merkmal Nr. 14:2).

- Der caudale Fortsatz des 2Ax ist stark verlängert. Bezüglich der Wertung dieses Merk-
mals als Autapomorphie oder Plesiomorphie sei auf die Diskussion von Merkmal Nr. 24:
1/2 weiter oben im Abschnitt zu den Coleoptera verwiesen.

- Das 3Ax ist nicht mehr als einheitlich sklerotisiertes Element ausgebildet, sondern in drei
Teilbereiche untergliedert (Merkmal Nr. 31:1).

- Das Fulcrum artikuliert sowohl mit dem 2Ax als auch mit dem 1Ax (Merkmal Nr. 41:1).
Für dieses Merkmal gibt es wieder zwei Interpretationsmöglichkeiten. Einerseits kann die
Verlagerung des Fulcrum unter das 1Ax bei Coleoptera und Neuropterida konvergent ent-
standen sein. Das würde bedeuten, daß die Archostemata, mit Ausnahme von Priacma, den
ursprünglichen Zustand der Gelenkung zwischen Fulcrum und 2Ax bewahrt haben. Ande-
rerseits ist denkbar, daß die zumindest teilweise Verlagerung des Fulcrum unter das 1Ax
ein abgeleitetes Merkmal der gemeinsamen Stammart von Coleoptera und Neuropterida ist.
Daraus ergäbe sich für die Archostemata, daß unter den untersuchten Taxa nur Priacma
den für die Coleoptera ursprünglichen Zustand bewahrt hat und die anderen Taxa sekundär
die Gelenkung wieder unter das 2Ax verlagert haben. Daß dies durchaus möglich ist, zei-
gen Ansätze zu einer solchen Rückverlagerung bei einigen Polyphaga, z.B. bei Cleridae
und Elateridae. Aus einer solchen Interpretation folgt zwangsläufig, daß Priacma die
Schwestergruppe der restlichen Archostemata ist.

Die Verwandtschaftsverhältnisse zwischen den Teilgruppen der Neuropterida werden in der
Computeranalyse nicht eindeutig aufgelöst. Zu den wahrscheinlichen Schwestergruppen-
verhältnissen sei auf die Rekonstruktion der Grundmuster weiter oben und die Abb.5 ver-
wiesen.

Strepsiptera

Die Position der Strepsiptera innerhalb der Neuropterida, wie sie in 112 der 130 kürzesten
Bäume angenommen wird, beruht ausschließlich auf Reduktionen der Gelenkfortsätze des
Notum (ANP, MNP und PNP) und der Position des Fulcrum unter dem 1Ax. Die Reduk-
tionen sind mit hoher Wahrscheinlichkeit als Konvergenzen erklärbar. Gleiches gilt auch
für die Lage des Fulcrum, zumal die Morphologie von Fulcrum, 1Ax und 2Ax deutlich von
der bei Coleoptera und Neuropterida verschieden ist. Die mehrmalige unabhängige Ver-
lagerung des Fulcrum unter das 1Ax ist durchaus nicht unwahrscheinlich, wie auch die
Existenz einer gemischten Gelenkung zwischen Fulcrum, 2Ax und 1Ax im Vorderflügel-
gelenk der Panorpidae (Abb.82) zeigt. Aufgrund der hier untersuchten Merkmale erscheint
eine nahe Verwandtschaft der Strepsiptera mit den Coleoptera und/oder den Neuropterida
als nicht sehr wahrscheinlich. Wie DNA-Analysen (Whiting et al. 1997) und Untersuchun-
gen der Flügeladerung (Whiting & Kathirithamby 1995) zeigen, kann ein Schwestergrup-
penverhältnis zwischen Coleoptera und Strepsiptera mit hoher Wahrscheinlichkeit ausge-
schlossen werden.

76

Neuropterida + Coleoptera

Das Taxon aus Neuropterida und Coleoptera wird durch vier bzw. fünf autapomorphe
Merkmale als Monophylum begründet:

- Der Kopf des 1Ax ist deutlich verbreitert und cranial abgestutzt (Merkmal Nr. 8:2).

- Der Quotient aus der Länge des Notum und der Länge des 1Ax liegt zwischen 3 und 3,8.
Dies bedeutet, daß das 1Ax, gegenüber dem plesiomorphen Zustand, relativ zum Notum
vergrößert ist (Merkmal Nr. 16:1).

- Das 2Ax trägt einen langen caudalen Fortsatz (Merkmal 24: 1/2).

- Wenn Priacma den Zustand aus dem Grundmuster der Coleoptera bewahrt hat, dann ist
die partielle Verlagerung des Fulcrum unter das 1Ax eine Synapomorphie der beiden Taxa
(siehe auch den Abschnitt zu Merkmal Nr. 41:1 unter Neuropterida).

- Das Subalare weist mindestens ein Drittel der Notumlange auf und ist damit gegentiber
dem plesiomorphen Zustand leicht vergrößert (Merkmal Nr. 46:1).

Holometabola

Als abgeleitete Merkmale der Holometabola konnen wenigstens drei Strukturen genannt
werden:

- Die Praealarbrücke ist so weit reduziert, daß sie keinen Kontakt mehr zum Episternum
hat (Merkmal Nr. 6:1).

- Das Basalare besitzt im hinteren dorsalen Bereich einen mittelgroßen, als einfache Beule
ausgebildeten Rastknopf (Merkmale Nr. 39:1 und 40:1).

- Im Zusammenhang mit dem zuvor genannten Merkmal steht die Ausbildung einer Auf-
nahme für den Rastknopf des Ba durch die Basis der Subcosta und den Humerus (Merkmal
NE Ae IP):

Einfluß der Körpergröße auf die Flügelgelenkstrukturen

Wie der Vergleich sehr unterschiedlich großer, aber wahrscheinlich relativ nah verwandter
Arten zeigt, hat die Körpergröße nur einen geringen Einfluß auf die Ausbildung der Ele-
mente des Flügelgelenks. Dies ist gut bei den Buprestidae (Abb.39 - 41B), Chrysomelidae
(Abb.62 A - 64B) und im Vergleich von Micromalthus (Abb.17) mit den anderen Archo-
stemata (Abb.8 bis 16B) zu sehen. Die hauptsächlichen Veränderungen des Flügelgelenks
bei Größenreduktion sind eine verringerte Sklerotisierung der Medianplatten, des Cau-
dalarms des 3Ax und des apikalen Bereichs des PNP. Ansonsten bleiben die Einzelele-
mente in der für das jeweilige Taxon charakteristischen Ausbildung erhalten.

Ursprung der Axillarsklerite

Die Untersuchung der Flügelgelenkelemente bei niederen Holometabola ergibt keinen
Widerspruch zu der z.B. schon von Snodgrass (1909) geäußerten Ansicht, daß die
Axillarsklerite einerseits aus einer Abspaltung des Notum (1Ax), andererseits aus
sekundären Sklerotisierungen der Flügelmembran (2Ax, 3Ax) hervorgehen. Dabei gibt es
bei rezenten Pterygoten (Brodskiy 1988, 1994) keine Hinweise auf eine Entstehung der drei
Axillarsklerite und der Medianplatten aus vielen kleineren Elementen, wie sie von
Kukalova-Peck (1991) für einen anzestralen Pterygotenflügel angenommen werden und die
teilweise auch innerhalb der Coleoptera noch erkennbar sein sollen (Kukalova-Peck &
Lawrence 1993). Auch Reduktionsstadien der Flügel und ihrer Gelenke (Abb.67) (Smith
1964, Geisthard 1974) liefern bei den Coleopteren keine Hinweise auf einen mehrteiligen

Hil

Ursprung der Axillarsklerite. Je nach Ausmaß der Reduktion verlieren die Medianplatten,
2Ax und 3Ax ihre typische Gestalt und Sklerotisierung, lösen sich aber nie in mehrere
Elemente auf. Das 1Ax verschmilzt schon bei relativ geringen Reduktionen mit dem
Notum, behält aber seine charakteristische Form auch bei extremer Rückbildung bei (Smith
1964, Geisthard 1974).

ZUSAMMENFASSUNG

In dieser Arbeit wurden die Skelettstrukturen der Basis der Hinterflügel verschiedener
Vertreter der Insekten untersucht. Zu diesem Zweck standen insgesamt 83 Arten der Coleo-
ptera, Neuropterida und weiterer Taxa der Holometabola und hemimetaboler Neoptera zur
Verfügung. Es wurden 67 Arten der Coleoptera bearbeitet. Davon entstammen fünf Arten
den Archostemata, sieben den Adephaga, zwei den Myxophaga und 53 den Polyphaga.
Neun Arten der Neuropterida konnten untersucht werden, von denen drei den Megaloptera,
zwei den Raphidioptera und vier den Planipennia angehören. Weitere vier Arten der
Holometabola und drei Arten der hemimetabolen Neoptera wurden zu Vergleichszwecken
herangezogen.

Um für die Rekonstruktion der Evolution des Flügelgelenks innerhalb der Holometabola
eine von abgeleiteten Merkmalen möglichst freie Außengruppe zur Verfügung zu haben,
wurde das Grundmuster des Flügelgelenks der Stammart der Neoptera aufgrund von eige-
nen Untersuchungen und anhand von Literaturangaben rekonstruiert. Darauf aufbauend
wurden die Entwicklung der Flügelbasis und die Verwandtschaftsverhältnisse bei den
Neuropterida und den Coleoptera sowie bei ihren Teilgruppen sowohl mental als auch mit
Hilfe des Computers analysiert.

Für die Computeranalyse wurde eine Matrix mit 36 höherrangigen Taxa und 47 Merkmalen
der Skelettelemente der Hinterflügelbasis zusammengestellt.

Sowohl aus der mentalen als auch aus der computergestützten Verwandtschaftsanalyse
ergaben sich einige durch Autapomorphien aus dem Bereich der Flügelbasis begründbare
monophyletische Taxa.

So sind die Holometabola durch drei abgeleitete Merkmale gekennzeichnet, von denen
besonders der Erwerb eines Arretierungsmechanismus zwischen den Basen von Costa und
Subcosta und dem Basalare zu nennen ist.

Ein monophyletisches Taxon aus Coleoptera und Neuropterida wird durch vier abgeleitete
Merkmale wahrscheinlich gemacht. Diese umfassen unter anderem ein gegenüber dem
Notum vergrößertes 1Ax und einen langen caudalen Fortsatz am 2Ax.

Die Coleoptera selbst sind, wie ihre vier Teilgruppen, die Archostemata, die Adephaga, die
Myxophaga und die Polyphaga, jeweils durch mehrere autapomorphe Merkmale als mono-
phyletisch ausgewiesen. Myxophaga und Polyphaga sind durch synapomorphe Merkmale
als Schwestergruppen gekennzeichnet. Ebenfalls durch Synapomorphien belegte
Schwestergruppenverhältnisse konnten für die Adephaga und das Taxon aus Polyphaga +
Myxophaga sowie für die Archostemata und das Monophylum aus Adephaga + Myxophaga
+ Polyphaga ermittelt werden. Micromalthus debilis ist durch die Struktur der Flügelbasis
als den Archostemata zugehörig charakterisiert.

Für die Neuropterida ergaben sich aus der Untersuchung ebenfalls mehrere abgeleitete
Merkmale. Als besonders auffällig sei die gegenüber dem Grundmuster der Holometabola

78

abgewandelte Gestalt des 1Ax genannt. Für die Teiltaxa der Neuropterida konnten durch
Autapomorphien begründbare Schwestergruppenverhältnisse zwischen den Megaloptera und
den Raphidioptera sowie zwischen den Planipennia und dem Taxon aus Megaloptera +
Raphidioptera ermittelt werden. Die beiden Teilgruppen der Megaloptera, die Sialidae und
die Corydalidae, sind ebenfalls durch abgeleitete Merkmale als Monophyla gekennzeichnet.
Für die Megaloptera insgesamt ließen sich allerdings keine Autapomorphien in den Struk-
turen der Flügelbasis ermitteln.

Da diese Untersuchung auf einer relativ geringen Anzahl von Taxa beruht, sind zur Über-
prüfung und Erweiterung der hier gewonnenen Ergebnisse weitergehende Arbeiten erfor-
derlich. Insbesondere aus den Neuropterida sollten möglichst viele Arten untersucht
werden. Auch eine Analyse der Verwandtschaftsverhältnisse innerhalb der Archostemata
würde helfen, noch offene Fragen zur Evolution des Flügelgelenks und zum hier postu-
lierten Schwestergruppenverhältnis zwischen Coleoptera und Neuroptera zu klären.

Aufgrund der in dieser Arbeit gewonnenen Erkenntnisse ist davon auszugehen, daß die
Untersuchung des Flügelgelenks bei den restlichen Taxa der Holometabola wesentliche
Beiträge zur Ermittlung der phylogenetischen Beziehungen zwischen diesen Gruppen leisten
kann.

ABSTRACT

This work deals with the skeletal structures of the hind wing base of insects. Eighty-three
species were available for investigation. Sixty-seven species of Coleoptera were examined:
five Archostemata, seven Adephaga, two Myxophaga, and 53 Polyphaga. Of the nine neu-
ropterid species, three belong to Megaloptera, two to Raphidioptera and four to Planipen-
nia. Four additional species of holometabolous taxa and three species of hemimetabolous
Neo-ptera were used for outgroup comparison.

The phylogenetic relationships between the taxa of Neuropterida and Coleoptera were ana-
lysed mentally as well as by computer (PAUP 3.1). For the cladistic analysis a matrix con-
sisting of 36 taxa and 47 characters of the hind wing base was composed.

For Holometabola three derived characters could be found. The most prominent of these
characters is a locking mechanism between the bases of costa and subcosta and the basa-
lare.

A sistergroup relationship between Coleoptera and Neuropterida is supported by four syn-
apomorphies. Among these are a first axillary which is greatly enlarged with respect to the
notum and a long posterior process of the second axillary.

The Coleoptera as a whole, as well as its four sub-groups Archostemata, Adephaga, Myxo-
phaga, and Polyphaga are each supported as monophyla by several autapomorphies.

Myxophaga and Polyphaga are classified as sistergroups. Sistergroup relationships are sup-
ported for Adephaga and the taxon consisting of Myxophaga + Polyphaga and for Archo-
stemata and the taxon comprising Adephaga + Myxophaga + Polyphaga. Micromalthus
debilis is characterized as a member of Archostemata by its wing base structures.

For Neuropterida a number of autapomorphies were found, too. Especially striking is the
form of the first axillary. Within the Neuropterida sistergroup relationships between
Megaloptera and Raphidioptera and between Planipennia and the taxon comprising Mega-
loptera + Raphidioptera are supported.

79

The size of the insect has only a subordinate influence on the structures of the wing base.
Investigation of closely related species of different size revealed that mainly the extent of
sclerotisation of the median plates, the caudal arm of the third axillary, and the apical area
of the posterior notal wing process is reduced in smaller animals. The form of the wing
base elements is not affected by body size.

The investigation of the wing base of holometabolous insects yielded no evidence against
the theory already mentioned by Snodgrass (1909), that the axillary sclerites originate as
split off from the notum (1Ax) and as newly formed elements of the wing membrane (2Ax,
3Ax). In recent pterygotes (Brodskiy 1988, 1994) nothing supports an origin of the three
axillaries and the median plates from a number of smaller elements, as postulated by
Kukalova-Peck (1991). According to Kukalova-Peck & Lawrence (1993) these original ele-
ments are recognizable in some extant Coleoptera. But even in Coleoptera with different
degrees of wing degeneration (Abb.67) (Smith 1964, Geisthard 1974) there is no indication
of an origin of the axillaries from smaller elements.

LITERATUR

Achtelig, M., & N.P. Kristensen (1973): A re-examination of the relationships of the Raphi-
dioptera (Insecta). - Z. zool. Syst. Evol.forsch. 11:268-274.

Adolph, G.E. (1879): Über Insektenflügel. - Nova Acta Leopoldina Carolinae 41:213-292.

Arnett, R.H. jr. (1967): Present and future systematics of the Coleoptera in North America. -
Ann. Ent. Soc. Am. 60:162-171.

Badonnel, A. (1934): Recherches sur l’anatomie des Psoques. - Bull. Biol. France Belgique, Suppl.
18, 241 S.

Baehr, M. (1975): Skelett und Muskulatur des Thorax von Priacma serrata Leconte (Col.:
Cupedidae). - Z. Morph. Tiere 81:55-101.

Bauer, A. (1910): Die Muskulatur von Dytiscus marginalis L. Ein Beitrag zur Morphologie des
Insektenkörpers. - Z. wiss. Zool., Abt. A 95:594-646.

Belkaceme, T. (1991): Skelet und Muskulatur des Kopfes und Thorax von Noterus laevis Sturm.
Ein Beitrag zur Morphologie und Phylogenie der Noteridae (Coleoptera: Adephaga). - Stuttgart.
Beitr. Naturk., Ser. A 462:1-94.

Betts, C.R. (1986): The comparative morphology of the wings and axillae of selected Heteroptera.
- J. Zool. (B) 1:255-282.

Beutel, R.G. (1986): Skelet und Muskulatur des Kopfes und Thorax von Hygrobia tarda (Herbst).
Ein Beitrag zur Klärung der phylogenetischen Beziehungen der Hydradephaga (Insecta: Coleoptera).
— Stuttgart. Beitr. Naturk. (A) 388:1-54.

- (1995): Phylogenetic analysis of Elateriformia (Coleoptera: Polyphaga) based on larval characters.
- J. Zool. Syst. Evol. Res. 33:145-171.

Brodskiy, A.K. (1979a): Evolution of the flight apparatus in Plecoptera. Part 1. Functional mor-
phology of the wings. - Ent. Rev. 58:31-36.

- (1979b): Evolution of the flight apparatus in Plecoptera. Part II. Functional morphology of the
axillary apparatus, the skeleton, and the musculature. - Ent. Rev. 58:16-26.

- (1986): Flight of the gigantic stonefly Allonarcys sachalina (Plecoptera, Pteronarcyidae) and the
study of the wing subination mechanism in insects. - Zool. Zhur. 65:349-360.

- (1987): Struktur und Funktion der Insektenflügel. - Trudy Vses. ent. Obsch., Akad. Nauk SSSR,
Leningrad 69:4-19 [russisch].

- (1988): Structure, functioning and evolution of the wing articulation in insects. - Tschtenuja Pam.
Nikolaja Aleksandreovitscha Cholodkovskovo 78:3-47 [russisch].

80

Brodskiy, A.K. (1992): Structure, function, and evolution of the terga of wing-bearing segments
of insects. II. Organizational features of the terga in different orders of insects. - Ent. Rev. 71:8-28.

- (1994): The evolution of insect flight. - Oxford Sci. Publ., Oxford University Press, Oxford, New

York, Tokyo, 229 S.

Browne, D.J. (1991): Wing structure of the genus Eucanthus Westwood; confirmation of the primi-
tive nature of the genus (Scarabaeoidea: Geotrupidae: Bolboceratinae). - J. ent. Soc. South. Africa
54:22 1-230.

Browne, D.J., & C.H. Scholtz (1994): The morphology and terminology of the hindwing
articulation and wing base of the Coleoptera, with specific reference to the Scarabaeoidea. - Syst.
Ent. 19:133-143.

- & - (1995): Phylogeny of the families of Scarabaeoidea (Coleoptera) based on characters of the

hindwing articulation, hindwing base and wing venation. - Syst. Ent. 20:145-173.
- & - (1996): The morphology of the hind wing articulation and wing base of the Scarabaeoidea
(Coleoptera) with some phylogenetic implications. - Bonn. zool. Monogr. 40; 200 S.
Browne, D.J., C.H. Scholtz & J. Kukalova-Peck (1993): Phylogenetic significance of
wing characters in the Trogidae (Coleoptera: Scarabaeoidea). - Afr. Ent. 1:195-206.
Comstock, J.H., & J.G. Needham (1898): The wings of insects. - Am. Nat. 32:43-48, 81-89,
231-257, 335-340, 413-424, 561-565, 769-777, 903-911.
- & - (1899): The wings of insects. - Am. Nat. 33:117-126, 573-582, 845-860.
Crampton, G.C. (1914): The ground plan of a typical thoracic segment in winged insects. - Zool.
Anz. 44:56-67.

- (1918): A phylogenetic study of the terga and wing bases in Embids, Plecoptera, Dermaptera, and
Coleoptera. - Psyche, Cambridge 25:4-12.

- (1919): Phylogenetic study of the mesothoracic terga and wing bases in Hymenoptera, Neuroptera,
Mecoptera, Diptera, Trichoptera and Lepidoptera. - Psyche, Cambridge 21:58-64.

Crowson, R.A. (1960): The phylogeny of Coleoptera. - Ann. Rev. Ent. 5:111-134.

- (1967): The natural classification of the families of Coleoptera. - Nathaniel Lloyd, London. 214 S.
- (1972): A review of the classification of Cantharoidea (Coleoptera), with the definition of two new

families, Cneoglossidae and Omethidae. - Rev. Univ. Madrid 21(82):35-77.

- (1975): The evolutionary history of Coleoptera, as documented by fossil and comparitive evidence.

- Atti X Congr. Nazionale Ent. Sassari, Italia: 47-90.

Czihak, G. (1953/1954): Beiträge zur Anatomie des Thorax von Sialis flaviventris L. - Österr. zool.
Z. 4:421-448.

Doyen, J.T. (1965): The skeletal anatomy of Tenebrio molitor (Coleoptera: Tenebrionidae). - Misc.
Publ. Ent. Soc. Am. 5:102-150.

Emelianov, A.F. (1977): Homologe Strukturen im Flügel von Zikaden und primitiven
Polyneoptera. - Trudy Vses. ent. Obsch., Akad. Nauk SSSR, Leningrad 58:3-48.

Ennos, A.R. (1987): A comparitive study of the flight mechanism of Diptera. - J. exp. Biol.
127235323272:

Farris, J.S. (1982): Outgroups and parsimony. - Syst. Zool. 31:328-334.

Ferris, G.F. (1940): The morphology of Plega signata (Hagen) (Neuroptera: Mantispidae). -
Micro-entomology 5:33-56.

Ferris, G.F., & P. Pennebaker (1939): The morphology of Agulla adnixa (Hagen)
(Neuroptera: Raphidiidae). - Microentomology 4:121-142.

Forbes, W.T.M. (1922): The wing-venation of Coleoptera. - Ann. Ent. Soc. Am. 15:328-352.
- (1926): The wing-folding patterns of the Coleoptera. - J. New York Ent. Soc. 34:42-68, 91-139.
- (1943): The origin of wings and venational types in insects. - Am. Midl. Nat. 29:381-405.

Geisthardt, M. (1974): Das thorakale Skelett von Lamprohiza splendidula (L.) unter besonderer

Berücksichtigung des Geschlechtsdimorphismus (Coleoptera: Lampyridae). - Zool. Jb. (Anat.)
93:299-334.

81

Hamilton, K.G.A. (1972a): The insect wing, Part III. Venation of the orders. - J. Kansas Ent.
Soc. 45:145-162.

- (1972b): The insect wing, Part IV. Venational trends and the phylogeny of the winged orders. - J.
Kansas Ent. Soc. 45:295-308.

Hennig, W. (1950): Grundzüge einer Theorie der Phylogenetischen Systematik. Deutscher Zentral-
verlag, Berlin, 370 S.

- (1966): Phylogenetic Systematics. University of Illinois Press, Urbana, 246 S.

- (1969): Die Stammesgeschichte der Insekten. W. Kramer, Frankfurt a.M., 436 S.

- (1982): Phylogenetische Systematik. P. Parey, Berlin, Hamburg, 246 S.

Ivanov, V.D. (1995): Comparative analysis of wing articulation in archaic Lepidoptera. - Ent. Rev.
74:32-53.

Keler, S. von (1963): Entomologisches Wörterbuch. Akademie verlag, Berlin, 774S.

Kelsey, C.P. (1957): The skeleto-motor mechanism of the dobsonfly Corydalus cornutus 2: Ptero-
thorax. - Mem. Cornell Univ. agr. Exp. Stn. 346:1-42.

Kinzelbach, R. (1971): Morphologische Befunde an Fächerflüglern und ihre phylogenetische
Bedeutung (Insecta: Strepsiptera). - Zoologica 41:1-256.

Klausnitzer, B. (1975): Probleme der Abgrenzung von Unterordnungen bei den Coleoptera. - Ent.
Abh. 40:269-275.

Kleinow, W. (1966): Untersuchungen zum Fliigelmechanismus der Dermapteren. - Z. Morph. Okol.
Tiere 56:363-416.

Korn, W. (1943): Die Muskulatur des Kopfes und des Thorax von Myrmeleon europaeus und ihre
Metamorphose. - Zool. Jb. (Anat.) 68:273-331.

Kristensen, N.P. (1981): Phylogeny of insect orders. - Ann. Rev. Ent. 26:135-157.

Kukalova-Peck, J. (1983): Origin of the insect wing and wing articulation from the arthropodan
leg. - Can. J. Zool. 61:1618-1669.

- (1991): Fossil history and the evolution of hexapod structures. - In: C.S.I.R.O (ed.), Insects of
Australia, 2nd ed., Vol. 1., Melbourne University Press, Melbourne, 141-179.

Kukalova-Peck, J., & J.F. Lawrence (1993): Evolution of the hind wing in Coleoptera. -
Can. Ent. 125:181-258.

Larsen, O. (1966): On the morphology and function of the locomotor organs of Gyrinidae and other
Coleoptera. - Opusc. Ent., Suppl. 30, 242 S.

Lawrence, J.F., & A.F. Newton (1982): Evolution and classification of beetles. - Ann. Rev.
Ecol. Syst. 13:261-290.

- & - (1995): Families and subfamilies of Coleoptera. - In: Pakaluk, J., & S.A. Slipinski: Biology,
phylogeny, and classification of Coleoptera: 779-1006.

Maki, T. (1936): Studies of the skeletal structure, musculature, and nervous system of the alderfly
Chauliodes formosanus Petersen. - Mem. Fac. Sci. Agric. Taihoku Imp. Univ. 16:117-243.

- (1938): Studies of the thoracic musculature of insects. - Mem. Fac. Sci. Agric. Taihoku Imp. Univ.
24: 1-343.

Malicky, H. (1973): Trichoptera (Köcherfliegen). - Handbuch der Zoologie 4, 2-2/29, 187 S.

Matsuda, R. (1970): Morphology and evolution of the insect throax. - Mem. Ent. Soc. Can. 76:3-
431.

Mickoleit, G. (1967): Das Thorakalskelett von Merope tuber Newman (Protomecoptera). - Zool.
Jb. (Anat.) 84:313-342.

- (1968): Zur Thoraxmuskulatur der Bittacidae. - Zool. Jb. (Anat.) 85:386-410.

- (1969): Vergleichend anatomische Untersuchungen an der pterothorakalen Pleurotergalmuskulatur
der Neuropteria und Mecopteria (Insecta: Holometabola). - Z. Morph. Tiere 64:151-178.

- (1971): Das Exoskelett von Notiothauma reedi MacLachlan, ein Beitrag zur Morphologie und
Phylogenie der Mecoptera (Insecta). - Z. Morph. Tiere 69:318-362.

- (1973a): Über den Ovipositor der Neuropteroidea und Coleoptera und seine phylogenetische
Bedeutung (Insecta: Holometabola). - Z. Morph. Tiere 74:37-64.

82

Mickoleit, G. (1973b): Zur Anatomie und Funktion des Raphidiopteren-Ovipositors (Insecta,
Neuropteroidea). - Z. Morph. Tiere 76:145-171.

Miller, F.W. (1933): Musculature of the lacewing Chrysopa plorabunda (Neuroptera). - J. Morph.
39.2951.

Nixon, K.C., & J.M. Carpenter (1993): On outgroups. - Cladistics 9:413-426.

- & - (1996): On consensus, collapsibility, and clade concordance. - Cladistics 12:305-321.

Onesto, E. (1959a): Morfologia della regione articolare delle ali di Anthocaris cardamines (L.)
(Lepi-doptera, Pieridae). - Ann. Inst. Mus. Zool. Univ. Napoli 11,1:1-40.

- (1959b): La regione articolare delle ali di Blatta orientalis L. (Insecta, Blattodea). - Ann. Inst. Mus.
Zool. Univ. Napoli 11,6:1-40.

- (1960): La regione articolare di Ameles decolor (Charp.) (Insecta, Mantoidea) e il suo significato
morfologico per la filogenesi di Blattotteroidei. - Ann. Inst. Mus. Zool. Univ. Napoli 12:1-27.

- (1961): Morfologia della regione articolare alare di Forficula auricularia L. (Insecta, Dermaptera).
- Ann. Inst. Mus. Zool. Univ. Napoli 13:1-31.

- (1963): Studio sulla morfologia comparata del dermascheletro dei Tettigonioidei. - Ann. Inst. Mus.
Zool. Univ. Napoli 15:1-38.

- (1965): Morfologia della regione articolare alare e delle pleure nei Plecotteri. - Boll. Soc. Nat.
Napoli 74:22-39.

Orchymont, A. d’ (1920): La nervation alaire des Coleopteres. - Ann. Soc. Ent. France 89: 1-50.

- (1921): Apercu de la nervation alaire des Coleopteres. - Ann. Soc. Ent. Belgique 61:256-278.

Paulus, H.F. (1986): Comparative morphology of the larval eyes of Neuropteroidea. - In: Gepp
et al.: Recent Research in Neuropterology, Graz: 157-164.

Pfau, H.K. (1977): Zur Morphologie und Funktion des Vorderflügels und Vorderflügelgelenks von
Locusta migratoria L. - Fort. Zool. 24:341-345.

- (1986): Untersuchungen zur Konstruktion, Funktion und Evolution des Flugapparates der Libellen
(Insecta, Odonata). - Tijdschrift Ent. 129:35-123.

- (1991): Contributions of functional morphology to the phylogenetic systematics of Odonata. - Adv.
Odonatology 5:109-141.

Ponomarenko, A.G. (1969): Historical development of the Coleoptera-Archostemata. - Trudy
Pal. Inst. 125:1-240.

- (1971): The geological history and evolution of beetles. - Proc. Int. Congr. Entomol., 13th, Moscow
1: 281.

- (1972): On the nomenclature of the wing-venation in Coleoptera. - Ent. Rev. 51:454-458.
Pringle, J.W.S. (1976): The muscles and sense organs involved in insect flight. 3-15 - In: Rainey,

R.C. (ed.): Insect Flight - Symposia of the Royal Entomological Society of London 7, Blackwell
Scientific Publications, Oxford u.a., 287 S.

Redtenbacher, J. (1886): Vergleichende Studien über das Flügelgeäder der Insekten. - Ann.
kaiserl.-königl. nat. Hofmus. 1:153-232.

Roger, O. (1875): Das Flügelgeäder der Kafer, Erlangen, 17 S.
Romeis, B. von (1968): Mikroskopische Technik. Oldenbourg, München, 757 S.

Schneider, P. (1978): Die Flug- und Faltungstypen der Käfer (Coleoptera). - Zool. Jb. (Anat.)
99:174-210.

- (1987): Mechanik des Auf- und Abschlages der Hinterflügel bei Kafern (Coleoptera). - Zool. Anz.
218:25-32.

Scholtz, C.H., & D.J. Browne (1996): Polyphyly in the Geotrupidae (Coleoptera:
Scarabaeoidea): a case for a new family. - J. Nat. Hist. 30:597-614.

Scholtz, C.H., D.J. Browne & J. Kukalova-Peck (1994): Glaresidae, archaeopteryx of
the Scarabaeoidea (Coleoptera). - Syst. Ent. 19:259-277.

Sharplin, J. (1963): Wing base structure in Lepidoptera. I. Fore Wing Base. - Can. Ent. 95:1024-
1050.

83

Smith, D.S. (1964): The structure and development of flightless Coleoptera: A light and electron
microscopic study of the wings, thoracic exoskeleton and rudimentary flight musculature. - J.
Morph. 114:107-184.

Snodgrass, R.E. (1908): A comparative study of the thorax in Orthoptera, Euplexoptera and
Coleo-ptera. - Proc. Ent. Soc. Washington 9:95-108.

- (1909): The thorax of insects and articulation of wings. - Proc. U.S. Nat. Mus. 36:511-595.

- (1911): The thorax of the Hymenoptera. - Proc. U.S. Nat. Mus. 39:37-91.

- (1927): Morphology and mechanism of the insect thorax. - Smith. Misc. Coll. 80:1-108.

- (1935): Principles of insect morphology. McGraw-Hill, New York, N.Y., 667 S.

Stellwaag, F. (1914): Der Flugapparat der Lamellicornier. - Z. wiss. Zool. 108:359-429.

Swofford, D.L. (1993): PAUP phylogenetic analysis using parsimony Version 3.1. - Laboratory
of Molecular Systematics, Smithsonian Institution.

Tannert, W. (1958): Die Flügelgelenkung bei Odonaten. - Dtsch. Ent. Z. (N.F.) 5:394-455.

Taylor, L.H. (1918): The thoracic sclerites of Hemiptera and Heteroptera. - Ann. Ent. Soc. Am.
11:225-254.

Wallace, F.L. (1971): A comparative morphological study of the wing venation of the order
Coleoptera. - Dissertation, Clemson University, 126 S.

Wallace, F.L., & R.C. Fox (1975): A comparative morphological study of the hindwing
venation of the order Coleoptera, Part I. - Proc. Ent. Soc. Washington 77:329-354.

Watrous, L.E., & Q.D. Wheeler (1981): The out-group comparison method of character ana-
lysis. - Syst. Zool. 30:1-11.

Weber, H. (1924a): Das Grundschema des Pterygotenthorax. - Zool. Anz. 60:17-37 + 57-83.

- (1924b): Das Thoraxskelett der Lepidopteren. - Z. Anat. Entw.gesch. 23:277-331.

- (1926): Der Thorax der Hornisse. - Zool. Jb. (Anat.) 47:1-100.

Waiting aM AL Me Carpenter, O©-D- Wheeler & W.-C, Wheeler (1997): The Strepsi-
ptera problem: Phylogeny of the holometabolous insect orders inferred from 18S and 28S ribosomal
DNA sequences and morphology. - Syst. Biol. 46:1-68.

Whiting, M.F., & J. Kathirithamby (1995): Strepsiptera do not share hind wing venational
syn-apomorphies with Coleoptera: a reply to Kukalova-Peck and Lawrence. - J. New York Ent. Soc.
103:1-14.

Wootton, R.J. (1979): Function, homology and terminology in insect wings. - Syst. Ent. 4:81-93.

1L0°0
OSTO
0LTO
LOT'0

910
ele 0
bilo
9S1°0
6700
8010
1110
0STO
CLI‘
860°0
$900
9€0°0
£s00

(e’qqV)

(e’gqV)
(q) pun (e)

windJo[n
sap dur

XVI Sap
ogue7

sııvaun SJosuaq
‘ds vixnyjuy
pupııpu DA0ydosjpyIJ

‘ds snysidg

D]091110y Dy110dojKyd
wuma11pun]KkD UOAPUapOUlY
‘ds smpon®

SNULINUL SAaJSaJOYIUE
sapiojpidsaa snsoydosrin
povdo vsvydojg
DIIIVAOY] DULOJdOIIAQ
‘ds snsoydojayy

snaoid snpiydospay
SYDUIBADIU SNISIAG

‘ds papuy

‘ds snjodivy

pjpjnun] DJopum1)

‘ds snsodsosolp

‘ds pydvosospay

SIigap SnyPWwoAasıy
AOJOIUOI VSAIUIOUAT
SUDLIDA SadndojSig
DIDAMIS DWOIDIAT

snjojidvs sadng

seplayelq
oepnsoidng
oephsoidng
SepıyuAg
Jeuoıny
DepIoegeieag
JepıuesnT
orptuAyderg
septuAyderg
sepiydiis
sepiydiis
oeprydis
oepriydoupAH
SepıjıydoipAH
JepIIsNAq
Sepıgeie)
aepiqeled
aepiqeied
JepL1odsonmı
oepiydessoipAH
SEPIYIEWOASIN
Srpıpadny
oepipodny
oepipednyg

aepipedny

egeydAjod
eseyddjog
eseyddjog
eseyddjog

eseyddjog
e3eydAjod
edeydAjod
eseyddAlog
eseyddjog
eseyddlog
edeyddjog
eseyddjog
edeydAjod
eseydapy
veseydopy
eseydapy
vseydopy
egeydoxAW
eseydoxAW
BILWO]SOYIIW
BILUDJSOYIIV
BILUIDJSOYIIY
BILUDJSOYIIV

EIEWOSISOUIIV

219Jd03]09
819Jd03]09
219Jd02]09

819Jd03]09

219Jd03]09
e19Jd03]09
e19Jd03[09
e19Jd03]09
e19Jd09[09
819Jd03[09
e.19)d03]09
e19Jd03]09
e19)d03]09
819Jd03[09
e19Jd03]09
e19)d02]09
e19Jd02]09
e19Jd03[09
2.19Jd03]09
e19)d09]09
219)d03]09
e19)d02]09
elajdoajop

e.19Jd09]09

84

(e’qqv)
0 [OYUN

x MPOHS Jop o8 | uasyoy usyas |/ x yI|/ osuery]
-uvy] / XY] osuReyT | -IMz g joyurM | Sop osueq | -wnjoKN

(€QqV) dNd Pun XV] UOA assıuypey1>Ausdur PUN [OYULA 2] Eq

85

(e’qqV)
x MOOUS Jop 98

-up] / xy] due

(e-qqv)
(q) pun (e)
uasyoy Usyos
-1mz J [OYUN

wngo[ng
Sop due
EN
sap due

XVI Sep
asuey
/ 23uej
-unJON

zw ds smgopkyd

[ds sniqojjdyd

‘ds ppisspg

DJvauyuioap DS.ıD10uUNd2T
yndod vjawosksy)

18DADASD S119I011J
suadsaplslaosojjia piyJupdpsy

sSmoL1D SNIÄI

4omow O1qaua]
DIAOJDIISAA DYÄT
pypjounduajdas DJJ2U1290J
pipyndundep-ıononb DIAIDI
Saplojsausap SNI2093JXH
‘ds sn1yovjvw

SNLIDINUAOL SNULISDUDY
‘ds sopoy>14]

SNIADPAD] SISOW.IAT
SUDIIUSIU SIADYJUDD
pjnpipuajds pzıyoAdupT
snyasopid sajoisay

4asiu snipidainuay

(€-qqy)
0 [ULM

seulysuAYIONO

:oepluolpNoing

dBUIYySUAYIONC

:9epIuol[n91n)y

SePIPWOSAIYI
depljawiosAIyD
aepljawosdAlyD
aepljawiosAlyD

OeULIWIe'T

roeproAquielag

aeuloXquielad

:aeproAquielag

JepIuonIgausL
SepIooWN
JepI[[PUII0J
JepI[[PUIII0J
SepIuoJAxaWwAT
VePLIAOW
ED)
UIID)
JePNSSULIOA
SepLıemue)
sepııAdue]
JepL1oje[g
SepLıojelg

egeydAjod

eseyddjog
eseyddjog
eseyddjog
eseyddjog
eseyddjog

eseyddjog

eseyddjog
edeydAjod
edeydAjod
edeydAjod
egeydAjod
edeydAjod
edeydAjod
edeydAjod
eseyddjog
eseyddjog
eseyddjog
eseyddjog
egeydAjoq
edeydAjod

819Jd03[09

e19Jd03[09
819Jd03]09
819Jd03[09
e19Jd03[09

819)d03[09
e19Jd02[09

819Jd09]09
219Jd03]09
819Jd03]09
219}d03]09
819Jd03[09
e19}d09]09
819Jd03[09
219Jd03[09
e19Jd02[09
e19Jd03[09
e19Jd03]09
819Jd03]09
819Jd03[09

e19Jd03]09

(£'qqV) dNd Pun xy] UOA 9ossıuypeyssausgue pun [oyurM :] BJoqey, Junzyosuog

86

(e’qqV)
x dydoNsS Joep 93 | uasyoy uayas

-ur] / XVI >3ue7

(e’gqV)
(q) pun (e)

-IMZ [OYUN

wnJojn
sop osury
Na]
Sop adue

XVI Sop
o3ue7
/ »3ue]

-WN}JON

sisdoiydo pıpıydpy
DXIUPD DIINSY

snjnusod snjppksog
SIUAOIIAJSDA SAPOINDBYD
DIUDIN] SIDI

14a1aq DJaND

‘ds snupydo.sojy)

(€-qqV)
0 [um

sepıpıydey
sepuıpıydey
JepıfepA107
JepıfepA107
epIfels
JePIUOSJSULIÄN
eumowAue]

soepluolnoing

elajdorpiydey
eıo}doipıydey
eısYdojegaW
elajdojesayy
eısydojedaw

eruusdiueig

egeydAjoq

epuisJdommsN
epusydomsaN
epusydomaN
epuioydommsN
epusydomsN
epliajdoinoan

elajydoajod

(€qqV) dNd pun xy] UOA 9ssıuyjeyssausdur] pun [OYUIM :] B[oqey, Funzyosuog

die Verwandtschaftsanalyse mit PAUP 3.1.

für

Datenmatrix

Tabelle 2

Einträge mit mehreren Merkmalszuständen werden als Polymorphien gewertet.

Chrysomelidae part.
Cerambycidae

Cleridae
Micromalthidae

Myrmeleonidae
Hydroscaphidae
Microsporidae
Cupedidae
Carabidae part.
Strepsiptera |

Lymexylonidae
Sialidae

Lampyridae
Hydrophilinae
Helophorinae
Curculionidae

Elateridae
Coccinellidae

Plecoptera
Mecoptera
Tenebrionidae
Staphylinidae
Silphidae
Meloidae
Melyridae
Dermestidae
Cantharidae
Buprestidae
Lucanidae
Scarabaeidae
Hispinae
Byrrhidae
Corydalidae
Raphidiidae
Dytiscidae
Cicindelinae

87

88

Fortsetzung Tabelle 2: Datenmatrix für die Verwandschaftsanalyse mit PAUP 3.1.

Einträge mit mehreren Merkmalszuständen werden als Polymorphien gewertet.

Ae

d

Chrysomelidae part.
siptera

Cerambycidae

Cleridae
Micromalthidae

Cupedidae
Carabidae part.

Myrmeleonidae
Hydroscaphidae
Microsporidae
Strep

Lymexylonidae
Sialidae

Plecoptera
Mecoptera
Tenebrionidae
Staphylinidae
Silphidae
Meloidae
Melyridae
Lampyridae
Hydrophilinae
Helophorinae
Elateridae

| Dermestidae
Buprestidae
Lucanidae
Scarabaeid
Coccinellidae
Hispinae
Byrrhidae
Curculionidae
Corydalidae
Raphidiidae
Dytiscidae
Cicindelinae

89

PRA

0.5 mm
8 i wm

as

Abb.8: Priacma serrata (Archostemata: Cupedidae). Linkes Hinterflügelgelenk von dorsal.

9

Abb.9: Priacma serrata (Archostemata: Cupedidae). Metathorax links, von lateral. Flügel und
Axillarsklerite entfernt.

90

0,25 mm

10

Abb.10: Priacma serrata (Archostemata: Cupedidae). Metathorax links, von latero-ventral.

11 0,25 mm

Abb.11: Distocupes varians (Archostemata: Cupedidae). Linkes Hinterflügelgelenk von dorsal.

91

12

0,25 mm

Abb.12: Distocupes varians (Archostemata: Cupedidae). Metathorax links, von dorsal. Flügel und
Axillarsklerite entfernt.

13 0.25 mm

Abb.13: Tenomerga concolor (Archostemata: Cupedidae). Linkes Hinterflügelgelenk von dorsal.

14 15

Abb.14-15: Tenomerga concolor (Archostemata: Cupedidae). 14: Metathorax links, lateral. Flügel und
Axillarsklerite entfernt. 15: Linkes Hinterflügelgelenk von dorso-lateral. Flügel und 3Ax entfernt.

0,25 mm

:
N \ —MNP

16
Abb. 16 A, B: Cupes capitatus (Archostemata: Cupedidae). A: Linkes Hinterflügelgelenk von dorsal.

B: Metathorax links, lateral. Flügel und Axillarsklerite entfernt.

Abb. 17: Micromalthus debilis (Archostemata: Micromalthi-
dae). Linkes Hinterflügelgelenk von dorsal.

93

1mm

18 B

Abb.18A,B: Dytiscus marginalis (Adephaga: Dytiscidae). A: Linkes Hinterfliigelgelenk von dorsal. B:
Metathorax rechts, von lateral. Fliigel und Axillarsklerite entfernt.

Abb.19: Cicindela lunulata (Adephaga: Carabidae). Linkes Hinterflügelgelenk von dorsal.

1Ax

2AXx

Abb.20: Cicindela lunulata (Adephaga:
Carabidae). Metathorax links, von
ventro-lateral. Flügel und 3Ax entfernt.

20

21

0,5 mm OLS mm

Abb.21A,B: Cicindela lunulata (Adephaga: Carabidae). A: Metathorax links, von lateral. Flügel und
Axillarsklerite entfernt. B: Rechtes Hinterflügelgelenk frontal. Flügel in Ruhelage.

95

Abb.23: Amara sp. (Adephaga: Carabidae).
Metathorax rechts, von lateral. Flügel und
23 Axillarsklerite entfernt.

Abb.24: Harpalus sp. (Adephaga: Carabidae). Linkes Hinter-
flügelgelenk von dorsal.

96

5 En un

Abb.25A,B: Rechtes Hinterflügelgelenk von Vertretern der Myxophaga von dorsal. A: Hydroscapha
sp. (Hydroscaphidae). B: Microsporus sp. (Microsporidae).

BR

1mm
A

26

Abb.26A-C: Hydrophilus piceus (Polyphaga: Hydrophilidae). A: Linker Hinterflügel von dorsal. B: |
Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax frontal. |

97

FE PNP

PWP

Ba

0,25 mm

Abb.27A,B: Helophorus sp. (Polyphaga: Hydrophilidae). A: Linkes Hinterflügelgelenk von dorsal. B:
Metathorax links von lateral. Flügel und Axillarsklerite entfernt.

28

Abb.28: Nicrophorus vespilloides (Polyphaga: Silphidae). Linkes Hinterflügelgelenk von dorsal.

98

F
AMD
Cc
eu a PN
PWP
BaRK 4
Ba Sb

0,5 mm mG

Abb.29A-D: Nicrophorus vespilloides (Polyphaga: Silphidae). A:
Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. B:
Linkes Hinterflügelgelenk von dorsal. Flügel und Axillarsklerite
entfernt. C: 1Ax frontal. D: Ausschnitt des Metathorax links, von
lateral. Flügel und 3Ax entfernt.

IE BSc > & bat
( x
BR
S | as
\ 2Ax N =
Se J
# Sa fy MNP
N
AMD
So \
A =
\
N
30 0,5 mm

Abb.30A-D: Oeceoptoma thoracica (Polyphaga: Silphidae). A: Linkes Hinterflügelgelenk von dorsal.
B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von ventral. D:
lAx frontal.

99

PRA 1Ax La

Abb.31A-D: Blitophaga opaca (Polyphaga: Silphidae). A: Linkes Hinterflügelgelenk von dorsal. B:
Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von ventral. D: 1Ax
frontal.

B

Abb.32A-D: Quedius sp. (Polyphaga: Staphylinidae). A: Linkes Hinterflügelgelenk von dorsal. B:
Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax frontal. D: 1Ax und 2Ax von
ventral.

100

3Ax

N
A PNP

33 0,5 mm

Abb.33A,B: Ontholestes murinus (Polyphaga: Staphylinidae). A: Linkes Hinterflügelgelenk
von dorsal. B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt.

1Ax

MNP

PNP

34

Abb.34: Sinodendron cylindricum (Polyphaga: Lucanidae). Linkes Hinterflügelgelenk von
dorsal.

101

ANP

0,5 mm

35 — MNP

Abb.35: Sinodendron cylindricum (Polyphaga: Lucanidae). Linkes Hinterflügelgelenk von dorsal.
Flügel und Axillarsklerite entfernt.

PNP PN

\

Sb

BaRK \
Ba

BR
1Ax

2AX

36 0,5 mm

Abb.36A-C: Sinodendron cylindricum (Polyphaga: Lucanidae). A: Metathorax links, von lateral. Flügel
und Axillarsklerite entfernt. B: 1Ax und 2Ax von ventral. C: 1Ax frontal.

102

Abb.37A-D: Phyllopertha horticola (Polyphaga: Scarabaeidae). A: Linkes Hinterflügelgelenk von
dorsal. B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von
ventral. D: 1Ax frontal.

BSc NS: SJ
Ai
\ 7 D
Dar ey / 8:
ae, 7 c
ae j

is PNP / |

( ANP \_ |
F

PN

Abb.38A-D: Byrrhus sp. (Polyphaga: Byrrhidae). A: Linkes Hinterflügelgelenk von dorsal. B:
Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von ventral. D: 1Ax
frontal.

103

39 \

Abb.39: Chalcophora mariana (Polyphaga: Buprestidae). Linkes Hinterflügelgelenk von dorsal.

Abb.40: Chalcophora mariana (Polyphaga: Buprestidae). Metathorax links, von lateral. Flügel und
Axillarsklerite entfernt.

104

Eps

Abb.41A,B: Anthaxia sp. (Polyphaga: Buprestidae). A: Linkes Hinterflügelgelenk von dorsal. B:
Metathorax links, von lateral. Flügel und Axillarsklerite entfernt.

Abb.42A,B: Argiotes pilosellus (Polyphaga: Elateridae). A: Linkes Hinterflügelgelenk von dorsal. B:
Metathorax links, von lateral. Flügel und Axillarsklerite entfernt.

105

Abb.43A,B: Denticollis linearis (Polyphaga: Elateridae). A: Linkes Hinterflügelgelenk von dorsal. B:
Metathorax links, von lateral. Flügel und Axillarsklerite entfernt.

0,5 mm

44 PN

Abb.44A-C: Hemicrepidius niger (Polyphaga: Elateridae). A: Linkes Hinterflügelgelenk von dorsal.
B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax frontal.

106

45

Abb.45: Lamprohiza splendidula (Polyphaga: Lampyridae). Linkes Hinterflügelgelenk von dorsal.

BaRK

Ba

= mL

Abb.46: Lamprohiza splendidula (Polyphaga: Lampyridae). Metathorax links, von lateral. Flügel und
Axillarsklerite entfernt.

107

Abb.47: Cantharis nigricans (Po-
lyphaga: Cantharidae). Linkes
Hinterflügelgelenk von dorsal.

48 Abb.48: Cantharis nigricans (Polyphaga:
Cantharidae). Metathorax links, von lateral.
0.5 mm Flügel, 2Ax und 3Ax entfernt.

108

Abb.49: Cantharis nigricans (Polyphaga:
Cantharidae). Metathorax links, von lateral.
Flügel und 3Ax entfernt.

0.25 mm %
49 Si
H
BSc PRA
ane
2Ax % F
€
AMD

A

0,5 mm

50

Abb.50A,B: Dermestes lardarius (Polyphaga: Dermestidae). A: Linkes Hinterflügelgelenk von
dorsal. B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt.

109

BaRK

05mm

Abb.51A-D: Trichodes sp. (Polyphaga: Cleridae). A: Linkes Hinterflügelgelenk von dorsal. B:
Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von ventral. D:
1Ax frontal.

52 0,25 mm

PEN
Abb.52A-D: Thanasimus formicarius (Polyphaga: Cleridae). A: Linkes Hinterflügelgelenk von
dorsal. B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von
ventral. D: 1Ax frontal.

110

PWP

0,5 mm

53

Abb.53A,B: Malachius sp. (Polyphaga: Melyridae). A: Linkes Hinterflügelgelenk von dorsal. B:
Metathorax links, von lateral. Flügel und Axillarsklerite entfernt.

Abb.54A,B: Hylecoetus dermestoides (Polyphaga: Lymexylonidae). A: Linkes Hinterflügelgelenk von
dorsal. B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt.

111

BSc H

«

Ky
RN PRA

BR——

2Ax MK
V7 1Ax
0
3Ax AM
y, PNP
A
PN
55
0,25 mm

Abb.55A-D: Calvia quatuordecimguttata (Polyphaga: Coccinellidae). A: Linkes Hinterflügelgelenk von
dorsal. B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von
ventral. D: 1Ax frontal.

Abb.56A-D: Coccinella septempunctata (Polyphaga: Coccinellidae). A: Linkes Hinterflügelgelenk von
dorsal. B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von
ventral. D: 1Ax frontal.

112

05mm

Abb.57A,B: Lytta vesicatoria (Polyphaga: Meloidae). A: Linkes Hinterflügelgelenk von dorsal.
Metathorax links, von lateral. Flügel und 3Ax entfernt.

0,5 mm ER

Ba

ANP

Abb.58A-C: Lytta vesicatoria (Polyphaga: Meloidae). A: Metathorax links, von latero-ventral. B:

Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: Linkes Hinterflügelgelenk von
dorsal. Flügel und Axillarsklerite entfernt.

113

114

BaRK

0,5 mm

Abb.59A,B: Tenebrio molitor (Polyphaga: Tenebrionidae). A: Linkes Hinterflügelgelenk von dorsal.
B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt.

60

Abb.60A-D: Clytus arietis (Polyphaga: Cerambycidae). A: Linkes Hinterflügelgelenk von dorsal. B:
Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von ventral. D: 1Ax
frontal.

115

NIE

ee, 1Ax
= 4 NES

A =
BaRK
Ba
Be" nn B

61 0,5 mm

Abb.61A-D: Agapanthia villosoviridescens (Polyphaga: Cerambycidae). A: Linkes Hinterflügelgelenk

von dorsal. B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von

: 5
D
2Ax
(G

ventral. D: 1Ax frontal.

0,25 mm

Abb.62A-D: Crioceris asparagi (Polyphaga: Chrysomelidae). A: Linkes Hinterflügelgelenk von dorsal.
B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von ventral. D:
1Ax frontal.

116

2Ax

0,5 mm

Abb.63A-D: Chrysomela populi (Polyphaga: Chrysomelidae). A: Linkes Hinterflügelgelenk von dorsal.
B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von ventral. D:
1Ax frontal.

Abb.64A-D: Leptinotarsa decimlineata (Polyphaga: Chrysomelidae). A: Linkes Hinterflügelgelenk von
dorsal. B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von
ventral. D: 1Ax frontal.

117

2Ax

Cc

F
PWP
PN
BaRK
Sb
Ba
B

0,25 mm

Abb.65A-D: Cassida sp. (Polyphaga: Chrysomelidae). A: Linkes Hinterflügelgelenk von dorsal. B:
Metathorax links, von lateral. Fliigel und Axillarsklerite entfernt. C: 1Ax und 2Ax von ventral. D: 1Ax

frontal.

| 2Ax
AN ANP. .

AMD

PN

Abb.66A-D: Phyllobius sp.1 (Polyphaga: Curculionidae). A: Linkes Hinterflügelgelenk von dorsal. B:
Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von ventral. D: 1Ax
frontal.

118

Abb.67: Phyllobius sp.2 (Polyphaga:
Curculionidae). Linkes Hinterflügelge-
lenk von dorsal.

0,5 mm PN

0,5 mm a
Abb.68A-D: Chlorophanus sp.1 (Polyphaga: Curculionidae). A: Linkes Hinterflügelgelenk von
dorsal. B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt. C: 1Ax und 2Ax von

ventral. D: 1Ax frontal.

NS)

69B Abb.69A,B: Cueta beieri (Planipennia:

Myrmeleonidae). A: Linkes Hinterfligelge-
0,1 mm lenk von dorsal. B: ANP, 1Ax und F von
dorso-lateral.

120

RN

0,5 mm 4 a
70 PYVVP

Abb.70: Cueta beieri (Planipennia: Myrmeleonidae). Metathorax links, von ventro-lateral.

2x
a Sb
u we
pwP
Epm

0,5 mm i

71

Abb.71: Cueta beieri (Planipennia: Myrmeleonidae). Metathorax links, von lateral. Flügel und 3Ax
entfernt.

73

Abb.73: Sialis lutaria (Megaloptera:

121

Abb.72: Osmylus fluvicephalus (Planipennia: Osmyliidae).

Metathorax links, von lateral. Flügel und 3Ax entfernt.

BSc

#8
Sialidae). Linkes Hinterflügelgelenk von dorsal.

i

Abb.74: Sialis lutaria (Megaloptera:
Sialidae). Linkes Hinterflügelgelenk
von dorsal. Flügel, 2Ax und 3Ax ent-
fernt.

Abb.75: Sialis lutaria
(Megaloptera: Sialidae).
Metathorax links, von
latero-ventral.

05mm

23

ANP

Gs. PWP

0,5 mm

76 Bis — u

Abb.76A,B: Chauliodes rastricornis (Megaloptera: Corydalidae). A: Linkes Hinterflügelgelenk von
dorsal. B: Metathorax links, von lateral. Flügel und Axillarsklerite entfernt.

G

Tg
BSc

ANP

rare 1mm

Abb.77: Corydalus cornutus (Megaloptera: Corydalidae). Linkes Hinterflügelgelenk von dorsal.

124

Abb.78: Corydalus

cornutus (Megalop-

tera: Corydalidae).
Metathorax links,

78

von latero-ventral.

gy Abb.79: Raphidia ophiopsis (Raphi-
0,25 mm dioptera: Raphidiidae). Linkes Hinter-
flügelgelenk von dorsal.

79

125

Abb.80: Raphidia ophiopsis (Raphidioptera:
Raphidiidae). Metathorax links, von latero-
ventral.

80 0,25 mm

Abb.81: Panorpa communis (Mecoptera: Panorpidae).
Metathorax links, von lateral. Flügel und Axillar-
sklerite entfernt.

Tg BSc

Abb.82: Panorpa communis
(Mecoptera: Panorpidae). Me-
sothorax von dorsal. Links sind
Flügel, 2Ax und 3Ax entfernt.

82 0,5 mm

Abb. 83: Perlodidae: Gen. sp. (Plecoptera).
Metathorax links, von lateral. Flügel und
Axillarsklerite entfernt.

Abb. 84: Elenchus koebelei (Strepsiptera:
Elenchidae). Rechte Flügelbasis von dorsal.

Nach Kinzelbach (1971: Abb. 69).

Abb. 85: Elenchus koebelei (Strepsiptera:
Elenchidae). Thorax rechts, von lateral.
Flügel und Axillarsklerite entfernt. |

Nach Kinzelbach (1971: Abb. 40). |

Anschrift des Verfassers:

Dr. Thomas Hörnschemeyer

Institut für Zoologie und Anthropologie
Abt. Morphologie und Systematik
Berliner Str. 28

37073 Göttingen

Deutschland/Germany

ch
da
ae

ies ann, c M.: fae zur Systematik und Phylogenese der holarktischen
Sesiiden (Insecta, Lepidoptera). 1971, 190 S., DM 48,—

iswiler, V., HR. Güttinger & H. Bregulla: Monographie der Gattung
| thrura. Swainson, 1837 (Aves, Passeres, Estrildidae). 1972, 158 S., 2 Tafeln, DM 40,—

pee #

nt. Se Die Wirbeltiere von Fernando Poo und Westkamerun. Unter besonde-

Bi H: Das en der Dolichopodiden (Diptera): Homologie und Grund-
=. 1974, 60 S., DM 15,—

nd 1 birds. 1976, 93 S.,.1.'Tafel, DM 23,—
J.: Secondary contact zones of birds in northern Iran. 1977, 64 S., 1 Falttafel,

€, J.: Les batraciens de Madagascar. 1978, 144 S., 82 Tafeln, DM 36,—
r, E.: Das Aktionssystem von Winter- und Sommergoldhähnchen (Regulus
A . ignicapillus) und deren ethologische Differenzierung. 1979, 151 S., DM 38,—

der, 8.0. A taxonomical study of the genus Apistogramma Regan, with a
x f Brazilian and Peruvian species (Teleostei: Percoidei: Cichlidae). 1980, 152

24.

ZI

26.

Dh

28.

29:

30.

31:

32.

33.

34.

35.

36.

3%

38.

38.

40.

41.

42.

43.

Arratia, G: Deserition of the primitive “amily Diplom

DM 30,— Ki
Nikolaus, G.: Distribution atlas of Sudan's birds with not
RT 322 S., DM ee — RUE

liche Re 1988, 175 S,DM44,— en Sa
Lang, M.: Phylogenetic and biogeographic patterns of asi |
Squamata: “Iguanidae”). 1989, 172 S., 43,— A
Hoi-Leitner, M.: Zur Veränderung der Stugetierfauna desi L
im Verlauf der letzten drei Jahrzehnte. 1989, 104 S., DM 26, ala pi
Bauer, A.M.: Phylogenetic systematics and Biogeography of t
(Reptilia: Gekkonidae). 1990, 220 S.,.DM55,— Son
Fiedler, K.: Systematic, evolutionary, and ecological instincts
within the Lycaenidae (Insecta: Lepidoptera: Papilionidea). 1991,
Arratia, G.: Development and variation of the suspensorium |
(Teleostei: Ostariophysi) and their phylogenetic relationships.
Kotrba, M.: Das Reproduktionssystem von Cyrtodiopsis w
Diptera) unter besonderer Berücksichtigung | der i inneren wena
1993, 115 S., DM 32,— Bar
Blaschke-Berthold, U.: Anatomie und Phylogenie der Bi
Diptera). 1993, 206 S, DM 52, — _

Hallermann, J.: Zur Morphologie der Einnpidelresient ae
eine vergleichend-anatomische Untersuchung. 1994, 133 S., DM
Arratia, G., & L. Huaquin: Morphology of the lateral li
of Diplomystid and certain primitive Loricarioid Catfishes 2 ne
considerations. 1995, 110 'S., DM 28,—

Lepidoptera, Zygaenidae). 1995, 224 S., DM 56,— we
Martens, J., & S. Eck: Towards an Ornithology of the Him
ecology and vocalisations of Nepal birds. 1995, 448 S., 3 Farbtafel
Chen, X.: Morphology, phylogeny, biogeography and eee (
Cyprinidae). 1996, 227 S., DM 57,— 2
Browne; DA, SCH’ScholHtz23 The morphology Be the hint
wing base of the Scarabaeoidea (Coleoptera) with some phylogen i
200 S., DM 50,— i
Bininda-Emonds, O.R.P., & AP. Russell: A morphologic
phylogenetic relationships of the extant phocid seals” a lia

1996, 256 S., DM 64,— HR
Klass, K.-D.: The external male genitalia and the phylog
Mantodea. 1997, 341 S., DM 85,—
Hörnschemeyer, T.: Morphologie und Evolution des Fligelge
und Neuropterida. 1998, 126 S., DM 32,— ;

AND CRANIOMETRIC APPROACH TO
SYSTEMATICS

a : ” Pal

/
f
;
en

he, F
x = f ¢
N Jf & 7
BR
ie |
nn RE NEE ur
by

ERNST-HERMANN SOLMSEN

1998 =

x Herausgeber:

_ ZOOLOGISCHES FORSCHUNGSINSTITUT
UND MUSEUM ALEXANDER KOENIG
BONN

l’editeur. Commandes et demandes pour échanges adresser a la b
CUL,..S.. VD. |

BONNER ZOOLOGISCHE MONOGRAPHIEN
Preis: 30,— DM
Schriftleitung/ Faller: G. Rheinwald

Zoologisches Forschungsinstitut il Muscunt Al
Adenauerallee 150—164, D-53113 Bonn, =
Druck: Je CARTHAUS, Bonn

ISBN 5925385488 ae
ISSN 0301-671 xe 2 ae

NEW WORLD NECTAR-FEEDING BATS: BIOLOGY,
MORPHOLOGY AND CRANIOMETRIC APPROACH TO
SYSTEMATICS

by
ERNST-HERMANN SOLMSEN

BONNER ZOOLOGISCHE MONOGRAPHIEN, Nr. 44
1998

Herausgeber:
ZOOLOGISCHES FORSCHUNGSINSTITUT
UND MUSEUM ALEXANDER KOENIG
BONN

Die Deutsche Bibliothek — CIP Einheitsaufnahme

Solmsen, Ernst-Hermann:

New world nectar feeding bats: biology, morphology and craniometric approach to
systematics / by Ernst-Hermann Solmsen. Hrsg.: Zoologisches Forschungsinstitut und
Museum Alexander Koenig, Bonn. - Bonn : Zoologisches Forschungsinst. und

Museum Alexander Koenig, 1998
(Bonner zoologische Monographien ; Nr. 44)
Zugl.: Hamburg, Univ., Diss., 1994
ISBN 3-925382-48-8

CONTENTS

Page

nnedinenan se ee ee Poe ake Ras Ee ahs HIS ROMO LE 5)
EIS SHORVISC TUNES SS ee echten ee en i
On the systematics, distribution and ecology of New World flower bats .......... 8
‘TP AOACIIONG. JOOISUTOI, We EN EB SE N a N am sic Ute eee a ea 8
GeoOrnaM MERMIS UMMM OM coe not om ek hate ee ew 2 10
Kantalsroosunesbehaysour mieratien 2. Zee es ee een 12
Achviy DISS ee el en EL SU re 15
Beedinesbioloes, FEeedinsgecolosy Siac ag. wen. ee nen. 16
Chiropteran adaptations to nectari- and pollenivory .................... 16
Bodszstzerandaveiehte rn oe re le a sees 16
ES@UyasMaMe ween ey ee Renae ny ete IRE ee SR Se V7
Balnas. one Sls a Se Se aa er Re a a 18

DGS ING RICE Sos See SURE RE SEE 18
dapianonsmmshost plantsito.chiropteroplily 2. 2... 4... .. 62 een. 20
IZERAMIC EC OUST OMe I Mets! caee) a stated I a MeN es 20
BICSSOMmmS Maem mee eet me a gin tn. so Ne ee ON RR 20
BlossomeespositHongr nr 2 0 a a 2 20

Diet of various New World nectar feeding bat genera .................. 2
SENSOBZsysiems// Onientatlone ne A: 25
Neouslichpercepulonsecholocationen een. ae 25
Oollcalsensee a 2 ee Ro re EN ISHIINGIETE 25
Ollasienygsense,/ KOltactorysperceptlonn RR. 26
eine MG UCR D1 OlO ON nr a i Rea Ro ee es ee 26
Syiolosy 5 oy Sab ee et ce ete aN Sie ar eas Lt Oa: One entree oe ee 31
yizyramall amd meines a A IO Re ROTM eI os 3
yA [nl PEE he ae oi ts ten ahs le. A uote OM as Ba $i

I Aes HUNG (0 Sri ernest hi N EPID 2
Descaiiang see AI ei a A EL N By

DPE WINES <2 ont Sle Bo alle eRe Se ream Antigo Pa er, SP ene ent oer ere Se 33
Sea sitet iitGw pire ee ee Ok Me th el moa oe Sin, aoe Gn AR ee 33
Shlenlations, Ra ro Se ee Ce oa EY oe ae ta 34

Omi Alicea AV SI Semen U ee ye Me NE eg nee ee en 34
MUO Whi CS a ee hee fie ei en) Al. 1 aie Esa haiti ass Gee ahs eene 34
[RGSS 2 1 3)g psa ae N ELEND 35
Monel OSveOmtmerspecies examned rn a, oe ie ee es 35
Exxkestimal| name NO) OCW AER ene oc oi a EN ge AER eesti: Gi tieaslieim Getue is ht wees: 35)
Skullkmospholesyare nr 7. CN Ne a Re A m einer ‚asp: 45

Skulledeseripuionszof tnessinsle sense een. 46

Morphomeity. «och Se N ee oe 62

Univariate analysis: „a... a nl a 62
Skull proportions’... esse re N Se 63
Allometriesyns „u.a Sarnen ern Sr See Sr 66
Allometicicomparisonyor individual’ generae Den 2a ene nee 66

Sex dimorphism: (0.5 es ae ey ee 1%)
Discussion... 3... usa ar re une See I 74
Morpholosicalladaptations to nectanivyory 2 en ee 74
Gross morpholosy of thehead, 22.2.2220 oe 74
Skull morphology 3.2.5. 2... 2.2.2.2: 22.2 22 ee eee 74
Dentition 2.2.2.2 RER Ee IB
Ineisivi! 28 a aie a aces & gk ay 5 Le). lee aap BE Sr a IS

Canini. su... Nr OS N 73
Premolars 2... 292. ee ee 76

Molar teeth: 0.28 5 ans ee oe ee 76

Rostrum 2... 2.2. 2 2 PE Ae N 76
Forehead). 4 cc oe 2 ee ee a Wi
Zygomatic arches >... ee es be ee a 2 BONN Ti
Braincase. 2... 2. ae en ee se ee 78

Skull base 2. ..: 2... aby PART eee a re 78
Mandible. 22 3.02 2.2.2... 82 2 se 2 N 80
Craniometry .. 2 sc ue) ee N ee 80
Univariate analysis .. ..... 2a Se a 80
Skull proportions. ...:2. 2... vias seo Rilke er SA 81
Allometries.... au... 2 0a ee oe ee N 83
Intergeneric allometrics ... 12... 2. 2 20 2 0 a 83
Allometrical sex comparison ....................2. 2 ee 2
Allometric conclusions... 36.0005 26 Na... 20 92
Systematic conclusions, .:.. I. Se da oe ee 93
SUMAMMATY .. sure u ig a ee sa ae eth WA, esr teste kG Sea 97
literature. cited. ass. eee eh ee 2 oe 98

Appendix. 4... el. ES RS el 2 a 108

INTRODUCTION

With more than 900 species, the Chiroptera represent the most comprehensive mammalian
order next to the rodents. By developing physical ability for continuous flight bats
succeeded in exploring habitats almost without vertebrate competitors - the nocturnal
airspace. Their morphological adaptations are just as consistent; having specialized their
anatomy in flight and a hanging resting position, most bats are hardly capable of qua-
drupedal locomotion, apart from crawling over short distances.

General body shape hardly varies within the Chiroptera - presumably due to strict
requirements for free flight. Differences are delicate and restricted to body size, wing

shape, development of the uropatagium and tail length. However, the chiropteran head is
one of the most striking characters in specific determination. Due to echolocation many
microchiropterans show an oddity of complex structures on ears and often also nose which
aid in directed emission and perception of sound.

Having successfully conquered the nocturnal skies, the primarily insectivorous Chiroptera
developed a variety of new diet preferences - camivory (Megaderma, Vampyrum,
Phyllostomus, Trachops), piscivory (Pizonyx, Noctilio) and even - unique among the
mammalia - sanguivory (Desmodus, Diaemus, Diphylla). Besides, many species also take
vegetable food: frugivorous and even few nectarivorous species prevail among the
Megachiroptera (Pteropodidae), but also the Microchiroptera developed a variety of fruit
and flower-feeding genera within the Phyllostomidae.

These various diet preferences within the Microchiroptera have their morphological
expression primarily in the shape of the head and in the dentition. Thus, skull morphology
of bats developed some remarkable diversity within the mammalia.

Fig.l: Skull shape depending
on diet specialization within
the Phyllostomatidae
(Centurio senex = fruit eater,
Choeronycteris harrisoni =
blossom feeder, adapted
from Koopman 1987)

In this respect, some members of the phyllostomid family achieved a striking diversity,
with the highly specialized flower bats representing an extreme (fig.1).

Especially the Glossophaginae succeeded in extensive specialization (prolongation of the
rostrum; long tongue with bristle-like papillae, weak and delicate teeth). The systematic
integration of all long-nosed Phyllostomids into a single monophyletic subfamily is still
discussed controversially. As the main characters used to define glossophagine bats are
derived characters influenced by function, they could well have developed independently
within the phyllostomids.

For more than 25 years, several approaches have been made to this problem (tab.1) and
various different systematic relationships of single genera were proposed from time to
time.

Table 2: Studies on the system of New World nectarfeeding bats

- morphological examinations
teeth, dentition (Phillips 1971)
gastrointestinal system (Forman 1971, 1979; Rouk & Glass 1970)
tongue, hyoid musculature (Wille 1954; Greenbaum & Pillios 1974; Griffiths 1982)
female sex apparatus (Smith & Hood 1982)
cerebral anatomy (McDaniel 1976)

- cytological examinations
cromosomal analysis (Baker et al. 1967; Patton & Gardner 1971; Stock 1975;
Baker & Bass 1979; Haiduk & Baker 1982)
hematology (Valdivieso & Tamsit 1971)

- immunological studies
serum proteins (Baker et al. 1981)
serum proteins (Gerber & Leone 1971)

Classification of Brachyphylla as an essentially basic form within its own subfamily
Brachyphyllinae remained undisputed (Baker et al. 1979) as well as everyone aggreed to
summarize the genera Erophylla and Phyllonycteris within the Phyllonycterinae (Flower
vampires), a subfamily closely related to the Brachyphyllinae (Silva Taboada & Pine
1969).

Contrary conclusions primarily concern the systematic position of the genera or
subfamilies having been summarized so far as Glossophaginae.

Some really revolutionary improvement was achieved by the works of Griffith published
1982. Analyzing the tongue and hyoid morphology and their musculature, respectively, he
was the first to seperate three genera (Lionycteris, Lonchophylla and Platalina) from the
Glossophaginae s.str., summarizing them within their own subfamily Lonchophyllinae.
These results led to a lively discussion between different research groups in the United
States (Haiduk & Baker 1982; Warner 1983; Hood & Smith 1982; Griffiths 1983; Smith
& Hood 1984). The point of this discussion is: do the New World nectarivorous bats
represent a monophyletic group, or did the glossophagine bats develop twice,
independently from each other?

Even within nectar feeders, among all morphological structures, the skull undoubtedly was
the main structure to undergo (possibly different) adaptations to intake of food. They are

numerous and sometimes extreme. According to the bounty of differently specialized
species modifications in skull structures vary with degree of specialization on nectar
feeding. Adaptations affect the visceral skull as well as the neurocranium. Dentition, man-
dibular and palatinal bones show modifications of remarkable extent. Even the braincase
changes its bony processus as a consequence in reduced masticating musculature.

So, subsequently, the skull morphology of the New World nectarivores and related
phyllostomid genera will be examined comparatively.

The aim is to understand and to describe the principles of skull construction in nectar
feeding phyllostomatids as an adaptation to nectar feeding. Allometrical comparison of
craniometric data will result in some evidence on suprageneric relationships of the genera.

As an introduction to this subject, chapter 2 will give an overview on systematics,
distribution and ecology of New World flower bats, mostly compiled from available
literature.

ACKNOWLEDGEMENTS

This work was facilitated by various suggestions and the support of many helpful people;
to all of whom I am indebted.

Several museums kindly lent me material from their collections: Natural History Museum
Copenhagen, Dr. Baagoe; American Museum of Natural History, New York, Dr. Walter
Fuchs; Royal Ontario Museum, Toronto; United States National Museum, Smithsonian
Institute, Washington, Dr. Charles O. Handley; Dr. Juliane Diller (geb. Koepcke), Munich.

A working place on the spot was generously provided in the collections by:

American Museum of National History New York, Dr. K. F. Koopman; British Museum
of Natural History, Dr. Janet Leigh, Dr. J. E. Hill; Rijksmuseum for Natuurlijke Historie
Leiden, Dr. C. Smeenk; Senckenberg Naturhistorisches Museum und Forschungsinstitut,
Dr. H. Felten, Dr. Kock, Mr O. Liitt; Naturhistorisches Museum Wien, Dr. F. Spitzen-
berger; Naturhistorisches Museum Basel, Mr. J. Gebhardt; Zoologisches Forschungsinstitut
und Museum Alexander Koenig, Bonn, Dr. R. Hutterer.

The organization of my journey to Ecuador I owe to Mr. Erwin Patzelt from Oldenburg /
Holstein; Dr. Luis Albuja-V. and Prof. Gustavo Orcés from the Escuela Politecnica
Nacional provided useful information on the Ecuadorian bat fauna.

At the Zoologic Museum of the University of Hamburg I was supported by Mrs. U.
Frerichs who prepared the skull drawings of the genera which had only been accessible
abroad and thus sometimes were represented by insufficient photographs.

Dr. Derek Vinyard was a help to me not only in solving various computer-related and
software problems but also in translation of the manuscript; Mr. Nelson Mascarenhas
kindly prepared loan specimens for dispatch; Mr. Preuss did some of the required skull
preparations. My wife Hannelore helped me recording the values during our stays at the
museums and ensured me to keep free of many other problems, too.

Last, but not least I like to thank Prof. Dr. Harald Schliemann for introducing me to the
world of these fascinating mammals.

ON THE SYSTEMATICS, DISTRIBUTION AND ECOLOGY OF NEW WORLD
FLOWER BATS

Taxonomic Position

All neotropic flower bats belong to the New World phyllostomids (Phyllostomatidae):
Order Chiroptera
Suborder Microchiroptera
Superfamily Phyllostomatoidea
Family Phyllostomatidae / (Phyllostomidae)
Subfamilies:
Lonchophyllinae
Brachyphyllinae
Phyllonycterinae
Glossophaginae
Phyllostominae
Stenoderminae
Carolliinae
Currently, the Phyllostomatidae are divided into seven subfamilies of which four
subfamilies contain 38 predominantly nectarivorous species in 15 genera:
Subfamily Lonchophyllinae Griffiths, 1982
Genus Lionycteris Thomas, 1913
L. spurrelli Thomas, 1913
Genus Lonchophylla Thomas, 1903
L. thomasi Allen, 1904
. dekeyseri Taddei, Vizotto & Sazima, 1983
. mordax Thomas, 1903 with subspecies
. robusta Miller, 1912
. handleyi Hill, 1980
. bokermanni Sazima, Vizotto & Taddei, 1978
L. hesperia Allen, 1908
Genus Platalina Thomas, 1928
P. genovensium Thomas, 1928

Salil Salil u Salli oe

Subfamily Brachyphyllinae Gray, 1866
Genus Brachyphylla Gray, 1834

B. nana Miller, 1902 (= B. pumila Miller, 1918)

B. cavernarum ssp. Gray, 1834 with subspecies

Subfamily Phyllonycterinae Miller, 1907
Genus Erophylla Miller, 1906
E. sezekorni (Gundlach, 1861) with subspecies

E. bombifrons (Miller, 1899) with subspecies
E. b. santacristobalensis: Hispaniola
E. b. bombifrons: Puerto Rico
Genus Phyllonycteris Gundlach, 1861
Subgenus Phyllonycteris Gundlach, 1861
Ph. (Ph.) poeyi Gundlach, 1861 with subspecies
Subgenus Reithronycteris Miller, 1898
Ph. (R.) aphylla (Miller, 1898)

Subfamily Glossophaginae Bonaparte, 1845
Genus Glossophaga Geoffroy St.Hilaire, 1818
G. soricina (Pallas, 1766) with subspecies
s. handleyi (= G. s. leachii): North America
s. mutica: Island population Tres Marias Is.

a 9) 9)

s. antillarum: Jamaica I.

s. valens: South America, Ecuador, Peru

an

s. soricina: South America (east of Andes)

. commissarisi Gardner, 1962 with subspecies

am 9

. longirostris Miller, 1898 with subspecies

. leachii (Gray, 1844) (= G. morenoi Martinez & Villa, 1938; =G.alticola
Davis, 1944)

G. mexicana Webster & Jones, 1980
Genus Monophyllus Leach, 1821

M. redmani Leach, 1821 with subspecies

M. plethodon Miller, 1900 with subspecies

QQ

Genus Leptonycteris Lydekker, 1891
L. nivalis (Saussure, 1860)
L. yerbabuenae Martinez & Villa-R., 1940 (= L. sanborni Hoffmeister, 1957)
L. curasoae Miller, 1900 with subspecies
Genus Lichonycteris Thomas, 1895
L. obscura Thomas, 1895 (= L. degener Miller, 1931)
Genus Anoura Gray, 1838
A. caudifer (Geoffroy St.Hilaire, 1818)

A. cultrata Handley, 1960 (= A. brevirostrum Carter, 1968; = A. werckleae
Starrett, 1969)

A. geoffroyi Gray, 1838 with subspecies

A. latidens Handley, 1984
Genus Hylonycteris Thomas, 1903

H. underwoodi Thomas, 1903 with subspecies
Genus Scleronycteris Thomas, 1912

10

5. ega Thomas, 1912
Genus Choeroniscus Thomas, 1928
Ch. godmani (Thomas, 1903)
Ch. intermedius (Allen & Chapman, 1893)
Ch. minor (Peters, 1869) (= C. inca Thomas, 1912)
Ch. periosus Handley, 1966 with subspecies
Genus Choeronycteris Tschudi, 1844
Subgenus Choeronycteris Tschudi, 1844
Ch. (Ch.) mexicana Tschudi, 1844
Subgenus Musonycteris Schaldach & McLaughlin, 1960
Ch. (M.) harrisoni (Schaldach & McLaughlin, 1960)

Geographie Distribution

The 13 genera of the subfamilies Glossophaginae and Lonchophyllinae are distributed
throughout the subtropical and tropical areas of the New World. Two further, very closely
related (Silva-Taboada & Pine 1969) subfamilies of the Phyllostomatidae, the
Brachyphyllinae and Phyllonycterinae, also comprising nectarivorous species, are restricted
to the islands of the Caribbean. Some genera, like Glossophaga soricina or Anoura
geoffroyi are widespread with distributional ranges as far from southern United States
down to southern Peru. Others have an extremely restricted distribution: the Banana bat
(Choeronycteris harrisoni) which probably shows the highest adaptation to nectar feeding
was not discovered before 1960, and only very few specimens were subsequently captured
near the same locality from Central Mexico.

Data on the geographic distribution predominantly refer to the locality of the collected
material; subsequently individual taxa are given in a detailed list of all currently known
distribution areas:

Subfamily Lonchophyllinae

Lionycteris

L. spurrelli: E Panama to E Peru and Brazilian Amazon region, west of the Andes,

however, not south of Colombia

Lonchophylla

L. thomasi: Eastern Panama to E Peru and Amazon region of Brazil, but west of the
Andes to the south not beyond Ecuador

. dekeyseri: E Brazil

. mordax: Costa Rica to W Ecuador (L. m. concava) and E Brazil (L. m. mordax)

. robusta: Nicaragua to N Peru; east of W Venezuela

. handleyi: Ecuador and Peru (east of the Andes)

. bokermanni: SE Brazil

. hesperia: Arid regions in SW Ecuador and NW Peru

ee

Platalina
P. genovensium: restricted to arid regions of Western Peru

Il

Subfamily Brachyphyllinae

Brachyphylla

B. nana: Cuba, Island Hispaniola, Cayman Islands and southern Bahama Islands (fossil
from Jamaica)

B. cavernarum: Puerto Rico, Virgin Islands except island St. Croix (B. c. intermedia), St.
Croix, Anguilla south to St. Vincent (B. c. cavernarum) and Barbados Islands (B. c.
minor)

Subfamily Phyllonycterinae

Erophylla

E. sezekorni: N and central Bahama Islands (E. s. planifrons), SE Bahama Islands (E. s.
mariguanensis), Cuba, Cayman Islands (E. s. sezekorni) and Jamaica (E. s. syops)

E. bombifrons: Island Hispaniola (E. b. santacristobalensis), Puerto Rico (E. b. bombi
jrons)

Phyllonycteris: Cuba, Hispaniola und Jamaica (rezent), fossil from Puerto Rico

Ph. poeyi: Cuba (Ph. p. poeyi), island Hispaniola (Ph. p. obtusa)

Ph. aphylla: Jamaica

Subfamily Glossophaginae

Glossophaga

G. soricina: North American Mainland (G. s. handleyi = leachii); Tres Marias Islands (G.
s. mutica); Jamaica (G. s. antillarum), South American mainland, Ecuador, Peru (G.
s. valens), South America east of the Andes (G. s. soricina)

G. commissarisi. Southern Mexico (G. c. commissarisi), northwestern Mexico (G. c.
hespera)

G. longirostris: NW Ecuador, N Colombia, NW Venezuela (G. I. longirostris); northern
South America, Caribbean (G. /. elongata); E Colombia to Trinidad I. (G. I. major);
S Venezuela and Guayanas (G. !. campestris); Central Colombia (G. I. reclusa);
Tobago and Grenada Islands to St. Vincent I. (G. I. rostrata)

G. leachii: Nicaragua

G. mexicana: Southern Mexico east of Oaxaca and W Chiapas (G. m. mexicana); western
Oaxaca to Michoacan (G. m. brevirostris)

Monophyllus

M. redmani: Jamaica (M. r. redmani); Cuba, island Hispaniola, southern Bahama Islands
(M. r. clinedaphus); Puerto Rico (M. r. portoricensis)

M. plethodon: Barbados Island (M. p. plethodon); Puerto Rico (subfossil) (M. p. frater);
Lesser Antilles from Anguilla south to St. Vincent (M. p. luciae).

Leptonycteris

L. nivalis: Texas to Guatemala

L. yerbabuenae: Arizona, NE Mexico to El Salvador

L. curasoae: South American mainland, Isla Margarita and Aruba Islands (L. c. tarlosti)
Curacao and Bonaire (L. c. curasoae)

Anoura

A. caudifer. Restricted to tropical South America east of the Andes, Colombia to the
Amazon delta, NW Argentina and SE Brazil

A. cultrata: Costa Rica to N Venezuela and Bolivia; however, not beyond west of the
Andes

12

A. geoffroyi: Tropical Mexico to W Ecuador (A. g. lasiopyga), central Colombia to central
Bolivia (A. g. peruana) Venezuela, Guayana, Trinidad, Grenada Islands, E Bolivia to
eastern Brazil (A. g. geoffroyi)

A. latidens: N Venezuela to E Peru

Hylonycteris

H. underwoodi: Western Mexico from Jalisco to Oaxaca (H. u. minor); Veracruz to
Panama, incl. Belize (H. u. underwoodi).

Scleronycteris

S. ega: Southern Venezuela, northwestern Brazil (Amazon region)

Choeroniscus

C. godmani: W Mexico and northern fringe of South America to Surinam

C. intermedius: Trinidad I., Guyana, Surinam, N Brazil and Peru (east of the Andes)

C. minor. South American tropics from W Ecuador to the Amazon delta, north to
E Venezuela and south to NW Bolivia

C. periosus: W Colombia (C. p. periosus), northwestern Venezuela (C. p. ponsi)

Choeronycteris

C. mexicana: SW USA to Honduras incl. Tres Marias Islands

C. harrisoni: SW Mexico (Colima, Guerrero u. Michoacan)

Habitat, roosting behaviour, migration

The ecology of bats is predominantly determined by two elements: finding food on one
hand and on the other - just as compelling - finding suitable day shelter. Thus, any
locality of collected material will only reveal half of the occupied habitat. An insight to
the ecological demands of a chiropteran species will only be gained by long-term field
observations or by comparative observations of captive animals in their roost and during
foraging. But the capturing data of most specimens allows - at least tentatively -
assessment to the habitat of the species. Accordingly, the members of some genera are
restricted to tropical rainforest (Choeroniscus, Hylonycteris, Lichonycteris), while others
occur almost everywhere (Anoura, Glossophaga, Lonchophylla). Some genera
(Leptonycteris, Platalina and Choeronycteris) are adapted to arid areas, where they
predominantly feed on cactus flowers.

These highly adapted flower visitors depend on pollen as a protein source all around the
year. Unless they cover larger distances they can only get it in the highly constant milieu
of the neotropical rainforest. This ecological request is mainly due to the fact that
hovering flight requires a relative high amount of energy in food intake (v. Helversen &
Reyer 1984). One strategy to succeed with limited sources is outrunning intraspecific
competion - this will only work in low population densities and within large distribution
areas (e.g. Amazon rain forest). Here, most nectarivorous bats inhabit territories in small
family groups all the year round (Choeroniscus, Koepcke 1987).

There are, however, areas, where seasonal peaks of food supply determine the amount of
food available to the bats, requiring a quite different strategy. Thus Leptonycteris, though
highly specialized on blossom food, lives in large colonies - all the host plants of their
arid habitats usually bloom simultaneously, providing a rich food supply over a short time.
In need of constant food supply, the animals are forced to visit their host species currently
flowering within various areas, and often cover considerable distances (Humphrey &
Bonaccorso 1979). Furthermore, climate changes require long migrations, as the bats are

13

incapable of surviving by means of prolonged lethargy phases. Onset of the rainy season
with the climate getting cooler sets off migration in Leptonycteris (Easterla 1973). During
summer, L. nivalis is found in the higher levels of Big Bend National Park, Texas, and
several areas of Northern Mexico; and in winter they go further south, passing down at
least to Jalisco and Morelos (Barbour & Davis 1969; Kunz 1982).

Where the supply of night-flowering plants does not support a minimum of individuals
required to sustain genetic diversity, these habitats will be compliant to less specialized
genera who also include quite a lot of small insects in their diet (Glossophaga, Anoura
and Zonchophylla). Correspondingly, most of them have more extensive distribution areas.

Habitat data on single taxa:

Lionycteris: Most specimens of L. spurrelli Handley (1976) recorded in Venezuela were
captured in humid forest, roosting in caves and rocky crevices during the day. In Peru,
Tuttle (1970) captured two bats at the edge of indigenous villages, one of them amongst
flowering Cashew trees.

Lonchophylla: Spends the day in hollow trees, sometimes in caves. In Venezuela, Handley
(1976) collected most of his L. robusta and L. thomasi in humid forested areas. Detailed
information on L. thomasi from the Peruvian rain forest was given by Koepcke (1987).
From six specimens, three were captured in open riverine woodland, two in a tall cassava
field and one at a river bank. Several months she observed these bats in their day shelters
beneath embankments and among the roots of hollow trees. Though they sometimes
moved to another roost, the species altogether proved sedentary during the mating season.

Platalina: There are no ecological data on Platalina genovensium yet.

Brachyphylla: These bats prefer caves, though there are some records from buildings and
one from a well (Novak & Paradiso 1983). They live in small groups (Beatty 1944) or in
large colonies (5000 - 10,000 individuals). Their day shelters are not always in the dark
(entrance areas of caves, well shafts, dense foliage). As observations in captive specimens
revealed that they do prefer the darkest area of their cages, staying within lighter areas
may be accepted as a temporary behaviour (Swanepoel & Genoways 1983).

Erophylla: Buden (1976) recorded E. sezekorni not only from the deeper, darker cave
areas but also from the lighter surroundings of the entrance. Koopman et al. (1957),
however, collected their specimens on several islands of the Bahamas exclusively in the
deeper cave areas.

Phyllonycteris: Roosts in caves during the day (Novak & Paradiso 1983).

Glossophaga: G. soricina exploits a variety of different roosts - natural hideouts like
caves, hollow trees and crevices, but also artificial hiding-places: drainage pipes, deserted
mines, cellars, roof framework or undersides of bridges (Tuttle 1976; Webster 1982).
According to this wide range of roost selection they show a considerable compliance to
other bat species: there are more than 30 other species which are known sometimes to
share the same roost (Webster et al. 1984). The strongest coincidence is found with
Carollia perspicillata: In Peru, more than 60% of all known day shelters have been
recorded for both genera (Graham 1988). Apparently, the members of both taxa even
share the same locations within a shelter forming mixed clusters - probably an evidence
of mutual benefit by means of socialization, such as less effort in thermoregulation and
water budget. According to Koepcke (1987), G. soricina prefers lighter vegetation or
densely covered cultivated land. One of the bats she observed flew between the dwelling

14

houses, another amidst a tall corn field. Various Glossophaga were recorded from a
banana plantation at a woodside and next to a river.

Monophyllus: M. redmani prefer humid caves, where they usually live in colonies of
considerable size. Sexes apart from each other the individuals cling to the walls, the
ceilings and chimneys in dense clusters (Homan & Jones 1975). M. plethodon is
exclusively known from netted individuals. As there is also one record of a dead specimen
in front of a cave entrance in Dominica Island (Schwartz & Jones 1967) these bats
presumably accept caves as day shelters.

Leptonycteris: L. nivalis is a colonial cave dweller which is also found in adits, deserted
buildings and hollow trees. Such caves are characterized by a musk-like scent resembling
that of Tadarida brasiliensis (Barbour & Davis 1969). The size of the colonies may
exceed 10 000 individuals: Easterla (1972) reported a population density of 1615 bats per
square meter!

Lichonycteris: All records are from dense rain forest areas; up to now no data on roost
selection.

Anoura: In Venezuela, Handley (1976) found Anoura predominantly in humid and
woodland areas, often at high altitude. They spend the day in crevices and caves.

A. cultrata reportedly inhabits higher levels (mountainous forests from 220 m to 2600 m)
(Tamsitt & Nagorsen 1982). Most often the localities are in humid rain forest areas where
the animals are caught with nets along rivers or streams or at the edge of clearings or
villages. The specimens I caught myself in Ecuador spent the day in a very humid cave
(San Antonio de Pichincha, height 2300 m above sea level).

A. caudifer has been reported by Koepcke (1987) from the Amazon area of Peru from
cultured areas next to woodland, but she did not detect any roost sites. - One specimen
collected in 1983 at Rio Cuyabeno, Ecuador, was also caught with a net at sunset at a
riverine woodside next to a banana plantation. Their preferred roost sites include tree
hollows, caves, drainage pipes, sewers and buildings. Up to 13 individuals were recorded
at one single location. In Manaus, Reis (1981) detected three individuals in a fallen hollow
log in the company of Micronycteris megalotis.

Hylonycteris: Phillips & Jones (1971) collected H. underwoodi in dense woodland of
Jalisco, Mexico. Some small groups of two and eight individuals were reported by Laval
(1977) under a wooden bridge and a hollow tree. This species, however, also accepts
caves and tunnels as roost sites (Allen 1942), but apparently in small groups of very few
individuals only. In Guatemala, one female was caught together with two specimens of
different species Glossophaga and one Lichonycteris obscura next to a night-blooming
tree. Currently, nothing is known about their socialization in the roost site.

Scleronycteris: One of the three known specimens Handley (1976) netted in Tamatama,
Rio Orinoco, T.F. Amazonas, Venezuela, at a riverine jungle clearing.

Choeroniscus: Several individuals of C. intermedius have been reported by Koepcke
(1987) from the Amazon basin of Peru roosting in small groups or in pairs under logs or
in hollow trees in riverine areas. Three of them were found separately among the roots of
fallen trees, one pair beneath the bark of a rotten log. All individuals were 50 - 70 cm
above the floor and occupied their gloomy roost sites for several months. Three C. in-
termedius she caught at the edge of a primary forest, one of them flying above a low field
and another at a dead water of Rio Llullapchis. One specimen of C. minor, captured at
Rio Cuyabeno in Ecuador 1980, was also found at a river bank (Patzelt pers. comm.).

15

There is also one Venezuelan report on eight specimens hanging beneath a log which had
fallen across a river (Sanborn 1954).

Choeronycteris: C. mexicana is known from various habitats, from arid brier to tropical
secondary forest and mixed oak wood (Arroyo-Cabrales et al. 1987). As day shelters they
prefer caves and deserted adits, usually clinging themselves at dim recesses next to the
entrance. So, they accept even very small caves. There is some controversy about whether
they congregate with other species: whereas Goodwin (1946) regarded C. mexicana as a
mostly solitary species merely moving about, there have been later reports on various
Vespertilionidae and Tadarida, also Glossophaga, sharing their roost sites with these bats.
They occupy both caves and artificial shelters. Davis & Russell (1954) found a group of
C. mexicana hanging beneath a tree. The individuals cling seperately 2 - 5 cm apart from
each other, usually holding grip with only one foot and thus capable of observing intruders
by rotating their body up to 360 degrees. C. mexicana is an extremely alert, easily startled
species, which will rather leave the roost immediately than move to darker sites (Barbour
& Davis 1969).

Activity Patterns

Brown (1968) pointed out how activity patterns depend on diet: correspondingly,
insectivorous species are most active in the early evening, whereas frugivorous and
piscivorous bats show almost equal activity patterns all over the night. In the sanguivorous
Desmodontidae, the activity pattern is mainly determined by darkness, as these bats are
most active at complete darkness. Nectar feeding bats leave their roosts soon after sunset
heading for their host plants according to certain patterns, so their activity pattern
sometimes may be bimodal. They are, however, certainly active during the first half of the
night.

Detailed information is available for only few species.

There are some observations on the food intake of Lonchophylla thomasi from east Peru
by Koepcke (1987) showing that these bats leave their roosts at complete darkness not
before 18.25 or 18.35. One specimen covered with pollen was caught around 9 p.m.; at
least one activity phase occurs during the first night hours.

As reported by Swanepoel & Genoways (1983), Brachyphylla cavernarum leave their day
shelter some time after nightfall, at least one hour after sunset and some 20 minutes later
than Artibeus. First, all individuals of a colony fly out synchronously, finishing their
activity almost as simultaneously within the very last minutes before sunrise.

Activity patterns of Glossophaga soricina were studied by Erkert & Kracht (1978),
revealing that this species is influenced by a quite inflexible circadian system which
synchronizes with light and is induced by sunset, with a free periodic length of just 23.4
to 25 hours merely adapting to external stimuli. In eastern Peru, Koepcke (1987) captured
foraging G. soricina shortly before midnight, and they were observed at banana blossoms
in the early morning as well. In a similar way Sazima & Sazima (1978) reported an
accumulation of foraging bats between 1.20 and 4.00 a.m., with activity maxima in the
evening and during the last night hours (La Val 1970; Bonaccorso 1979).

According to Fleming et al. (1972), Sazima & Sazima (1978), Bonaccorso (1979) and
Koepcke (1987), spatial distribution of food supply determines the flight routes in G.
soricina. Depending on the pollen suppliers available, the species heads for higher or

16

lower vegetation levels, approaching individual plants in a trap-lining way keeping a
certain sequence and sometimes covering considerable distances. On their way the bats
regularly visit night shelters for about half an hour. G. soricina approaches flowering
plants both individually and in groups, the size of the latter depending on the number of
open blossoms per night and tree.

As reported by Barbour & Davis (1969), Leptonycteris nivalis leave their roost
comparatively late in the evening, but detailed information on their activity is still to
come.

Chronological shifts in activity rhythms within the same habitat were reported by Koepcke
(1987) in Panguana (Peru) in three sympatric nectar feeding genera: Lonchophylla thomasi
always flew into the nets before 9 p.m., Choeroniscus intermedius between 8 and 11 p.m.
and Anoura caudifer never before midnight.

Feeding biology / Feeding ecology

Chiropteran adaptations to nectari- and pollenivory

The Glossophaginae represent small to mid-sized Phyllostomatidae with a reduced
dentition, a distinctly elongated nose and a widely protrusible tongue - all adaptations to
a feeding specialization on nectar and soft fruit. In ecological respect they represent
nocturnal equivalents of hummingbirds, and their development may partially have been
influenced by similar parallel evolutionary constrains. This is shown in many similarities
(weight limit, ability of hovering, elongated tongue, prolongation of the rostral skull).
They usually feed during hovering, but sometimes the bats will go down onto the
blossom, thereby impairing further development of the fruit with the claws.

Body size and weight

In contrast to frugivorous Phyllostomids which often grow quite large, the highly
specialized nectar feeding bats range at the lower level of body size and weight (cf. tab.2).
This is partly explained by ecological aspects of the flowers, as size and structure of “bat
blossoms” must be sufficiently resistant to bear the weight of approaching and often even
landing bats. On the other hand, body size of these animals will be essentially limited by

Table 2: Body length and weight of blossom-feeding Microchiroptera (adapted from Dobat & Peikert
1985)

body length weight (g)

(mm)
Vampyrum spectrum 125-135 145-190
Phyllostomus hastatus 100-130 52,2-101,1
Phyllostomus discolor 75-91 22,2-40,0
Choeronycteris harrisoni 80-89 ca. 25
Choeronycteris mexicana 60-86 10-20
Leptonycteris nivalis 76-78 18-30
Glossophaga soricina 48-84 ST
Anoura geoffroyi 60 11,3-17,7
Choeroniscus godmani 53-58 7,6

Lichonycteris obscura 46-55 7,1-8,1

17

the energy balance which can be achieved. The more a bat specializes on limited plant
food species, the more its body size will be restricted by comparatively expensive
approaches to single flowers.

Body shape

Wing aspect ratio; phalanges

As in other flying vertebrates, the geometry of wing surface related to body weight gives
some insight into flight conditions and flight demands of bats, respectively (Smith &
Starrett 1979). Thus, relative length of the wing bones participating in flight activity will
be determined by aerophysical demands rather than by systematic relationship - large or
stocky species have longer 2. phalanges in their 3rd digit. So, differentiation
corresponding to relative length of the phalanges in digit HI gives evidence of wing shape:
the longer the metacarpals, the narrower the wing (in fast, tenacious flyers). On the other
hand, bats with comparatively stout metacarpals have broader wings (slow, astute flyers,
foliage gleaners).

Early as 1943, Sanborn classified the Glossophaginae into two groups, referring to relative
length of metacarpals and phalanges:

1) First phalanx III longer than 1/3 of metacarpal length II] and second phalanx of 3rd
finger shorter than 1,5 times the length of first phalanx III: Glossophaga, Lichonycteris,
Scleronycteris, Choeroniscus, Hylonycteris, Choeronycteris, Platalina

2) First phalanx III shorter than 1/3 of metacarpal length III and

a) Second phalanx III shorter than 1,5 times the length of first phalanx III: Lonchophylla
and Leptonycteris (meaning the smaller species within each genus), Monophyllus

b) Second phalanx III longer than 1,5 times the length of first phalanx III: Lionycteris,
Anoura, Lonchophylla and Leptonycteris (meaning the large species within each genus).

All these results correspond to what we currently know on bat ecology. The genera
mentioned first with comparatively short metacarpals usually represent highly specialized
nectar feeders requiring astuteness rather than velocity when patrolling among the
blossoms. Though Glossophaga feeds a good deal on insects, these bats presumably
capture them on the substrate, not in the air. In case of Leptonycteris, predominantly a
hovering nectar feeder, the unusual long-winged profile may not be explained by its
feeding ecology alone. Here, the “wing geometry of fast and long-range flight” may have
been of evolutionary significance for seasonal migrations (Sahley et al. 1995). Anoura
reportedly takes a large proportion of insects in their diet (Gardner 1977); it is, however,
questionable whether the relativly elongated wings could be explained as a device for
capturing insects in flight (obviously, the morphology of the uropatagial region seems to
oppose this opinion, see below) and requires further observation.

Uropatagium, tail

The degree of tail membrane development and the presence or absence of a bony tail may
be interpreted as an ecological adaptation leading to selective advantages both in foraging
and roosting behaviour. Within the primarily insectivorous microchiroptera, a well
developed uropatagium with a long bony tail and long cartilaginous calcars is regarded as
a plesiomorph condition. Within the Phyllostomatidae there are numerous variations,
including a lacking tail, calcar or uropatagium (Sturnira, Anoura), various intermediate
stages and extreme conditions as a short uropatagium combined with a long, projecting

18

tail (Phyllonycteris, Monophyllus) or an extensive tail membrane with the tail remaining
very short or lacking at all (Stenoderminae, Choeroniscus). Generally, the frugivores tend
to develop a reduced uropatagium, as they need to climb about on their host plants,
especially when the bats even have their roosts among the branchwork or the foliage.

Most nectar feeding bats possess reduced or entirely lacking tails (Leptonycteris, Anoura).
The degree of development does, however, not necessarily correspond to the degree of
specialization on nectar feeding (cf. figures of interfemoral membranes in ‘Results:
morphology of the species examined’), but might also have been influenced by some
additional ecological demands.

Pelage

An overview on adaptations of hair structure to pollen intake was given by Howell &
Hodgkin (1976): even visible to the naked eye, living specimens have their nape hairs
standing up like the bristles of a bottle brush rather than recumbent as in other bats.
Further differences are revealed in their fine structure: whereas many Chiroptera usually
have a smooth hair shaft under the microscope, pollinating bats possess scales standing up
from the shaft, thus facilitating pollen fixing in the pelage.

Fixing pollen is not only advantagous for the host plants to be pollinated - as nectar is
predominantly an energy supplier, the intake of pollen serves as an essential and
sometimes exclusive protein source. According to Howell (1974), analysis of stomach
contents in bats having been caught at their feeding plants always revealed nectar
exclusively in the stomach (and pollen only in the fur). Very probably the bats take up
pollen afterwards, grooming in their roosts. Howell (1974) described that the bats ingest
the pollen combed from the fur with the claws by constantly licking their feet. This
behaviour is also supported by faeces analysis (Harris 1959), showing that in no case there
was any anther material in the faeces of nectar feeders. Using the fur for pollen transport,
the animals keep full stomach capacity for their “fuel” (the nectar). Considering the
narrow limits of their energy balance, this may be a pre-condition for efficient exploitation
of these resources (v. Helversen & Reyer 1984).

Dobat & Peikert (1985:110) point to the fact that the chiropteran fur generally is well
suited for pollen transport, thus doubting the significance of Glossophagine fine hair
structure in allowing pollen transfer. Comparable conditions to those described by Howell
& Hodgkin (1976) were found independently in some non-pollenivorous bats. Thus, the
fine strucure of the hair shaft enlarging the surface may be determined by different eco-
logical demands; this characteristic seems to occur widely among bats. Obviously, it
seems impossible to prove any anthophile specializations in pollinating bats compared
with insectivorous species, the scales on the hair shaft which are arranged like keratinized
cones - as found in the pelage of all nectar feeders - certainly represent ideal devices for
embedding and fixing pollen grains.

Digestive tract
Tongue

The long, tapering, very protrusible and highly mobile tongue represents the characteristic
feature of specialized nectar feeding bats. Its tip is covered with brush-like papillae
directed backward and thus enabling efficient nectar intake (Griffiths 1978). The protrac-
tility of the tongue is incredible - Glossophaga can extend its tongue up to three times the
snout length. In the extremely long-nosed banana bat Choeronycteris harrisoni the
extended tongue is up to 76mm long - corresponding body length of 80mm (v. Helversen
1993).

19

Early authors supposed the long tongue to be folded in an s-shape within the closed mouth
(Moller 1932). Next, the tongue was disproved to be inserted within a dermal pouch as in
woodpeckers or Pholidota. In fact, the tongue, when not in use, shortens to an extent that
it fits into the mouth cavity. Later on, anatomical studies revealed an extremely complex
morphology of the tongue musculature.

The M. genioglossus is broadened into an extrinsic tongue muscle, the M. sternohyoideus
is integrated into the tongue as a retractive muscle (“In functional terms, it could be called
Sternoglossus” Wille 1954; “tunnel insertion” in Griffiths 1983) and considerably leng-
thened - its origin at the sternum being shifted back from the manubrium to the base of
the xiphoid process. Simultaneously, the insertion of the M. stylohyoideus at the tongue is
shifted from ventral to lateral, thus enabling to support the M. sternohyoideus when
retracting the tongue.

In Glossophaga, the tongue is passed through by one comparatively enormous central
artery (Lonchophyllinae: two arteries) and two lateral large veins (Griffiths 1978). The
latter are covered by muscle bundles which contract and press the stemmed flow of blood
up to the tongue tip thus elongating and stiffening the tongue additionally (vasohydraulic
tongue). All the time the entire tongue remains entirely flexible and can be bent in all
directions. Furthermore, it reacts with a reflex on contact with sugar, thereby moving into
nectar droplets without need of visual control.

Thus, the nectar supply is exploited within very short time. It is, anyway, still unknown
how the rapid in- and efflux of the blood necessary for the high frequent licking
movements - 12 movements per second (v. Helversen & v. Helversen 1975) is achieved.

Dorsally and laterally the tongue is covered with papillae, while its underside is
completely smooth. A detailed description of various papillae was given by Griffiths
(1982). Essentially, effective nectar intake is achieved by means of the hair papillae
(Papillae filiformes) of the tongue situated at the anterior third and laterally (brush-like
tongue). Aided by specialized lateral (Lonchophyllinae) or median (Glossophaginae)
grooves, these structures retain considerable amounts of fluid which is set free by
compression of the tongue at the palate during retraction. Nevertheless, the detailed
process and the coordination of tongue motoricity and swallowing are not yet sufficiently
known.

Esophagus

All chiropterans have a quite narrow esophagus, as they usually chew up their food
thoroughly before swallowing. This is the same in nectar feeders which take in fluid food
or very small particles. Compared to insect-eating chiropterans, the esophagus epithelium
is much thinner in nectar feeding bats (Dobat & Peikert 1985) and not keratinized as in
pure insect-eaters.

Stomach

The stomach of blossom bats is designed to take large quantities of fluid within a rather
short time. According to Howell (1979), Leptonycteris will absorb 4g - corresponding to
22% of its body weight - within just 20 minutes. Interestingly, the muscle layer of the
stomach is very thin. In macroscopical respect there is a conspicously oversized blind sack
and an enlarged pylorus area, both features contributing to the necessary volume capacity.
Furthermore, the low proportion of pepsinogen producing cells within the fundus glands
of the mucosa (Rouk & Glass 1970, Forman 1971) correllates with the diet being
comparatively poor in proteins.

20

Intestines

Generally, Chiroptera have quite short intestines, presumably a concession to their flight
ability. Thus, indigestable bits are expelled after a surprisingly short time. Among bats,
the frugivorous species have the longest intestines, whereas the latter are very short in
insect-eaters and flower bats - probably due to their diet containing more energy. This
may also explain their astuteness in flight.

Adaptations in host plants to Chiropterophily
Size and Constitution

Despite of their low weight - even in terms of chiroptera - nectar feeding bats represent
heavyweights compared to other pollinators. This influenced both the structure of the
blossoms and the entire constitution of chiropterophilous plants. Though the plants belong
to completely different taxa they do share some common features: according to Baker
(1961), plants which are supposed to be pollinated by bats must be strong and thus are
usually tree-shaped. So, we find the following form types in order of frequency (Dobat &
Peikert 1985):

. Trees (e.g. Ceiba, Crescentia, Parkia)

. Shrubs (e.g. Symbolanthus)

. Pillar cacti (e.g. Carnegia)

. Lianas (e.g. Mucuna)

. Epiphytes (e.g. Capanea, Markea, Trienaea, Vriesea)

. Herbaceous plants (e.g. Agave, Musa)

. Herbs (e.g. Lisianthus)

oY NR

aA Nn

Undoubtedly, most of them are woody, tree-shaped plants. Though the existence of quite
low, ground growing chiropterophilous herbaceous plants and herbs seems to be incom-
patible with these physical demands, Baker’s view as cited is nevertheless supported in
two respects: first, these plants are quite rare, and second, they tend to gigantism. All in
common their blossoms are shifted as high as possible, thus facilitating orientation for the
bats approaching them and at the same time reducing access for unspecific and thus less
efficient nectar consumers.

Blossom shape

There is an enormous variety in adaptations which cannot be discussed in detail here
(review in Dobat & Peikert 1985). It is, however, interesting how some chiropterophil
plants adapted their blossom shape to the head morphology of selected pollinating bat
species. Despite their various shapes (bellflowers, funnel-shaped blossoms, dish-like
flowers, tubular blossoms, labiate flowers, papilionaceous flowers, capitulum flowers,
spadiciform flowers, spadiciform brush-like flowers, brush-like flowers, brush-like
bellflowers) there is a common feature: the anthers always extend beyond the corolla, so
the blossom shape forces the pollinator into a position allowing the pollen to be fixed
within the fur (face, neck and shoulders) guaranteeing that any contaminated fur area will
most probably touch the stigma of the next blossom to be visited.

Blossom exposition

Blossom exposition represents an important further characteristic of chiropterophilous
flowers, facilitating access to the blossoms or inflorescences by shifting them out of the

2

range of disturbing foliage. Additional aspects are mentioned by v. Helversen (1993): open
exposure of flowers give the glossophagine pollinators space for wing movements during
hovering flight thereby also minimizing the risk of encountering predators (better visual
control and shorter stay). This is achieved in several ways (van der Pijl 1957):

Flagelliflory or Penduliflory

Here, the host plants develop one or more long thin stalks of the inflorescence which is
usually pendulous but may sometimes point off the stem almost horizontally. The length
of the flagellae varies between 0,6 and 5 m!

Cauliflory
The blossoms are arranged along the stem or along the main branches, also facilitating
approaches of the bats (van der Pijl 1936). Example: Ceiba pentandra.

Pincushion blossoms

In this configuration, the inflorescences are arranged spherically emerging everywhere
from the foliage.

Towering individual inflorescences

As the herbaceous plant does not grow very tall (e.g. Agave), it develops a (tree-shaped!)
inflorescence and increases probability of becoming exclusively pollinated by bats.

Developing storeys
By arranging the leaves in distinct storeys, the inflorescences are separated from the
remaining vegetation area (Ceiba).

Diet of various New World nectar feeding bat genera
Analysis of stomach contents in captured specimens revealed some information on their
diet. Additionally, many captured bats still carry pollen in their fur (especially around the
muzzle, but also on the shoulder or on the neck) allowing to identify or at least to draw
conclusions about the plant species they visit.
Lionycteris: Although the diet of L. spurrelli is still unknown, it may resemble that of
Lonchophylla (Gardner 1977). In Peru, one specimen of L. spurrelli was captured among
blooming Cashew trees (Tuttle 1970).
Lonchophylla: According to Walker et al. (1964), Lonchophylla feeds on blossoms, taking
in nectar, pollen, but also insects and fruit. Similar reports on Panamanian species of
Lonchophylla were given by Duke (1967) who reported of a diet consisting of nectar,
over-ripe fruit, pollen and insects. This may also apply to the remaining species of the
genus (Gardner 1977).
In L. thomasi, the faeces and the contents of stomach and intestines were analyzed: in five
specimens from east Peru Koepcke (1987) found larger amounts of pollen (2 specimens),
pulp and seeds of Piper sp. (1 specimen), remnants of unidentified fruit (3 specimens) as
well as vaious thoroughly chewed and indeterminable insects (4 specimens). One of the
bats contained a yellowish fluid, probably nectar or fruit juice. Another specimen was
covered with pollen on its head, breast and on the wing membranes; its faeces also
revealed nothing than pollen. Gardner (1977) observed L. thomasi at banana blossoms
(also in east Peru), and these bats were covered with pollen on their head and shoulders,
too.
For L. mordax Gardner (1977) reported insects, fruit, nectar and pollen, without, however,
specifying the plant diet. For six L. m. concava caught in Costa Rica, Howell & Burch
(1974) identified the following particles: nectar and pollen of Mucuna sp.(1 ind.), nectar
and pollen of Musa sp.(2 ind.), remnants of lepidopterans (3 ind.).

22

L. robusta: Pollen, nectar, fruit and insects (Gardner 1977). Wille (1954) considered L.
robusta as a nectar feeding bat, though stomach analysis of 17 specimens from Costa Rica
and Panama by Fleming et al. (1972) revealed 90% insect remnants (unfortunately, only
one analysis was usable at all). Howell & Burch (1974) failed to detect any plant material
in three specimens of L. robusta from Costa Rica, instead they found remnants of
Lepidoptera, Coleoptera and Streblidae (= ectoparasites on bats).

Platalina: The diet of P. genovensium is unknown; it probably consists of pollen, nectar
and insects (Gardner 1977).

Brachyphylla. B. nana consumes fruit, pollen, nectar and insects (Gardner 1977;
Swanepoel & Genoways 1983). Stomach contents of 43 specimens from Cuba consisted
of partly digested pollen grains. One stomach contained butterfly scales, another one
fragments of a fly (Silva Taboada & Pine 1969). Furthermore, these authors regularly
found individuals whose head, breast and shoulders were powdered with pollen.
Consequently, Silva Taboada & Pine (1969) classified B. nana as chiefly pollen feeders,
probably adding soft fruit and nectar to their diet.

According to Gardner (1977), B. cavernarum feeds on fruit and insects, the fruit
predominantly being taken from Manilkara zapota (Nellis 1971), papaya (Carica papaya),
mango (Mangifera indica), almond (Terminalia catappa), royal palm (Roystonea
boringuena) and Cordia sp. (Nellis & Ehle 1977). These authors also reported on captive
individuals which took bananas, apples, pears, peaches and melons - but never citrus fruit
- apart from the blossoms of Ceiba pentandra, Thespesia populnea, royal palm and
Hymenaea courbaril. During field observations, Nellis & Ehle (1977) failed to distinguish
between pollen and nectar intake, but most of the faeces beneath their roosts contained
pollen (Swanepoel & Genoways 1983).

Erophylla: E. sezekorni takes various fruit, pollen, nectar and insects (Gardner 1977). The
earliest descriptions on food intake of this species date from the second half of the 19th
century (Osborn 1865): fruit of Cordia collococca, whose soft parts are licked up. Hall &
Kelson (1959) called this species “Buffy Fruit Bat”. Silva Taboada & Pine (1969)
analyzed the stomach contents of 30 E. sezekorni from Cuba: in all individuals they found
partly digested pollen grains. Three of them contained seeds of Hohenbergia
(Bromeliaceae); in four specimens they detected insect remnants, including parts of an
elaterid beetle (Conoderus, Elateridae), of a cockroach (Blattidae, Orthoptera) and various
undetermined Diptera and Lepidoptera.

Hall & Kelson (1959) called E. bombifrons "Brown flower bat””; Tamsitt & Valdivieso
(1970), however, reported this species as frugivorous (Gardner 1977).

Phyllonycteris: P. poeyi probably feeds on a variety of fruit, pollen, nectar and insects
(Gardner 1977). With respect to the tongue anatomy, Allen (1942) supposed P. poeyi to
eat pulp, fruit juice, pollen and nectar. Silva Taboada & Pine (1969) analyzed the stomach
contents of 42 individuals from Cuba and found partly digested pollen masses. Only one
stomach contained lepidopteran scales.

Glossophaga: Presumably due to its conspicuously elongate tongue, G. soricina formerly
was considered a blood feeder. Later it was supposed to eats insects, until Goodwin &
Greenhall (1961) revealed that it feeds on nectar, soft fruit and possibly on seeds (Husson
1962). Gardner’s (1977) substantial information on the diet of this genus did not only
mention nectar, flower parts (blossom constituents) and fruit, but also insects. In captivity
(large flight cages) Glossophaga hunted and ate insects deliberately; and insects were also
the favourite food of captive individuals having been kept for 14 months in El Salvador,

23

showing, by the way, some interesting shifts in diet preference: prior to the rainy season,
the animals preferred honey water, durig the rest of the time they liked insects most. Cap-
tive Glossophaga accepted honey water or fruit juice taking it from a shallow bowl during
hovering flight (Novak & Paradiso 1983). Having analyzed the stomach contents of 217
individuals from Costa Rica and Panama, Fleming (1972) described Glossophaga as an
omnivorous genus. Only 38 stomachs were completely empty, the remaining contained
34% plant material and 66% insect remnants. This also corresponds to the results of
Alvarez & Gonzalez (1970) from Mexico where 61% of 174 stomachs examined did not
contain any pollen at all. It is, nevertheless, wogth mentioning that from all
Glossophaginae studied so far, G. soricina showed the greatest variety of different pollen
grains (deriving from at least 34 plant families).

For G. commissarisi from Costa Rica, Howell & Burch (1974) reported remnants of
lepidopterans, fruit (Acnistus) as well as pollen and nectar of Musa and Mucuna. Insects,
fruit, pollen and nectar are noted by Gardner (1977).

G. longirostris: Insects, fruit, pollen, nectar and probably other blossom parts (Gardner
1977). Wille (1954) and Valdivieso & Tamsitt (1970) considered G. longirostris a
nectarivorous species. Goodwin & Greenhall (1961) reported a diet of nectar and pulp,
fruit juice and, occasionally, insects.

Monophyllus: Up to now no reliable reports. McNab (1971), Phillips (1971) and other
authors supposed Monophyllus to feed on soft fruit or nectar, possibly on insects, too.
Tamsitt & Valdivieso (1970) failed to sustain captive specimens of M. redmani by means
of banana pulp and sugar water, as the bats refused any food. For M. plethodon there is no
information available.

Lichonycteris: Up to now hardly reliable reports. According to Tuttle (1970) and Handley
(1976), the development of snout, tongue and molars support to assume that the members
of this genus feed on nectar, pollen and probably fruit. Tamsitt & Valdivieso (1961)
classified Lichonycteris as fruit- and nectar feeders. Carter et al. (1966) reported two
specimens they captured next to a night-blooming plant in Guatemala carrying pollen on
their fur and on the tail membrane.

Leptonycteris feeds on nectar, pollen, fruit and insects (Novak & Paradiso 1983), the latter
comprising only a small proportion and thus may have been eaten accidently along with
the nectar and pollen (Hoffmeister 1957). On the other hand, Rasweiler (1977) pointed out
the significance of insect consuming for a healthy diet. As Walker (1964) assumed, the
long snout reaches the spine-free parts of cactus fruit; the canines are used to rip the
pericarp, and the juice is licked up with the tongue. This genus is characterized by
accumulations of yellow or red faeces beneath the roosts, pointing to a diet of pollen,
nectar and fruit juice. Correspondingly, Dalquest (1953) reported on L. nivalis he captured
in San Luis Potosi, Mexico, their stomachs filled with viscous, bright red fruit juice ..
“almost certainly the juice of the fruit of the organ cactus”. Blossoms of Agave scabra, A.
chisosensis and A. lechugilla (Easterla 1972), Agave schotti and Carnegia gigantea
(Cockrum & Hayward 1962) are also reported. The stomachs of 13 L. nivalis from
Michoacan and Hidalgo, Mexico, contained pollen grains of 22 identified plant species
from the genera Agave, Ipomoea, Ceiba and Myrtillcactus (Alvarez & Gonzalez Q. 1970).
L. yerbabuenae has been observed on the blossoms of Malvaviscus, on blossoms and fruit
of cactus and presumably also on the blossoms of Datura stramonium (jimsonweed)
(Novak & Paradiso 1983; Davis 1974; Schober 1984).

On L. curasoae confirmed reports are not given yet, but this species probably feeds
similar to other species of its genus (Gardner 1977).

24

Anoura: According to Nagorsen & Tamsitt (1981), this genus is characterized by
opportunistic insectivory, additionally feeding on pollen and nectar.

A. caudifer takes fruit, nectar, pollen and insects (Gardner 1977). As Sazima (1976)
reported, A. cultrata picks insects from the substrate (foliage gleaner, Wilson 1973). The
stomachs of four individuals from Venezuela contained both insect fragments and a
creamy fluid. Eight specimens from Panama contained yellow, white and greenish masses,
respectively and in two cases unidentified insects. 18 individuals from Colombia had
pollen and plant fibres in their stomachs (Tamsitt & Nagorsen 1982). Both individuals
Starrett (1969) based on his description of Anoura werckleae carried Hibiscus (Wercklea
lutea) pollen in their fur. A. cultrata from Costa Rica was described to eat pollen and
nectar (Laval & Fitch 1977), whereas Howell & Burch (1974) found lepidopterans in the
stomachs.

Gardner (1977) gave a list of various plants whose blossoms were known by several
authors to have been visited by Anoura: Vochysia, Symbolanthus latifolius, Purpurella
grossa. Additionally, he emphasized the high percentage of insect food in A. geoffroyi,
pointing to the fact that some of these blossoms do not give any nectar at all so that the
bats probably visite them just because of the insects which are attracted by the scent
(Goodwin 1946). This is supported by Alvarez & Gonzalez Q. (1970) who found pollen
in more than the half of 69 specimens from Mexico; most of this pollen came from
entomophile plants. So, they considered A. geoffroyi an insectivorous species with
occasional pollen intake.

Up to now there are no reports on food intake of Anoura latidens available.
Hylonycteris: Insects, pollen and nectar (Gardner 1977). Goodwin (1946) supposed AH.
underwoodi to visit flowers; Hall & Kelson (1959) reported on nectar consuming, and fruit
remnants of the jobo plum (Spondias lutea) they detected beneath a day shelter in
Veracruz, Mexico, gave evidence of frugivory (Hall & Dalquest 1963). Carter (1966)
found pollen grains on rump and uropatagium of a specimen he caught in Guatemala next
to night-blooming flowers. There is a description from Tabasco, Mexico, by Villa-R.
(1967) of one specimen with cocoa pollen (Theobroma cacao) on its whiskers and head
fur. Analysis of stomach contents by Alvarez & Gonzalez Q. (1970) revealed exclusively
pollen (Lonchocarpus 99,8%, only 0,2% Agave and Pinus) for two H. underwoodi from
Chiapas, Mexico. Early reports on insectivory were given by Howell & Burch (1974) who
found remnants of lepidopterans in one specimen from Costa Rica.

Scleronycteris: Most probably fruit, pollen, nectar and insects; up to now no information
on the feeding ecology (Gardner 1977).

Choeroniscus: Presumably pollen, nectar and insects (Gardner 1977); no valid information
available yet. In his description of C. godmani, Villa-R. (1967) relied on analysis of
stomach contents by Goodwin & Greenhall (1961) for C. intermedius from Trinidad
Island: “Microscopical examination of the stomach contents of one specimen, however,
revealed some minute particles that are possibly honey or fruit juice, many fragments of
a coleopterous insect, and numerous brown and white, hair-like strands, probably either
from insects or from fruit. This specimen, at least, had fed to a large extent on insects”.”
Having examined four individuals from east Peru, Koepcke (1987) detected nectar in the
intestines of two specimens, pollen in one of them and in three cases some Coleoptera and
Hymenoptera as well as indeterminable plant material in two C. intermedius.
Choeronycteris: Fruit, pollen. Nectar and probably insects (Gardner 1977). Several authors
described C. mexicana as a flower-feeding bat (Dalquest 1953; Park & Hall 1951; Wille

22)

1954; Hall & Kelson 1959). Its host plants are reportedly Lemaireocereus, Myrtillocactus
and /pomea arborea (Villa-R 1967) as well as Ceiba and Agave (Alvarez & Gonzalez
1970). All the results on stomach contents and the host plants identified so far (all of them
are specialized chiropterophilous night-blooming plants) convinced Alvarez & Gonzalez
(1970) of the fact that C. mexicana is an obligate nectar feeder. Until now, there has been
no evidence on insectivory. Schaldach & McLaughlin (1960) detected C. harrisoni at
banana blossoms (Musa sp., therefore named the genus Musonycteris). Gardner (1977)
mentioned some pollen at the head and muzzle in some of the individuals having been
captured at a small banana plantation in Colima, Mexico, and which had been included in
the first description by Schaldach & McLaughlin.

As a conclusion, all taxa mentioned here have been either proved to feed on flowers or
they are most probably nectar feeders. As already stated in the introduction, the short-
skulled forms (Glossophaga, Lionycteris) but also Anoura frequently take insects,
predominantly beetles and moths. On the other hand, there is no evidence yet for
insectivory in taxa with an extremely elongate skull (Choeronycteris).

Sensory systems / Orientation

Acoustic perception; echolocation
Like all Microchiroptera, the nectarivorous phyllostomatids perform an efficient
echolocation. Especially the nose leaf certainly contributes to sound emission. Presumably
the lancet (upper part of the nose leaf) is necessary to focus the emitted sound bundles
vertically (Hartley & Suthers 1987).

Analyzing the sounds of various phyllostomid species, Griffin & Novick (1955) managed
to prove that echo location is also essential in orientation of nectar feeding bats. Further
investigation revealed the orientation pulses of the flower bats to be frequency-modulated
signals of 1-5 ms length (FM-sounds of the vespertilionid type).

Experimental investigation on the significance of acoustic perception in foraging was
performed by Howell (1974): the flower visiting species e.g. Glossophaga soricina,
Anoura geoffroyi and Choeronycteris mexicana emit 5-10 short searching pulses per
second, each of them lasting 0,5-2 ms. When approaching an obstacle (or aiming at a
blossom) the number of emitted orientation pulses increases to 30 signals per second, thus
enabling to assess distances precisely even at flight velocities of several meters per
second. When the bats were further tested on their ability to avoid obstacles, the
predominantly insectivorous species complied with the abilities of other Microchiroptera
(Myotis), whereas the species mainly interacting with chiropterophileous plants percepted
only much stronger wires. Determination of acoustic perception thresholds by means of
shunting off the cochlea potential did not reveal any diet specific differences but indicated
a polyphyletic origin of the subfamily (Howell 1974).

The importance of echolocation in pollinating bats is also documented by the development
of the acoustic cerebral areas (Baron & Jolicoeur 1980). Their progression indices come
quite close to those of insectivorous Microchiroptera.

Optical sense
Though in all microchiropterans a highly developed echolocation apparatus proves
dominance of the acoustic system over the remaining senses, in certain situations it may
be replaced or complemented by optical perception. So, visual orientation becomes

26

important beyond range of sound, for instance in order to identify large, far objects, land
marks or the horizon (Suthers 1966; 1970).

All phyllostomids have well developed eyes with efficient differentiation of brightness and
shapes. Flower bats always keep their eyes open, when active. Some nectar feeding
species (e.g., Anoura caudifer) are reported to have a tapetum (v. Helversen 1993) and
perform a well developed ability for pattern recognition. It is, by the way, interesting for
this respect that some bat flowers developed conspicous patches for “close range guiding”
the bats in approach (Dobat & Peikert 1985). But the absence of retinal cones gives no
evidence for colours to be discriminated (Suthers 1970). Anatomically however, the
optical areas in brain cortex are clearly less developed than the acoustic centers.

Olfactory sense / Olfactory perception

In fruit feeding bats, the leading role of food detection by olfactory sense has been
satisfactorily documented (for both Megachiroptera and frugivorous phyllostomids). This
is also confirmed in brain anatomy by relative size of the Bulbus olfactorius. Although
this structure turns out smaller in nectar feeding New World Microchiroptera, it still
remains considerably larger than in species which exclusively feed on insects having the
smallest Bulbi olfactorii among all Chiroptera (Dobat & Peikert 1985). Chiropterophile
blossoms are often characterized by a slightly sour, musty scent which is apparently
responsible for attracting pollinators. According to observations by Vogel (1958) a sudden
breeze finished pollinating activity immediately, which also gives evidence of the well
developed olfactory abilities of the nectar feeders. Olfaction does not only serve for long-
distance orientation but is also important in short-distance target discrimination - detection
of the nectarbearing flowers (v. Helversen 1993).

Reproductive Biology

Reproductive data of nectar feeding bat species is mainly based on comments on the
sexual status of captured specimens. Pregnant females give information about size and
weight of fetus; lactation periods are easily recorded from the dates of netted females
carrying juveniles. Development of youngsters, but also relative weight and measures of
gonades (enlargement of uterus, ovarian follicles, appearance of corpora lutea in females;
size of testicles in males) allow conclusions on seasonal breeding patterns by comparing
the different information to the date of capture.

So far, we still have poor knowledge on the reproductive behavior of nectarivorous
phyllostomid chiropteres: among the species of the tropical rain forest, breeding all over
the year without marked seasonal periods is common, whereas those inhabitating subtropic
(more arid) zones or andine mountain forests show one definitely seasonal or two separate
(bimodal) reproductive periods per year.

Lionycteris: Tuttle (1970) reported a pregnant female of L. spurrelli containing one single
embryo captured in Peru on August 5th.

Lonchophylla: Wilson (1979) took pregnant L. mordax in Costa Rica as well in March as
in August. Also in Costa rica LaVal & Fitch (1977) found pregnant L. robusta in
February, May, August and October; one lactating female in January. According to
Koepcke (1987) the reproductive period of L. thomasi in amazonian Peru occurs during
the dry season. She netted sexual inactive bats in June, October, November and December.

247

In contrast a female collected in July was pregnant. In September, a family, watched in
field by the same author, nursed a nearly full grown juvenile which still stayed with its
parents during following January. At Manaus, Reis (1981) found sexual active males
during dry season and at the beginning of the rainy season.

Platalina: No data on reproductive biology yet available for this very rare endemic
peruvian genus.

Brachyphylla: Twelve female B. nana trapped on Middle Caicos Island in March all were
pregnant, with crown rump length of fetuses between 24 and 34 mm (Buden, 1977). In
contrast females collected on Hispaniola in December and late August were not pregnant,
but one of the August females was lactating. The testes of one male netted during the
same time were only 3 mm long (Klingener et al. 1978). On Cuba, female B. nana carried
embryos from December through May, lactation ocurred from May to August; the
diameter of the testes of males varied from 5 to 9 mm in specimens caught in December
(Silva-Taboada 1979).

On Puerto Rico, nursing females of B. cavernarum have been collected on 5th July, but
there was no information about the young (Anthony 1918). Later studies on 25 females
(small uteri, no suspicious ovarian follicles) and males (testes 46 mm ) from St.Croix
gave no evidence for reproductive activities in December (Bond & Seaman 1958). Walker
et al. (1964) mentioned nursing females from Puerto Rico in July; later reports of the
same authors (1975) stated pregnant females in February and a lactating female in April.
On St. Croix, pregnant females were observed in March, and it was here that Nellis (1971)
collected a nursing female in April. Detailed observations by Nellis & Ehle (1977) on a
colony on St. Croix in the time between May and June showed the colony consisting of
pregnant females only, which give birth to their young during that time.

Baker et al. (1978) collected 15 adult females on Guadeloupe in July; none of them was
pregnant but three were obviously nursing. Males netted at the same time showed testes
of 4-6 mm length. Thus B. cavernarum probably has a more synchronized reproductive
cycle than, for instance, Artibeus. Also Wilson (1979) suggested a synchronized, probably
bimodale reproductive cycle for B. cavernarum, a second period of parturition occuring
annually at least in some populations.

Erophylla: Eleven (of approximately twenty) female E. sezekorni taken in Cuba on 26th
and 28th February contained small embryos (Anthony 1919). Buden (1976) summarized
the reproductive behavior of this species: “Most prenatal development takes place during
the first part of the year and parturition probably occurs in early summer.” Females
bearing young embryos were collected in early and late February. Individuals with well
developed fetuses were obtained in April and May. Lactating females were collected in
June and many immatures in July. Nearly adult youngsters were found in August. Thus E.
sezekorni seems to be a seasonal breeder possibly bearing only one single offspring per
year.

Pregnant E. bombifrons were captured on Puerto Rico by Valdivieso & Tamsitt (1971) in
June and July.

Phyllonycteris: Parturition in P. poeyi takes place mainly in June (Novak & Paradiso
1983). Goodwin (1970) trapped a pregnant female in January; Baker et al. (1978) reported
three gravid specimens from Haiti on 17th December.

Glossophaga: Nursing colonies containing several hundreds of female G. soricina and
their young were found in San Luis Potosi (México) during midsummer; in Guerrero (also

Mexico) in midsummer the studied colonies were formed of both sexes. My own ob-
servations in Ecuador during July 1983 dealed with both sexes within the same large
colony including pregnant and lactating females. In Veracruz (Mexico) Hall & Dalquest
(1963) found mixed colonies with more than 1000 individuals.

According to Wilson (1974) who compiled data of pregnant females for all months over
the year Glossophaga occured to be polyestric at most of the collecting sites.

Reproductive patterns and ontogeny of G. soricina have been studied extensively (Bleier
1979, Rasweiler 1972, 1974, 1979, Wilson 1979) and are reported by Alvarez, Willig,
Knox Jones & Webster (1991): ovulation is spontaneous and usually one ovum is released
per cycle. Ovulation may occur from either ovary, but tends to alternate between both.
Menstruation and ovulation take place at approximately the same time. The two-cell
development stage is achieved on day 2 or 3 after fertilization, the 8-cell stage by days 5
to 7, the 32-cell stage by day 8, and the blastocyst stage by day 10. The embryo is
contained within the ampulla of the oviduct until day 12 or 13, by which time the Zona
pellicucida usually is lost. Implantation occurs in the uterotubal junction on days 12 to 14.
Rasweiler (1974) divided the process of implantation into eight stages and Hamlett
(1935)described the embryonic growth thereafter. Glossophaga shows a discoidal and
haemochorial placenta. The occurance of menstruation and interstitial implantation
suggests that Glossophaga might posess considerable potential for development as an
animal model in human reproductive research (Rasweiler 1974).

Wilson (1979) found pregnant G. commissarisi in January, February, April, July and
September. This indicates a bimodal polyestrus. LaVal & Fitch (1977) report a seasonal
polyestrus on G. commissarisi in Costa Rica, their data on pregnant females refer to
February/March and October.

According to Wilson (1979) G. longirostris nurses its youngsters during rain period; the
data of capture show pregnant or lactating females from February to September.

Webster (1983) collected pregnant G. leachii (containing one single fetus each) in
February, April, June, July, August, September and November. Nursing mothers were
obtained in February, March, June and November.

G. mexicana is supposed to be monestric, the duration of breeding season remains unclear:
a pregnant female was collected in March, a lactating specimen in May. Other females
caught in February, March, April, May and August gave no evidence of reproductive
activity. Four males taken in June had testicle diameters of 4 mm; the testes of another
male captured in July measured 8x6 mm (Webster & Jones 1985).

Monophyllus: Buden (1975) reports pregnant M. redmani (each with a single fetus): on
28th January he obtained one female on Middle Caicos (Bahamas), its fetus with crown
rump length of 20 mm. On 3rd December and 24th February (on Hispaniola) three
specimens containing fetuses of 16-19 mm length. One from Puerto Rico was gravid on
5th February.

Pregnant females of M. plethodon were taken on Dominica between 24th March and 22nd
April. Crown-rump-length of fetuses variied from 17 to 24 mm; the larger ones were
caught later. Males captured at the same time had testes 4-4,5 mm long (Homan & Jones
1975).

Lichonycteris: On Costa Rica, Gardner, LaVal & Wilson (1970) reported a nursing female
collected together with a male young on 9th January. Another specimen taken in March

contained a 14 mm embryo. In Guatemala pregnant females are also dated in February
(Wilson 1979).

29

Leptonycteris: These bats form large colonies homing more than 1000 individuals. In their
northern habitats nursing females aggregate during springtime into breeding colonies
numbering thousands of animals; Smith & Genoways (1974) reported a colony of L.
curasoae on Isla Margarita (Venezuela), containing almost 4000 females and nearly adult
juveniles. In November no more juveniles but pregnant females and reproductive males
were found.

In Texas and Mexico, young L. nivalis appear to be born during summer (Davis 1974). In
contrast Wilson (1979) caught pregnant L. sanborni in Mexico as well in February, March,
April as in July, September und November.

Anoura: Pregnant and lactating A. caudifer were collected in January, February, May, June
and November by Carter & Jones (1978). Gardner (1970) reported on A. cultrata in
Columbia: a female taken in August carried a fetus of 28.5 mm length. Two specimens
captured in west central Colombia on 17th July aborted well developed fetuses ( 20 and
21 mm long); and lactating females were found on 30th and 31st July (Lemke & Tamsitt
1979). In southwestern Colombia the same authors collected three females on 10 August,
each contained a single embryo (11-14 mm crown-rump length). In Peru Carter (1968)
took lactating females on 16th und 21st August. Usually female A. cultrata bear a single
offspring, but there is also a report on twins (Tamsitt & Nagorsen 1982). The data
obtained of captive males in Costa Rica revealed sexually active individuals (testes > 6
mm) in February, May and July; in Panama in February; and in Columbia in May, July
and early August. Testes of males collected in March and April in Venezuela and in late
August in Columbia and Peru were smaller (1-4 mm) than those of specimens taken in
other months (Tamsitt & Nagorsen 1982).

The data for A. geoffroyi compiled by Wilson (1979) suggest this species on Trinidad to
form colonies of separated sexes within the same caves during particular seasons. In June
there were 20 males and 25 females in one cave; in October 29 males and only one
female; in November 32 male and 56 female bats. In this region A. geoffroyi obviously
give birth to its offspring at the end of raining season, so pregnant females were found in
November. In Nicaragua pregnant females were taken in July, in Costa Rica in March and
in Peru in June and July. In Mexico nursing mothers were found im July, November and
December (Carter & Jones 1978; Wilson 1979).

Hylonycteris: Carter (1966) mentioned a lactating female from Guatemala, captured on
2nd March. For Jalisco (Mexico) there are data by Phillips & Jones (1971) on three
pregnant female H. underwoodi collected in early September each bearing a single fetus
of 14, 18 und 21 mm crown-rump length. In December Hall & Dalquest (1963) took a
male with “small testes’; Gardner (1970) describes the testes of three males caught in
Costa Rica in February and one April and July as “moderately enlarged, averaging in
2503 TNs

Choeroniscus: Pregnant females of C. godmani were netted in Mexico during May, in
Sinaloa (Mexico) in July, in Nicaragua during March and in Costa Rica in December,
January, February and March (LaVal & Fitch 1977; Wilson 1979).

During her field work Koepcke (1987) watched a female C. intermedius with a newborn
baby in the amazonian rainforest in Peru in late June. Animals captured in the months of
August, November und December showed no reproductive activity. On Trinidad a
pregnant female is noted in August, probably this species is bimodal polyestric (Tuttle
1970).

30

Table 3: Chromosomal data on New World nectar-feeding bats (adapted from Baker 1979)

Notes
Genera and species are given in alphabetical order (inclusive species without informations)

Key to abbreviations: 2n = diploid chromosome set; FN = number of chromatids; M = metacentric; SM =
submetacentric; ST = subtelocentric; A = acrocentric.

Taxon - 2n FN x Y NG, Autor no

Anoura brevirostrum - - -

Musonycteris harrisoni - - -

Platalina genovensium - - -

A. caudifer 30 - - - - Yonenaga 1968 -
30 56 SM A - Baker 1973 -
A. cultrata 30 56 SM A - Baker 1979 1
A. geoffroyi 30 56 SM A - Baker 1967; Hsu et al.1968 -
30 - SM A - Baker & Hsu 1970 3
- - SM - - Pathak & Stock 1974 -
A. werckleae - - - - - - -
Choeroniscus godmani 19 32 SM Sit A Baker 1967 5)
19 - - - - Hsu et al. 1968 5
19 32 SM A A Baker 1970a 1
20 36 SM - - Patton & Gardner 1971 1
20 36 - - - Baker 1979 -
Ch. inca - - - - - - -
Ch. intermedius 20 36 - - - Baker 1970a -
20 - - - - Baker 1973 -
- - SM - - Pathak & Stock 1974 1
20 36 SM A - Stock 1975 1
Ch. minor - - - - - - -
Ch. periosus - - - - - - -
Ch. mexicana 16 24 - - - Baker 1967; Hsu et al.1968 1
16 24 SM SM - Baker 1973 -
Glossophaga alticola 32 60 M A - Baker 1967 4
G. commissarisi 32 60 M A - Baker 1967; Hsu et al.1968 5
G. longirostris 32 60 M A - Baker 1979 =
G. soricina 32 60 M A - Baker 1967; Hsu et al.1968 14
32 60 M A - Baker & Hsu 1970 4
32 60 SM A - Baker 1970a 1
Hylonycteris underwoodi 16 24 - - - Baker 1973 -
Leptonycteris curasoae - - - - - - -
L. sanborni 32 60 M A - Baker 1967; Hsu et al.1968 5)
L. nivalis 32 60 - : - Baker 1973 -
Lichonycteris degener = = = - = = 2
L. obscura 28 50 SM A - Baker 1973 1
24 44 - - - Baker 1979 22
Lionycteris spurelli 28 50 SM A - Baker 1979 1
Lonchophylly concava - - - - - - -
L. hesperia - - - - - - -
L. mordax - - - - - - -
L. robusta 28 50 SM A - Baker 1973 -
L. thomasi 30 34 - - - Baker 1973 -
32 38 - - - Gardner 1977 -
Monophyllus plethodon 32 60 SM A - Baker 1979 3
M. redmani 32 60 SM A - Baker & Lopez 1970b 7

Scleronycteris ega - - -

3

Choeronycteris: There is an outline by Wilson (1979) on C. mexicana: in Mexico females
are pregnant in spring. Those which migrate to Arizona and New Mexico there give birth
to their young during June/July. This species is monestric, but may have a second
breeding season per year, for in Jalisco a pregnant female has been caught in September
(Watkins & al., 1972). According to Barbour & Davis (1969) parturition in C. mexicana
takes place within 15 min. Newborn young are well developed and also furred.

Cytology

The New World Phyllostomatidae have been subject to thorough cytological examination.
Above all, the team of R.J. Baker, Texas Tech University in Lubbock, Texas, published
numerous caryological and cytogenetic papers on this subject. There are also detailed
chromosomal data on nectarivorous genera (cf. tab.3, from Baker 1979).

It is striking that even species within the same genus often show considerable differences
in their caryotype, Warner (1983) referred to this phenomenon as “Caryotypic
megaevolution”. It is, therefore, hardly surprising that by means of cytogenetic analysis
completely contradictory relationships were postulated by different authors, one example
being parallel evolution of a multiple sex chromosome system (Patton & Gardner 1971)
in Carollia and Choeroniscus: as the males in both genera have a XYY-configuration,
they were supposed to be related (Hsu et al. 1968). Further studies emphasized the weak
points of the “G-Banding Patterns”, thus preferring the C-banding analysis (hete-
rochromatin technique). Here, anyway, specimens from both genera showed the original
XY type, so chromosomal configuration seems to undergo comprehensive evolutionary
changes.

MATERIAL AND METHODS
Material

This study is based on skulls and specimens preserved in alcohol. The material comprises
29 genera from the subfamilies Phyllostominae, Carolliinae, Lonchophyllinae,
Brachyphyllinae, Phyllonycterinae and Glossophaginae.

Some of the individuals examined were captured during a three-week study visit to
Ecuador (July 1983), visiting locations in the surroundings of Quito (San Antonio de ~
Pichincha, 2100 m above mean sea level), in the secondary forest of West Ecuador
(Chontillal) and in the rain forest area east of the Andes (Rio Cuyabeno, Amazon
headwater region, Cueva de Jumandi). In the course of this journey, the available bat
collection of the Museum of the Escuela Politecnica Nacional (MEPN), Quito, could be
accessed and studied.

Some of the genera worked on here are known only by very few specimens. Thus it was
necessary to examine some of the extremely rare material personally in the collections.
Consequently, the following museums were visited:

- Zoologische Staatssammlung Miinchen

- Naturhistorisches Museum Wien

- Rijksmuseum vor Natuurlijke Historie, Leiden

- Zoologisches Forschungsinstitut und Museum Alexander Koenig, Bonn

32

- Naturhistorisches Museum Basel
- British Museum of Natural History, London
- American Museum of Natural History, New York.

Additional material was kindly lent by the following museums and collections:

- American Museum of Natural History, New York, (AMNH)

- British Museum (Natural Hiatory), London, (BMNH)

- Collection Dr. Juliane Diller, geb.Koepcke, München, (JK)

- Musée d’Histoire Naturelle, Geneve, (MHNG)

- Naturhistorisches Museum, Basel, (NHMB)

- Naturhistorisches Museum der Alexander v. Humboldt Universität zu Berlin (MNHUB)
- Naturhistorisches Museum, Wien, (NHMW)

- Museum Kopenhagen

- Rijksmuseum voor Naturlijke Historie, Leiden, (RMNH)

- Royal Ontario Museum, Toronto, (ROM)

- Senckenberg Forschungsinstitut und Naturhistorisches Museum, Frankfurt, (SMF)
- Smithsonian Institute, United States National Museum, Washington, D.C. (USNM)
- Zoologisches Forschungsinstitut und Museum Alexander Koenig, Bonn, (ZFMK)

- Zoologisches Institut und Zoologisches Museum der Universität Hamburg, (ZIM)

A complete list of all material is given in the appendix.

Methods

Descriptions
Gross morphology

Both alcohol preserved specimens or skins, were examined using a stereomicroscope. All
descriptions comprised also body size, characteristics of head morphology (nose leaf,
auricles, tragus, lower lip) and features of the pelage (colouring, areas of the body, hair
shaft) as well as the bones of the wings (metacarpalia, phalanges) and tail length. The
insertions of the plagiopatagium at the hind extremities and the shape of the uropatagium
are of special diagnostic value.

Skull morphology

Skulls were examined with a stereomicroscope using various magnifications (5x; 12.5x;
30x). For most overviews, five times magnification was sufficient, whereas features of the
dentition often required some higher magnification. If possible, the description considered
rostrum length compared to total skull length, the skull contours in lateral view as well as
the arrangement of the skull basis towards the level of the palate. Development of the
zygomatic arches was equally mentioned as were the features of the skull base: pterygoid
processes, convexity of pre- and basisphenoid (,,basisphenoid pits”) and basioccipitale.
In all genera, the dentition was documented by dental formulas.The upper incisivi, canini,
premolars, molars, lower incisivi, lower canini, lower premolars and lower molars were
briefly described referring to specific peculiarities (contour of edges, height of crowns,
interdental distances, development of the masticatory surfaces, relative dimensions).
More detailed descriptions of the glossophagine dentition are given by Phillips (1971).

eS)
195)

Drawings
All specimens borrowed could be documented by drawings of the skull in dorsal, lateral
and basal view. Using a stereo microscope with a drawing projector, the mandibles were
drawn in top view and in lateral view. The scale is always 10 mm.
The material examined during museum stays was photographed by means of a macro lens
(135 mm) on fine-grain black and white negative film; the drawings were done after these
pictures.

Measurements

The preserved skulls were measured by means of a slide caliper (0.02 mm). Any skull
measure was taken three times and the mean value was recorded on prepared record
forms. The measures are given in fig.2. A table with all values measured can be ordered
from the author.

Fig. 2: Measurements

BL = basal length MB = mastoid width

CBL = condylobasal length M-M = width over molars

C-C = width over incisivi OZR = upper tooth row (C-M’)

CCL = condylocanine length POB = postorbital width

CH = height of coronoid process SGL = total skull length

GL = length of palate UZR = lower tooth row (C-M,)

HSB = width of braincase UKL = mandible length

HSH = height of braincase ZB = zygomatic width; (distance between prs.

IOB = interorbital width zygomata)

34

To see how differences in cranial measures (i.e. ”y”) depend on skull size (i.e. ”x”), I
decided to calculate an artificial volume quantity called SIZE. In order to exclude
influence of rostral length on the chosen reference quantity, SIZE (=x) is determined by
neurocranial measures only (cf. fig.3).

Fig.3: SIZE = (Condylocanine Length minus Upper Tooth Row) * Height of Braincase * Width of
Braincase

Calculations

Univariate Analysis

For each skull measures in each species, the mean value and standard deviation was
calculated. If possible, males and females were compared to each other as seperate random
samples (F-test, t-test).

Allometrics

In individuals of different size, allometric calculations allow to distinguish proportional
shifts exclusively based on size increase from differences in proportion caused by other
reasons. Often, such differences in proportion reveal deviating construction principles
which can also be evaluated in a taxonomical respect.

Mathematically, allometrics are recorded by means of the allometric formula:
y = b * x’
log y = log b + a * log x

This causes comparision of measures representing different dimensions: units of length
against volume measures. Thus, for allometric analysis instead of the usual regression
lines the reduced elliptic major axis of distribution (Rempe 1962) was referred to.

All necessary calculations were performed on a IBM compatible personal computer by
means of a regression calculating software written in BASIC (D. Vinyard, after
Miiller/Kick 1983, see appendix 9.4) as well as DIVA (Ver.2.O), a statistics software by
D. Plogmann. All calculations were based on the works of Rempe (1962).

33

RESULTS

Morphology of the species examined

External morphology

Many of the bats described here resemble each other to an enormous extent and may only
be identified by delicate characters. External characters relevant for identification are the
development of the uropatagium, the insertion of the wing membrane at the foot or at the
tibia, presence and development of a calcar, the length relations of the bones supporting
the wings, shape and relative size of the nose leaf or the auricle and the tragus,
respectively.

Generally, in all phyllostomid flower bats, the lower lip shows a distinctly V-shaped
median notch laterally lined by small warts. The elongate tongue covered with brush-like
hairy papillae is clearly visible also in living oder undissected specimens. In all long-faced
species an elongate rostrum will attract attention.

Lionycteris

L. spurrelli: Very small bat; forearm length <35 mm. Muzzle imperceptibly elongated
compared to non-nectar-feeding phyllostomids. Relatively large nose leaf (wide and short).
Auricles small and rounded, dark. Lower lip showing a very small median V-shaped notch
of warty bulges with a tip ending ventrally in an unpaired wart (fig.4).

Tragus short with small serrated basal wings (ref. fig.4). Pelage reddish brown to black
brown, lighter ventrally (greyish brown); wing membranes darker than fur.

Forearm sparsely haired, very short thumb. Metacarpalia III>IV>V, 1. phalanx III>IV=V.
Plagiopatagium inserts distally at the tibia. Uropatagium present. Short tail projecting
beyond the wing membrane in its last third and ending with a perceptible stump (fig.4).

sv wy

Th»

) Loe
rae |
aS NS

Fig.4: Lionycteris spurelli, head, tragus, and interfemoral membrane

36

Lonchophylla
Resembling in appearance Lionycteris, but with distinctly stretched head. Compared to
Lionycteris, the nose leaf is long and narrow. Auricle rounded, tragus with smooth

contours (fig.5).

Fig. 5: Lonchophylla robusta, head Fig. 6: Lonchophylla spec., interfemoral membrane

Extensive uropatagium present, short tail, dorsally projecting beyond the wing membrane.
Calcar does not reach foot length, plagiopatagium inserts at the ankle joint (fig.6).

L. thomasi: Smallest species within the genus, forearm length 31-32 mm.

L. mordax: Forearm length 34-35 mm.

L. robusta: Forearm length 41-46 mm.

L. handleyi: Largest known species within genus, forearm length 44-48 mm.

Platalina
P. genovensium: Relatively large flower bat, forearm length 48-49 mm. Externally visible
elongated head, the smooth nose leaf lacking a median “suture”; two distinct narrow

grooves parallel with both basal edges. Forearm and distal part of upper arm naked.

x
N

Y Ay

Ss
=
N
S
ES
Y L, WI
Sy

i hy, N
st N UN

Fig.7: Platalina genovensium, interfemoral
membrane Fig.8: Brachyphylla cavernarum, head

37

Well developed uropatagium, the median extension reaching the lower third of tibia.
Calcar as long as foot including claws. Plagiopatagium inserts at tarsus. Tail (ca. 8 mm)
present, projects dorsally in first quarter of the uropatagium as a touchable stump (fig.7).

Brachyphylla

Comparativlely very large bats, head not visibly elongated. Nose leaf without spear but
showing two concentrically lining, deep circular wrinkles around the nostrils. Thus, the
face (fig.8) resembles that of the Desmodontinae (true vampires). Strikingly strong thumbs
and large, strong feet. The uropatagium forms a well developed interfemoral membrane.
No externally visible tail, no calcar, the plagiopatagium inserts at the tarsus. Pelage colour
varies, with head and back usually light brown, yellowish to ivory, ventral side most often
brown (lighter than dorsally). The (naked) plagiopatagium is darker, almost black.

B. nana: Forearm length 56-59 mm.

B. cavernarum: Forearm length 63-69 mm.

Erophylla

Large blossom bat with conical, moderately elongate muzzle. Nose leaf without spear,
similar to Brachyphylla, but more delicate and with small central tip. Resembles the nose
of Old World Rhinopomatidae ("pig snout”). Uropatagium restricted to a narrow fringe
being distinctly projected by the tail. Very short calcar.

E. sezekorni (fig.9): Forearm length 45-49 mm. Long thumb (1. phalanx I = 7mm).
Metacarpalia: II>IV<V (III >=V); 1. phalanges: I] > IV >=V. Uropatagium short,
marginal outline wedge-shaped, running from very short calcars to the tail tip.
Plagiopatagium inserts at distal tibia immediately above the ankle joint. Pelage coloration

pale yellow brown to reddish grey, slightly blotchy (”frosty”, “mangy”), ventral view and
face lighter. Short tragus with smooth margin.

Fig. 9: Erophylla sezekorni, head, tragus, and interfemoral membrane

38

E. bombifrons (fig.10): Forearm length 46-50 mm, thumb very long (1. phalanx I =
8.5mm); metacarpalia: II>IV>V; 1. phalanges: III>IV; tragus only terminally pointed and
with a blunt margin. Short uropatagium, marginal outline semicircular between very short
calcars (3.4mm). Tail projects half over this margin. Plagiopatagium inserts at distal tibia
immediately above ankle joint. Pelage dorsally reddish light brown to dark brown; slightly
lighter ventrally, beige grey; hairs dark only at distal third of the shaft, basal portion light
for two thirds. |

Fig. 10: Erophylla bombifrons, tragus and interfemoral membrane

Phyllonycteris

P. poeyi (fig.11): Rather large flower bat, similar to Erophylla;, forearm length 46-49 mm;
with reduced nose leaf lacking a spear; but contours more similar to Brachyphylla. Very
short uropatagium; tail projects far beyond the margin of uropatagium; no calcar.
Plagiopatagium inserts at distal tibia distinctly above the ankle joint. Pelage coloured with
a shimmering light grey/beige.

U7,
ile ONG

ie!

Fig.11: Phyllonycteris poeyi, head

39

Fig.12: Glossophaga soricina, head and interfemoral membrane

Glossophaga

Small to medium sized nectarfeeding bats, forearm length 32-42 mm; head with
moderately elongate muzzle, small nose leaf. Well developed uropatagium with
semicircular fringe; short calcar (<foot length). Short tail, embedded only in first half of
uropatagium. Plagiopatagium inserts at ankle joint.

G. soricina (fig.12): Forearm length 34-37 mm.

G. commissarisi. Forearm length 31-35 mm.

G. longirostris: With a forearm length of 36-42 mm largest species of the genus; with
rather elongate face (name!). Ears short and rounded; small tragus with smooth contours
(Als):

G. leachii: Forearm length 35-38 mm (largest Central American species)

G. mexicana. Forearm length 32-36.5 mm.

Fig.13: Glossophaga longirostris, tragus

Monophyllus

M. redmani: Medium sized glossophagine bat, elongate muzzle hardly recognizable.
Forearm length 34.8-43 mm. Long thumb. Forearm haired in proximal third. Pelage dark
reddish brown, “frosty grey” coloured ventrally. Ears short and comparatively pointed;
tragus short, with strongly undulated edge and also undulating, notched basal lobe (fig.14).

40

Fig. 14: Monophyllus redmani, tragus

Uropatagium narrow, inner edge running at an acute angle; tail relativley long, distinctly
projecting beyond the naked fringe of the tail membrane (fig.15a).

M. plethodon: Usually larger than M. redmani. Forearm length 38.8-45.7 mm. Uropatagial
fringe semicircular (fig.15b).

Fig.15: Interfemoral membranes of Monophyllus: a: M. redmani, b: M. plethodon

Leptonycteris

Comparatively large flower bats with moderately elongate muzzle appearing blunt in
living specimens compared to other glossophagines. Nose leaf with short triangular spear,
tragus with smooth contours. Forearm naked; pelage coloration usually light brown.
Uropatagium reduced to a 3-4 mm thin fringe running at an acute angle. No externally
visible tail (but nevertheless present with three vertebrae).

L. nivalis: Forearm length 55.4-59.5 mm. Metacarpale III shorter than phalanges III;
uropatagial fringe covered with conspicuous hairs, 3-4 mm long.

L. yerbabuenae (fig.16): Forearm length 50-54.8 mm. Metacarpale III as long as phalanges
III; uropatagial fringe almost naked.

L. curasoae: Same size as L. yerbabuenae, uropatagial fringe sparsely haired.

41

Fig.16: Leptonycteris yerbabuenae, head, tragus, and interfemoral membrane

Lichonycteris

L. obscura (fig.17): Very small nectar feeding bat, forearm length 32.4 mm; with short
pointed muzzle, very small nose leaf with smooth, triangular, short spear. Ears and tragus
short, tragus with undulating, serrated contours. Forearm haired in its proximal third; coat
colour deeply brown; hair on the head light at the basis, only hair tips dark. Extensive
uropatagium with flat semicircular outer edge between long calcars. Calcar longer than
foot. Distinctly developed tail (1/3 of uropatagial width); terminal with small stump
projecting dorsally beyond the wing membrane.

(eK As al Oy,

a
dl

Fig. 17: Lichonycteris obscura, head, tragus, and interfemoral membrane

Anoura

Small to mid-sized Glossophaginae with long thick fur. Long muzzle with small nose leaf,
auricle and tragus short. Uropatagium sparsely developed, externally no tail visible
(name!). Forearm proximally haired, coat colouring uniform with dark greyish brown to
blackish brown.

42

A. caudifer (fig.18): Smallest species within this genus. Forearm length 35-38 mm, haired
up to the half, coat very dark, almost black. Tragus carrying small protuberances at the
outer margin. Uropatagium narrow with acute-angled fringe (semicircular in literature!),
densely covered with hair on the dorsal side.

Fig. 18: Anoura caudifer, tragus and interfemoral membrane

A. cultrata: No data of my own referring to gross morphology.
A. geoffroyi (fig.19): Middle-sized, forearm length 40-44 mm. Contours of tragus smoothly

rounded. Uropatagial fringe very narrow, with pointed parabolic contours (triangular
Husson 1962). Forearm only hairy in its proximal third. Coat colour dark grey.

Fig. 19: Anoura geoffroyi, head, tragus, and interfemoral membrane

Hylonycteris
H. underwoodi: Very small glossophagine bat, forearm length 31-36 mm; externally
resembling Choeroniscus godmani;, no data of my own referring to outer morphology.

Scleronycteris

S. ega: No data of my own; forearm length about 35 mm.

Choeroniscus

Small to middle-sized flower bats with clearly stretched head; nose leaf very small, in
frontal view equilaterally triangular. Short auricle with small tragus. Uropatagium large.
The short tail remains embedded within the proximal third of the tail membrane. Calcar
present, about the length of the foot. Pelage usually dark brown to blackish brown, colour
mostly uniform, gradually turning into lighter shades ventrally and to darker ones dorsally.
Hairs at the tip darker than at the basis, thus the “underwool” appears yellowish. Wing
membranes dark brown.

C. godmani (fig.20): Very small bat, forearm length 32-34.8 mm. Head stretched, but less
conspicuously than in the larger species of the genus. Tragus smooth. Metacarpalia:
II>IV>V; 1. phalanges: III>IV >V. Plagiopatagium inserts at the tarsus. Coat dark brown
to blackish brown.

Fig. 20: Choeroniscus godmani, tragus and interfemoral membrane

C. minor (fig.21): Forearm length 35-36 mm. Tragus “granular” at the tip. Metacarpalia:
III>IV>V; 1. phalanges: III>IV>V. Plagiopatagium as in C. intermedius. Coat colouring
dark grey brown. Hairs darker at the tip than in the basal 3/4 of the shaft; thus the lower
hair layer appears yellowish brown.

PAYOR u.
em V¥ony
wt IE)

N 17, Ä
at f/f
x z.
= ZED
\ N} a) Ee
S un 2
s SS | \ ER
Se I
Maler S
en = i SS TI
ee Nas N
ay IN N
\

Fig.21: Choeroniscus minor, head and tragus

44

C. intermedius: Forearm length 34-36 mm. Metacarpalia: II>IV>V; 1. phalanges:
II>IV>V. Plagiopatagium inserts at the proximal metatarsus.

u
Fig.22: Choeroniscus intermedius, interfemoral
membrane

C. periosus: Largest species of the genus. Forearm length 40.4 mm. No data of my own
referring to gross morphology.

Choeronycteris

C. mexicana (fig.22): Middle-sized microchiropteran (as a nectar feeder quite large) with
distinctly elongated muzzle. Nose leaf and auricles small, short ears (rounded-triangular
and in dorsal view looking like dolphin flippers). Small tragus, with its basis merely
profiled.

| SS
ER. IN 3 \\
Ane , » yy ) ya) \)

IN

Fig. 23: Choeronycteris mexicana, head, tragus, and interfemoral membrane

Forearm length 42-47 mm. Metacarpalia and 1. phalanx III>IV>V. Plagiopatagium inserts
at the tarsus, uropatagium present, entirely including the short tail (only first third of the
tail membrane). Calcar strong and straight, about same size as foot.

Pelage brownish grey to brown, ventrally brightened. Hairs only dark at the distal tip, the
shaft is light. Wing membranes brown, but not as dark as in Choeroniscus.

45

Fig.24: Choeronycteris
harrisoni, head

C. harrisoni (fig.23): Quite large blossom feeding bat with conspicous, extremely long
muzzle. Forearm length 40-43 mm. In other respect resembling C. mexicana.

Skull morphology

Differing from other mammals the chiropteran skull bones are delicate in order to reduce
weight for flight capability. In adult animals the individual bones are mostly fused
together leaving the sutures essential for bone discrimination not to be recognized. So, for
proper orientation, it is certainly useful to refer to the juvenile skull yet lacking
obliterations.

Within the chiroptera, the highly variable skull of phyllostomids is characterized by the
following features:
Postorbital process lacking; the posterior rim of the orbita is, if at all, distiguishable only
by lateral convexities of the frontalia.
Premaxillaria are completely present. Both the nasal and the palatinal parts of the
premaxillaries are completely developed and in adult individuals they are tightly
connected to each other and to the maxillaria. Both palatine bones mark the boundary of
two Foramina palatinae laterally.
The perioticum is largely separated from the rest of the skull (well visible in basal view).
There are no more than 34 teeth:

23,172 455,67

1210 23,445,67,

Sometimes, however, the number is reduced, down to 26:
-231--3456-
---1-23456-

There are no more than two incisivi on each side of a jaw, the canini being always well
developed. In the upper jaw only the genus Anoura bears three premolars, usually only
two of them are present in one jaw side (P° and P'); in most cases, the mandible bears
only three premolars per half (P,, P;, P,). If the premolars are limited to two, it is always
P, which is missing. Considerable modifications predominantly affect the molar teeth
(crown morphology); when the number is reduced, always M, is missing.

46

10 mm

10 mm

Fig.25: Lionycteris spurelli, skull a: lateral view, b: basal view, c: dorsal view, d: frontal view

Skull descriptions of the single genera

Lionycteris (fig.25)

L. spurrelli: Skull: rostrum relatively short, approximately limited to one third of total
length; distally flat and tapering. In lateral view, the upper incisivi continue the outline
straightly.

In basal view the palate is pointedly trapezoid, width over the last molars considerably
exceeds the distance of the canini to each other. Choanae situated quite anterior with the
palate reaching only half the basal length. Praesphenoid outlined but without ridge,
basisphenoid with strong convexities (basisphenoid pits). Pterygoids with hamuli ending
shortly and pointedly, no pterygoid wings. No zygomatic arches.

Teeth: dental formula BG) BMS 3) 4S) (6 7
Vee S 8) ahs) (6) 9)

Incisivi present in the upper and lower jaw; the upper ones being orientated anteriorly,
inner incisivi considerably larger than outer ones, long and spatular, with smooth edge, the
outer ones only one third the size, pointedly trapezoid (“tomahawk”), producing an oblique
edge which complements with those of the inner incisivi to an edge formed like a
horseshoe.

Canini strong, with cingulum. Premolars caniniform, anterior premolar separated from
canine tooth by distinct diastema. Molar teeth quite strong with broad crowns and
dilambdodont ectoloph; anterior molar larger than the middle one, the posterior one being
the smallest. Lower incisivi of same size, forming a closed row between the canini.

Lonchophylla (figs. 26-28)

Skull: similar to Lionycteris; in smaller forms the rostrum is distinctly less than half of the
total skull length, in larger species almost half of it. Skull laterally stretched, only weak

47

lacrimal inflation. Zygomatic arches lacking in all species. Basisphenoid more or less
arched, skull base almost at the same level as palate (in German literature: orthocran).

Teeth: dental formula 2 Sl ns ab By (6) 7
12-1223 a5 67

both in the upper and in the lower jaw two pairs of incisivi present; upper incisors similar
to Lionycteris (I11+I2), the lower ones equal in size, with trifid edge, the width of a small
gap between I2 and lower canine tooth corresponds to the crown width of the 12.
Premolars with three cusps (not monocuspid as in Lionycteris). Anterior premolar
separated from caninus by a diastema of its own length. Molar teeth still with functional
masticatory surfaces.

L. thomasi: Posterior palate edge V-shaped; incisors and anterior edge of the premaxillaria
in top view flat and almost straight. Basisphenoid pits vaulted. Palate length less than half
the total length of skull. Last upper premolar with lingual cusp, giving the tooth a T-
shaped appearance from top view. Anterior lower premolar with posterior (hooked) cusp.

L. mordax (fig.26): Posterior palate edge U-shaped; incisors and front edge of the
premaxillaria pointed anteriorly in top view. Basisphenoid with deeply vaulted pits. Palate
approximately half the total skull length. Last upper premolar without lingual cusp, thus
appearing narrow from above. Anterior lower premolar without posterior cusp.

Fig.26: Lonchophylla mordax,
skull lateral view

10 mm

L. robusta (fig.27): Posterior margin of the palate U-shaped; anterior edge of the
premaxillaria similar to L. thomasi, but a little more pointed. Basisphenoid only flatly
vaulted. Palate approximately half the total skull length. Last upper premolar with a well
developed lingual cusp, thus appearing triangular in top view. Anterior lower premolar
with well developed hooked posterior cusp.

Fig.27: Lonchophylla robusta,
skull lateral view

48

L. handleyi (fig.28): Like L. robusta, but the palate length clearly exceeds half the total
skull length. Last upper premolar with insignificant lingual cusp; the posterior cusp of the
anterior lower premolars is present and hooked.

Fig.28: Lonchophylla handleyi,
Bi ul E10, DI AR skull lateral view

Platalina (fig.29)

P. genovensium: Skull similar to Lonchophylla, very long rostrum, almost half the total
skull length; transition premaxillaria to nasalia as well as of nasalia to frontalia very flat
in lateral view (generic name). Zygomata lacking; braincase flat, pterygoids do not reach
the Bullae tympanicae.

Fig.29: Platalina genovensium,
=——— skull lateral view

10 mm

Teeth: dental formula -2531-234567
Da hea Deas 67

Inner upper incisivi strong, broad spatula form (frontal view), flat directed anteriorly; in
lateral view, the incisors run out along the prolonged premaxillaria (the name Platalina
also refers to the arrangement of the upper incisivi, which continue the flat line of outer
contours of skull seen in lateral view). Lower incisivi present, with trifid edge. Canini
strong, premolars with long narrow base, all postcanine teeth separated by wide gaps.
Molar teeth narrow, without W-shaped ectolophs; triangular with simplified tricuspid
pattern. Crowns in top view shorter than those of the premolars.

Brachyphylla (fig.30)
Skull: rostrum relatively short, braincase > viscerocranium. Solid zygomatic arches; length
of palate < half the total skull length, posterior edge of the palate V-shaped with normally
developed hamuli. Base moderately vaulted; strong joint processes (Proc. glenoidales).
Teeth: dental formula 2131 B14 506.7

72272234567

Fig.30: Brachyphylla nana, a: skull dorsal view, b: skull lateral view, c: skull basal view, d: mandible
top view

Inner upper incisivi invigorates with pointed edge, outer incisors very small “squeezed into
the gap between the large incisivi and the canine tooth”; caninus strong but only slightly
longer than incisivi, anterior upper premolar very small, closely adjacent to canine tooth
and to posterior premolars. Posterior premolar very strong, pointed and almost the same
size as canine tooth. Upper molars broad with multiple cusps, lacking dilambdodont
ectolophs (dentition of a fruit eater). Lower incisivi small, the inner ones equal in size to
the outer ones, completely filling the gap between the canine teeth; lower premolars
approximately equal in size with a pointed cusp (caniniform); last lower molar tooth
narrower than the front molars; thus the outline of the tooth row becomes convex.

B. nana: Total skull length 27-29 mm.
B. cavernarum: Total skull length 30-32 mm.

Erophylla (figs.3 1-32)
Skull: rostrum moderately elongate; zygomatic arches present; palates with two distinct
crosswise ridges anterior to each of the pterygoids; posterior edge of palate sharpened in
a V-shaped way. Base initially flat at the choanae, but basisphenoid with well vaulted
"basisphenoid pits”. Mandibula in lateral view flat and straight, with flat coronoid process
and strong angular process.
Teeth: dental formula -231--34567

12-1--34567
Upper incisivi: inner ones larger than outer ones; inner ones divided by a distinct middle
gap, at the upper edge broader than at the root, crowns with comb-like edge (”ravioli
edge”). Canines broad and strong; second upper premolar double the size of the first one,
both with a broad crown. First upper molar almost double as long as broad, the outer edge

50

6 ER men Ä

Fig.31: Erophylla bombifrons, a: skull lateral view, b: mandible lateral view, c: skull basal view, d:
skull dorsal view, e: mandible top view

(buccal) flat and straight. The second upper molar is in basal view almost triangular,
showing an incision at the outer edge. Third upper molar small, triangular with smooth
outer edge. Lower incisivi small, about the same size and forming a continuous line
between the broad canines. Second lower premolar slightly larger than the first; lower
molars with flat crowns and sharp outer edge.

E. sezekorni: In lateral view, the skull shows a slightly concave transition from rostrum to
braincase (fig.32).

E. bombifrons: Abrupt transition from rostrum to neurocranium with a distinct indentation
at the suture separating the nasalia and frontalia (in lateral view). Behind, the cranium
rises abruptly; the braincase being vaulted beginning from the root of the nose (fig.32).

Fig.32: Skull contours of E.
bombifrons (continuous line)
and E. sezekorni (dotted line)
adapted from Buden (1976)

51

Phyllonycteris (fig.33)

P. poeyi: Skull more elongate and flat; rostrum moderately elongate; continuous transition
to the cranium in lateral view. Zygomatic arches lacking in the specimens examined.
Palatinal area slightly lowered against skull base (in German literature: klinorhynch).
Premaxillaria shifted anteriorly, Foramina palatina minora lying in front of the canine
alveoli.

Fig.33: Phyllonycteris poeyi, skull a: lateral view, b: dorsal view, c: basal view

Teeth: dental formula DRS 125304916 9
2-12 456 7

Upper incisivi separated from canine tooth by broad diastema; inner ones larger than outer
ones, orientated posteriorly, jointly forming an almost straight edge (teeth somewhat
horselike). Canini strong and broad; first upper premolar small, the second one at least
double the size with lingual cusp. Upper molars flat with broad crowns lacking a
dilambdodont ectoloph.

Glossophaga (fig.34)

Skull: braincase longer than rostrum (only in G. longirostris almost of the same length).
Zygomatic arches developed quite well and preserved in most specimens. Skull base
vaulted to different extent within the species. G. commissarisi is separated from G.
soricina by contours of the presphenoid. Mandibula with flat coronoid and distinct
mentum.

Teeth: dental formula = 3 leo 3 AS 6 7
eS eA S67:

Upper incisivi uniform, sometimes slightly projecting forward, forming an almost
complete curve between the canini (gaps between the outer incisivi and the canine teeth
narrower than width of the outer incisivus), the inner incisivus is broader than high with
straight edge; the outer one is shorter, but thus with its edge in one level with the inner
one. Lower incisivi well developed, completely filling the space between the canini, with
flat crowns (width = height) and rounded to roundly rectangular profile. Strong canine
teeth with weak cingulum, the upper ones with distinct anterior and posterior edge. Upper

52

10 mm

Fig.34: Glossophaga commussarisi, a: Skull lateral view, b: mandible top view, c: skull dorsal view,
d: skull basal view

premolars triangular in lateral view, with narrow cusps without styli. Upper molars with
flattened W-pattern, third molar smaller than second, about half the surface of M2. The
inside of all three molars with distinct convexities. The lower molars are of similar shape
showing all five usual cusps, only the last molar slightly smaller.

G. soricina: Point of lower jaw with ridge on symphysis (“mentum”); pterygoids with
lateral widenings ("pterygoid wings”, but not the hamuli); presphenoid with distinct ridge;
basisphenoid just slightly vaulted.

Upper incisivi projecting anteriorly, the inner ones larger than the outer ones, anterior
edge of the premaxillare elongate (= in view from above incisivi well visible); lower
incisivi contiguous and uniform in size.

G. commissarist. Upper incisivi do not project forward, the inner ones about the same size
as the outer ones, the anterior edge of the premaxillaria evenly rounded (incisivi in top
view hardly visible); lower incisivi very small and comb-shaped; presphenoid ridge
flattened, pterygoids lacking lateral widenings.

G. longirostris: Upper incisivi projecting anteriorly, the inner ones about the same size as
the outer ones; lower incisivi large, forming a complete curve between the lower canini;
symphyseal ridge (mentum) only weak, pterygoids bulging only to low extent.

G. leachii: Upper incisivi not projecting forward, the inner ones about the same size as the
outer ones; anterior edge of the praemaxillare evenly rounded; complete presphenoid ridge
present.

G. mexicana: Lower incisivi tiny, separated by distinct gaps, upper incisivi projecting
anteriorly, the inner ones distinctly larger than the outer ones, anterior edge of the
premaxillaria pointedly elongate. Pterygoid wings lacking, presphenoid ridge subterminally
flattened.

Monophyllus (figs.35-36)

Skull: rostrum not quite half the total length of skull, skull appears more stretched than in
Glossophaga, similar to Anoura.

Teeth: dental formula 213456],
| Be Ts Ay

Teeth essentially like Glossophaga, the incisivi, however, remaining much smaller. The
upper ones with distinct gaps between each other as well as to the canini and different in
form: the inner ones with flat edge, the outer ones pointed. The lower incisivi are very
small, with flat, rounded crowns, arranged in two pairs which are separated by a broad
median gap.

10 mm

Fig.35: Monophyllus redmani, a: skull lateral view, b: skull basal view, c: skull dorsal view, d:
incisivi, e: upper tooth row

M. redmani (fig.35): The upper premolars are separated from each other by a conspicous
gap, occupying more than half the crown length of the anterior upper premolar.

M. plethodon (fig.36): Upper premolars separated from each other just by a small diastema
(<1/2 the length of the anterior upper premolars).

mI Fig.36: Monophyllus plethodon, upper tooth row

Leptonycteris (fig.37)

L. nivalis: Skull long and flat, rostrum almost half of total length, braincase wider than
high. Skull base well vaulted (basisphenoid), presphenoid ridge ending bulging; Pterygoid
processes flat, slightly club-shaped (in contrast to Choeroniscus), Fossa mandibularis
“shadowed” by Processus glenoidalis - i.e. comparatively solid mandibular joint.
Mandibula long and narrow, Processus coronoideus only slightly higher than Proc.
articularis. Curvature of the Ramus mandibularis inserts yet caudal of the Proc.
coronoideus.

Fig.37: Leptonycteris nivalis, a: skull lateral view, b: skull dorsal view, c: basal view

Teeth: dental formula -23172--3256-
12=-1=-23456=

Incisivi comparatively strong, forming a line, distance I? to C' longer than I’ to 'I. Lower
incisivi of equal shape ”droplet-spatula shape”, the inner ones larger than the outer ones,
medial spearated by a diastema. Lower incisivi present, with low, flat, rounded crown.
Well developed canini, the upper ones lacking a cingulum but with two secondary crowns,
the one situated at the base of the main shaft being more conspicuous. Lower canini with
distinct cingulum. Premolar teeth long and narrow with tall edges and distinct, but small
styli. Seen from above, the anterior lower premolar (P,) is vaulted outwards. Molars still
with masticatory surface (W-Pattern) but already reduced, M, with very long narrow base.

Lichonycteris (fig.38)
L. obscura: Skull: braincase clearly longer than rostrum (comparatively short total length;
cf. allometric data). Zygomatic arches very delicate and thus in most cases destroyed
during preparation; skull base with presphenoid “ridge”; basisphenoid vaulted towards the
basioccipitale.
Lower jaw with distinct mentum, Proc. coronoideus only slightly higher than articular
process.
Teeth: dental formula -231--3456-

---1-23456-

Upper incisivi equally arranged between the canine teeth, showing gaps between the
individual incisivi. Shape similar to Choeroniscus (I' droplet-spatula-shaped, I’ dagger-

Fig.38: Lichonycteris obscura,
a: lateral view, b: frontal view

10 mm

55

shaped, pointed with flat outer base). Lower incisors missing. Canini simple, with a slight
cingulum. Premolars comparatively short and wide, compared to other Glossophaginae.
Upper molars without conceivable masticatory surface due to reduction of the commis-
surae.

Anoura (figs. 39-41)
A. caudifer (fig.39): Skull: rostrum not quite reaching half the total length of skull. Base
moderately vaulted, no angular deviation between palatinal and basal level (in German
literature: orthocran). Choanae at about the same level with Fossa glenoidalis (palate
comparatively long).
Teeth: dental formula a2 eS Sas O 7

---1-234567
Upper incisors very small, the outer ones (dagger-shaped) twice the size of the inner ones
(droplet-shaped), medially separated by wide gap (about four times the width of the
incisivi). Anterior premolar very small and caniniform, clearly visible distance to canine
tooth; also a diastema to the second premolar. Third upper premolar with three cusps, the
second one more or less forming a two-cusped transitional form. Molars flat, but all three
of them with functional masticatory surface (dilambdodont crown by top view). Row of
teeth in basal view almost rectangular; width over canine teeth only slightly narrower than
molar width of the palate. Lower incisivi missing; Proc. coronoideus very flat.

Fig.39: Anoura caudifer, a: skull lateral view, b: skull dorsal view, c: skull basal view, d: mandible
lateral view, e: mandibel top view

A. cultrata (fig.40): Similar to A. caudifer, but differences in dentition: upper canines
strong with a sharp ridge running along the anterior edge; anterior lower premolar
enlarged to a long, narrow blade.

A. geoffroyi (fig.41): Like A. caudifer, but considerably larger; last upper premolar with a
median lingual cusp, projecting beyond the narrow base of the tooth.

56

Fig.40: Anoura cultrata,
mandibel lateral view with P,

A. latidens: Similar to A. geoffroyi, of about similar size; last upper premolar with a
median lingual cusp, enclosed within the wide triangular base of the tooth.

Fig.41: Anoura geoffroyi, a: skull lateral view, b: skull dorsal view, c: skull basal view, d: mandible
top view

Scleronycteris (fig.42)

S. ega: Skull: due to incomplete specimen (Typus BMNH) no records on braincase;
rostrum comparatively long (like Choeroniscus and Hylonycteris), no zygomatic arches
preserved.

The pterygoid processes (hamuli) are short and do not reach the Bullae tympanicae.

Teeth: dental formula 23013 Arc]
Soe bo 28 4b 5 6 7

Upper premolars with distinct distances from each other and from the canine tooth; molars

in upper jaw with mesostyli, lower molar teeth only with just slightly narrowing crowns,
too.

Fig.42: Scleronycteris ega,
a: skull frontal view, b: lateral
pus 102mm 4 view (occiput missing)

a

Choeroniscus (figs. 43-45)

C. godmani (fig.43): Skull: rostrum long and narrow, but less than half the skull length.
Braincase raised against rostrum, palatinal area elevated against skull base (in German
literature: airorhynch).If the skull is placed on the palate plane, the highest point of the
skull is not reached by the frontalia - as in the remaining glossophagines - but instead
rather by the parietalia being separated by a distinct Fossa parietalis. Interorbital width
hard to determine, as no postorbital processes visible. Foramen infraorbitale situated
rostrad (within the anterior eighth of the total skull length, above the first upper premolar).

Fig.43: Choeroniscus godmani skull, a: lateral view, b: dorsal view, c: basal view

Skull base with conspicously elongated rectangular palate surface (width over canini
equals width over molars), roof of the palate almost reaching the Fossae mandibulares.
Pterygoids form long, shovel-like widened processes coming in contact with the
tympanohyoid bones at the bullae. Presphenoid smoothly adjoining the well vaulted
basisphenoid. Contiguous ridge between presphenoid and basioccipitale comparatively
broad, prolonged in two wings curving around the Foramen occipitale.

Teeth: dental formula NY le=3A5 64
21.234587

Upper incisors very small, the outer ones being double the size of the inner ones,
separated by a large medial diastema, very different in form: I' is flat, with stamp-shaped
crown and definitely tiny. I’ almost caniniform, but though of twice the size of I’ so small
that it does not match the height of the cingulum of the C’. Canines delicate and narrow
with cingulum, length less than height of the maxillare at this level of the rostrum,
seperated from the anterior premolar (P*; Miller 1907) by a large diastema. Premolars
tricuspid, the middle conus being the highest, slightly exceeding cingulum C'. Base in
basal view elongated and narrow. Molars almost lacking a masticatory surface. There is
only a talon surrounding the hypoconus with the distal tip (metaconus) being the highest.

C. minor (fig.44): Skull with comparatively long rostrum, only slightly shorter than
braincase; zygomatic arches very delicate, in most cases destroyed by preparation, but

58

visible in X-ray examination. Palate elongate, in basal view rectangular, similar to C.
godmani, but considerably larger.

Fig.44: Choeroniscus minor, a: skull basal view, b: skull dorsal view, c: mandible top view, d: skull
lateral view

Skull base characterized by conspicously elongate pterygoid processes (hamuli), together
with the tympanohyoideum reaching below the Bullae tympanicae. In connection with the
alisphenoid they cover the vomer as well as one third of the presphenoid. Basisphenoid
distinctly structured, with ridge development towards the basioccipitale, the latter parting
into two wings which flank the Foramen magnum. Orbitae without distinguishable orbital
processes - thus, in most cases, the interorbital width ist no distinguishable measure from
the postorbital width.

Palatal level elevated against skull base, in lateral view, the rostrum appears very straight.
The contours of the braincase well vaulted, but without distinct indentation.

Mandibula elongate and narrow, lower incisivi missing - a median V-shaped notch
between the canini, allowing the long tongue to pass without opening the mandible joint.
Mentum with well developed symphyseal ridge; Proc. coronoideus flattened, merely
projecting beyond the Proc. articularis. Angular process forms the proximal mandibular
edge.
Teeth: dental formula re ee

---1-234567

Teeth very delicate showing wide gaps between individual premolars and molars.
Premolars and molars tricuspid in lateral view lacking relevant masticatory surfaces.
Incisivi only present in the upper jaw, very small, the outer ones exceeding the inner ones
in size and grouped in two pairs by a wide medial gap. Canini thin and pointed with
cingulum, premolars separated from the canini by distinctly developed diastema. Elongate,
narrow base (without masticatory surface), protoconus about 3/4 the height of the
cingulum of the canini, metaconus higher than cingulum.

C. intermedius (fig.45): Like C. minor.

59

10 mm Fig.45: Choeroniscus inter-
medius, skull lateral view
C. periosus: Similar to C. minor, but considerably larger. Clearly discernable from
Choeronycteris by its lateral skull contours - lateral shape of the nose rather concave than
convex.

Hylonycteris (fig.46)

H. underwoodi: Skull: resembling Choeroniscus in most characteristics, but lacking the
conspicously elongated pterygoid processes. Pterygoids developed normally, as in almost
all genera of this subfamily.

Ramus mandibularis stronger compared to Choeroniscus, with a more distinctly developed
Proc. coronoideus (cf. data of coronoid height).

Teeth: dental formula = 2S hs Beals 6} 7
ssn Jl =22 45 67

The dental features of Hylonycteris correspond to those of Choeroniscus.

10 mm
> 4

Fig.46: Hylonycteris underwoodi skull, a: basal view, b: dorsal view, c: lateral view, d: frontal view

Choeronycteris (figs. 47-48)

C. mexicana (fig.47): Skull: rostrum longer than braincase, in lateral view convex in its
proximal third up to half the length, flat angle between the nasalia and frontalia (distinct

60

bend); orbital processes lacking; zygomatic arches highly reduced, usually indetectable at
prepared skull. Skull base with elongate bony palate , the posterior edge reaching the
alisphenoid canal. Palate trapezoid in basal view (caninal width of the palate smaller than
molar width of the palate).

Fig.47: Choeronycteris mexicana, a: skull dorsal view, b: mandible top view, c: skull lateral view

Vomer covered by palatinum, septum continues as a presphenoid ridge up to the end of
the presphenoid. Alltogether, the base is stronger vaulted than in Choeroniscus. Pterygoids
lead into elongate, shovel-like widened hamuli, diverging concavely and, together with the
tympanohyale, forming a bony contact with the bullae.
Teeth: dental formula = 2,3 l= 3/4 5.67

---1-234567

Similar to Choeroniscus, but postcanine tooth row with larger interdental distances.
C. harrisoni (fig.48): Skull: like C. mexicana, but its rostrum is even more elongate. Thus
both, the ridge of the nose and the mandibula, are still more convex and show a distinct
angle to the braincase (lacrimal inflation). In basal view, the palate appears elongate-
rectangular compared to C. mexicana.
Teeth: dental formula -231--34567

---1-234567

Similar to C. mexicana, but the distances between the single teeth are even more
distinctly; the last of the molars is situated far anterior of the mandibular joint.

10 mm

Fig.48: Choeronycteris harrisoni skull, a: lateral view, b: frontal view

Table 4a: Univariate analysis of skull measurements

SGL = total skull length, GL = length of palate, CH = coronoid process

N = sample size, MW = arithmetic mean, SD = standard deviation

Species

YER) DY eh el la] teal eel teh 3) tes) er ee

hastatus
elongatus
perspicillata
castanea
subrufa
nana
cavernarum
sezekorni
bombifrons
poeyt
spurrelli
thomasi
mordax
robusta
handleyi
genovensium
soricina
commissarisi
longirostris

M. redmani
M. plethodon

© ale ae are hehe ee

caudifer
geoffroyi
brevirostrum
curasoae
nivalis
yerbabuenae
obscura
underwoodi
godmani
minor
indermedius
inca
mexicana
harrisoni
ega

No) Gy Zi) ES Ay es SH TS On WS

is US iy 8 0 Se

SGL
MW

38.19
29.98
23315
19,53
21.68
28.38
30.88
24.58
24.38
23:05
1979
20.92
23.62
26.09
28.96
32.33
20.90
20.24
23.14
23.05
23.25
22.38
253.33
23.66
27.91
26.99
2712
18.50
21.85
1953
22.78
22R02
23.43
29.67
34.18

SD

1.66
0.74
0.76
0.08
0.23
0.64
0.43
0.64
0.45
0.65
0.59
0.43
0.49
0.44
0.70
0.50
0.68
0.39
0.90
0.31
0.66
0.42
0.51
0.50
0.30
0.39
0.57
0.50
0.66
0.93
0.74
0.34
1.35
0.61
0.37

ER Ve) a) CS IN SCH

—
Dam WY Ee

GL
MW SD

WSS) WLS
13.432.030
10.71 0.40
8.75 0.44
9:992.0:077
LEST O35
14.09 0.26
LOTS. 02377
122037
10.85 0.49
9.63 0.40
12.12 0.69
12,832.072
14.49 0.28
177347052
1944155020
11.46 0.51
10.76 0.43
12:672.0.63
11.97 0.42
11.68 0.77
12.43 0.58
14.11 0.64
11.88 0.24
16.09 0.39
14.84 0.39
15.20 0.48
10.09 0.46
13.68 0.68
2722077
14.99 1.13
14.95 0.34
16.16

28.20 0.42
PLETE Val

Ty Co Cn Cy 0) RES Gy I) S Ci oy 72

aD

AnMnaPwWrAnY ODF + + I 0

m W OD

CH
MW

10.43
TROY
9.37]
3.96
4.75
6.99
8.39
4.39
4.47
4.71
3.66
3.42
4.18
4.39
5.00

22.68
3.61
Zn!
4.25
3.68
4.00
3.38
4.43
4.00
4.83
4.45
4.44
3.48
3.50
2.45
3.05

3.07
4.07
4.23
333

61

SD

0.68
051
0.19
0.03
Gl
0.27
0.30
0.30
0.25
0.27
0.17
0.13
0.17
0.14
0.44
0.18
0.28
0.30
0.30
0.29
0.31
0.21
0837
0.13
0.38
0.24
0.19
0.28
0.21
0.07
0.20

0.16
0.29
0.01

62

Table 4b: Univariate analysis of skull measurements

OZR = upper tooth row, UKL = mandible length, CC = width over canini

N = sample size, MW arithmetic mean, SD = standard deviation

OZR UKL CE

Species N MW SD N MW SD N MW SD
P. hastatus I MEOW Oey! ZA SO es 2 6 9:37 20:56
P. elongatus > NOS O35 5219415023 5 723039
C. perspicillata 2027778149080 20 14.23 0.69 AQ As). (OAS)
C. castanea 2 ese 20811 2220027 274200
C. subrufa OF 1685) 02.0 Oy ales ONY 9 443 One
B. nana qT SER O20 GE NSM O40 76102024
B. cavernarum OWI O38 4 20.69 0.39 ANG SOs OZ
E. sezekorni 15,27 7289") O26 la Ilyesy O.57/ 1542902028
E. bombifrons US WSS) OAD Ley Silko O83 IB SIO. O30
P. poeyi SAO O23 fa NOT O40 $3.26 0836
L. spurrelli HEY Cra OWS 1252031 1073.03 BOOMS
L. thomasi 142 6:6 O28 12 NS 8 O53 1473255 RONG
L. mordax Gy 27.995021 3 15.827024 6 336.039
L. robusta 1100295427033 10 17.60 0.49 JUL Bass, Os!
L. handleyi US AOS, 0333} 720227960 19 4.20 0.10
P. genovensium 2 NOS) O27 2 22.68 0.18 2042605022,
G. soricina 3612 7741,80 .0.26 36,.13.612047] 36) 32022.0907
G. commissarisi 8776.89 Ons 713205047 TS SOMO
G. longirostris gs OT OTT AMS alk OLS Mp, ALOE OIG
M. redmani 4 8.38 0.25 4 14.85 0.24 By 3.962008
M. plethodon a) Sela 024 5 14.66 0.46 5 4.06 0.44
A. caudifer Ne NS. OLB We S50) - O87 AAO OS
A. geoffroyi 1679537022 17 17.89 0.54 17 457.2024
A. brevirostrum = oles OA! 92162172058 OFA GG Only
L. curasoae 8 945 0.18 8 18.69 0.22 84820021
L. nivalis 8 8.84 0.18 811831.20837 6 4.44 0.15
L. yerbabuenae 4) 8383" 025 4218287022 4 444 0.43
L. obscura Saad OMS 6 12.48 0.46 73287206
H. underwoodi 77.602049 7 14.93 0.70 PG Sets) ON
C. godmani Si LS). O39) PME Oo” Sl. (028
C. minor LOR 7-90" O62 9 16.42 0.68 II Sesto. OLE
C. indermedius Seon OLOS 3 ISO) Oeil B83) ONS
C. inca 2228538 0.03 10217268 2736272082
C. mexicana 1S ae Zo R035 15.21.31 048 13742482048
C. harrisoni 37712807028 325968023 3 As 0920
S. ega In re 1574 1 3.84
Morphometry

Univariate analysis

For the taxa examined the results of univariate analysis of some selected skull measures
are given in tab.4a and 4b. The results show that the range of the data may overlap among

individual species of a genus, even if the means differ from each other significantly.

63

Statistically, only those random samples comprising numerous individuals can be
processed, whereas rare species with little material available may only be referred to by
individual records.

Sex-specific differences

Within the genus Phyllostomus (examined as outgroup), in the extremely large species P.
hastatus, most of the male skull measures examined rank over the female ones (see
appendix).

In all species of the genus Choeroniscus the cranial length values of the females proved
statistically significantly larger than those of the corresponding males (p < 0.05, t-test).

In Lonchophylla robusta the only available male was distinctly larger than the females
examined - in contrast to the smaller species of the genus. In L. handleyi which is even
larger, in 12 of 17 parameters the male values on average exceed the female ones, too,
although in the given random samples this could not be secured statistically. In all the
remaining taxa the sexes did not show any significant differences in their skull measures.

Skull proportions

The extent of adaptation to nectarivory may be expressed by the relation of rostrum length
to total skull length, as the following measure relations give evidence:

- Total length of skull to length of palate.

- Total length of skull to upper tooth row.

- Height of braincase cranium to length of palate.

- Total skull length to coronoid height.

- Height of braincase cranium to coronoid height.

- Mandible length to coronoid height.

Total length of skull to length of palate: the quotient of the means for the total skull
length (MWx) and the length of the bony palate (MWy) gives a good idea of relative
muzzle length in the species examined (the larger the quotient, the shorter the palate)
(tab.5, fig.49).

Table 5: Proportion of mean total skull length (MWx) / mean palate length (MWy)

Phyllonycteris Anoura 0.07
Erophylla Leptonycteris 0.05
Carollia Lonchophylla 0.08
Lionycteris Choeronycteris 0.05
Monophyllus Hylonycteris 0.02
Glossophaga Choeroniscus 0.08

Lichonycteris Musonycteris

64

length in mm
35

30

25

I
il:
UM

Lichonycteris
Choeroniscus

Lionycteris

il

=
=
a
OQ,
ie)
hes
iQ

Glossophaga
Hylonycteris
Monophyllus
Phyllonycteris
Lonchophylla
Leptonycteris
Choeronycteris

Carollia
Anoura

Fig.49: Skull proportions: total skull length (SGL) to palate (GL) of nectarivorous bat genera
(arranged to absolute body size)

Total length of skull to upper tooth row: the length of the upper tooth row between
caninus and last molar (MWy) is suitable as an additional measure for characterizing the
muzzle length, especially useful in those species which show a V-shaped palate (i.e.
Erophylla, Brachyphylla): tab.6.

Table 6: Proportion of mean total skull length (MWx) / mean length of the upper tooth row (MWy)

Lionycteris
Lichonycteris
Erophylla
Brachyphylla
Carollia
Leptonycteris
Glossophaga
Lonchophylla

xo) ay

Buy | SD |

Platalina
Hylonycteris

Choeroniscus

Monophyllus

Anoura

Choeronycteris

Musonycteris

65

Height of braincase to palate length: in order to compare measures being as independent
as possible of each other, the height of the braincase (MWx) is related to the length of the
palate (MWy): tab.7.

Table 7: Proportion of mean height of braincase (MWx) / mean length of palate (MWy)

x/y | SD |

x/y

Brachyphylla Anoura
Carollia Leptonycteris
Phyllonycteris Lonchophylla
Erophylla Choeroniscus
Phyllostomus Hylonycteris
Lionycteris Choeronycteris
Lichonycteris Platalina
Monophyllus Musonycteris
Glossophaga

Total skull length to coronoid height: with relative size of the coronoid process, the force
of the M. temporalis to move and hold the mandible against the upper jaw increases.
Thus, coronoid height is a morphological indicator for masticatory pressure achieved by
the contracting M. temporalis. The larger the Proc. coronoideus, the stronger the bite.
With respect to a functional shift to nectar feeding, also the proportions of coronoid height
(MWx) and total skull length (MWy) may be interesting: tab. 8.

Table 8: Proportion of mean coronoid height (MWx) / mean skull length (MWy)

Key PY, | SD |

Phyllostomus Leptonycteris
Brachyphylla Anoura
Phyllonycteris Monophyllus
Erophylla Platalina
Lionycteris Hylonycteris
Lichonycteris Choeronycteris
Glossophaga Choeroniscus
Lonchophylla Musonycteris

Height of braincase to coronoid height: to obtain a measure for the functional reduction of
the jaw apparatus, apart from the relationship to total length it may also be useful to
compare a quantity independent of rostrum prolongation, e.g. braincase height (Mwx) with
coronoid height (MWy), as with increasing jaw length the coronoid process does not
necessarily grow at the same ratio. Only if a Crista sagittalis is present, the height of the
braincase gives evidence of a functional relation to the Ramus mandibularis. Otherwise,
this measure is affected essentially by the current morphology (and volume) of the brain
(tab. 9).

66

Table 9: Proportion of mean height of braincase (MWx) / mean coronoid length (MWy)

Brachyphylla Glossophaga
Phyllostomus Phyllonycteris
Platalina Musonycteris
Lonchophylla Erophylla
Carollia Monophyllus
Leptonycteris Lichonycteris
Hylonycteris Choeronycteris
Lionycteris Choeroniscus
Anoura

Mandible length to coronoid height: as a reliable measure for rostrum length the length of
the mandible may be quoted. Relating mandibular length (MWx) to coronoid height
(Mwy), the quotients appear in the following distribution: tab. 10.

Table 10: Proportion of mean mandibula length (MWx) / mean coronoid heigth (MWy)

SD

| Ru: xe SY | SD

Carollia Platalina
Phyllonycteris Hylonycteris
Lionycteris Anoura
Erophylla Scleronycteris
Glossophaga Choeronycteris
Lonchophylla Choeroniscus
Leptonycteris Musonycteris
Monophyllus

These quotients clearly reflect different levels of specialization within the group but do
not necessarily allow statements on the construction principles they are based on.
Although size differences within nectar feeding bats are moderate, size-dependent shifts in
proportion cannot be definitely ruled out. After all, forearm length between Lichonycteris
obscura (30.3 mm) and Leptonycteris nivalis (58.2 mm) differs by approximately 100
percent.

So, the allometric relations between the measures mentioned within the subfamily and the
neighbouring phyllostomid genera Carollia (Carolliinae) and Phyllostomus (Phyllosto-
minae) will be referred to.

Allometrics
Allometric comparison of individual genera

In this study, a large portion of the species to be examined was represented by very few
individuals. Some of them show an intraspecific variability proving larger than
interspecific distances. Furthermore, measured differences sometimes are within the error
range pre-determined by the methods of measuring. Thus, a comprehensive comparison of
intraspecific allometries for the entire group may be impossible, and any examination will

67

have to be restricted to intra- and intergeneric allometries. These, however, are quite
suitable to give evidence of structural morphologic peculiarities and systematic relations
in different genera.

Regression (represented by the reduced major axis) was calculated with reference to the
measure SIZE for the following skull measures:

- Width over canini (CC)

- Upper tooth row (OZR)

- Length of palate (GL)

- Mandible length (UKL)

- Coronoid height (CH)

The formulae are given in detail in Rempe (1962).

The following tabular summaries show arithmetic mean values of both parameters as
logarithms, the gradient (tan alpha) of the reduced major axis in a double logarithmic
system of coordinates, the correlation coefficient (r) and the distance of the log. mean
value to the reduced major axis of the outgroup (“Lot”) for each genus examined (tab.11-
15). In the graphs the reduced major axis for the outgroup (Carollia and Phyllostomus
calculated as a unified common sample) is drawn as a broken line. The coordinate
resulting from the mean values for both parameters is marked as a point. In case the
correlation can be secured by the size of the random sample, the reduced major axis of the
respective nectarivorous genus is represented as a straight line scoring the coordinates of
the mean value (fig.50-54).

The generic names are placed aside to the mean value coordinates (An = Anoura, Bp =
Brachyphylla, Ca = Carollia, Cs = Choeroniscus, Ct = Choeronycteris, Ep = Erophylla,
Gp = Glossophaga, Hn = Hylonycteris, Le = Lichonycteris, Le = Leptonycteris, Li =
Lionycteris, Ps = Phyllostomus, Pn = Phyllonycteris, Lp = Lonchophylla, Mp =
Monophyllus, Ms = Musonycteris (C. harrisoni), Pn = Phyllonycteris, Ps = Phyllostomus,
Pt = Platalina).

A synopsis of the calculated distances to the outgroup is given in tab.16. Diagrams with
individual values and - if available - distribution ellipses for each genus do exist and can
be ordered from the author.

68

Table 11: Allometric proportions: width over canini to SIZE

MW SIZE MW CC tan o r Lot
Carollia 3.0448 0.6640 0.5384 0.7832 0.0043
Phyllostomus 3.4483 0.8966 0.4093 0.9364 0.0016
Brachyphylla 33195 0.7942 0.0306
Erophylla 3.1089 0.7045 0.0003
Phyllonycteris 3.1374 0.7200 0.0010
Lionycteris 2.8499 0.4884 0.8913 0.6971 0.0688
Lonchophylla 3.0317 0.5811 0.2907 0.8792 0.0797
Platalina 32127] 0.6236 0.1359
Glossophaga 2.321 0.5785 0.0193
Monophyllus 2.9833 0.6050 0.0282
Lichonycteris 2.8387 0.5154 0.0354
Leptonycteris 3.1805 0.6629 1.4206 0.7664 0.0827
Anoura 3.0543 0.6396 0.6239 0.7455 0.0341
Choeroniscus 2.9186 0.5364 0.5709 0.8843 0.0600
Hylonycteris 2.9045 0.5347 0.7449 0.7511 0.0536
Choeronycteris 3.1753 0.6253 0.1173
Musonycteris 3.2070 0.7607
Outgroup 3.1540 0.7305 0.5698 0.6900 0.0000
6
De
LOG er
CC ew
Bie as @Bp
pee Pn
07 i Ep
SS ele
2 ; An
oe ct @ Pt
Dh Mpe

3.0 3.3 LOG SIZE

Fig.50: Allometric width over canini to SIZE

69

Table 12: Allometric proportions: upper tooth row to SIZE

MW SIZE MW OZR tan & r Lot
Carollia 3.0448 0.8689 0.6390 0.7406 0.002 1
Phyllostomus 3.4483 1.0673 02822 0.9507 -0.0066
Brachyphylla 351.95 0.9923 0.5977 0.7862 0.0078
Erophylla 3.1089 0.8971 0.0040
Phyllonycteris 3.1474 0.8846 0.0346
Lionycteris 2.8499 0.7997 0.6090 0.5469 -0.0203
Lonchophylla 3.0317 0.9321 0.7183 0.9708 -0.0673
Platalina 371371 1.0401 -0.0897
Glossophaga 29199 0.8711 0.6522 0.6390 -0.0588
Monophyllus 2.9833 0.9188 -0.0767
Lichonycteris 2.8420 0.7682 0.0075
Leptonycteris 3.1805 0.9579 0.7942 0.6015 -0.0231
Anoura 3.0543 0.9434 0.6900 0.6673 -0.0679
Choeroniscus 2.9186 0.8842 0.5966 0.7884 -0.0725
Hylonycteris 2.9041 0.8798 -0.0749
Choeronycteris 3417783 1.0630 1.3399 0.8281 -0.1307
Musonycteris 3.2070 1.4076 -0.4603
Outgroup 341533 0.9221 0.4701 0.6149 0.0000

LOG
OZR

1,0

3,0 3,3 LOG SIZE

Fig.51: Allometric proportions upper tooth row to SIZE

70

Table 13: Allometric proportions: length of palate to SIZE

MW SIZE
Carollia 3.0448
Phyllostomus 3.4483
Brachyphylla 331195
Erophylla 3.1089
Phyllonycteris 3.1474
Lionycteris 2.8508
Lonchophylla 3.0317
Platalina 32437,
Glossophaga 2.9211
Monophyllus 2.9833
Lichonycteris 2.8414
Leptonycteris 3.1805
Anoura 3.0543
Choeroniscus 2.9186
Hylonycteris 2.9045
Choeronycteris 3.1153
Musonycteris 3.2070
Outgroup 3.1565

LOG
GL

1,2

MW GL

1.0152
1.1725
1.097711
1.0385
1.0351
0.9834
1.1499
1.2813
1.0724
1.0726
1.0033
1.1882
1133
1.1415
1.1358
2,
1.3554

1.0607

Fig.52: Allometric length of palate to SIZE

tan &

0.5367

08315

0.4277

0.5489

0.7368
0.4512

0.9549
0.6515
0.6781

1.6174

0.4039

3,3

Lot

0.0004
0.0061
0.0294
0.0030
0.0179
-0.0462
-0.1396
-0.1975
-0.1068
-0.0819
-0.0699
-0.1178
-0.0939
-0.1769
-0.1769
-0.2094
-0.2743

0.0000

LOG SIZE

all

Table 14: Allometric proportions: mandible length to SIZE

MW SIZE MW UKL tan & r Lot
Carollia 3.0448 1392 0.9468 0.9939 0.0043
Phyllostomus 3.4483 1.3360 0.3388 0.9833 -0.0062
Brachyphylla 353195 12755 0.5744 0.9456 -0.0052
Erophylla 32.0197 1.1958 -0.0638
Phyllonycteris 3.1474 1.1922 0.2800 0.7937 -0.0013
Lionycteris 2.8499 1.0972 0.6295 0.7994 -0.0436
Lonchophylla 3.053177 239 0.5743 0.9543 -0.0835
Platalina 3037] 1.3533 -0.1340
Glossophaga 2.9199 PANS) 0.6753 0.7311 -0.0856
Monophyllus 2.9833 1917,02 -0.0550
Lichonycteris 2.8420 1.0961 -0.0462
Leptonycteris 3.1805 1.2660 0.3848 0.5247 -0.0598
Anoura 3.0543 1.2244 0.4959 0.8573 -0.0765
Choeroniscus 2.9186 1.1863 0.6004 0.8904 -0.1010
Hylonycteris 2.9045 1.1691 -0.0903
Choeronycteris 31733 15297 1,3399 0.6252 -0.1259
Musonycteris 3.2070 1.4076 -0.1892
Outgroup 3.1533 1.1937 0.4615 0.9709 0.0000

LOG
UK

30 3.3 LOG SIZE

Fig.53: Allometric proportions: mandible length to SIZE

12

Table 15: Allometric proportions: coronoid height to SIZE

Carollia
Phyllostomus
Brachyphylla
Erophylla
Phyllonycteris
Lionycteris
Lonchophylla
Platalina
Glossophaga
Monophyllus
Lichonycteris
Leptonycteris
Anoura
Choeroniscus
Hylonycterts
Choeronycteris
Musonycteris

Out group

LOG
CH

08

MW SIZE

3.0448
3.4483
33195
3.1089
3.1374
2.8499
3.0317
321377
2.9211
2.9833
2.8387
3.1805
3.0543
2.9186
2.9045
3.1753
3.2070

3.1803

30

MW CH

0.6952
0.9285
0.8720
0.6486
0.6644
0.5633
0.6494
0.7539
0.5792
0.5819
0.5407
0.6536
0.6076
0.4568
0.5398
0.6129
0.6267

0.7807

0.5574

Fig.54: Allometric proportions: coronoid height to SIZE

0.6799

3,3

Lot

0.0097
0.0013
-0.0147
0.0920
0.0921
0.0330
0.0482
0.0451
0.0567
0.0837
0.0493
0.1269
0.1026
0.1777
0.0869
0.1647
0.1686

0.0000

LOG SIZE

Table 16: Distances to outgroup

Carollia
Phyllostomus
Brachyphylla
Erophylla
Phyllonycteris
Lionycteris
Lonchophylla
Platalina
Glossophaga
Monophyllus
Lichonycteris
Leptonycteris
Anoura
Choeroniscus
Hylonycteris
Choeronycteris
Musonycteris

Sex dimorphism

GL

0.0004
0.0061

0.0294

0.0030

0.0179
-0.0462
-0.1396
-0.1975
-0.1068
-0.0819
-0.0699
-0.1178
-0.0939
-0.1769
-0.1769
-0.2094
-0.2743

OZR

0.0021
-0.0066
0.0078
0.0040
0.0346
-0.0203
-0.0673
-0.0897
-0.0588
-0.0767
0.0075
-0.0231
-0.0679
-0.0725
-0.0749
-0.1307
-0.4603

UKL

0.0043
-0.0062
-0.0052
-0.0638
-0.0013
-0.0436
-0.0835
-0.1340
-0.0856
-0.0550
-0.0462
-0.0598
-0.0765
-0.1010
-0.0903
-0.1259
-0.1892

CH

0.0097
0.0013
-0.0147
0.0920
0.0921
0.0330
0.0482
0.0451
0.0567
0.0887
0.0493
0.1269
0.1026
O17 77
0.0869
0.1647
0.1686

CE

0.0043
0.0016
0.0306
0.0003
0.0010
0.0688
0.0797
0.1359
0.0193
0.0282
0.0354
0.0827
0.0341
0.0600
0.0536
0.1173
0.7607

Furthermore, it is of interest whether and how sexes differ in these criteria. In this study,
this aspect could only be realized to some extent in very common species, as there were
only few sufficiently comprehensive samples comprising both sexes. Thus, this aspect is
presented only as an example comparing the genera Anoura, Glossophaga and
Lonchophylla: tab.17.

Table 17: Allometric proportions: upper tooth row to SIZE (¢-2

Genus

Anoura
Glossophaga

Lonchophylla

SEX

1) Q +O Q,

vor OY,

tan &

0.61
0.76
0.67
0.60
0.726
0.756

test on

difference

N.S

74

DISCUSSION

Morphological adaptations to nectarivory

The Glossophaginae represent small to medium-sized phyllostomid bats with a reduced
dentition, distinctly prolonged rostrum and an extremely protrusible tongue - all this
indicates adaptations to a diet on nectar and fruit. In ecological terms, these bats may be
regarded as nocturnal equivalents of hummingbirds; their evolution having been influenced
by several parallel selective forces. There are numerous convergences, such as limits of
body weight, hovering ability, elongate tongue, prolongation of the rostral components of
the skull.

Gross morphology of the head

Since body shape as a whole hardly varies within the Chiroptera - presumably due to strict
requirements on undiminished flight ability (conspicous differences merely refer to body
size, wing area profile, development of the uropatagium and tail length) - their head shape
and thus their skull morphology did develop a remarkable morphological variety within
the mammals.

As already explained in chapter 1, this considerable variety is, above all, an expression of
successful utilization of various food sources with a variety of forms presumably unique
on the level of a mammalian order: starting from insectivory, the superfamily Phyllosto-
matoidea developed any conceivable specialized feeders compatible with flight behaviour:

carnivory (Phyllostominae), sanguivory (Desmodontinae), piscivory (Noctilionidae), frugi-
vory (Stenoderminae, Carolliinae, Phyllostominae) and nectarivory (Carolliinae, Phyllo-
stominae, Brachyphyllinae, Phyllonycterinae, Glossophaginae and Lonchophyllinae).
Merely browsers and grazers do not exist as an equivalent specialization would have
certainly required a digestive system incompatible with flight.

Accordingly, above all, nectarivory is marked by morphological adaptations of the head:
- Prolongation of the muzzle with size diminution of the nose leaf.

- Prolongation and specialization of the tongue (bristle-like papillae, lateral or median
nectar groove, musculature of the tongue base) for optimal nectar intake.

- Lower lip with median notch guiding the tongue (lower incisivi very small or missing,
the resulting gap allows nectar intake with the tongue with closed jaws).

- Pinnae and tragus are also reduced in size, the ears are rounded, equally broad as long
and, compared to other Phyllostomids using echolocation, quite short.

Skull morphology

In osteological terms, this morphological variety of head shape is expressed in a bounty of
different skull shapes within the family of the leaf nosed bats (Phyllostomatidae). Espe-
cially the visceral skull is an important substrate for evolutionary changes in order to
optimize utilization of various food sources. In this respect, the highly specialized flower
bats take an extreme position. Besides the prolongation of the viscerocranium and
variation of the dentition (referring to form, number and arrangement of the individual
tooth types), the adaptations comprise the masticatory and pharynx musculature as well as
corresponding bony components of the jaws, their attachment at the braincase and at the
mandible.

Glossophaga Monophyllus Lichonycteris Scleronycteris Hylonycteris Musonycteris
commissarisi redmani obscura ega underwoodi harrisoni

Lionycteris Lonchophylla Platalina
spurelli handleyi genovensium

Fig.55: Incisivi within Glossophaginae (a) and Lonchophyllinae (b)

Dentition

Incisivi (fig.55)

The incisors are rudimentary or in some genera the lower ones are lacking completely
(“tongue guiding channel”). Comparatively primitive genera with a short rostrum
(Glossophaga, Lionycteris) still have two pairs of incisors in the upper and lower jaw.

In the Lonchophyllinae, the inner upper incisivi are clearly larger than the outer ones. This
corresponds to the configuration which apperars to be primitive for the entire group. Both,
the Phyllostominae and the Carolliinae as outgroups, show this pattern of incisivi in the
upper jaw.

Within the Glossophaginae, however, this condition remained merely in two Glossophaga
species, as all other species and genera developed quite considerable modifications:
starting from a very uniform incisor pattern (e.g. Glossophaga commissarisi) the tooth
pairs move apart. In Leptonycteris there is just a little median gap between the big but flat
inner incisors; Lichonycteris developed several gaps distributed evenly between the
differently formed upper incisors (the inner ones are flat, the outer ones pointed). This
situation with the inner upper incisivi becoming distinctly smaller (Anoura) and addi-
tionally flattened (Choeroniscus, Hylonycteris, Scleronycteris, Choeronycteris) is found in
other genera, too. In this evolutionary trend the space for the tongue between the canine
teeth also in the upper jaw increases.

Lonchophyllinae, Brachyphyllinae and Phyllonycterinae have two pairs of incisors equal
in size in their lower jaw. Though this plesiomorph condition is preserved in Glossophaga,
in Monophyllus, a closely related genus, the lower incisors have become very small and
moved apart, both pairs separated by a wide median gap. In Leptonycteris there is also a
gap between the inner lower incisors, but the inner ones are clearly larger than the outer
ones. In Lichonycteris, Anoura, Scleronycteris, Hylonycteris, Choeroniscus and Choero-
nycteris, the lower incisivi are completely absent, and the lower canine teeth are separated
from each other by a deep V-shaped groove above the mandibular symphysis (tongue
guiding channel).

Canini

Canines are distinctly developed in all genera, though they remain slim and delicate in the
highly specialized nectar feeders. Thus, although the canine teeth seem quite useless for
nectar intake, they must have some biological value. Their function, however, is not
necessarily corresponding to foraging rather for other biological purposes like grooming

76

or possibly just stabilizing the closed jaws against distortion when the canines interlock.
Freeman (1995) states a considerable tooth-on-tooth wear (thegosis) on the anterior surface
of upper canines in nectarivorous microchiropterans. This is interpreted as indicator for
tight embracement of lower canines by the upper ones to support the jaw during the rapid
movement of the tongue during feeding.

Premolars

Within the Lonchophyllinae, Lionycteris still bears very strong premolars resembling
canine teeth. In Brachyphylla, the posterior upper premolars also are of similar relative
size to the canine tooth, the anterior one, however, remaining very small as specific for
this genus. In the course of further specialization the premolars become tricuspid, moving
apart widely as the jaw bone prolongates. Only in Anoura - unique among phyllostomids -
there are three premolars on each side of the upper jaw. This situation represents a
virtually reverse evolutionary trend and possibly originated from a secondary doubling of
the anterior premolars (Phillips cited by Koopman, pers. comm. 1991). This secondary
replenishment of the gap between the upper canine tooth and the upper P° having
developed in the course of rostrum prolongation may derive its ecological significance
from the high percentage of insect food which has been reported for this genus.

Molar teeth

Considerable changes occurred in the molar teeth: whereas Glossophaga and Lionycteris
still show the dilambdodont profile of the masticatory surface typical of insectivorous bats,
Choeroniscus, Hylonycteris and Choeronycteris hardly have any crown cusps nor W-
shaped ectolophs.

With increasing jaw length, the molar teeth increasingly moved apart, too. Thus, in the
very long-muzzled forms there is apparently no masticory function any more. Furthermore,
the crowns are not only narrow in top view, but also hardly project beyond the gums. The
masticatory surfaces are largely reduced; leaving only few pointed cusps (premolariform)
with a physiological value on which we only can speculate. They appear of very little use
in food preparation and do not seem to be essential for a nectarivorous way of life.

Rostrum

Within the Phyllostomatidae, the specialized nectar feeding bats of the New World are
characterized by a prolongation of the visceral cranium involving the maxillaria, nasalia,
palatina, the vomer and the mandibulae.

The degree of this prolongation varies considerably between single species and may thus
be considered an evidence of the degree of feeding specialization. Species preferring
varied diet like Glossophaga or Lionycteris have shorter rostra than highly specialized
nectar feeders like Choeronycteris.

Within a genus these proportions are also influenced by body size of the species. Smaller
species showing comparatively larger braincases and shorter jaw lengths, respectively. In
this case, the construction principles are just the same despite their differring proportions.
Differences in proportion are found in closely related taxa (subspecies, species) whereas
different construction principles occur beyond the species level.

In contrast to the short viscerocranium of the predominantly frugivorous genus
Brachyphylla, the Phyllonycterinae (Erophylla and Phyllonycteris) were subject to a
moderate muzzle prolongation as they specialized in nectarivory. The situation is the same

71

for Lionycteris within the Lonchophyllinae and for Glossophaga within the Glossopha-
ginae.

Within the genus Lonchophylla there are already some species with relative long rostra; L.
thomasi has a comparatively short rostrum; this feature, however, depending on body size:
within each genus, relative rostrum length increases allometrically with body size. The
longest rostrum of all Lonchophyllinae is found in the genus Platalina with the palate
length comprising more than half of the total skull length.

Also within the Glossophaginae, the smaller species show relatively short rostra
(Lichonycteris obscura, Choeroniscus godmani). There are, however, considerable
differences between genera of equal size (Leptonycteris and Choeronycteris). As an
extreme, the Mexican banana bat Choeronycteris harrisoni has the longest rostrum (in
relative and absolute terms) among all nectar feeding bats.

Forehead

Since rostrum length is emphasized, the development of the forehead is increasingly
restricted: interorbital and postorbital width are relatively small and, in the extreme case,
no longer distinguishable from each other any more. Similar to other extremly long-jawed
mammals (cf. Myrmecophaga) the Glossophaginae lack a demarcation from the orbitae by
visible orbital processes.

Zygomatic arches

Reduction of the zygomatic arches corresponds to the significance of the masticatory
musculature which is reduced with increasing specialization. These structures are often
damaged during preparation or inadvertently totally removed (in Husson 1965, the missing
zygomatic arches were erroneously recorded as a determination feature within the
Glossophaginae sensu strictu). Whereas neither the Carolliinae nor the Lonchophyllinae

Brachyphylla
(frugivorous)

Fig.56: Skulls of bats with
different diet.

(nectarivorous) JOB = interorbital width

POB = postorbital width

HKB = width of braincase

MB = mastoid width

C.s. = crista sagittalis

Choeroniscus

78

have any bony zygomatic arches, these structures remain detectable at least by X-ray prior
to preparation (Choeroniscus). Only in the extremely long-nosed species of the genus
Choeronycteris no zygomatic arches could be found.

Braincase (fig.56)

Although cranial modifications with reference to nectarivory predominantly manifest
themselves in the viscerocranium, in some respects also the braincase of nectar feeding
bats is clearly distinguishable from that of other phyllostomids. This is principally due to
receding significance of the masticatory muscles; as the origin of the M. temporalis does
not express itself at the vertex of the bone surface any more.

In the predominantly frugivorous genus Brachyphylla, a Crista sagittalis is just present to
some extent. In all genera of the remaining nectarivorous subfamilies this characteristic
lacks in both sexes.

The Pars petrosa is merely developed (i.e. = in specialized nectar feeders like
Choeroniscus, the width of the braincase exceeds mastoid width).

Skull base (fig.57)

In most species, the base of the neurocranium and the palate are arranged in an almost
parallel way, i.e. there are no angles between both areas. On the other hand, the outgroup
genera of Phyllostominae and Carolliinae, but also Brachyphyllinae and Phyllonycterinae
show an inclination of the palate plane towards the skull base. In Choeroniscus and
Hylonycteris, there is a striking elevation of the brain capsule against the palate plane
(fig.57). This affects measurements of skull height in so far as the slide caliper rests on
the hamuli rather than the Bullae tympanicae. Whether this arrangement is possibly
significant for some special form of nectarivory cannot be assessed at the moment. In
terms of functional morphology, there is a reference to the position of the mandibular
joint. In comparison, the even longer skull of Choeronycteris tends to the opposite angular
inclination. This might be due to mechanical constraints which require the elongate
rostrum bent down in ventral direction to gain stability.

The convexities of the basisphenoid vary considerably between the species of a genus, too
the “basisphenoid pits” represent an essential determination feature within the genus
Glossophaga (Webster & Jones 1980).

One of the strangest features within the family is the extreme prolongation of the
pterygoid processes (hamuli). In Choeroniscus and Choeronycteris they are extending to
the Bullae tympanicae with lateral widenings. Functional significance may not have been
interpreted properly on the basis of the preserved material. The soft palate is, however,
probably prolonged in occipital direction: the hamuli function as an abutment (roller
bearing) against the tendons of the M. levator velum palatini. Functionally, this might
improve efficiency of the swallowing motor apparatus during high-frequency tongue
movements.

>

Phyllonycteris

clinorhynchic

Monophyllus a

orthocranic

Choeroniscus

airorhynchic

Choeronycteris

clinorhynchic

Musonycteris

re ee

Fig.57: Orientation of the skull
base in Phyllonycterinae and

clinorhynchic Glossophaginae

80

Mandible (fig.58)

Whereas Carollia and Phyllostomus, but also Brachyphylla have strong mandibles,
increasing specialization brings about a flattening of the Ramus mandibularis, the
Processus coronoideus becomes relatively smaller with increasing length of the mandible.
The highly specialized nectar bats show a characteristic ridge (symphysal ridge) at the
mandibular symphysis, giving the lower jaw a “bulb bow profile” in lateral view -
probably due to reduction of the lower incisivi and the development of a V-shaped
notching of the mandibular tip and of the median lower lip forming a tongue guiding
channel. None of the basic genera shows a comparable feature. It obviously may serve to
stabilize the fused mandibular tip as the notching between the lower canini has increased
to an extent that afflicts the mandibular suture. According to Freeman (1995) all nectari-
vorous bats have fused mandibulae, which probably increases stability of the jaws but also
could result from less need for minute adjustments at the symphysis in order to precise
registration of cheek teeth.

Phyllonycterinae Brachyphyllinae
Phyllostominae
N |?

Erophylla Brachyphylla

Glossophaginae ef ?

Anoura

ps = = nes — ET
2 <= Hylonycteris = Non ame FR \
Dee

i Monophyllus IN

Choeronycteris

en & 7 Glossophaga
ene 5 7

Musonycteris Scleronycteris

Phyllostomus

Carolliinae

I

le

Carollia

Lonchophylinae fleet

. :
a ed dan 3
Platalina Lionycteris

Lonchophylla

Fig.58: Mandibles of the phyllostomid families examined in lateral view. Specialization on nectar
feeding increases from right to left. For clarifying the construction principles, the jaws were drawn in
the same size, ingnoring the scale. The arrows indicate evolutionary trends of mandible construction
as derivations of common features

Craniometry

Univariate Analysis

The results of univariate analysis of the skull measures are given in tab.4a, b (p. 64-66)
for all taxa examined. Only for individual species the mean has been worked out.

These data reveal, however, that their range may overlap between individual species of a
genus, even if the mean values significantly differ from each other.

Only those taxa comprising sufficient numbers of specimen were processed statistically,
the rare species allow nothing than individual records.

81

Sexual dimorphism:

- Within the genus Phyllostomus which was examined as an outgroup, almost all skull
measures of male P. hastatus - a very large species - range above the female ones.
Swanepoel & Genoways (1979) recorded the forearm lengths and seven skull measures of
eight individuals collected at different locations showing that the relations between the
sexes appeared similar in forearm length. In biological terms, this difference in size may
be explained by the social system of these bats; in Phyllostomus hastatus, the males
establish harem groups within territories (Koepcke 1987).

- In all species of Choeroniscus examined so far, the female length values of the skull
exceed those of the males in a statistically significant way. Probably, this difference in
size allows these bats to utilize their resources in a more efficient way, as they live in
very small family groups (Koepcke 1987). Though the rain forest teems with chiroptero-
phile flowering plant species, they occur in quite low density. Probably it is only due to
slightly diverging niches that both sexes may survive sharing the same habitat. Appro-
priate evidence is, however, still to come. It is quite as probable that body size is deter-
mined by social factors rather involving the females.

- The only measured male of Lonchophylla robusta is definitely larger than all the female
individuals - unlike smaller species of the genus where the females are larger. In L.
handleyi which grows even a bit larger, the data of the males exceed those of the females
in 12 of 17 parameters on the average, but based on the available sample this could not be
secured statistically.

Females larger than males are not uncommon among bats. This may be a response to the
increased energetic demands of flight in pregnant females (Myers 1978; Williams &
Findley 1978). Williams & Findley (1978) however suggested other factors like thermore-
gulation and fat storage might be more important determinants of body size of females
than adaptations to flight (Sahley et al. 1995).

Skull proportions

An account of both, the absolute size relations and adaptive degree of the species
examined, gives the summary of measurement relations of the ‘Results’: this refers to
rostrum length and gives evidence of decreasing functional significance of the masticatory
apparatus.

1) Proportions of total skull length to length of the palatea: Although palate length is a
part of total skull length, it remains a useful measure for recording the size of the visceral
skull.

Both phyllonycterine genera (Phyllonycteris and Erophylla) perform an even higher
quotient (= shorter bony palate) in tab. 5 than the outgroup (Carollia). Here, there must be
considered that both genera have a V-shaped edge of the palate running at a very acute
angle and thus affecting the value in this parameter (but not total rostrum length).
Furthermore, this ranking essentially represents the degree of specialization within the
nectar feeders: Lionycteris and Glossophaga / Monophyllus proved to be basal forms in
this respect, Choeroniscus and Choeronycteris the most specialized ones.

2) Total skull length to length of upper tooth row: Some aspects in dentition which cannot
be explained exclusively by rostrum prolongation prove inconvenient: the number of teeth
varies considerably among the genera (cf. dental formulas). In the course of rostrum
prolongation only the distance between the postcanine teeth is enlarged in all long-headed
species, but the measured length may consist of just four to six teeth.

82

Thus, within the resulting order of relative lengths of the maxillar tooth row (the smallest
value for quotients represents the longest tooth row) the dental formula which differs from
the original pattern of the subfamily (five postcanine teeth on one half of the upper jaw)
must be taken into consideration: in Leptonycteris and Lichonycteris there are two
premolars and molars each on one half of the upper jaw, and in Anoura there are 6 teeth
in total forming the postcanine tooth row. In comparison, there is an extreme long tooth
row in Monophyllus. Though considered primitive by other characteristics, in this respect,
this genus is in no way inferior to the highly specialized species.

3) Height of the braincase to palate length: Whereas total skull length in 1) certainly
shows a trend to muzzle prolongation, this will not be expected for height of the
braincase. Size and shape of the braincase are determined primarily by brain morphology,
and in forms with strong masticatory musculature there is an additional influence by the
musculature originating at the braincase (Crista sagittalis). As the latter is weakly
developed in all Glossophaginae (revealing a decrease of masticatory function in food
intake), it does not influence the height of the braincase at all.

The position of Phyllostomus within the order of the quotients remains remarkable,
suggesting that the palate is relativly longer compared to braincase height - in this case,
the quotient is determined by the Crista sagittalis with an incomparable brain case height.

4) Total skull length to coronoid height: The height of the coronoid process at the Ramus
mandibularis refers to masticatory functions of the species examined: here the M.
temporalis which is essential for efficient snapping, inserts at the mandible. As
specialization on nectarivory increases, this musculature becomes less important. The
quotient of total skull length and coronoid height indicates which of the genera
experienced the most distinctive prolongation of the visceral skull with a simultaneous
reduction of the masticatory function.

5) Height of the braincase to coronoid height: Apart from the relation to total length, the
comparison of a quantity independent of rostrum prolongation as braincase height and
coronoid height does make sense as a measure reflecting the functional reduction of the
jaw apparatus: in this context, the otherwise very long-headed Platalina takes a basal rank,
and with a quotient of 2.12, Hylonycteris shares the same level with the unspecialized
genus Lionycteris (2.13). In contrast to the comparison SGL to CH, also Musonycteris has
shifted to a middle position, and Erophylla stands out with a quite flat mandible compared
to the hight of the braincase (2.17). These partly different orders in both quotients are
strongly determined by absolute skull size - the denominator - braincase size - is
comparatively larger in smaller individuals. This supports once more the necessity of
performing allometric comparisons, too.

6) Mandible length to coronoid height: The length of the mandible may be referred to as
a reliable measure for the length of the viscerocranium. Setting mandible length to
coronoid height also results in a representative perspective for the extent of specialization
in the genera examined. The larger the quotient, the longer the mandible related to
coronoid height. In addition to its significance, the value of mandible length is reinforced
by the negative trend of coronoid height in the course of higher specialization on nectar
and pollen feeding.

Allometrics
Intergeneric allometrics

Although allometric approaches usually are chosen in different size relations of the
systematic sample taxa (in individuals of equal size, proportion differences would stand
out already by simple comparison of the values), it may nevertheless be useful to analyse
allometries even in individuals of almost same size, as allometries sometimes do differ in
individual features even in species of equal dimensions.

Referring to nectar intake, some specialized flower bats show an extremely prolonged
rostrum. Relative length to total skull length varies considerably between individual
species. So, there are differences in relative length of the visceral skull both between the
species of a genus and between inidivual genera. Furthermore, the species show
considerable differences in body size (factor up to 1.8). Thus, it was to examine whether
different proportions can be explained by allometry - and whether the rostra of different
size follow the same principles of construction.

From the total of 17 skull measures involved in univariate analysis, five allometric
parameters which are especially affected by rostrum prolongation were related to SIZE:

1. Width over canini (CC) (figs. 61-62): This measure gives information on terminal
rostrum width and affects the geometry of the visceral skull with respect to its relative
size.

2. Upper tooth row (OZR) (figs. 63-65): Distinct assignment of the canine, premolar and
molar alveoli to the maxillaria allows to determine relative length of these maxillar bones.
It remains, however, important to consider the dental formula and the number of
postcanine teeth, respectively, in the upper jaw.

3. Length of the palate (GL) (figs. 66-70): Apart from the maxillaria, primarily the
palatinae become stretched in the course of rostrum prolongation. Thus, this measure
compared to the braincase (SIZE) also gives information on relative prolongation of the
visceral skull. The conspicous contours of the proximal palate edge variy in individual
species between sharply V-shaped over smoothly U-shaped up to a weakly rounded,
almost straight form. Thus, this measure is considerably determined by the construction of
the bony palate.

4. Length of mandible (UKL) (fig.71): This measure reflects a very useful correlate of the
respective rostrum length. Compared to OZR and GL it has the advantage that a) the
measure also comprises the rostral part anterior of the canines, b) none of the
constructions not originally associated with the visceral skull will affect the measure, and
c) the measures are easily accessible.

5. Coronoid height (CH) (fig.72): Representing an expression for the functional
significance of the M. temporalis for the interlocking jaws, the relation of coronoid height
to braincase (SIZE) gives information on relative importance of the chewing process, thus
considering an important aspect of functional morphology in the analysis of rostrum
prolongation.

The data given earlier (p.66) lead to the following statements:

Width over canini (C-C): In the genera Carollia and Phyllostomus summarized in the
outgroup, the canine distances are comparatively wider than in the Glossophaginae and
Lonchophyllinae (fig.50, tab.5).

84

|
440 .546 840.100 1602.032 3055 .000 5825.74

Fig.59: Relation SIZE to width over canini: comparison between Phyllonycteris (4) and outgroup (+)

With increasing skull size, the anterior width of the palate increases to a proportionally
less extent, the gradient of the reduced elliptic major axis clearly remaining below 45
degrees (tan alpha = 0,57). It is, however, remarkable that the genera Erophylla and
Phyllonycteris having been summarized within the subfamily Phyllonycterinae do not
differ from the outgroup in this respect (fig.59).

In the Glossophaginae, size-dependent shift in proportion follows the same pattern; the
distances between the canine teeth are, however, smaller at equal body size in a genus-

14.452

8.952

5.546

L Eee
ey Et
3.435 on HT
Sin
2.128
439.827 875.913 1744.375 3473.913 6918.278

Fig.60: Relation SIZE to width over canini: comparison between Glossophaga (U), Phyllostomus (+)

85

6.340

ha

-B29

1.889

225.184 402.365 718.955 1284 .646 2295.43

Fig.61: Relation SIZE to width over canini: comparison bewteen Glossophaga (2) and Brachyphylla (+)

specific way, e.g. the rostra are narrower at their tip than in the outgroup (fig.60). This
also applies to Brachyphylla, although in many other respects this genus approaches the
outgroup much closer than other nectar feeding bats. This is partly due to their
considerably differring skull geometry; in basal view, the palate appears rounded in a
trapezoid way (the width of the palate over the second molar tooth ist much larger than
over the canines, fig.61).

9 220
6.139
cet
4.088 Pe
BT
BE

2.722
1.812

372.464 668 622 1200 265 2154 _633 3867 -£

Fig.62: Relation SIZE to width over canini: comparison between Glossophaga (2) and Lonchophiylla (+)

86

15.827

10.018

6.341

175.229 917.535 1771.504 3420 .282 6603.61

Fig.63: Relation SIZE to upper tooth row: comparison between Brachyphylla (+) and outgroup (©)

Still, the lower gradient of the reduced major axis in Lonchophylla combined with a high
correlation coefficient remains to be interpreted. Also here, an acute-angled geometry of
the palate weakens proportional shifts in size alterations (figs.50, 62).

Upper tooth row (OZR): Referring to the length of the postcanine tooth row in the upper
jaw, the basic subfamilies Brachyphyllinae and Phyllonycterinae do not differ from the
outgroup (cf. figs.51, 63).

In size-dependent comparison of proportions, also the glossophagine genera Lichonycteris
(fig.51) and Leptonycteris (figs.51, 64) closely approach the allometric line characterizing
the outgroup - despite a condition of higher specialization in other features, their maxillar
tooth row does not exceed that of Carollia, Phyllostomus, Brachyphylla, Erophylla and
Phyllonycteris in length. The reason is quite simple: in both genera, there are only two

12.065

9.997

6.863

794.895 1041 .789 1365 .367 1789 .449 2345.27

Fig.64: Relation SIZE to upper tooth row for Leptonycteris

87

10.689

8.855

7.336

6.078

535.767 01.650 919.778 1205 .718 1580 ..E

Fig.65: Relation SIZE to upper tooth row for Anoura

molars on each half of the upper jaw (figs.63 and 64). With respect to this feature, the
remaining Glossophaginae and Lonchophyllinae equipped with five postcanine teeth, are
subject to the same allometric conditions. It is, however, remarkable that Anoura (figs.51,
65) does not show a larger integration constant although it has more teeth (3 premolars in
the upper jaw), as in the other genera, with increasing size the large number is
superimposed by larger interdental distances within the length of this measure.

Length of the palate (GL): With respect to statistically securable allometric conditions of
this skull measure, Brachyphylla, Erophylla and Phyllonycteris correspond to the outgroup
in this parameter, too. Although the random samples of Brachyphylla and Phyllonycteris

24.487
oP ae
16.311 e
- ne

ca
10.865 7
7.238

252.689 453 310 813 213 1458 858 2617 108

Fig.66: Lonchophylla: relation between SIZE and palate length

88

24.880

16.540

10.995

7.309

255.697 460 .042 827.694 1489 .162 2679.254

Fig.67: Relation SIZE to palate length: Lonchophylla (+) and Platalina (MM)

differ with p<0,05 from those of the outgroup, the slopes of the reduced major axes of the
distribution do not.

The proportional shift of relative palate length remains inconsistent in the Glossophaginae
and Lonchophyllinae. There is no doubt that the gradient of the reduced major axes in all

genera, where correlation could be secured, is distinctly below 45 degrees and thus not to
be secured against the outgroup, but their integration constants do show considerable
differences.

Lionycteris, Lichonycteris, Leptonycteris, Monophyllus and Anoura have comparatively
short palates which nevertheless exceed those of the outgroup in length. The genera to
which the respective subfamilies (Glossophaga and Lonchophylla) owe their name show
a slightly higher level (longer palates).

Compared to the outgroup, the palate of Platalina is relatively longer (0.1975) than in the
smaller Lonchophylla (0.1396), but the position of the mean value in the double
logarithmic coordinate system closely approaches the reduced elliptic major axis of
Lonchophylla (fig.66, 67). This means the relatively more elongate palate may be
explained by allometry; the principles of construction are, thus, the same.

The extremely elongate palates of Choeroniscus, Hylonycteris and Choeronycteris (C.
mexicana and C. (=Musonycteris) harrisoni) are still to be interpreted: in Choeroniscus,
there is a transposition of the reduced major axis with respect to the remaining
Glossophaginae (figs.52, 68). Here we find a different genus-specific structural feature:
even very small C. godmani have a much longer palate than a Lichonycteris or Lionycteris
of the same size. Besides the angular elevation of the palatinal area in relation to skull
base, Hylonycteris and Choeroniscus share a bony palate which is considerably
prolongated in occipital direction. This functional significance of this apparently derived
feature cannot immediately be releated to nectarivory. I would regard this as
systematically useful evidence on close relationship of both genera.

Referring to this structural feature, the position of Musonycteris in the double logarithmic
coordinate system gives evidence on close relationship to Choeroniscus. Interpretation of

89

=1.790

14.976

10.293

7.075

354.632 608.170 1042.9693 1788.619 306

Fig.68: Relation SIZE to palate length: comparison between Choeroniscus (+), Glossophaga (0)

the allometries is, however, hindered by a relatively shorter palate in C. mexicana: here
the reduced major axis of C. mexicana - though statistically not securable due to the small
random samples with a correlation coefficient of 0.5810 - has an extremely steep slope
(with a tan a of 1.6174) so with increasing size that the position of Musonycteris - despite

88.775
41.761
19.645 a
9.242
4.347 /
341.678 1010 .958 2991 .229 8850 .467 26186

Fig.69: Relation SIZE to palate length: comparison between Choeroniscus (2) and Choeronycteris (+)

90

of its different construction - comes up to values which would have been expected for
Choeroniscus. This manifests itself by intrageneric allometrical calculation of a common
random sample comprising C. mexicana and C. harrisoni (fig.69): they significantly differ
from Choeroniscus both in the gradient and in the distribution ellipse (p<0.05).

Length of mandible (UKL): Allometric regression of this measure to the reference
quantity SIZE results in similar groupings: once more, Brachyphylla and Phyllonycteris
approach the allometric line (reduced major axis) of the outgroup (Carollia and
Phyllostomus). In this parameter, the mean values of Erophylla differ from Phyllonycteris
by a mandible length with proportions resembling those of the remaining basic
glossophagines and lonchophyllines.

Furthermore, the integration constants of Choeroniscus, Hylonycteris and Choeronycteris
are more distant from the outgroup. In this parameter even Platalina matches distances to
the outgroup usually applying to extremely long-skulled bats (fig.70, tab.8), subsequently
following the allometries calculated for the genus Lonchophylla.

36.843
24.492
me
Ang
16.282 =
ae eee alia
met Ber
so
eae =
ige eu
10.824
7.195 Be
284.067 511.084 919.527 1654 .385 2976 .51

Fig.70: Relation SIZE to mandible length: comparison between Choeroniscus (+) and Lonchophylla
+ Platalina (0)

Once more, Musonycteris sticks out within this comparison. Though it is about the size of
Platalina, it shows a distictly more elongate mandible, in this respect even exceeding
Choeronycteris mexicana by far.

Coronoid height (CH): In the double logarithmic representation of regression of coronoid
height to SIZE, the situation is reversed: the Ramus mandibularis is strongest within the
outgroup members and in the individuals with the relativly longest rostra, the rami of the
mandible are more flattened. Thus, the reduced major axis of the outgroup runs above,
and deviations of individual genera are represented by positive values in the table (fig.54).

91

In allometrical respect of this feature, Brachyphylla does not even differ from Phyllo-
stomus or Carollia (F-value = 1.82 referring to a table value of 3.23 for 95%
probability).It is quite interesting that Lionycteris, Lonchophylla and Platalina have the
least distances from the outgroup (they all differ with p<0.05 from the outgroup), and
Platalina remains exactly on the allometric line calculated for Lonchophylla. In other
words: Platalina genovensium represents a large species of Lonchophylla also with respect
to mandible proportions.

Compared to body size, the phyllonycterines Erophylla and Phyllonycteris have flattened
coronoid processes just as in Glossophaga, Monophyllus, Hylonycteris or Anoura. In
Choeroniscus and Choeronycteris (including Musonycteris) the flattening of the mandibles
is even intensified, though their proportions do match in allometrical respect. There is,
however, a striking difference between Hylonycteris and Choeroniscus (fig.71). As the
skull morphology of Choeroniscus differs from Hylonycteris only by its shovel-like,
widened hamuli reaching up to the bullae, there is to be considered whether their function
may be related to the extreme flattening of the mandible.

In the course of this study, many species were represented by few specimens. Thus it is
impossible to give a comprehensive comparison of intraspecific allometries for the entire
group. It does, however, make sense to give an intergeneric comparison of allometric lines
between genera comprising numerous individuals and species of distinctly different size.
In this respect, the previous systematic reviews suggest to refer to Lonchophylla, Glosso-
phaga, Anoura and Choeroniscus apart from the reference genera Carollia and Phyllosto-
mus. Erophylla, Phyllonycteris and Leptonycteris proved less appropriate, although there
were enough specimens available. But the values measured overlap in individual species
i.e. in these genera respective intraspecific variability exceeds interspecific distances.
Furthermore, differences obtained by measuring are partly found within the error range
given by the measuring method.

6.949
4.734
=
par ag
© DE
3.225 a sie ts
ft oe en
ee +
é er
ee
2.197
1.497
277 .230 481.497 836 272 1452 452 25227 64?

Fig.71: Relation SIZE to coronoid height: comparison between Hylonycteris (DO) and Choeroniscus (+)

En

Allometrical sex comparison

As already mentioned this comparison can only be performed by random samples of those
genera comprising many specimens available. In tab.11 “upper tooth row” (as a regression
to the volume measure SIZE) shows for the three genera Anoura, Glossophaga and
Lonchophylla that no sex difference could be secured at all.

Instead, there were differerences calculated in position and/or gradient in

- Anoura: Molar width of the palate (p<0.05/p<0.05); width of brain case (p<0.05/n.s.);
mandible length (p<0.05/n.s.)

- Choeroniscus: Molar width of the palate (p<0.05/p<0.05); width over canini
(p<0.05/p<0.05)

- Glossophaga: Coronoid height (p<0.05/n.s.)

- Lonchophylla: Mandible length (p<0.05/n.s.)

These arithmetical differences refer to random samples of genera comprising many
individuals - unfortunately, samples of some species are represented by one sex only, for
example in Anoura brevirostrum (females only). Thus, by means of the material available,
no statements can be made on sex-related differences in the studied allometrics.

Allometric conclusions
All this considered, my data give evidence of the following:

In the nectarivorous bats studied here, the proportions of the visceral skull compared to
the braincase do not all follow the same allometrics. Five rostrum-related skull measures
of the viscerocranium compared to the proportions of the calculated neurocranium
volumes, give evidence of different construction principles.

In relative terminal rostal width - measured at the distance of the upper canine teeth - all
nectarivorous genera show narrower rostra compared to the genera of the outgroups. The
palate surfaces of the Glossophaginae are more (Choeroniscus, Hylonycteris) or less
(Glossophaga, Monophyllus) rectangular; in Lionycteris, Lonchophylla but also in
Platalina the geometry of palate surface tends to be trapezoid. Referring to their allome-
tric gradient only the Lonchophyllinae differ from the outgroup genera.

Three measures (length of the palate, upper tooth row, mandible length) represent rostrum
length to a considerable extent. Allometric analysis of all three parameters in proportion
to the volume quantity SIZE does not reveal any difference of the Brachyphyllinae
(Brachyphylla) and Phyllonycterinae (Erophylla, Phyllonycteris) - classified as basic
groups compared to the outgroup (Carollia, Phyllostomus). Comparing the upper tooth
row to SIZE and of the mandible length to SIZE, the genera of the Lonchophyllinae
(Lionycteris, Lonchophylla, Platalina) and Glossophaginae (Anoura, Choeroniscus,
Choeronycteris, Glossophaga, Hylonycteris, Leptonycteris, Lichonycteris, Monophyllus,
Musonycteris) do differ significantly from the outgroup, but they do not differ among each
other. There is a remarkably precise correspondence of the measures taken in Platalina
with the allometrics calculated for Lonchophylla.

A clear difference between the genera Choeroniscus, Choeronycteris and Hylonycteris
towards the remaining Glossophaginae was revealed by allometric comparison of palate
length to the SIZE measure. The allometric regression line (reduced elliptic major axis)
shows a transposition (towards a relativly longer palate at equal size), revealing principle
differences in skull morphology.

Systematic Conclusions

The systematic relationships in the twelve “glossophagine” genera are discussed for long
time:

Based on relative length of metacarpals and phalanges, even Sanborn (1943) classified the
Glossophaginae into two groups - Glossophaga, Lichonycteris, Scleronycteris,
Choeroniscus, Hylonycteris, Choeronycteris and Platalina against Lonchophylla,
Leptonycteris, Monophyllus, Lionycteris and Anoura. These data have to be interpreted,
however, mainly in a functional context.

Baker (1967) refered to similarities in the caryotype between Leptonycteris and
Glossophaga with Phyllostomus, Trachops and Macrotus on one hand and Choeroniscus,
Choeronycteris and Carollia on the other.

In contrast, Gerber (1968) using immunologic and electrophoretic comparisons of serum
proteines, proposed Choeronycteris to be closer related to Phyllostomus than to Anoura,
Glossophaga and Leptonycteris. The latter genera he supposed to be closer relatives of
Carollia, Artibeus and Sturnira.

Walton & Walton (1968) who examined shoulder and pelvic girdles, did not find any
dichotomy within the Glossophaginae.

Having studied the dentition of the glossophagines, Phillips (1971) presumed that these
bats apparently comprised several lines. One of these groups consisted of Glossophaga,
Monophyllus, Leptonycteris, Anoura, Lonchophylla, Lichonycteris, Lionycteris,
Hylonycteris, Scleronycteris and Platalina. This group seemed to be closely related to the
phyllostomatine line of Micronycteris-Macrotus (Smith 1976). The second group of Phil-
lips’s comprised Choeroniscus, Choeronycteris and Musonycteris, apparently showing
some relationship with the line of Phyllostomus within the Phyllostomatinae (Smith 1976).
Both groups Phillips characterized by features of the dentition and the skull (c.f. fig.72).

Choeronycteris

Choeroniscus

Platalina

Lichonycteris

Hylonycteris
Lonchophylla
Lionycteris
Anoura

Leptonycteris

Fig.72: Cladogram of New
World nectar-feeding bats
adapted from Phillips (1971)

Glossophaga

Monophyllus

94

Lonchophylla
Lionycteris

Platalina
Brachyphylla

Phyllonycteris
Erophylla

Glossophaga

Monophyllus
Lichonycteris
Leptonycteris

Anoura

Hylonycteris

Choeroniscus
Fig.73: Cladogram of New
World nectar-feeding bats

& hoeronycteri s adapted from Griffiths (1982)

Based on comparative morphology of the tongue anatomy and the hyoid region (“tunnel
insertion of the geniohyoideus, posterior shift of the styloglossus insertion” in the true
Glossophagines; two lingual arteries in the Lonchophyllinae) Griffiths (1982) proposed
that the genera Lonchophylla, Lionycteris, and Platalina must have separated at a very
early stage from a line of the Brachyphyllinae / Glossophaginae.

At the same time Baker and Hayduk (1982) concluded from chromosome examinations
(G-Banding patterns) that the genera separated by Griffiths represent a closely related
group within the Glossophaginae. Warner (1983) argued with both points of view,
emphasizing the difficulties:

In addition to both hypotheses, he proposed a (deliberately artificial) even more
“economical” cladistic arrangement. Whereas Griffiths’ cladogram (fig.73) required at least
seven convergences, Hayduk & Baker (c.f. fig.74) confined themselves to one convergent
new development, instead, however, requiring five “retrogressive developments” - all
within the group of the “Lonchophyllinae”. Warner's cladogram (c.f. fig.75) required three
reversions (i.e. features developing retrogressively towards the original condition - all
within the “Lonchophyllinae”) as well as only two convergences. Thus it would require -
theoretically - a minimum of auxiliary hypotheses. Though Griffiths’ convergences appear
to be the most expensive in numerical terms (7 events), they refer without exception to
modifications of the tongue musculature. Similar modifications of the tongue base are also
known in other mammalian groups having elongate, protrusible tongues - thus convergent
development in both bat groups may be quite probable.

In contrast, the reversions required in Warner’s, and especially in Hayduk’s & Baker's
models, refer to three very special adaptations to nectarivory: “brush-tip of the tongue

95

Brachyphylla

Phyllonycteris

Erophylla

3 Glossophaga

Monophyllus

Lichonycteris

Leptonycteris

Anoura
Lonchophylla

Lionycteris
Platalina

Choeronycteris

Fig.74: Cladogram of New Choeroniscus
World nectar-feeding bats
adapted from Hayduk & Baker

(1982) Hylonycteris

covered with hairlike papillae” (1), one single tongue artery present (2) and enlarged
tongue venes (3). It is quite improbable that these features - once having developed - were
lost again within a group of nectar feeders.

Furthermore, the cladogram of Hayduk & Baker requires additional reversions in the
derived features “tunnel insertion of the geniohyoideus” and “back shifting of the insertion
of the styloglossus”. Presumably the changes at the insertion of these two muscles
correspond to the extreme agility of the tongue. As this is a very characteristic feature of
nectar feeders, too, it also seems illogical that these characterisctics should have been
reverted to their original condition within a nectarivorous species group. Exactly matching
the point, the authors found that Lonchophylla and Lionycteris show a wide immunolo-
gical distance from other nectarivorous New World bats. So, they apparently deviated
from the line leading to the Glossophaginae prior to separation of Brachyphylla. This
supports Griffiths’ classification as subfamiliy “Lonchophyllinae”.

Based on this study submitted on skull morphology and allometrics of the measures which
characterize rostrum prolongation and which may be recorded craniometrically, the
following additional points turned out: 1. Lionycteris, Lonchophylla and Platalina have

96

; Brachyphylia
Phyllonycteris

Erophylla

Lonchophylla
Lionycteris

Platalina

Lichonycteris

Monophyllus

Glossophaga

Leptonycteris

Anoura

Choeroniscus
Fig.75: Cladogram of New Hylonycteris
World nectar-feeding bats

adapted from Wamer (1983) Ch a
oeronycteris

some remarkable features in common: independent from rostral length, all genera lack the
zygomatic arches. This derived feature links them with the familiy Carolliinae (genera
Carollia and Rhinophylla) which also always lack bony zygomatic arches.

The dentition shows a strikingly strong development of inner upper incisivi, and the lower
ones are always completely preserved. Even here, their features match those of the
Carolliinae, though probably representing a symplesiomorph condition. The Ramus
mandibularis proves quite strong in the allometric analysis (all three genera have the
relatively least distance to the outgroup and to Brachyphylla) and marks a plesiomorph
condition common to all three genera, too.

Referring to size-dependent proportional shift, Lionycteris, Lonchophylla and Platalina
correspond to the same allometric constant in the measures examined (width over the
canine teeth, coronoid height, length of the maxillar tooth row C' - M’, palate length and
mandible length). Thus, the relativly longer rostrum of Platalina is determined by body
size and constructed in accordance with the same principle as Lionycteris and Loncho-
phylla.

Considering the cranial characteristics having become available I suggest to include
Platalina genovensium as the largest species in the genus Lonchophylla. The original, very
wide gap in size difference compared to the so far known species of Lonchophylla which

97

I suppose will mainly have encouraged Thomas in classifying this species in a genus of its
own, is meanwhile linked by Z. handleyi which Hill described in 1983.

2. Choeroniscus, Hylonycteris and Choeronycteris differ from the remaining
Glossophaginae in essential construction principles: with reference to skull morphology,
Choeroniscus resembles Hylonycteris by its airorhynchic skull; on the other hand (in con-
trast to Hyloncteris) Choeroniscus and Choeronycteris share extremely elongate pterygoid
processes as well as considerably flatter ramus mandibulares. By allometrical comparison,
all three genera give evidence of a transposition of the allometric line in the ratio of SIZE
to palate length: although palate length increases with growing total skull length by the
same factor as in other Glossophaginae, the allometric line (reduced major axis) runs
parallel but on a higher level. So, even in the smaller Choeroniscus godmani, relative
length of the palate clearly exceeds that of an equal-sized Lichonycteris.

Whereas Hylonycteris in allometric comparison behaves almost identical to Choeroniscus
in four of the measures examined (CC, GL, OZR, UKL), allometric analysis of the
telation of SIZE to coronoid height reveals a different development of the Ramus
mandibularis which would have been overlooked by merely examining the skulls
individually. Although, like other Glossophaginae even Hylonycteris has a comparatively
flat mandible, in Choeroniscus and Choeronycteris this flattening developed to a more
advanced level, distinctly demarcated by the integration constant.

These data on skull morphology and allometry allow to differentiate four predominantly
nectarivorous subfamilies within the Phyllostomatidae: these are the Brachyphyllinae und
Phyllonycterinae endemic to the Antilles Islands as well as the Lonchophyllinae and
Glossophaginae (having been summarized as Glossophaginae by other authors). This
systematic subdivision is in accordance with Griffiths (1982).

Within the Glossophaginae sensu strictu the tribe Choeronycterini comprising the genera
Hylonycteris, Choeroniscus and Choeronycteris (incl. subgenus Musonycteris) represents
an extremely high specialized group of nectar feeders.

SUMMARY

Skull morphology of 13 New World nectarivorous bat genera was analyzed under
functional aspects and compared with individuals from systematically neighbouring
subfamilies of the Phyllostomatidae (Carolliinae, Phyllostominae).

The nectarivorous flower bats are characterized by special adaptations to this diet,
primarily imposing cranially by rostral prolongation to various extent. The degree of this
prolongation varies considerably within and between individual genera and can be judged
as an evidence of the extent of feeding specialization. Species taking varied diet
(Glossophaga, Lionycteris) have shorter jaws than highly specialized nectar feeders
(Choeronycteris) whose palate length reaches half the total skull length. These proportions
are, however, also influenced by body size: small species possess relatively larger
braincases and shorter jaw lengths, larger specimens relatively smaller braincases and
longer jaws, respectively.

By means of 17 measures, skull proportions as well as their allometric conditions
(gradient of their reduced major axes, integration constants) were compared among the
genera examined (total number of specimens: 265).

98

As a reference quantity for individual measures, a calculated volume measure - SIZE -
was chosen, representing the braincase.

By allometrical comparison of the measures related to feeding apparatus (width of the
palate over the canini, palate length, upper tooth row, mandible length, height of the
coronoid process) the following points turned out:

Generally, the various development of rostrum prolongation cannot be characterized by
distinctly distinguishable construction principles. It is, rather, affected by the allometric
and integration constants which are established for the entire group.

In contrast, in the genera Choeroniscus, Hylonycteris and Choeronycteris the length of the
palate shows a transposition of the allometric line deviating from the condition of the
remaining Glossophaginae. This represents another principle of construction proving the
close relationship of all three genera. The systematic separation of the subfamily -
formerly summarized in a single taxon - into Glossophaginae and Lonchophyllinae
according to morphological criteria of the soft parts (Griffiths 1983) is also supported by
the cranial data of this study (lacking zygomatic arches, incisors, mandibular shape).
Furthermore, the allometric data allow to classify Platalina genovensium as a large species
within the the genus Lonchophylla.

LITERATURE CITED

Allen, H. (1898): On the Glossophaginae. — Trans. Amer. Philos. Soc. 19: 237-266.

Allen, J.A., & F.M. Chapman (1893): On a collection of mammals from the
island of Trinidad, with description of new species. — Bull. Amer. Mus. Nat. Hist.
13:203-234.

Allen, G.M. (1908): Notes on Chiroptera. — Bull. Mus. Comp. Zool., Harvard Coll.
52:25-62.

— (1942): Hylonycteris underwoodi in Mexico. — J. Mamm. 23:97.

Alvarez, T. & L. Gonzalez Q (1970): Analisis polinico del contenido gastrico de
murcielagos Glossophaginae de México. — An. Escuela Nac. Cienc. Biol., Mexico
18:137-165.

Arata, A.A., J.B. Vaughn & M.E. Thomas (1967): Food Habits of certain
Colombian Bats. — J. Mamm. 48:653-655.

Arroyo-Cabrales, J., RR. Hollander & I. Knox) Jones Ara Im:
Choeronycteris mexicana. — Mammalian Species 291:1-5.

Ayala, S.C. & A. D’Allessandro (1973): Insect feeding behavior of some
Colombian fruit-eating bats. — J. Mamm. 54:266-267.

Baker, H.G. (1961): The adaptation of flowering plants to nocturnal and crepuscular
pollinators. — Quart. Rev. Biol., Washington 36:64-73.

Baker, R.J. (1967): Karyotypes of the Familiy Phyllostomatidae and their taxonomic
implications. — Southwest. Nat. 12:407-428.

Baker, R.J., J.K. Jones, jr. & D.C. Carter (Eds, 1976): Biology of Bats of the
New World Family Phyllostomatidae. Part I. — Spec. Publ., Mus. Texas Tech. Univ.
LOP2TS pp.

—, — & — (Eds, 1977): Biology of Bats of the New World Family Phyllostomatidae. Part
II. — Spec. Publ., Mus. Texas Tech. Univ. 13, 364pp.

99

Balken RZ TIER Jones, jr c& DC. Carter (Eds; 1979): Biology of Bats of the
New World Family Phyllostomatidae. Part III. — Spec. Publ., Mus. Texas Tech. Univ.
16, 441pp.

Baker, R.J., P.V. August & A.A. Steuter (1978): Erophylla sezekorni. —
Mammalian Species 115:1-5.

Baker, R.J. & R.A. Bass (1979): Evolutionary Relationship of the Brachyphyllinae
to the Glossophagine Genera Glossophaga and Monophyllus. — J. Mamm. 60:364-372.

Bear ER a Honeycutt, Mal. Arnold, 5V¥.M. Sarich & HH.
Genoways (1981): Electrophoretic and immunological studies on the relationship of
the Brachyphyllinae and the Glossophaginae. — J. Mamm. 62:665-672.

Barbour, R.W., & W.H. Davis (1969): Bats of America. — Univ. Press Kentucky,
Lexington. 286 pp.

Baron, G., & P. Jolicoeur (1980): Brain structure in Chiroptera: some multivariate
trends. — Evolution 34:386-393.

Beatty, H.A. (1944): Fauna of St.Croix. — Virgin Islands J. Agric., Univ. Puerto Rico
28:181-185.

Bonaccorso, F.J. (1979): Foraging and reproductive ecology in a Panamanian bat
community. — Bull. Florida State Mus., Biol. Sci. 24:359-408.

Brown, J.H. (1968): Actvity patterns of some Neotropical Bats. — J. Mamm.
49:754-757.

Buden, D.W. (1975): Monophyllus redmani Leach (Chiroptera) from the Bahamas,
with notes on variation in the species. — J. Mamm. 56:369-377.

— (1976): A review of the bats of the endemic West Indian genus Erophylla. — Proc. Biol.
Soc. Washington 89:1-16.

— (1977): First records of bats of the Genus Brachyphylla from the Caicos Islands with
notes on geographic variation. — J. Mamm. 58:221-225.

Carter, D.C. (1968): A new species of Anoura (Mammalia, Chiroptera,
Phyllostomatidae) from South America. — Proc. Biol. Soc. Washington 81:427-430.

Carter, D.C., R.H. Pine & W.B. Davis (1966): Notes on Middle American bats.
— Southwest. Nat. 11:488-499.

Carvalho, C.T. (1961): Sobre os habitos alimentares de Phyllostomideos (Mammalia,
Chiroptera). — Rev. Biol. Trop. 9:53-60.

Choate, J.R., & E.C. Birney (1968): Sub-recent Insectivora and Chiroptera from
Puerto Rico, with the description of a new bat of the genus Stenoderma. — J. Mamm.
49: 400-412.

Cockrum, E.L., & B.J. Hayward (1962): Hummingbird bats. — Nat. Hist., New
York 71:38-43.

Dalquest, W.W. (1953): Mammals of the Mexican state of San Luis Potosi. —
Louisiana State Univ. Stud., Biol. Sci. Ser. 1:1-229.

Davis, W.B. (1944): Notes on Mexican Mammals. — J. Mamm. 25:370-403.

— (1974): The Mammals of Texas. — Bull. Texas Parks Wildl. Dept. 41:1-294.

Davis, W.B., & D.C. Carter (1962): Review of the Genus Leptonycteris
(Mammalia, Chiroptera). — Proc. Biol. Soc. Washington 75:193-198.

100

Davis, W.B., & J.B. Dixon (1976): Activity of bats in a small village clearing near
Iquitos, Peru. — J. Mamm. 57:747-749.

Davis, W.B., & R.J. Russell (1954): Mammals of the Mexican State of Morelos.
— J. Mamm. 35:63-80.

Dobat, K., & T. Peikert-Holle (1985): Blüten und Fledermäuse. —
Senckenberg-Buch 60. 370pp. Waldemar Kramer, Frankfurt/M.

Duke, J.A. (1967): Mammal dietary. Bioenvironmental and radiological-safety
feasibility studies, Atlantic-Pacific interoceanic canal. — Batelle Mem. Inst., Suppl.
33pp.

Easterla, D.A. (1972): Status of Leptonycteris nivalis (Phyllostomatidae) in Big Bend
National Park, Texas. — Southwest. Nat. :287-292.

— (1973): Ecology of the 18 species of Chiroptera at Big Bend National Park, Texas. —
Northwest Missouri State Univ. Stud. 34:1-165.

Erkert, H.G., & S. Kracht (1978): Evidence for ecological adaptation of circadian
systems. Circadian activity rhythms of neo-tropical bats and their re-entrainment after
phase shifts of the Zeitgeber-LD. — Oecologica 32:71-78.

Fleming, T.H., E.T. Hooper & D.E. Wilson (1972): Three Central American
bat communities: structure, reproductive cycles, and movement patterns. — Ecology
53:655-670.

Forman, G.L. (1971): Gastric morphology in selected Mormoopid and Glossophagine
bats. — Trans. Illinois Acad. Sci. 64:273-284.

Forman, G.L., C.J. Phillips & C.S. Rouk (1979): Alimentary tract: 205-227.
In: Baker, R.J., J. Knox Jones & D.C. Carter (Eds): Biology of the bats of the New
World Family Phyllostomatidae, Part III. — Spec. Publ., Mus. Texas Tech. Univ. 16,
441 pp.

Freeman, P.W. (1984): Functional cranial analysis of large animalivorous bats. —
Biol. J. Linnean Soc. 21:387-408.

— (1995): Nectarivorous feeding mechanisms in bats. — Biol. J. Linnean Soc. 56:439-464.

Gardner, A.L. (1962a): Bat records from the Mexican states of Colima and Nayarit.
— J. Mamm. 43:102-103.

— (1962b): A New Bat of the Genus Glossophaga from Mexico. — Contr. Sci., Los
Angeles County Mus. 54:1-7.

— (1977): Feeding habits: 293-350. In: Baker, R.J., J.K. Jones & D.C. Carter (Eds):
Biology of the bats of the New World Family Phyllostomatidae, Part II — Spec. Publ.,
Mus. Texas Tech. Univ. 13, 364pp.

— (1986): The taxonomic status of Glossophaga morenoi Martinez & Villa, 1938
(Mammalia: Chiroptera: Phyllostomidae). — Proc. Biol. Soc. Washington 99:489-492.

Geoffroy St. Hilaire, E. (1818): Sur de nouvelles chauve-souris, sous de nom de
Glossophages. — Mém. Mus. hist. Nat., Paris 4:411-418.

Genoways, H.H., RJ. Baker & W.B. Wyatt (1973): Nongeographic variation in the
long nosed bat Choeroniscus intermedius. — Bull. South. Calif. Acad. Los Angeles 72:
106-107.

Gerber, J.D., & C.A. Leone (1971): Immunologic comparisons of the sera of
certain phyllostomatid bats. — Syst. Zool. 20:160-166.

101

Goodwin, G.G. (1946): Mammals of Costa Rica. — Bull. Amer. Mus. nat. Hist.
87:273-473.

Goodwin, G.G., & A.M. Greenhall (1961): Review of the bats of Trinidad and
Tobago. — Bull. Amer. Mus. Nat. Hist. 122:187-302.

— & — (1964): New records of bats from Trinidad with comments on the status of
Molossus trinitatus Goodwin. — Amer. Mus. Nov. 2195:1-23.

Graham, G.L. (1988): Interspecific associations among Peruvian bats at diurnal roosts
and roost sites. — J. Mamm. 69: 711-720.

Gray, J.E. (1833): Characters of a new genus of bats (Brachyphylla) obtained by the
Society from the collection of the late Rev. Lansdown Guilding. — Proc. Zool. Soc.
London 1833:122-123.

— (1838): A description of the bats (Vespertilionidae) and the description of some new
genera and species. — Mag. Zool. Bot. 2:483-505.

— (1844): Mammalia: 7-36. In: The zoology of the voyage of H.M.S Sulfur under the
command of Captain Sir Edward Belcher...during the years 1836-42 (R.B. Hinds, ed.);
Smith, Elder & Co., London.

— (1866): Revision of the genera of Phyllostomidae, or leafnosed bats. — Proc. Zool. Soc.
London 1866:111-118.

Gundlach, J.C. (1861): A new genus of Chiroptera of Cuba, Phyllonycteris, described
by Dr. Gundlach. — Aus. Monats. Ko. Akad. Wiss. Berlin: 817-819.

Greenbaum, I.F., & C.J. Phillips (1974): Comparative anatomy and general
histology of tongues of long nosed bats (Leptonycteris sanborni and L. nivalis) with
reference to infestation of oral mits. — J. Mamm. 55:489-504.

Griffin, D.R., & A. Novick (1955): Acoustic orientation of Neotropical bats. — J.
Exp. Zool. 130:151-199.

Griffiths, T.A. (1978): Muscular and vascular adaptations for nectar-feeding in the
glossophagine bats Monophyllus and Glossophaga. — J. Mamm. 59:414-418.

— (1982): Systematics of the New World nectar-feeding bats (Mammalia, Phyllostomidae)
based on the morphology of the hyoid and lingual regions. — Am. Mus. Nov.
2742:1-45.

— (1983): On the phylogeny of the Glossophaginae and the proper use of outgroup
analysis. — Syst. Zool. 32:283-285.

Haiduk, M.W., & R.J. Baker (1982): Cladistical analysis of G-banded
chromosomes of nectar feeding bats (Glossophaginae: Phyllostomidae). — Syst. Zool.
31:252-265.

— & — (1984): Scientific method, opinion, phylogenetic reconstruction, and nectar-feeding
bats: a response to Griffiths and Warner. — Syst. Zool. 33:343-350.

Hall, E.R. (1981): The Mammals of North America. Wiley, New York, Chichester,
Brisbane & Toronto.

Hall, E.R., & W.W. Dalquest (1963): The mammals of Veracruz. — Univ. Kansas
Publ., Mus. nat. Hist. 14:165-362.

Hall, E.R., & K.R. Kelson (1959): The mammals of North America. 1,
XXX+546+79 pp.; New York.

Handley, C.O., jr. (1960): Descriptions of new bats from Panama. — Proc. U.S. Natl.

102

Mus. 112:459-479.

Handley, C.O., jr. (1966): Description of new bats (Choeroniscus and Rhinophylla)
from Colombia. — Proc. Biol. Soc. Washington, DC 79:83-88.

Handley, C.O., jr., & W.D. Webster (1987): The supposed occurence of
Glossophaga longirostris Miller on Dominica and problems with the type series of
Glossophaga rostrata Miller. — Occ. Pap., Mus. Texas Tech. Univ. 108:1-10.

Harris, B.J. (1959): Pollination of flowers by bats in the tropics. — Abstracts IX. Int.
Bot. Congr. 2 (Montreal): 151-152.

Hartley, D.J., & R.A. Suthers (1987): The sound emission pattern and the
acoustical role of the noseleaf in the echolocating bat, Carollia perspicillata. — J.
Acoust. Soc. Am. 82:1892-1900.

Heithaus, E.R., et al. (1975): Foraging patterns and resource utilization in seven
species of bats in a seasonal tropical rain forest. — Ecology 56:841-854.

Helversen, D..v., & ©. v. Helversen, (1975): Glossophacamsonianna
(Phyllostomatidae) - Nahrungsaufnahme. — Publ. Wiss. Film E 1837: 3-10.

Helversen, O. v., & H.V. Reyer (1984): Nectar intake and energy expenditure in
a flower visiting bat. — Oecologie 63:178-184.

Helversen, O. v. (1993): Adaptations of Flowers to the Pollination by Glossophagine
Bats. In: Animal-Plant Interactions in Tropical Environments (eds. W. Barthlott et al.).
Museum Koenig, Bonn, 1993.

Hill, J.E. (1980): A note on Lonchophylla (Chiroptera, Phyllostomatidae) from Ecuador
and Peru, with the description of a new.species. — Bull. Brit. Mus. Nat. Hist. (Zool.)
38:233-236.

— (1985): The status of Lichonycteris degener Miller, 1931 (Chiroptera: Phyllostomidae).
— Mammalia 49:579-582.

Hoffmeister, D.F. (1957): Review of the long-nosed bats of the genus Leptonycteris.
— J. Mamm. 38:454-461.

— (1986): Mammals of Arizona. Univ. Arizona Press, XX+602 pp.

Homan, J.A., & J.K. Jones, jr. (1975): Monophyllus plethodon. — Mammalian
Species 58:1-2.

— & — (1975): Monophyllus redmani. — Mammalian Species 57:1-3.

Hood, C.S., & J.D. Smith (1982): Cladistical analysis of female reproductive
histomorphology in phyllostomatoid bats. — Syst. Zool. 31:241-251.

Howell, D.J. (1974a): Acoustic behavior and feeding in Glossophagine bats. — J.
Mamm. 55:293-308.

— (1974b): Bats and pollen: physiological aspects of the syndrome of chiropterophily. —
Comp. Biochem. Physiol.(A) 48:263-276.

— (1979): Flock foraging in nectarfeeding bats: Advantages to the bats and to the host
plants. — Amer. Natural. 114:23-49.

Howell, D.J., & D. Burch (1974): Food habits of some Costa Rican bats. — Rev.
Biol. Trop. 21:281-294.

Howell, D.J., & N. Hodgkin (1976): Feeding adaptations in the hairs and tongues
of nectar-feeding bats. — J. Morph. 148:329-336.

Hsu, T.C., R.J. Baker et al. (1968): The multipe sex chromosome system of

103

American leaf-nosed bats (Chiroptera, Phyllostomatidae). — Cytogenetics 7:27-38.

Humphrey, S.R., & F.J. Bonaccorso (1979): Population and community ecology
(409-441). In: R.J. Baker, J.K. Jones, jr., & D.C. Carter (eds.): Biology of the bats of
the New World Family Phyllostomatidae, Part III. — Spec. Publ., Texas Tech. Univ.
16.

Husson, A.M. (1962): The bats of Suriname. — Zool. Verh. 58, 282+XXX pp.

Jones, J.K. (1964): Bats from Western and Southern Mexico. — Transac. Kansas Acad.
Sci. 67:509-516.

Koepcke, J. (1987): Okologische Studien an einer Fledermaus-Artengemeinschaft im
tropischen Regenwald von Peru. Dissertation, Miinchen; 439 pp.

Koopman, K.F. (1981): The distributional patterns of New World nectar-feeding bats.
— Ann. Missouri Bot. Gard. 68:352-369.

— (1983): Biogeography of the bats of South America. Pymat. Press, New York.

Koopman, K.E., M.K. Hecht & E. Ledecky-Janecek (1957): Notes on the
mammals of the Bahamas with special reference to the bats. — J. Mamm. 38:164-174.

Koopman, K.F., & G.T. McIntyre (1980): Phylogenetic analysis of chiropteran
dentition. — Proc. 5th Intl. Bat Res. Conference: 279-288.

Kunz, T.H. (1982): Ecology of Bats. Plenum Press, New York, London: 425 pp.

LaVal, R.L. (1977): Banding returns and activity periods of some Costa Rican bats. —
Southwest. Nat. 15:1-10.

LaVal, R.L., & H.S. Fitch (1977): Structure, movements and reproduction in three
Costa Rican bat communities. — Occ. Pap., Mus. Nat. Hist. Univ. Kansas 69:1-28.
Leach, W.E. (1821): The characters of seven genera of bats with foliaceus appendages

to the nose. — Trans. Linn. Soc. London 13:73-82.

ecrmulkcorn Once TER Damisitt(1979):: Anoura ‘culitrata’ (Chiroptera:
Phyllostomidae) from Colombia. — Mammalia 43:579-581.

Linares, O.J. (1967): Extension de distribucion para Lonchophylla robusta con algunas
notas sobre las espécies Venezuelanos del genero Lonchophylla. — Bol. Soc. Venez.
Espeleol. 1:53-59.

Lydekker, R. (1891): In: Flower, W.H. & R. Lydekker: An Introduction to the study
of mammals living and extinct. Adam & Charles Black, London, 763pp.

Marinkelle, C.J., & A. Cadena (1972): Notes on Bats new to the Fauna of
Colombia. — Mammalia 36:50-58.

Martinez, L., & B. Villa-R (1938): Contribuciönes al conocimiento de los
murcielagos de México. — An. Inst. Biol., Mexico 9:339-360.

Mc Daniel, V.R. (1976): Brain anatomy (147-200). In: Baker, R.J., J.K. Jones, jr., &
D.C. Carter (eds.): Biology of bats of the New World family Phyllostomatidae, Part I.
— Spec. Publ., Mus., Texas Tech. Univ. 10.

Mc Nab, B.K. (1971): The structure of tropical bat faunas. — Ecology 52:352-358.

Miller, G.S. (1898): Descriptions of five new phyllostome bats. — Proc. Acad. Nat.
Sci. Philadelphia 50:326-337.

— (1899): Two new glossophagine bats from the West Indies. — Proc. Biol. Soc.
Washington 8:33-37.

104

Miller, G.S. (1900): The bats of the genus Monophyllus. — Proc. Washington Acad.
Sci. 2:31-38.

— (1902): Twenty new American bats. — Proc. Acad. Nat. Sci. Philadelphia 54:389-412.

— (1906): Twelve new genera of bats. — Proc. Biol. Soc. Washington 19:83-86.

— (1907): The families and genera of bats. — Bull. U.S. Nat. Mus. 57:1-282.

— (1913): Five new mammals from tropical America. — Proc. Biol. Soc. Washington
26:31-34.

— (1918): Three new bats from Haiti and Santo Domingo. — Proc. Biol. Soc. Washington
31:39-40.

Moller, W. (1932): Das Epithel der Speiseröhrenschleimhaut der blütenbesuchenden
Fledermaus Glossophaga soricina im Vergleich zu insektenfressenden Chiropteren. —
Z. mikr.-anat. Forsch. 29:637-653.

Nagorsen, D., & J.R. Tamsitt (1981): Systematics of Anoura cultrata, A.
brevirostrum and A. werckleae. — J. Mamm. 62:82-100.

Nellis, D.W. (1971): Additions to the natural history of Brachyphylla (Chiroptera). —
Carribean J. Sci. 11:91.

Nellis, D.W., & C.P. Ehle (1977): Observations on the behavior of Brachyphylla
cavernarum (Chiroptera) in Virgin Islands. — Mammalia 41:403-409.

Nowak, R.M., & J.L. Paradiso (1983): Walker’s Mammals of the World. 4th
Edition, 1983; Vol.I, Johns Hopkins University Press, Baltimore & London, LXI+568
PP-

Osburn, W. (1865): Notes on the Cheiroptera of Jamaica. — Proc. Zool. Soc.
1865:61-85.

Pallas, P.S. (1766): Miscellanea zoologica. Hagae Comitum, xii + 224pp.

Park, H., & E.R. Hall (1951): The gross anatomy of the tongues and stomachs of
eight New World bats. — Trans. Kansas Acad. Sci. 54:64-72.

Patton, J.L., & A.L. Gardner (1971): Parallel’ evolufonozmnkıple
sex-chromosome systems in the phyllostomatid bats, Carollia and Choeroniscus. —
Experentia 27:105-106.

Peters, W. (1869): Uber die zu den Glossophagae gehörigen Flederthiere und über eine
neue Art der Gattung Coleura. — Monatsber. preuss. Akad. Wiss.:361-368.

Pfrimmer Hensley, A., & K.T. Wilkins (1988): Leptonycteris nivalis. —
Mammalian Species 307: 1-4.

Phillips, C.J. (1971): The dentition of glossophagine bats: development,
morphological characteristics, variation, pathology and evolution. — Misc. Publ. Mus.
Nat. Hist. Univ. Kansas 54:1-138.

Phillips, C.J., & J.K. Jones, jr. (1971): A new subspecies of the Long-nosed
Bat, Hylonycteris underwoodi, from Mexico. — J. Mamm. 52:77-80.

Pijl, L. van der (1936): Fledermäuse und Blumen. — Flora 31:1-40 N.F., Jena
1936/37.

— (1957): The dispersal of plants by bats. — Acta Bot. Neerl. 6:291-315.

Pine, R.H. (1972): The bats of the genus Carollia. — Techn. Monogr. Texas agric.
Exp, Stn6821-125,

Rasweiler, J.J. (1977): The care and management of bats as laboratory animals: 519-

105

617. In: W.A. Wimsatt (ed.): Biology of bats, Vol.3. Academic Press, New York, 651
pp.

Reis, N.R. dos (1981): Estudo ecolögico dos quiröpteros de matas primärias e
capoeiras da regiao de Manaus, Amazonas. — Tese-doutorado, 242 pp.; Manaus.
Rempe, U. (1962): Uber einige statistische Hilfsmittel moderner

zoologisch-systematischer Untersuchungen. — Zool. Anzeiger 169:93-140.

Ross, A. (1961): Notes on Food Habits of Bats. — J. Mamm. 42:66-71.

Rouk, C.S., & B.P. Glass (1970): Comparative Gastric Histology of five North and
Central American Bats. — J. Mamm. 51:455-472.

Sarnlkeyzs Col.) MA. Horner & T.Hr Fleming (1993): Flight speeds and
mechanical power outputs of the Nectar-Feeding Bat, Leptonycteris curasoae
(Phyllostomidae: Glossophaginae). — J. Mamm. 74:594-600.

Sanborn, C.C. (1933): Bats of the genera Anoura and Lonchoglossa. — Field. Mus.
Publ. Zool. Chicago 20:23-27.

— (1943): External characters of the bats of the subfamily Glossophaginae. — Field Mus.
Nat. Hist. Chicago Zool. Ser. 24:271-277.

— (1954): Bats from Chimanta-Tepui, Venezuela, with remarks on Choeroniscus. —
Fieldiana Zoology 34:289-293.

Saussure, M.H. (1860): Note sur quelques mammiféres du Mexique. — Rev. Mag.
Zool. Paris, sér. 2,12:479-494.

Sazima, M., & I. Sazima (1978): Bat pollination of the passion flower, Passiflora
mucronata, in southeastern Brazil. — Biotropica 10:100-109.

Salman D Vizotto & V.A. Taddeıi (1978); Una nova especie de
Lonchophylla da Serra do Cipo, Minas Gerais, Brasil. — Rev. Bras. Biol. 38:81-89.

Schaldach, W.J., & C.A. Mc Laughlin (1960): A new genus and species of
Glossophagine bat from Colima, Mexico. — Contr. Sci., Los Angeles County Mus.
37:1-8.

Schwartz, A., & J.K. Jones, jr (1967): Review of bats of the endemic Antillean
genus Monophyllus. — Proc. U.S. Natl. Mus. 124(3635):1-20.

Silva Taboada, G. (1976): Historia y Actualizacion Taxonomica de algunas Especies
Antillanas de Murciélagos. — Poeyana 153:1-24.

Silva Taboada, G., & R.H. Pine (1969): Morphological and behavioral evidence
for the relationship between the bat genus Brachyphylla and the Phyllonycterinae. —
Biotropica 1:10-19.

Smith, J.D. (1976): Chiropteran evolution (49-69). In: R.J. Baker, J.K. Jones, Jr &
D.C. Carter (eds): Biology of bats of the New World family Phyllostomatidae. Part I.
— Spec. Publ., Texas Tech. Univ. 10, 218pp.

iii DE CxS, Hood (1984)> Genealogy of the New World
nectar-feeding bats re-examined: A reply to Griffiths. — Syst.
Zool. 33:435-460.

Smith, J.D., & A. Starrett (1979): Morphometric analysis of chiropteran wings
(229-316). In: R.J. Baker, J.K. Jones, jr. & D.C. Carter (eds): Biology of Bats of the
New World Family Phyllostomatidae. Part III. — Spec. Publ., Texas Tech. Univ. 16,
441 pp.

106

Starrett, A. (1969): A new species of Anoura (Chiroptera, Phyllostomatidae) from
Costa Rica. — Contr. Sci., Los Angeles County Mus. 157:1-9.

Stock, D. (1975): Chromosome banding pattern homology and its phylogenetic
implications in the bat genera Carollia and Choeroniscus. — Cytogenet. Cell Genet.
14:34-41.

Strickler, T.L. (1978): Functional osteology and myology of the shoulder in the
Chiroptera. — Contr. Vert. Evol. 4:1-198.

Studholme, K.M., & C.J. Phillips (1987): Interspecific comparisons of
immunohistochemical localization of retinal neurotransmitters in four species of bats.
— Brain Behav. Evol. 30:160-173.

Suthers, R.A. (1970) Vision, -olfaction. taste 269.309 SB
Wimsatt (ed.): Biology of bats. Vol.2. Academic Press, New York
& London, 477 pp.

Swanepoel, P., & H.H. Genoways (1983a): Brachyphylla cavernarum. —
Mammalian Species 205:1-6.

— & — (1983b): Brachyphylla nana. — Mammalian Species 206:1-3.

Taddei, V.A., L.D. Vizotto & 1, Sazima (1983): Uma niovares pieemendc
Lonchophylla do Brasil e chave para identificacao daszespeeies
do genero (Chiroptera, Phyllostomatidae). — Cilene tag Om liastzre
Banılo235262526297

Tamsitt, J.R., & D. Nagorsen (1982): Anoura cultrata. — Mammalian Species
179123,

Tamsitt, J.R., & D. Valdivieso (1966): Bats from Colombia in the Swedish
Museum of Natural History, Stockholm. — Mammalia 30:97-104.

Thomas, O. (1913): A new genus of Glossophagine bat from Colombia. — Ann. Mag.
Nat. Hist. London Ser.8,12:270-271.

— (1928): A new genus and species of glossophagine bat, with a subdivision of the genus
Choeronycteris. — Ann. Mag. Nat. Hist. London Ser. 10,1:120-123.

Tschudi, J.J. von (1844): Therologie (p.1-6). In: Untersuchungen über die Fauna
Peruana, St. Gallen. Teil 1: 262 pp.

Tuttle, M.D. (1970): Distribution and zoogeography of Peruvian bats, with comments
on natural history. — Univ. Kansas Sci. Bull. 49:45-86.

— (1976): Collecting techniques (71-88). In: R.J. Baker, J.K. Jones, jr. & D.C. Carter
(Eds.): Biology of bats of the New World family Phyllostomatidae. Part I. — Spec.
Publ., Texas Tech. Univ. 10, 218pp.

Valdivieso, D.J. (1964): La fauna quiroptera del Departamento de Cundinamarca,
Colombia. — Rev. Biol. Trop. 12:19-45.

Valdivieso, D.J., & J.R. Tamsitt (1971): Hematological data from tropical
American bats. — Can. J. Zool. 49:31-36.

Villa-R, B. (1967): Los murciélagos de Mexico. — Inst. Biol. Univ. Nac. Autonoma
Mexico, 491 pp.

Vogel, S. (1958): Fledermausblumen und Blumenfledermäuse im tropischen Amerika.
— Umschau 58:761-763.

Walton, D.W. (1963): A collection of the Bat Lonchophylla robusta Miller from Costa

107

Rica. — Tulane Stud. Zool. 10:87-90.

Walton, D.W., & G.M. Walton (1968): Comparative osteology of the pelvic and
pectoral girdles of the Phyllostomatidae (Chiroptera, Mammalia). — J. Grad. Res. Cent.
South. Methodist Univ. 37:1-35.

Warner, R.M. (1983): Karyotypic megaevolution and phylogenetic analysis: New
World nectar-feeding bats. — Rev. Syst. Zool. 32:279-282.

MacbistemuWoD.. «& € ©. Handley, jr. (i986): Systematics of Millers
Long-Tongued Bat, Glossophaga longirostris with description of two new subspecies.
— Occ. Pap., Mus. Texas Tech. Univ. 100:1-26.

Webster, W.D., & J.K. Jones, jr. (1980): Taxonomic and nomenclatorial notes
on bats of the genus Glossophaga in North America, with description of a new
species. — Occ. Pap., Mus. Texas Tech. Univ. 71:1-12.

— & — (1982): A new subspecies of Glossophaga commissarisi (Chiroptera:
Phyllostomidae) from Western Mexico. — Occ. Pap., Mus. Texas Tech. Univ. 76:1-6.

— & — (1984): Glossophaga leachii. — Mammalian Species 226:1-3.

— & — (1985): Glossophaga mexicana. — Mammalian Species 245:1-2.

— & — (1987): A new subspecies of Glossophaga commissarisi (Chiroptera:
Phyllostomidae) from South America. — Occ. Pap., Mus. Texas Tech. Univ. 109:1-6.

MWiebster wD. Ll WeeRobbins, ROE Robbins & RJ. Baker (1982):
Comments on the status of Musonycteris harrisoni (Chiroptera: Phyllostomidae). —
Occ. Pap., Mus. Texas Tech. Univ. 78:1-5.

Wille, A. (1954): Muscular adaption of the nectareating bats (subfamily
Glossophaginae). — Trans. Kansas Acad. Sci. 57:315-325.

Wilson, D.E. (1973): Bat faunas: A trophic comparision. — Syst.Zool. 22:14-29.

Author's address

Dr. Ernst-Hermann Solmsen

c/o Zoologisches Institut und

Zoologisches Museum der Universitat Hamburg
Martin-Luther-King-Platz 3

20146 Hamburg

Germany

108

APPENDIX
Material examined

Phyllostomatinae
Phyllosstomus discolor:

BMNH 5.5.22.1, -, Ecuador

Phyllostomus elongatus:

MNHUB 3217, ?, Suriname, leg. A. Kappler
MNHUB 3359, &, Suriname, leg. A. Kappler
MNHUB 3985, 5, Suriname, leg. A. Kappler
MNHUB 3985(a), 7, Suriname, leg. A. Kappler
MNHUB 4185, 5, Berg an Dal, Suriname, leg. Mösche

Phyllostomus hastatus:

BMNH 14.9.1.17, ?, Sierra de Carabobo, Venezuela
MNHUB "--”, <, St. Pablo, leg. G. Hopke, 6.3.1897
MNHUB 59, -, Bolivia, leg. Steinbach A.11.09
MNHUB 158, -, Bolivia, leg. Steinbach A.11.09
MNHUB 2592, °, Brazil, Parreys

MNHUB 10025, -, Brazil, leg. Posnansky

MNHUB 37387, -, Para, Brazil, leg. Otto Bertram
SME 25475, no data

Carolliinae

Carollia castanea:

SME 41994, 2, Villavicencio, Colombia, leg. H. Stephan, 1971

SMF 41995, 7, Villavicencio, Colombia, leg. H. Stephan, 1971

SMF 43027, ©, Finca el Buque, Villavicencio, Colombia, leg. E. Thiery, 1979
SMF 45690, °, Lagao del Calado, Manaus, Brazil, leg. F. Reiss, 1971

SMF 54902, 9, Apulo, Cundinamarca, Colombia, leg. E. Patzelt, 1982

ZIM (SO) 5, 2, Rio Cuyabeno, Ecuador, leg. E. Patzelt, 1982

Carollia perspicillata:

MEPN 1106, ©, Rio Pastaza, Ecuador, leg Spillmann

MEPN 1176, -, leg Spillmann

MEPN 3024, ©, Cerro Guataraco, Ecuador, leg. Spillmann, 24.11.1930
MEPN 3284, 0, Sto. Domingo de los Colorados, Ecuador, leg. Spillmann
MEPN 3285, °, Esmeraldas, Ecuador, leg. Spillmann

MEPN 3292, ?, Cerro Guataraco, Ecuador, leg. Spillmann, 12.9.1932
MEPN 8027, -, San Lorenzo, Esmeraldas, Ecuador, leg. L. Albuja, 16./17.2.1980
MEPN 8142, ?, Rio Palenque, Ecuador, leg. L. Albuja, 28.4.1981

MEPN 8144, -, Rio Palenque, Ecuador, leg. L. Albuja, 28.4.1981

MEPN 8147, -, Rio Palenque, Ecuador, leg. L. Albuja, 29.10.1981
MEPN 33101, <, Avila, Mangayacu, Ecuador, leg Spillmann, 1939

SMF 15655, °, Panama, leg. I. Eibl-Eibesfeld

SMF 15656, -, Panama, leg. I. Eibl-Eibesfeld

SMF 15657, -, Panama, leg. I. Eibl-Eibesfeld

SMF 18748, ?, Cueva de Ganango, Venezuela, 27.1.1952

SME 18751, no data

SMF 43022, =‘, Finca el Buque, Villavicencio, Meta, Colombia, leg. E. Thiery, 15.5.1979
SMF 43023, , Finca el Buque, Villavicencio, Meta, Colombia, leg. E. Thiery, 1979
SMF 43024, 2, Finca el Buque, Villavicencio, Meta, Colombia, leg. E. Thiery, 1979
SMF 43025, ©, Finca el Buque, Villavicencio, Meta, Colombia, leg. E. Thiery, 1979
SME 43026, ?, Finca el Buque, Villavicencio, Meta, Colombia, leg. E. Thiery, 15.5.1979

SMF 66381, ©, Hacienda la Pacifica Cabas, Guanacaste, Costa Rica, leg. Küsten & Joermann,

8.10.1931

Carollia subrufa:

AMNH 97516, -, Barillos, Guatemala

AMNH 235304,
AMNH 235716,
AMNH 235717,
AMNH 235721,
AMNH 235725,
AMNH 235726,

AMNH 243759,
AMNH 230500,

c', Santa Rosa, 6 km N Avellana, Guatemala
2°, Santa Rosa, Guatemala
7, Santa Rosa, Guatemala
7, Santa Rosa, Guatemala
JS, Santa Rosa, Guatemala
$, Santa Rosa, Guatemala

*, Tutiapa, El Paraiso, Guatemala
cd‘, Oxapampa, Peru, leg. M.G. Tuttle, 25.7.1954

Rhinophylla pumila:

AMNH 262470,

-, Aqua dulce, Pando, Bolivia, leg. A.W. Dickermann, 22.7.1986

Rhinophylla fischerae:

AMNH 230481,
AMNH 230482,
AMNH 230483,
AMNH 230496,
AMNH 230498,
AMNH 230499,
AMNH 230500,

Brachyphyllinae

JS, Dept. Pasco, Prov. Oxapampa, Peru, leg, D.L. Knowlton, 23.7.1964
os, Oxapmapa, Peru, leg. K.R. Stringer, 7.7.1964

2, Oxapmapa, Peru, leg. K.R. Stringer, 7.7.1964

?, Oxapampa, Peru, leg, D.L. Knowlton, 25.7.1964

JS, Oxapampa, Peru, leg, D.L. Knowlton, 25.7.1964

?, Oxapmapa, Peru, leg. K.R. Stringer, 22.7.1964

7, Oxapampa, Peru, leg. M.D. Tuttle, 25.7.1954

Brachyphylla cavernarum (= B. minor):

AMNH 149366,
AMNH 149367,
AMNH 213898,

cd‘, Barbados
c', Barbados
2, Barbados

BMNH 18.4.1.11, -, Antigua

Brachyphylla nana (= B. pumila):
AMNH 97597, ©, Dominican Republic
AMNH 214390, ©, Dominican Republic
AMNH 244914, 2, Dominican Republic
BMNH 52.588, °, Haiti

BMNH -, <, Cuba

ROM 45708, -, Los Patos, Dominican Republic

ZFMK 77.651, -, Cuba

109

110

Phyllonycterinae

Erophylla bombifrons:

AMNH 39339, <, Pueblo Viejo, Porto Rico, 5.7.1916

AMNH 39340, <, Pueblo Viejo, Porto Rico, 5.7.1916

AMNH 39341, 2, Pueblo Viejo, Porto Rico, 5.7.1916

AMNH 39345, 5, Pueblo Viejo, Porto Rico, 6.7.1916

ROM 42754, 2, Corozal, Puerto Rico

ROM 42755, °, Corozäl, Puerto Rico

ROM 44552, ©, Aguas Buenas, Puerto Rico

ROM 45709, ?, Fantina, Dominican Republic

ROM 45710, 2, Fantina, Dominican Republic

ROM 45711, ?, Fantina, Dominican Republic

USNM 252618, 7, San Michel, Haiti, leg. A.J. Poole, 11.3.1928
USNM 252619, ?, San Michel, Haiti, leg. A.J. Poole, 11.3.1928
USNM 253634, 2, San Michel, Haiti, leg. Poole & Perrygo, 23.12.1928

Erophylla sezekorni:

AMNH 41056, , Siboney, Cuba, 26.2.1917

AMNH 41057, , Siboney, Cuba, 26.2.1917

AMNH 41059, , Siboney, Cuba, 26.2.1917

AMNH 41066, ?, Siboney, Cuba, 26.2.1917

AMNH 41067, ¢, Siboney, Cuba, 26.2.1917

AMNH 41069, 7, Siboney, Cuba, 26.2.1917

ROM 63164, 7, Cueva de Gabairo, C39, Cuba

ROM 63165, 7, Cueva de Gabairo, Cuba

ROM 63166, S, Cueva de Gabairo, Cuba

USNM 538180, ©, Cayman Brac West End, 3.5km NE near South East Bay, Cayman Is., leg.
G.S. Morgan, 30.7.1979

USNM 538181, ©, Cayman Is., leg. G.S. Morgan, 30.7.1979

USNM 538182, 3, Cayman Is., leg. G.S. Morgan, 30.7.1979

USNM 538183, ?, Cayman Is., leg. G.S. Morgan, 30.7.1979

USNM 538184, 7, Cayman Is., leg. G.S. Morgan, 30.7.1979

USNM 538185, ?, Cayman Is., leg. G.S. Morgan, 30.7.1979

Phyllonycteris poeyi:

ROM 63170, 2, Cueva de los Majaes, Siboney, 14km C39, Cuba

ROM 63171, 2, Cueva de los Majaes, Siboney, 14km C39, Cuba

ROM 63172, 5, Cueva de los Majaes, Siboney, 14km C39, Cuba

ROM 37069, FEM

ROM 37070, M

ROM 37071, 2, St. Claire Cave, Ewarton, 2mls. S St. Catherine Parish, Jamaica
SMF 12139, ¢, Cuba, don. Berlin

ZFMK 77.649, Cuba, leg. Gundlach

Lonchophyllinae

Lionycteris spurrelli:

AMNH 97220, 9, Mocajuba, Rio Tocatins, Brazil, leg. Ollalla Bros., 15.11.1931
AMNH 97222, 5, Mocajuba, Rio Tocatins, Brazil, leg. Ollalla Bros., 15.11.1931
AMNH 97224, °, Mocajuba, Rio Tocatins, Brazil, leg. Lalla Bros., 16.11.1931
AMNH 97261, ©, Mocajuba, Rio Tocatins, Brazil

AMNH 230207, °, Peru

AMNH 230209, °, Peru

AMNH 239891, , Peru

AMNH 260004, °, Venezuela

BMNH 69.392, 5, Guayana

BMNH 13.8.10.1 (Type), , Condoto Choko, Colombia, leg. Spurrell

MHNG 1682.83, 7, Jumandi, 10km N Archidona, Napo, Ecuador, leg. J. Garzoni, 1982

Lonchophylla thomast:

AMNH 97272, ?, Brazil

AMNH 210688, ?, Bolivia

AMNH 230281, 2, Peru

AMNH 230282, 2, Peru

BMNH 65.633, 7, British Guayana

BMNH 69.1278, 5, Araenga, Para, Brazil

BMNH 69.1280, ?, Araenga, Para, Brazil

MEPN 2018, -, Sinolotor, Ecuador, leg. Spillmann

MEPN 80265, <, Urbina, Esmeraldas, Ecudor, leg. L. Albuja, 16.2.1980
JK 30, -, Peru, leg. J. Koepke

JK 66, -, Peru, leg. J. Koepke

JK 278, -, Peru, leg. J. Koepke

RMNH 30, <, leg. J.A. Allen, 1904

RMNH 69, <, leg. J.A. Allen, 1904

RMNH 17346, 2, Marowijne, Nassaugebergte, Suriname, leg. Surinam Exp., 25.2.1949
RMNH 17347, FEM. Marowijne, Nassaugebergte, Suriname, 25.2.1949

Lonchophylla mordax:

BMNH 3.9.5.31, 2, Lamarao, Bahia, leg. ©. Thomas

BMNH 3.9.5.32, ©, Lamarao, Bahia, leg. ©. Thomas

BMNH 3.9.5.33, 5, Lamaräo, Bahia, leg. ©. Thomas

BMNH 3.9.5.34, (Type), 2, Lamarao, Bahia, leg. ©. Thomas

BMNH 3.9.5.35, 7, Lamaräo, Bahia, leg. ©. Thomas

BMNH 3.9.5.36, -, Lamaräo, Bahia, leg. ©. Thomas

MEPN 7463, °, Boca del Rio Lito, Ecuador, leg. G. Herrera, 3.6.197?

Lonchophylla robusta:

BMNH 78.1354, ?, Yaupi, Morona, Prov. Santiago, Ecuador

BMNH 78.1355, °, Yaupi, Morona, Prov. Santiago, Ecuador

BMNH 78.1356, ?, Yaupi, Morona, Prov. Santiago, Ecuador

BMNH 78.1357, ?, Yaupi, Morona, Prov. Santiago, Ecuador

BMNH 78.1358, *, Yaupi, Morona, Prov. Santiago, Ecuador

BMNH 78.1359, ?, Yaupi, Morona, Prov. Santiago, Ecuador

BMNH 78.1360, 2°, Yaupi, Morona, Prov. Santiago, Ecuador

BMNH 78.1361, °, Yaupi, Morona, Prov. Santiago, Ecuador

BMNH 78.1362, ?, Yaupi, Morona, Prov. Santiago, Ecuador (2°93'S, 77°54'W)
MEPN 80214, ©, San Lorenzo, Prov. Esmeraldas, Ecuador, leg. L. Albuja
MEPN 80464, -, Barragantete, Ecuador, leg. L. Albuja

Lonchophylla handleyi:
BMNH 81.174, ?, Los Tayos, Morona, Prov. Santiago, Ecuador (3° 07'S, 78° 12'W), leg. J.E. Hill
BMNH 78.1363, 5, Los Tayos, Morona, Prov. Santiago, Ecuador, leg. J.E. Hill

i?

BMNH 78.1364, ©, Los Tayos, Morona, Prov. Santiago, Ecuador, leg. J.E. Hill
BMNH 78.1365, ?, Los Tayos, Morona, Prov. Santiago, Ecuador, leg. J.E. Hill
BMNH 78.1366, ©, Los Tayos, Morona, Prov. Santiago, Ecuador, leg. J.E. Hill
BMNH 78.1367, 2, Los Tayos, Morona, Prov. Santiago, Ecuador, leg. J.E. Hill
BMNH 78.1368, (Type), ?, Los Tayos, Morona, Prov. Santiago, Ecuador, leg. J.E. Hill
BMNH 78.1369, 5, Los Tayos, Morona, Prov. Santiago, Ecuador, leg. J.E. Hill
BMNH 78.1370, ©, Los Tayos, Morona, Prov. Santiago, Ecuador, leg. J.E. Hill
BMNH 78.1371, ?, Los Tayos, Morona, Prov. Santiago, Ecuador, leg. J.E. Hill
BMNH 78.1372, ©, Los Tayos, Morona, Prov. Santiago, Ecuador, leg. J.E. Hill
BMNH 78.1374, ?, Los Tayos, Morona, Prov. Santiago, Ecuador, leg. J.E. Hill
BMNH 78.1375, ?, Los Tayos, Morona, Prov. Santiago, Ecuador, leg. J.E. Hill
BMNH 78.1376, ?, Los Tayos, Morona, Prov. Santiago, Ecuador, leg. J.E. Hill
BMNH 78.1377, <, Yaupi, Morona, Ecuador

BMNH 78.1378, 7, Yaupi, Morona, Ecuador

Platalina genovensium:

AMNH 257108, 2, Peru

BMNH 27.11.1938 (Type), ~, Lima, Peru, leg. Esposto
NHMB 10.623, 2, Angolo, Peru, leg. W. Markl, 1957

Glossophaginae

Glossophaga commissarist:

BMNH 67.799, °, Bonanza, Nicaragua, leg. J. KnoxJones jr., 29.2.1964

BMNH 94.3.30.3, -, Nicaragua

SMF 11931, -, Marajo, Brazil, leg. W. Ehrhardt, 1.10.1923

SMF 12185, 2, Marajo, Brazil, leg. W. Ehrhardt, 16.11.1923

SMF 12192, 3, Marajo, Brazil, leg. W. Ehrhardt, 10.10.1923

SMF 13396, 2, Hacienda Chilata, Sonsonate, El Salvador, leg. H. Felten, 28.6.1953

MEPN 79852, -, Limoncocha, Napo, Ecuador, leg. L. Albuja, 8.1979

MHNG 1682.88, ©, Momotombo, cite geothermique, Nicaragua, leg. Chambrier-Jaccoud, 4.2.1983

Glossophaga longirostris:

BMNH 11.5.25.83, ©, San Estevan, Venezuela, leg. S.M. Klages
BMNH 11.5.25.84, 0, San Estevan, Venezuela, leg. S.M. Klages
BMNH 11.5.25.85, <, San Estevan, Venezuela, leg. S.M. Klages
BMNH 11.5.25.86, °, San Estevan, Venezuela, leg. S.M. Klages
BMNH 11.5.25.87, 2, San Estevan, Venezuela, leg. S.M. Klages
BMNH 27.11.19.25, 2, Curacao

BMNH 52.8.12.11, ?, St. Vincent

BMNH 91.5.15.8, 5, Mustique

BMNH 96.11.8.5, ?, Grenada

MEPN 3294, -, Esmeraldas, Ecuador, leg. Spillmann

MEPN 3295, -, Esmeraldas, Ecuador, leg. Spillmann

MEPN 3296, -, Esmeraldas, Ecuador, leg. Spillmann

MEPN 3297, -, Esmeraldas, Ecuador, leg. Spillmann

MEPN 80417, S, Borraganete, Ecuador, leg. L. Albuja

MEPN 80419, -, Borraganete, Ecuador, leg. L. Albuja, 1.4.1980
MHNG 1057.24, ?, Girardot, Cundinamarca, Columbia, leg. Valdivieso, 1961
MHNG 1682.90, 2, Tolu, Bolivar, Columbia, leg. Mechler, 16.4.1965
RMNH 12227, °, Tobago, Trinidad, leg. G.F. Mees, 13.2.1954

RMNH 12228, ?, Tobago, Trinidad, leg. G.F. Mees, 22.3.1954

RMNH 12239, °, Grafton, Tobago, Trinidad, leg. G.F. Mees, 1954
RMNH 12242, 3, Grafton, Tobago, Trinidad, leg. G.F. Mees, 13.2.1954
RMNH 14359, °, Los Testigos, leg. P.W. Hummelinck, 1936

RMNH 14388, 0, Aruba, leg. P.W. Hummelinck, 4.10.1945

RMNH 14389, %, Curacao, leg. P.W. Hummelinck, 26.9.1948

RMNH (14390), 2, Curacao, leg. P.W. Hummelinck, 26.9.1948

RMNH 14722, -, Aruba, 1.10.1945, don. Van Pijl, 1945
RMNH 14724, ?, Curacao, 10.1.1946

Glossophaga leachi ("G. alticola”):
BMNH 67.800, 5, Managua, Nicaragua, leg. A.A. Alcom, 14.6.1956

SMF 13457, 2, Finca Raquelina, Ahuachapan, El Salvador, leg. H. Felten, 2.7.1953

SMF 15078, 5, Hacienda San Antonio, Sonsonate, leg. H. Felten, 12.11.1953

Glossophaga morenoi (= mexicana):
AMNH 167474, -, Mexico
AMNH 171259, -, Mexico
AMNH 167481, -, Mexico
AMNH 185083, -, Mexico
AMNH 189210, -, Mexico

Glossophaga soricina:

ZIM (SO) 10,
ZIM (SO) 11,
ZIM (SO) 12,
ZIM (SO) 13,
ZIM (SO) 14,
ZIM (SO) 15,
ZIM (SO) 16,
ZIM (SO) 17,
ZIM (SO) 18,
ZIM (SO) 19,
ZIM (SO) 20,
ZIM (SO) 21,
ZIM (SO) 22,
ZIM (SO) 26,
ZIM (SO) 28,
ZIM (SO) 29,
ZIM (SO) 30,
ZIM (SO) 31,
ZIM (SO) 33,
ZIM (SO) 34,
ZIM (SO) 35,
ZIM (SO) 36,
ZIM (SO) 39,
ZIM (SO) 42,
ZIM (SO) 80,

-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.
-, Hacienda el Timbre, Esmeraldas, Ecuador, leg.

BMNH 39119, no data
BMNH 1.11.3.14, 2, Rio Jordao, Brazil
BMNH 1.11.3.15, 2, Rio Jordao, Brazil

MMMM M MIMO MOM OMe

. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1982
. Patzelt, 1932
. Patzelt, 1982

113

114

BMNH 1.11.3.18, 2, Brazil
BMNH 3.7.1.162, 7, Ipanema, SaoPaulo, Brazil
BMNH 11.12.22.5, 2, Sn. Amazons., Brazil
BMNH 11.12.22.6, ?, Rio Yumunda, Sn. Amazons., Brazil
BMNH 24.2.4.5, ?, Caldeirao, Marajo, Amazon., Peru
BMNH 28.5.2.130, 2, Chicosa Loreto, Peru
BMNH 66.4394, °, San Jose, Costa Rica
BMNH 67.798, ©, San Antonio, Chinandega, Nicaragua
BMNH 88.8.8.27, -, Tabasco
MEPN 1445, ?, Isla Silva, Prov. Guayas, Ecuador
MEPN 8024, 7, San Lorenzo, Esmeraldas, Ecuador, leg. L. Albuja, 16.2.1980
MEPN 47109, -, Chontillal, Ecuador, leg. E. Patzelt, 27.11.1974
MEPN 78104, °, San Pedro de los Catanes, 28km via Lago Agrio, Ecuador, leg. R. Nevarrete,
5.10.1978
MEPN 791123, &, Jumandi, Prov. Napo, Ecuador, leg. L. Albuja, 3./5.12.1979
MHNG 1061.61, 5, Mallares/Sullana, Peru, leg. Kramer & Markl, 3.1957
MHNG 1682.79, -, Villarica, Paraguay
MHNG 1682.84, <, Atacames, Sua, Ecuador, leg. J. Garzoni, 1982
MHNG 1682.85, 7, Santa Rosa, Mun. Vigia, Para, Brazil, leg. Novaes Souza (M. Goeldi), 1972
MHNG 1682.86, 7, Arroyo Tagatya-mi, 25km E de Puerto Max, Conception, Paraguay, leg
Weber, 1983
MHNG 1682.87, ?, Momotombo, cite geothermique, Nicaragua, leg. Chambrier Jaccoud, 1983
MHNG 1682.89, -, Escazu, Costa Rica, leg. C.F. Underwood, 23.7.1898
M781 (Kopenhagen), 7, Realejo, Oersted, 1949
M782 (Kopenhagen), 2, no data
M2983 (Kopenhagen), ©, Caldurao, Marajo, 1923
M2984 (Kopenhagen), ?, Aruba, ded. 13.8.1968
L13 (Kopenhagen), no data
RMNH 34373, 5, Matta, Suriname, 11.1.1963
RMNH 34374, 2, Nickerie, Suriname, leg. D.G. Reeder, 19/20.5.1981
ZFMK 80.11, <, Orocue, 5.1897, ex Coll. Mus. Göttingen, Bürger coll. 1978
ZFMK 8075, ?, Puerto Maldonado, Rio Madre de Dios, Peru, leg. E. Lenkenhoff, 8.1980
ZFMK 67194, 2, Mato Grosso, Brazil
ZFMK 67195, 2, Mato Grosso, Brazil
ZFMK 80865, ©, Puerto Maldonado, Rio Madre de Dios, Peru, leg. E
ZFMK 80868, <, Puerto Maldonado, Rio Madre de Dios, Peru, leg. E. Lenkenhoff, 1980
ZFMK 80869, <, Puerto Maldonado, Rio Madre de Dios, Peru, leg. E. Lenkenhoff, 1980
ZFMK 80870, 2, Puerto Maldonado, Rio Madre de Dios, Peru, leg. E. Lenkenhoff, 1980
E
E

. Lenkenhoff, 1980

ZFMK 80876, ¢, Puerto Maldonado, Rio Madre de Dios, Peru, leg. E. Lenkenhoff, 1980
ZFMK 80877, 2, Puerto Maldonado, Rio Madre de Dios, Peru, leg. E. Lenkenhoff, 1980
ZFMK 811502, 2, Rio Tambopata, Peru, leg. E. Lenkenhoff, 1981

NHMW 21362, 2, Huanaco, Peru, leg. G. Paetzmann, 1976

Monophyllus plethodon (incl. M. luciae):

BMNH 18.4.1.7, 3, Antigua

BMNH 18.4.1.8, ©, Antigua

BMNH 18.4.1.9, ©, Antigua

BMNH 32.4.1.11, 2, Domenica

RMNH 17854, °, Dark Cave, Barbuda, Lesser Antilles, leg. P.W. Hummelinck, 6.7.1955

Monophyllus redmani:

BMNH--, -, Jamaica

BMNH 65.3996, -, Jamaica

BMNH 7.1.1.666, -, Jamaica

BMNH 7.1.1.667, -, Jamaica

BMNH 7.1.1.668, -, Jamaica

BMNH 75594, ©, Ewarton, St. Cath. Parish, (Googwin), Jamaica

MHNG, 982.93, °, Jamaica, leg. C.F. Underwood

SMF 57976, ?, St. Claire Cave, Ewarton, St. Catharine Parish, Jamaica, leg. Goodwin,
29.12.1965

SMF 57977, °, St. Claire Cave, Ewarton, St. Catharine Parish, Jamaica, leg. Goodwin,
29.12.1965

SMF 57978, <, St. Claire Cave, Ewarton, St. Catharine Parish, Jamaica, leg. Goodwin,
29.12.1965

ZFMK 82270, &, Green Hills, Blue Mountains (1200m asl), Jamaica, leg. H.H. Wii, 1982

Lichonycteris degener (incl. L. obscura):

AMNH 95118, °, Brazil

AMNH 95485, ?, Brazil

AMNH 131769, -, Costa Rica

BMNH 3.4.5.36, ?, Cayenne, Fr.-Guayana, leg. O. Thomas

BMNH 95.4.29.1 (Type), °, Managua, Nicaragua, leg. D. Rothschuh

BMNH 96.10.1.20, -, San Jose, Costa Rica, leg. C.F. Underwood

MEPN 741050, ?, Rio Palenque, Ecuador, leg. B. Stott & K. Mioyota

MHNG 1682.82, 5, Santarem, Rio Solimoens, Amazonas, Brazil, leg. P. Pictet, 1957

Leptonycteris nivalis (incl. L. curasoae):

BMNH 631811, 5, Jalisco, Mexico, leg. A.C. Buller

BMNH 66.6040, -, Jalisco, Mexico, leg. A.C. Buller

BMNH 70.2057, °, Cuevas del Guano, Falcon, Venezuela

BMNH 70.2058, 5, Cuevas del Guano, Falcon, Venezuela

BMNH 93.5.7.9, <, Jalisco, Mexico

BMNH 93.5.7.10, M. Tizapan el Alto, Jalisco, Mexico, leg. A.C. Buller
BMNH 93.5.7.11, M. Tizapan el Alto, Jalisco, Mexico, leg. A.C. Buller
MHNG 1682.78, -, Columbia

RMNH 14394, 5, Cueba Bosa, leg. P.W. Hummelinck, 1949

RMNH 14395, 5, Cueba Bosa, leg. P.W. Hummelinck, 7.3.1949

RMNH 14396, 5, Cueba Bosa, leg. P.W. Hummelinck, 1949

RMNH 14397, 5, Cueba Bosa, leg. P.W. Hummelinck, 7.3.1949

RMNH 14398, 2, Cueba Bosa, leg. P.W. Hummelinck, 7.3.1949

RMNH 14717, ©, Cueba di Watapana, Lima, Bonaire, leg. P.W. Hummelinck, 1954
1068 (Museum Kopenhagen). 2‘, Jalisco, Mexico

1069 (Museum Kopenhagen). ©, Jalisco, Mexico, leg. Buller, 1893, ex Coll. Brit.Mus.
SMF 37780, ?, Rancho las Margaritas, Mexico, leg. Greenhall & Schmidt, 25.4.1969
SMF 37781, -, Rancho las Margaritas, Mexico, leg. Greenhall & Schmidt, 25.4.1969
SMF 37782, -, Rancho las Margaritas, Mexico, leg. Greenhall & Schmidt, 1969

Leptonycteris yerbabuenae (= sanborni):
AMNH 208226, 5, Oaxaca, Mexico
AMNH 208227, 5, Oaxaca, Mexico

116

AMNH 213763, 5, Oaxaca, Mexico
MHNG 1184.11, ©, Chiapas, 7mls WSW Ocozocautle, Mexico, leg. Carter, 1962

Anoura caudifer:

SMF 69747, ?, 15km SE St. Laurent, Fr.-Guayana, leg. D. Kock & H. Stephan, 29.10.1985
SMF 69749, 3, 15km SE St. Laurent, Fr.-Guayana, leg. D. Kock & H. Stephan, 29.10.1985
SMF 69750, 2, 15km SE St. Laurent, Fr.-Guayana, leg. D. Kock & H. Stephan, 29.10.1985
MHNG 1682.80, 5, Mera, Pastaza, Ecuador, 12.3.1981

RMNH 34379, ©, Brownsberg, Distr. Brokopondo, Suriname, leg. G.F. Mees, 23.2.1972
RMNH 17269, 7, San Miguel Paulista, Sao Paulo, Brazil, leg. E. Deute, 2.10.1960

RMNH 13487, 2, Jodensavanne, Suriname, leg. J. Lindenau, 1954

ZFMK 59.55, °, Rio Bobonaza, Ecuador, 10.1.1959

ZFMK 59.56, 2, Rio Bobonaza, Ecuador, 10.1.1959

MEPN 791127, ?, Cueva di Archidona, Jumandi, Ecuador, leg. L. Albuja, 4.10.1979
IKD2IITF Benunleo I aKoepke

"Lonchoglossa”, Juli 1982

L-17 (Museum Kopenhagen), no data

L-18 (Museum Kopenhagen), no data

Anoura cultrata (=brevirostrum, =wercklae):
AMNH 214324, %, Peru
AMNH 233250, $, Peru
AMNH 233251, 7, Peru
AMNH 233252, ?, Peru
AMNH 233253, 2, Peru
AMNH 233254, 2, Peru
AMNH 233255, %, Peru
AMNH 233262, 7, Peru
AMNH 233263, 7, Peru
AMNH 233268, °, Peru

Anoura geoffroyi:

RMNH 16416, =, Tafelberg, Suriname, leg. D.C. Geyskes, 1958, det. Husson

RMNH 17851, 7, Tamana Bat Cave, Trinidad, leg. P.W. Hummelinck, 8.1.1955

RMNH 17853, ©, Tamana Bat Cave, 230m asl, Trinidad, leg. P.W. Hummelinck, 8.1.1955
RMNH 34375, 2, Katalebo, Nickerie, Suriname, leg. D.G. Reeder, 3.5.1981

SMF 69766, ©, Kourou, Fr.-Guayana, leg. D. Kock & H. Stephan, 13.11.1985

SME 69767, 5, Savanne le Gallion, 23km S Cayenne, Fr.-Guayana, leg. D. Kock & H. Stephan,
13.11.1985

MEPN 1126, -, Bocas del Cerro, Saloya, Ecuador, leg. Spillmann, 4. 1939

MEPN 1127, <, Bocas del Cerro, Saloya, Ecuador, leg. Spillmann, 4. 1939

MEPN 7941, ©, Piedro Blanca, Rumirahui, Ecuador, leg. L. Albuja, 15.4.1979

MEPN 7942, 5, Piedro Blanca, Rumirahui, Ecuador, leg. L. Albuja, 15.4.1979

MEPN 3948, 5, Rio Saloya, Bocas del Cerro, Saloya, Ecuador, leg. Spillmann, 4. 1939
MEPN 3949, 5, Rio Saloya, Bocas del Cerro, Saloya, Ecuador, leg. Spillmann, 4. 1939
MEPN 6762, -, Esmeraldas, Ecuador, -

MEPN 39411, -, Rio Saloya, Bocas del Cerro, Saloya, Ecuador, leg. Spillmann, 1939
NHMW 30720, °, Campinas, 22 54'S/47 06'W, Sao Paulo, Brazil, leg. C. Vanzolini, 1906
MHNG 1682.81, 2, Mera, Pastaza, Ecuador, 18.11.1981

ZIM (SO) 187, 2, San Antonio de Pichincha, Quito, Ecuador, leg. E.H. Solmsen, 7. 1983

1077

ZIM (SO) 198, ?, San Antonio de Pichincha, Quito, Ecuador, leg. E.H. Solmsen, 7. 1983
ZIM (SO) 199, -, San Antonio de Pichincha, Quito, Ecuador, leg. E.H. Solmsen, 7. 1983
ZIM (SO) 201, ?, San Antonio de Pichincha, Quito, Ecuador, leg. E.H. Solmsen, 7. 1983

Anoura latidens:
AMNH 261230, 2, Terr. Fed. Amazonas, Venezuela, leg. R.W. Dickerman, 14.4.1984

Anoura wiedi:

BMNH 27.11.1928, ©, Campinas, Brazil

BMNH 27.11.1929, <, Brazil

BMNH 27.11.1930, 9, Brazil

BMNH 27.11.1931, °, Brazil

BMNH 27.11.1932, °, Brazil

RMNH 25482, ©, Campina, Estado de Sao Paulo, Brazil, leg. Vangolini & G. Doria, 1906
793 (Museum Kopenhagen), 7, Lagoa Santa, Brazil, leg. Reinhardt, 27.8.1955

792 (Museum Kopenhagen), 7, Lagoa Santa, 1866

Hylonycteris underwood:

AMNH 178904, -, Panama

AMNH 189688, -, Mexico

AMNH 238199, ?, Panama

AMNH 256826, °, Belize

BMNH 3.2.1.3, -, Taibaca, Costa Rica, leg. C.F. Underwood

BMNH 3.2.1.4, -, Taibaca, Costa Rica, leg. C.F. Underwood

BMNH 3.2.1.5 (Type), -, Rancho Redondo, Costa Rica, leg. C.F. Underwood

Scleronycteris ega:
BMNH 7.1.1.671 (Type), -, Ega, Amazonas, Brazil, James Collection Bates

Choeroniscus godmani:

AMNH 131765, 5, San Jose, Costa Rica, leg. C.F. Underwood, 18.6.1938

AMNH 172778, -, Tapanatepec, Oaxaca, Mexico, leg. A. Johnson, 27.1.1954

AMNH 172779, 5, Tapanatepec, Oaxaca, Mexico, leg. A. Johnson, 27.1.1954

BMNH 79.12.24.1, 5, Guatemala, leg. D. Godman

SME 43028, , Finca el Buque, Villavicencio, Dept. Meta, Colombia, leg. E. Thiery, 15.5.1979
SMF 41990, 2, Finca el Buque, Villavicencio, Dept. Meta, Colombia, leg. H. Stephan, 29.6.1991
USNM 385917, °, Merida, 59km SE El Dorado, Venezuela, 9.6.1966, SUP

USNM 385919, ?, Bolivar, 59km SE El Dorado, Venezuela, 13.6.1966, SUP

USNM 385920, ?, Merida, 59km SE El Dorado, Venezuela, June 1966, SUP

Choeroniscus minor (= C. inca, = C. intermedius):

AMNH 67625, 5, Los Pozos, Ecuador

AMNH 140471, ?, Kamakusa, British Guayana, leg. H. Lang, 5.2.1923

AMNH 142901, -, British Guayana

AMNH 230285, 2, San Pablo (900m), Prov. Oxpampa, Dept. Pasco, Peru, leg. J.C. Kelly,
12.7.1964

BMNH 53.3.19, -, Rio Cupari, Bates Collection

BMNH 69.1275A, ?, Araenga, Para, Brazil, leg. R. Lainson, 18.10.1969

BMNH 69.1275B, ¢, Araenga, Para, Brazil, leg. R. Lainson, 17.10.1969

MEPN 7551, ?, Plan Piloto, via Quininde, Esmeraldas, Ecuador, leg. L. Albuja, 3.5.1975

118

JK 189, -, Peru, leg. J. Koepke

IKT Berugles2]aKoepke

SMF 54044, °, Manau, Brazil, leg. U. Schnitzler, 2.1977

SME 69883, «, Camp Caiman, Montagnes de Kourou, Fr.-Guayana, leg. D. Kock & H. Stephan,
10.1985

SMF 69802, ?, Camp Caiman, Montagnes de Kourou, Fr.-Guayana, leg. D. Kock & H. Stephan,
10.1985

ZIM (OS) 1982, -, Rio Cuyabeno, Ecuador, leg. E. Patzelt, 1982

USNM 361573, ?, Belem, Fazenda Velho, Para, Brazil, leg. C.O. Handley, 1965

USNM 361574, 7, Belem, Fazenda Velho, Para, Brazil, leg. C.O. Handley, 1965

USNM 361575, ©, Belem, Fazenda Velho, Para, Brazil, leg. C.O. Handley, 1965

Choeroniscus periosus:

AMNH 217038, ?, Colombia

Choeronycteris mexicana:

AMNH 212358, 2, Cerro de San Felipe, Oaxaca, Mexico, leg. T. MacDougall, (2815), 1965
AMNH 212359, =, Cerro de San Felipe, Oaxaca, Mexico, leg. T. MacDougall, 1965
AMNH 212360, %, Cerro de San Felipe, Oaxaca, Mexico, leg. T. MacDougall, 1965 (7/12)
AMNH 212361, 5, Cerro de San Felipe, Oaxaca, Mexico, leg. T. MacDougall, (2818), 1965
AMNH 212362, °, Cerro de San Felipe, Oaxaca, Mexico, leg. T. MacDougall, (2819), 1965
AMNH 212365, °, Cerro de San Felipe, Oaxaca, Mexico, leg. T. MacDougall, (2822), 1965
BMNH 60.449, 2, Sonora (8mls NE Imuris), Exchange with Univ. Kansas

BMNH 75.2.27.60, -, Duenas, Guatemala

BMNH 27.11.1935, 7, Los Masos, Jalisco, Mexico, Exchange with Genua Museum

MHNG 1177.16, no data

ZEMK 77.652, -, Sierra Mixteca, Mexico

USNM 50800, ©, Querendaro, Michoacan, Mexico, 5.8.1892

USNM 50801, 2, Querendaro, Mexico, 5.8.1892

USNM 50802, 7, Querendaro, Michoacan, Mexico, 5.8.1892

USNM 50803, 2, Querendaro, Michoacan, Mexico

USNM 50804, °, Querendaro, Michoacan, Mexico

Choeronycteris (=Musonycteris) harrisoni:

AMNH 235179, 5, Mexico

BMNH 61.1612, ©, Colima, Mexico, ded. Univ of Arizona

SMF 22500, ©, Pueblo Juarez. Colima, Mexico, leg. A.L. Gardner, 1.4.1960

‚ler, Vv. HR. Ser 1. Breculta: Rinne der Gattung
rura en 1837 (Ayes, Passeres, Estrildidae). 1972, 158 S., 2 Tafeln,

poe Die Wirbeltierfauna von Fernando Poo und Westkamerun.

1ost,.O.: ar nm hs cinclus) mit besonderer Berück-
‘ sun ihrer Ernährung. 1975, 183 S., DM 46,—

: s ie birds. 1976, 92.87 1. Talel, M23. —
2 en) contact zones of Sn in northern Iran. 1977, 64 S., 1 Falt-

en von ae ind Somme aideahachen (Rees
1 ignicapillus) und deren ethologische Differenzierung. 197915 ESS

r, D.G.: Funktionell- morphologische Untersuchungen zur Radiation
nES- ‚und A ode der Papageien (Psittaci). 1980, 192 S.,

er, SO: A ree teal study of the genus Apistogramma Regan, with
Bi 3 ilian age Peruvian species (Ieleostei: Percoidei: Cichlidae). 1980,

a Bibliographie der Säugetiere es Vögel der Türkei. 1986,

Bi en zur Ste cnenik der rezenten Cases ee
au cm. cramiomen ischier zen 1987, 96 S., DM 24,—

24.

25.

26.

ER

28.

29:

30.

31.

32:

33.

34.

3%

36.

37:

38.

39:

40.

Al.

42.

43.

44.

Teleostei, pa a Re and phylogenetic
a ig DM I ee ze

Böhme, W.: Zur Genitalmorphologie der Se Funktionelle
geschichtliche Aspekte. 1988, 175 S., DM 44,— |

Lang, M.: Phylogenetic and biogeographic patterns of nee
(Reptilia: Squamata: “Iguanidae”). 1989, 172 S., DM 43,—

Hoi-Leitner, M.: Zur Veränderung der Säugetierfauna des
Gebietes im yore der letzten drei Jahrzehnte. 1989, 104 Se DM

Bauer, A. M.: Phylogenetic systematics and Biogeography of th
lini (Reptilia: Gekkonidae). 1990, 220 S., DM 55,— | 5

Fiedler, K.: Systematic, evolutionary, and ecological iipliean
phily within the Lycaenidae (Insecta: ree Prous
DM 53,—

Arratia, (=: en and variation of he suspensorium ¢

sidae, Dh) unter besonderer BR Foe I inneren.
schlechtsorgane. 1993, 115 S., DM 32,— Er

Blaschke-Berthold, U.: Anatomie und Phylogenie der Bibi
secta, Diptera). 1993, 206 S., DM 52,—

Hallermann, J.: Zur Morphologie der Hine aes der |
— eine vergleichend- anatomische Untersuchung. 1994, eS S.

skin of Diplomystid and certain Bre Loricarioid Catfis
eee seen considerations. 0: 110 S., DM 28, —

struktur ES geographischen Variation im Zagora i 1 -Su
Komplex (Insecta, Be Fer 1995, 224 Oe M

and wing base of the Scarabaeoidea (Coleoptera) with some p
cations. 1996, 200 S., DM 50,— RRS.

Bininda-Emonds, O.R.P,&A.P. Russell: A morpholo,
the phylogenetic relationships of the extant phocid seals (Me 1
Phocidae). 1996, 256 S., DM 64,—

Klass, K.-D.: The external male ee and the Dosen f

optera und Nee 1998, 126 S., DM 32,—

Solmsen, E.-H.: New World nectar-feeding bats: Bilde we iC
craniometric approach to systematics, 1998, 118 S., DM ei, Sr ae |

en % y
AG te hal
DR a } 5
by
) >
H h; ri
- } k a
‘ : 1
4 :
i 4 }
: % (

FÜR

ATIK IN DEUTSCHLAND:

„HÄUSER & K.-H. LAMPE

OOLOGISCHE MONOGRAPHIEN, Nr. 45

a ie

BRone.

x x
ART,
x x
SUN i
rs ; : nS
re
er

BONNER ZOOLOGISCHE MONOGRAPHIEN

Die Serie wird vom Zoologischen Forschungsinstitut und Museum Alexander Koenig
herausgegeben und bringt Originalarbeiten, die für eine Unterbringung in den „Bonner
zoologischen Beiträgen” zu lang sind und eine Veröffentlichung als Monographie
rechtfertigen.

_ Anfragen bezüglich der Vorlage von Manuskripten sind an die Schriftleitung zu richten;
Bestellungen und Tauschangebote bitte an die Bibliothek des Instituts.

This series of monographs, published by the Zoological Research Institute and Museum
Alexander Koenig, has been established for original contributions too long for inclu-
sion in „Bonner zoologische Beiträge”.

Correspondence concerning manuscripts for publication should be addressed to the
editor. Purchase orders and requests for exchange please address to the library of the
institute.

LInstitut de Recherches Zoologiques et Muséum Alexander Koenig a établi cette série
de monographies pour pouvoir publier des travaux zoologiques trop longs pour étre
inclus dans les „Bonner zoologische Beiträge”.

Toute correspondance concernante des manuscrits pour cette série doit étre adressée a
l’editeur. Commandes et demandes pour échanges adresser a la bibliothéque de l’insti-
tut, Ss. Vv. p.

BONNER ZOOLOGISCHE MONOGRAPHIEN, Nr. 45, 1999
Preis: 16,- DM
Schriftleitung/Editor: G. Rheinwald

Zoologisches Forschungsinstitut und Museum Alexander Koenig
Adenauerallee 150-164, D-53113 Bonn, Germany
Druck: JF CARTHAUS, Bonn

ISBN 3-925382-49-6
ISSN 0302-671 X

BIODIVERSITÄTSINFORMATIK IN DEUTSCHLAND:
BESTANDSAUFNAHME UND PERSPEKTIVEN

von

W.G. BERENDSOHN, C.L. HAUSER & K.-H. LAMPE

BONNER ZOOLOGISCHE MONOGRAPHIEN, Nr. 45
1999

HERAUSGEBER:
ZOOLOGISCHES FORSCHUNGSINSTITUT
UND MUSEUM ALEXANDER KOENIG
BONN

Die Deutsche Bibliothek - CIP-Einheitsaufnahme

Berendsohn, W. G.:
Biodiversitätsinformatik in Deutschland: Bestandsaufnahme und Perspektiven / von W.
G. Berendsohn, C. L. Häuser & K.-H. Lampe. Hrsg.: Zoologisches Forschungsinstitut

und Museum Alexander Koenig, Bonn. - Bonn : Zoologisches Forschungsinst. und
Museum Alexander Koenig, 1999

(Bonner zoologische Monographien ; Nr. 45)

ISBN 3-925382-49-6

INHALT

Seite

SE OGIVEHSIGALSIIMKOMMILATL OIG. er. een ae) SIR 5
|! Jl, TEVUMUSIGDNAVE "St gS eo ee re RE END IAUNE, 5
BO CIVERSILAD Byer tice arte ee eter ee ah. Wt LS MIEN EOE Se da 5
Niolekuilanerl! Dene jie Mats 0.2 eR COUPE LEHRE N 5
OnsanismischiEDenermt tt REIN het SN BEM 5
Okosystemaneisbenchs as. CON) I ENR i t.. 6
benenidem Mi OdiversitalsimlOnmatlOMy in. ¢ yaar ee. 6
Bioinformatik (molekulare Biodiversitätsinformation) ............ i
Umweltinformatik (Biodiversitätsinformation auf der Okosystemebene)

Re Bee re oe se TEEN SEES NM Cate alae 4 ae 7
Orsanısmische Biodiversitätsmtormatier nn. een n....n. 8

Poms cdeuvumeder Artinformaiion......2... mn... UR ee cle 8

1. 3. Bedeutung der Information in biologischen Sammlungen ........... 9

2. Biodtivelälsıı OS e0 0 10 ee ee ere ri eee rr ne 9
2,1, Deiimiitonent ee ee cn ie al RR SE RUE AR a LEER RIESE 9
PRESS CUGUIN Smee er. N IL WG NINE BEE Phe ed A D 10

Se Erusulerende internationale Strukturen. . 2.2.2.2... lern. I]
PalevlicenmavonalempolitischemRahmen -.. nr. 2. nn. 11
Dicsbitodiversitatskomvention nn a ne. I]
Weitere wichtige Abkommen und Konventionen .................. 12

Py UO mseizune aufinternauonalenEbene rn. een... 12
OECD Vesascience Forum GBIF TI. NE eo Pe 13

G7 Environment and Natural Resources Management Project (ENRM) 14
iinniatnvenider Veremten Nationen... 2.2.5 ee ae ee... 14
United Nations Food and Agriculture Organization (FAO) ....... 14
WininedINations Environment Brosammer mr nme. 14

United Nations Educational, Scientific and Cultural Organization .. 14

IDINYEIESIUAS. 7 REN En Ines ae u An u 15
U Prostzines en Ate Oe, ME a SS a Cr 15
Europäische Umweltagenur EVA). 2.2.2.0... eee oe nennen. 16
ResierunssunabhansiseOrsanısauonen Yan. nn... 18
PAR SAMMI INAS SUI OCR eee te ee N LE 18

3.3. Globale und europäische Initiativen auf der organısmischen Ebene ... 19
OrSanismenteeistenr Aa Fld tgs She dee dias ekg sen 19

Biolosısche Sammlungen, ce 2 ME ee 22

Nomenklatur :.43 2. S23 a ee a RE 2D

3.4. Standardisierung: u... ee ee en eee 24

4. Strukturen in Deutschland, |. =... 2 A eee eee 23
4.1. Umsetzung internationaler Übereinkommen ..................... 25
4.2. Umweltinformationssysteme 2.2022 ene eee 2
Organisation von Umweltinformationssystemen .................. 25
Floren- und Faunenkartierung und Listen, Beringung .............. 28

4.3. Informationssysteme zu genetischen Ressourcen in Deutschland ..... 30

4.4. Deutsche Informationssysteme zur globalen Biodiversität auf der
organismischen Ebene... u... 2 srl 2. ee 3h

Taxonbezogene Datenbanken und Informationssysteme in Deutschland 32

Wirbeltiere.. +... seat Ba i a eee eee

Insekten... ............2.-.2.:.- ee Se ee Sel ee

Marine. Organismen». 1020. 12.22. onthe eel es Be eee 33

Pflanzen. .........-..- 22... 22-22 22 oto Sa 33
Sammlungsdatenbankprojekte ............. 22.0222 nee 34
Verknüpfte Information. ............ 0... 22.2.2... 22.22.22 eee 31

4.5. Zusammenfassung ......... kee lee cl ae eee 89

5. Strategie und Prioritäten im Bereich Biodiversitätsinformatik ........... 40
5.1. Strategie für eine national koordinierte Forschungsförderung ....... 40
5.2. Verbesserung der Infrastruktur I... . 2.2222 442. 2 oe eee 4]
Koordination . ........:..,.. ene sro Ee ee ee eee +]
Schaffung von - oder Anbindung an - Standarddatenkataloge ........ 42

3.3. Informationserschließung +..2=. 2.222 I: eee). occ oe eee 43
Organismenregister ... 2... . ... waere ae eee do
Erschließung der Sammilungsinformation= = 4. 4-22 ren 45
Schaffung von verknüpften biologischen Informationssystemen ..... 47
Danksagung: 1.3.2252 I. aan ah cee ee 48
Zitierte Literatur u... 0.20 ee en See 48

1. BIODIVERSITÄTSINFORMATION
1.1. Einleitung

Diese Schrift soll Biologen aller Fachrichtungen, Informatiker und wissenschafts-
politische Entscheidungsträger gleichermaßen ansprechen. In interdisziplinären
Artikeln, die an ein solchermaßen breites Publikum gerichtet sind, sind bisweilen
Einführungen in Themengebiete notwendig, die dem Spezialisten als unnötige
Längen vorkommen werden, dem anderen Spezialisten aber erst den Zugang zum
Gesamtthema ermöglichen. Solche Längen bitten wir den Leser, uns nachzusehen.

Grundlage dieser Arbeit war ein von den Autoren im August 1998 ım Auftrag des
Bundesministeriums für Bildung, Wissenschaft, Forschung und Technologie
(BMBE, Referat 422) fertiggestelltes Gutachten zu Prioritäten in der Biodiversitäts-
informatik unter Berücksichtigung vorhandener internationaler und nationaler
Strukturen.

Biodiversität

Die gesamte Vielfalt organismischen Lebens (“Biodiversität”) stellt die für die
Sicherung der menschlichen Existenz bei weitem wichtigste und zugleich die am
kompliziertesten strukturierte, natürliche Ressource unseres Planeten dar. Verfüg-
barkeit und allgemeiner Zugang zu grundlegenden Informationen über die globale
Biodiversität sind daher von entscheidender Bedeutung für die zukünftige Entwick-
lung der Menschheit und werden zunehmend von politischer Seite gefordert, so
z.B. ım Rahmen der Biodiversitätskonvention, durch das OECD Megascience
Forum und verschiedene Initiativen der G7-Staaten.

Vereinfachend lassen sich fast alle biologischen Daten und Kenntnisse zur Biodi-
versität drei auch intuitiv greifbaren Ebenen zuordnen, die sowohl von ihrer wis-
senschaftlichen Erforschung her als auch hinsichtlich ihrer gegenwärtigen informa-
tionstechnischen Betreuung gut zu trennen sind:

Molekulare Ebene

Hier werden Nukleinsäuren und andere zelluläre Verbindungen und die dazwischen
wirkenden Steuerungsmechanismen untersucht. Auf diesem Gebiet liegen die
Hauptaufgaben der Fachgebiete der molekularen Genetik, Gentechnologie, Bio-
chemie und Physiologie. Große Datenmengen sind vor allem im Bereich der
Genom- und Proteinsequenzierung und Modellierung entstanden.

Organismische Ebene

Der ganze Organismus, seine Interaktion mit anderen gleichartigen Organismen
(Populationen) und die Klassifikation der Organismen in ein von Verwandschafts-
verhältnissen bestimmtes System steht im Fokus der Forschung in diesem Bereich
(die aber durchaus zunehmend auch mit molekularbiologischen Methoden betrie-
ben wird). Die Gebiete der Systematik und Taxonomie definieren sich tiber diese
Aufgaben, aber auch die Populationsgenetik (einschl. ecological genetics) ist hier
anzusiedeln. Die Ziichtungsforschung, der Artenschutz, sowie wichtige Teilgebiete

6

der Land- und Forstwirtschaft und des Fischereiwesens stellen angewandte Fach-
gebiete auf dieser Ebene dar.

Ökosystemare Ebene

Hier geht es um das Zusammenwirken verschiedenartiger Organismen und Popula-
tionen mit ihrer Umwelt (Klima, Hydrologie, Boden, andere Organismen) und ihre
Organisation in Form von differenzierbaren Systemen. Als Wissenschaftszweige
sind vor allem die Ökologie und eine sich herausbildende, über den Bereich der
Biologie hinausreichende Umweltforschung zu nennen. Im angewandten Bereich
sınd große Teile der Land-, Forst- und Fischereiwirtschaft sowie die meisten
Aufgaben des Natur- bzw. Umweltschutzes hier angesiedelt.

Ebenen der Biodiversitätsinformation

Letztendlich ist eine Überwindung dieser Trennung im Rahmen eines umfassenden
Systems der Biodiversitätsinformation anzustreben, welches die Biodiversität
selbst in ihrem hierarchischen Aufbau (Moleküle, Zellen, Gewebe, Individuen,
Populationen, Arten, Gesellschaften, Ökosysteme) abbildet und modelliert. Wir
stehen aber erst am Anfang der Biodiversitätsforschung und das Feld der Biodi-
versitätsinformatik ist noch jünger. Die Trennung in die drei Bereiche bleibt daher
bis auf weiteres sinnvoll. So ist z.B. eine unmittelbare Herleitung aller Eigen-
schaften und Leistungen bestimmter Organismen allein auf Grundlage der Kenntnis
ihrer konstitutiven Moleküle derzeit nicht absehbar; ebenso lassen sich komplexe
organismische Funktionen nach wie vor meist nicht direkt einzelnen Molekülen
zuordnen (obwohl wir zunehmend Gene kennen, welche bestimmte Eigenschaften
kodieren, z.B. Wuchsform, Farbe oder Substratabbau). Die Vorhersagbarkeit
wichtiger Ökosystemeigenschaften ist dagegen teilweise bereits heute auf der
Grundlage der Kenntnis physiologischer Leistungen einzelner Organismen zu-
mindest grob möglich. Die Tragfähigkeit bzw. Vorhersagekraft der hier vorhande-
nen Modelle leidet jedoch oft an der unzureichenden Informationsbasis auf der
Ebene der Organismen, also dem Fehlen von auf das Taxon bezogener Information
zu eben diesen Leistungen. Die Beseitigung dieses Informationsdefizits, also die
unmittelbare Verknüpfung der vorliegenden Informationen zur organismischen und
synökologischen Ebene ist daher drängend, wobei vor allem die Ökosystemfor-
schung auf (neue) Daten aus dem organismischen Bereich angewiesen ist.

Die Informatik ist “die Wissenschaft von der systematischen Verarbeitung von
Informationen, besonders der automatischen Verarbeitung mit Hilfe von Digital-
rechnern” (Duden Informatik 1993). Die Verarbeitung von Biodiversitätsinformati-
on sollte daher als Biodiversitätsinformatik (engl. biodiversity informatics) be-
zeichnet werden (siehe 2.1). Der Begriff Bioinformatik (bioinformatics) ist als
Terminus von der molekularen Biodiversitätsinformatik belegt. Die ökosystemare
Ebene findet sich in der Umweltinformatik (environmental informatics), ebenfalls
ein bereits geprägter Begriff. Für die organismische Ebene wird teilweise der
neuere Begriff Biodiversitätsinformatik direkt verwendet, man sollte hier aber

besser von organismischer Biodiversitätsinformatik sprechen. Entwicklungsstand
und Datenverfügbarkeit in diesen drei Bereichen stellen sich wie folgt dar.

Bioinformatik (molekulare Biodiversitätsinformation)

Für den molekularen Bereich existieren in Deutschland bereits umfangreiche,
international meist gut eingebundene Datenbankprojekte. Der Informationszugang
bzw. -austausch ist hier vergleichsweise gut organisiert, bis hin zur Einbindung
privatwirtschaftlicher Sektoren. In diesem Zusammenhang sei das European
Molecular Biology Laboratory (EMBL) genannt', mit zentralem Sitz in Heidelberg
und Außenstellen in Hamburg, ın Grenoble (Frankreich) und schließlich mit dem
European Bioinformatics Institute (EBI) in Hinxton (England). Nach einer Ein-
schätzung der OECD Megascience Forum Working Group on Biological Informa-
tics sind im molekularen Bereich bereits heute mehr als 95% der Daten digitalisiert,
im organismischen Bereich hingegen weniger als 5%°. Die 50jährige Geschichte
der Molekularbiologie verlief parallel mit der Entwicklung der Informatik. Die
enge Verzahnung beider Wissenschaftsbereiche wird z.B. an den im mehrjährigen
Rhythmus aktualisierten und immer weitreichenderen Zielen des Human Genome
Project deutlich (Collins et Galas 1993, Collins et al. 1998). Die informations- und
labortechnische Entwicklung, aber vor allem auch die gut funktionierende interna-
tionale Zusammenarbeit führte zu einer exponentiellen Informationszunahme im
molekularen Bereich. So war z.B. letztlich in GenBank, einer der großen interna-
tionalen Sammelstellen von Daten aus dem molekularen Bereich, der Datenzu-
wachs in 10 Wochen größer als in den ersten 10 Jahren des Projekts (Robbins
1998). 1998 wurde die Marke von 2 Milliarden Basenpaaren überschritten, im
August 1999 waren es bereits über 3,4 Milliarden in 4,6 Millionen Sequenzen
(NCBI 1999a, b).

Umweltinformatik (Biodiversitätsinformation auf der Ökosystemebene)

Große Datenmengen werden hier in Form von Umweltinformationssystemen
zusammengetragen. Für die Ökosystemebene existieren auf nationaler Ebene
bedeutende Datenbank- und Informationssysteme (s. Abschnitt 4.2) und erfolgver-
sprechende Ansätze zu einer Koordinierung zeichnen sich zumindest auf europai-
scher Ebene erkennbar ab.

' Das EMBL wird von 15 Mitgliedsländern finanziert, der Jahresetat 1997 lag bei ca. 75
Mio. DM; der mittlere deutsche Finanzierungsanteil (1975-1996) betrug ca. 25%
(http://www.embl-heidelberg.de/Externallnfo/ public_relations/Facts.html).

“ Meredith Lane (Vortrag): Informatics in the service of biodiversity: Overcoming the
barriers. - Conference on Biological Informatics, 6-8 July 1998, Australian Academy of
Sciences, Canberra/Australia.

Organismische Biodiversitätsinformatik

Dagegen ist der Datenzugang und die Integration von Informationen auf der orga-
nısmischen Ebene mit wenigen Ausnahmen als defizitär zu kennzeichnen, obwohl
(oder gerade weil) in diesem Bereich schon seit ca. 250 Jahren biodiversitäts-
bezogene Daten erhoben und gespeichert werden, so z.B. in den weltweit auf etwa
3 Milliarden Exemplare (Lane 1998) geschätzten Beständen naturkundlicher
Forschungssammlungen. Daher erscheint eine Konzentration neuer Förderungs-
maßnahmen auf diesen Bereich dringend geboten, was im Einklang mit kürzlich
erhobenen Forderungen auf internationaler Ebene steht. (z.B. COP-4, SA2000,
OECD Mesascience Forum, Diversitas; siehe Abschnitt 3.1).

1.2. Bedeutung der Artinformation

Innerhalb dieses Bereichs dominiert vordergründig das Problem der enormen
Vielfalt existierender Lebensformen, die, von einzelnen Individuen ausgehend, eine
sıchere Zuordnung bzw. Verknüpfung von Information und Erkenntnissen massiv
zu erschweren scheint. Die als Folge der biologischen Evolution in der Abstam-
mungsgeschichte (Phylogenese) entstandene, natürliche hierarchische Ordnung
aller Organismen bietet jedoch einen hervorragenden Schlüssel, diese ansonsten
unüberschaubare Vielfalt der Lebensformen zu ordnen und damit auch nutzbar zu
machen (vergl. Steininger 1996). Die Charakterisierung und Benennung dieser
natürlichen Einheiten der Organismen (Taxa: Varietäten, Unterarten, Arten, Gat-
tungen, Familien, etc.) ist Aufgabe der Taxonomie, die mit Hilfe international
verbindlicher Regeln für die biologische Nomenklatur (siehe unter Abschnitt 3.3)
den verschiedenen Taxa eindeutige Namen zuordnet und diese in ein hierarchisches
Klassifikationssystem stellt. Dieses universelle Referenzsystem in der organis-
mischen Biologie bietet ideale Voraussetzungen für die Verknüpfung getrennt
vorliegender, qualitativ unterschiedlicher Informationen und Daten zu einzelnen
Organismen, wie auch zur Überprüfung der Gültigkeit bzw. des Wertebereichs
bestimmter Erkenntnisse und Hypothesen. In dieser Funktion hat sich das be-
stehende System der Organismen einerseits seit langem bewährt, andererseits
besteht ein erheblicher Forschungsbedarf zur weiteren Verfeinerung und Vervoll-
ständigung des Systems (Taxonomic Impediment, vergl. Darwin Declaration,
Environment Australia 1998). Aber auch für die vorhandenen Erkenntnisse sind
bisher die Speicherungs-, Organisations-, und Analysemöglichkeiten, die sich aus
der Entwicklung der modernen Informationstechnik ergeben, nur ansatzweise
ausgeschöpft worden. Eine Entwicklung der Biodiversitätsinformatik auf der
organismischen Ebene, unter Beteiligung taxonomischer, informatischer und
geographischer Kompetenz, ist daher eine vordringliche Aufgabe.

9

1.3. Bedeutung der Information in biologischen Sammlungen

Biologische Sammlungen umfassen sowohl Lebendsammlungen wie Botanische
oder Zoologische Gärten und Kultursammlungen (Bakterien, Pilze, Protisten,
Algen), als auch die konservierten Präparatesammlungen in Naturkundemuseen,
Universitäten und anderen ökologischen Forschungsstellen. Sie bilden einerseits
die materielle Arbeitsgrundlage der biologischen Systematik, andererseits sichern
sie (zumeist als konservierte Belege) die wissenschaftliche Überprüfbarkeit von
Forschungsergebnissen verschiedenster Teilbereiche der Biologie bzw. ermögli-
chen die Reproduzierbarkeit einzelner Befunde. Besonders die Lebendsammlungen
stellen daneben ein beträchtliches Reservoir genetischer Ressourcen dar, das z.B.
für medizinische oder biotechnologisch ausgerichtete Forschungen eingesetzt wird.
Aber die biodiversitätsinformatische Bedeutung der Sammlungen geht weit über
diese Verwendungen hinaus. Die Belege selbst und die mit ihnen assoziierten
Daten (Etiketten, Veröffentlichungen) sind zugleich Träger wesentlicher primärer
Information über Aufbau und Beschaffenheit, geographische Verbreitung und
Lebensweise einzelner Organismen sowie der Zusammensetzung der Ökosysteme,
denen sie angehören bzw. angehörten; und dies in einer sich über mehrere Jahr-
hunderte erstreckenden zeitlichen Dimension. So bilden diese Sammlungsbelege,
sowohl als Informationsträger als auch materiell, einen wesentlichen nationalen
Beitrag zur Bewältigung der im Rahmen des Globalen Wandels anstehenden
Probleme der Erhaltung und nachhaltigen Nutzung der natürlichen Biodiversität im
internationalen Rahmen.

Dieser generellen Bedeutung der primären Belege unseres Biodiversitäts-Wissens
wird in neuerer Zeit zunehmend Rechnung getragen. Im Rahmen der von der
Biodiversitätsinformatik-Arbeitsgruppe des OECD-Megascience Forum vor-
geschlagenen Global Biodiversity Information Facility (GBIF) nimmt Sammlungs-
information eine zentrale Position ein (Anonym 1999). Hier sollten auch die
Darwin Declaration (Environment Australia 1998) und die Beschlüsse der Ver-
tragsstaatenkonferenz der Biodiversitätskonvention (COP 1998) Erwähnung
finden. In den USA wird seit Jahren mit einem speziellen Férderprogramm der
National Science Foundation die ErschlieBung derartiger, in Institutionen der USA
vorhandener Belege gezielt vorangetrieben (vgl. NSF 1998). Analog hierzu wird
heute immer nachdrücklicher auch eine biodiversitätsinformatische Erschließung
der in Deutschland besonders umfangreich vorhandenen Belege (vgl. Biologische
Sammlungen unter 3.3) gefordert, so z.B. von der Direktorenkonferenz Natur-
kundlicher Forschungssammlungen, DNFS (Naumann et Greuter 1997).

2. BIODIVERSITÄTSINFORMATIK
2.1. Definitionen

Biodiversität ist die gesamte Vielfalt und Variabilität organismischen Lebens im
terrestrischen, marinen und limnischen Bereich. Sie beinhaltet die Mannigfaltigkeit

10

unter und zwischen Genen, Individuen, Populationen, Arten, Gesellschaften und
Ökosystemen. Biodiversitätsinformatik ist die Anwendung einer informatischen
Analysemethode oder einer Informationstechnologie auf Daten über biologische
Diversität und deren Verknüpfungen mit anderen Daten, z.B. mit abiotischen und
geographischen Daten. Im Zentrum dieses noch jungen Wissenschaftsgebiets
werden technologische und organisatorische Hilfsmittel entwickelt, um digital
erfaßte Biodiversitätsdaten mit Informationstechnologie zu verwalten (speichern,
indizieren, abfragen, analysieren, integrieren, visualisieren, publizieren usw.) und
um sie potentiellen Nutzern aus allen Bereichen der Wissenschaft und der Gesell-
schaft elektronisch zugänglich zu machen.

2.2. Bedeutung

Das Verständnis der Biodiversität hängt in hohem Maße von der Verfügbarkeit
relevanter Informationen ab. Während früher ein Wissenschaftler sich noch durch
Lesen der Literatur einen Überblick über ein Fachgebiet verschaffen konnte, ist das
heute aufgrund der rasanten Zunahme biologischer Informationen ohne entspre-
chende Informationstechnologie nicht mehr möglich. Darüber hinaus werden
elektronische Informationssysteme und Strukturen wie das Internet künftig Fra-
gestellungen zulassen, die heute noch visionär erscheinen. Es wird erwartet, dass
die Biologische Informatik als ein sich dynamisch entwickelndes, eigenständiges
Fach die Grundlage der Biologie des 21. Jahrhunderts bilden wird (Robbins 1998).
Sıe wırd vermutlich zu einem Paradigmenwandel in der Biologie führen (vergl.
Gilbert 1991 für den molekularen Teilbereich).

Trotz aktuell stark gesteigerter datentechnischer Möglichkeiten stellt jedoch die
generell hohe Komplexität biologischer Systeme eine effektive Verknüpfung und
Integration selbst der bereits vorliegenden, allerdings oft heterogenen und dezentral
organisierten Datenbestände vor besondere Probleme. Mit teilweisen Ausnahmen
im molekularen Bereich wird die Sachlage weiter kompliziert durch große be-
stehende Lücken ım Informationsbestand, fehlende Standards bei der Informations-
erfassung, unzureichende Bezugsdaten und fehlende klare Strukturierung der
vorhandenen wie zukünftig zu erwartenden Informationen zur globalen Biodi-
versität. Hier sollten weitere Untersuchungen auf dem Gebiet der Datenstrukturfor-
schung durchgeführt werden, die sich bemüht, die elementaren Komponenten der
Biodiversitätsinformation offen zu legen (vergl. Berendsohn 1998).

Wie bereits erwähnt, ist die biodiversitätsinformatische Erschließung im Bereich
der molekularen Biologie weit fortgeschritten. Dass der Begriff Bioinformatik, wie
er z.B. im Namen der EMBL-Außenstelle in Hinxton/UK, dem European Bio-
informatics Institute, Verwendung findet, meist ausschließlich auf den molekularen
Bereich bezogen wird, verdeutlicht grundlegende Defizite in den darüber liegenden
Ebenen, und hier besonders im organismischen Bereich. Ein Grund für das unter-
schiedliche Niveau der Biodiversitätsinformatik in den drei Ebenen molekular,
organismisch und 6kosystemar mag darin begründet sein, dass die erste und die
letzte Ebene stark angewandte Aspekte besitzen. So ist die molekulare Ebene für
medizinische und biotechnologische Industrien von Interesse, während auf der

11

Okosystemebene die Biodiversitätsinformatik zu einer Komponente der Umwelt-
informatik wird. Hingegen wurde es in der Vergangenheit weitgehend versäumt,
die anwendungsbezogenen Aspekte der organismischen Ebene gebührend argu-
mentativ zu vertreten.

In der Schaffung effektiver Strukturen für die Erfassung der Biodiversität und in
der Entwicklung von Verfahren für ihre Analyse liegt die Herausforderung für die
Biodiversitätsinformatik.

3. EXISTIERENDE INTERNATIONALE STRUKTUREN

3.1. Internationaler politischer Rahmen
Die Biodiversitätskonvention

Der globale politische Rahmen für die Biodiversitätsinformatik wird heute weit-
gehend vom Übereinkommen über die biologische Vielfalt (kurz “Biodiversitäts-
konvention”, UN 1992) bestimmt. Die Global Environmental Facility (GEF) als
der Finanzierungsmechanismus der Biodiversitätskonvention wird vom
Entwicklungsprogramm der Vereinten Nationen (UNDP) zusammen mit dem
Umweltprogramm (UNEP) und der Weltbank verwaltet. Zugang zu den Mitteln
haben Entwicklungsländer, Länder Mittel- und Osteuropas und GUS-Staaten.
Deutschland trägt etwa 12% des Gesamtvolumens des Fonds (2 Mrd. US$ für
1995-97). Im Förderbereich Biodiversität wurden zwar bis 1997 bereits rund 450
Mio. US$ bereitgestellt (BMU 1998), eine direkte Finanzierung von dem Gebiet
der Biodiversitätsinformatik zuzurechnenden Projekten fand dabei aber nicht statt.
Viele der genehmigten Projekte haben aber eine Datenverwaltungskomponente.

Der Einfluß der Biodiversitätskonvention ist eher in der Verpflichtung zur Ver-
waltung und Bereitstellung von Biodiversitätsdaten zu sehen, wie sie sich aus
mehreren Artikeln der Konvention und darauf beruhenden Beschlüssen der Ver-
tragsstaatenkonferenz (Conference of the Parties, COP) ergibt. Hierbei lassen sich
zwei Komponenten unterscheiden: Einerseits kommt den Nationalen Beiträgen
zum Clearing House Mechanism eine große Bedeutung zu, die die “wesentlichen
Informationen mit entsprechenden Quellenangaben zum Stand der Umsetzung der
Artikel der Konvention, ihrer Themen sowie den nationalen Rahmenbedingungen”
im jeweiligen Staat dokumentieren sollen (siehe z.B. ZADI 1999). Es handelt sich
hier also zumeist nicht um einen direkten Zugriff auf Biodiversitätsdaten, mit
Ausnahme der Information über als “genetische Ressourcen” im Sinne der Konven-
tion zu klassifizierende Lebendsammlungen (Kultursammlungen von Mikroorga-
nismen, Botanische und Zoologische Gärten, Saatgut- und Sortensammlungen,
Haustierrassen, etc.). Die zweite Komponente ist allgemeiner in der Bereitstellung
von Informationen zur Biodiversität zu sehen, wobei der in der Konvention ge-
forderte Technologietransfer (Nutzung von Biodiversitätsinformation oder von
Komponenten der Biodiversität) zwar oft im Vordergrund steht, es aber auch und
vor allem um die in den entwickelten Ländern vorhandenen Informationen und
Ressourcen zur Inventarisierung und Charakterisierung der Biodiversität ın anderen
Weltteilen geht. Dabei handelt es sich vor allem um Informationen auf der Ebene
der Organismen (siehe 1.1), also die von der systematischen Forschung erarbeiteten

2

Ergebnisse und die in den Forschungssammlungen vorhandenen Belegexemplare,
daneben aber auch um Forschungsergebnisse 1m Natur- und Artenschutz, in der
Biogeographie und in der Okologie (v. a. Tropenökologie).

Im abschlieBenden Bericht der 4. Konferenz der Vertragsstaaten wird erneut die

Bedeutung der elektronischen Erfassung von Biodiversitätsinformation auf der
Ebene der Organismen ausdrücklich hervorgehoben (COP 1998):

... 6. Parties and authorities should utilize information systems to maximum
effect in taxonomic institutions. In developing priority-setting criteria for informa-
tion products, taxonomic institutions should consider the needs of the wide range
of users of that information, including biological diversity managers. In particular,
taxonomic information, literature and checklists should be put into electronic form.

... 9. Government members of the Organization for Economic Cooperation and
Development should endorse and support the recommendations from the OECD
Megascience Forum's Biodiversity Informatics Subgroup, regarding the develop-
ment of a Global Biodiversity Information Facility (GBIF) to allow people in all
countries to share biological diversity information and to provide access to critical
authority files.

Weitere wichtige Abkommen und Konventionen

Neben der Biodiversitätskonvention sind internationale Übereinkommen zum
Natur- und Artenschutz und Übereinkommen zur Erhaltung der genetischen Res-
sourcen zu nennen, für deren Umsetzung globale Netzwerke und Datenbanken
ebenfalls immer mehr als essentielles Instrumentarium begriffen und eingesetzt
werden. Auf Deutschland bezogen sind hier vor allem zu nennen:

Zum Natur- und Artenschutz: Konvention von Ramsar über die Erhaltung der
Feuchtgebiete (UNESCO 1994), Konvention von Helsinki für das Ostseegebiet
(HELCOM 1992), Konvention von Barcelona zum Schutz des Mittelmeers (UNEP
1976), EU Vogelschutzrichtlinie (EU 1979), Berner Konvention über die Erhaltung
der europäischen wildlebenden Pflanzen und Tiere und ihrer natürlichen Lebens-
räume (Council of Europe 1979), Konvention über die Erhaltung der wandernden
wildlebenden Tierarten (UNEP 1979), Alpenkonvention (Anonym 1991), die EU
Flora-Fauna-Habitatrichtlinie (EU 1992) und das Übereinkommen über den inter-
nationalen Handel mit gefährdeten Arten freilebender Tiere und Pflanzen (Was-
hingtoner Artenschutzabkommen, CITES 1973/1979).

Zur Erhaltung genetischer Ressourcen: EG Verordnung vom 20. Juni 1994 über die
Erhaltung, Beschreibung, Sammlung und Nutzung der genetischen Ressourcen der
Landwirtschaft (EU 1994).

3.2. Umsetzung auf internationaler Ebene

Für die Ökosystemebene der Biodiversität existieren bereits zahlreiche interna-
tionale bzw. globale Initiativen und Programme, die eine Datenverarbeitungs-
komponente aufweisen. Diese beziehen sich im wesentlichen auf allgemeine

18

Umwelt- und Naturschutzaspekte und werden im folgenden nur beispielhaft be-
handelt. Ebenso sind internationale IT-Strukturen im molekularbiologischen Gebiet
bereits weit entwickelt (vergl. 1.1). Dieser Bereich wird daher hier nur gestreift,
während der organismischen Ebene besondere Aufmerksamkeit gewidmet wird.

OECD Megascience Forum GBIF

Das Megascience Forum der Organisation für Wirtschaftliche Zusammenarbeit und
Kooperation (OECD) beschloß 1996, eine Expertengruppe’ zur biologischen
Informatik einzurichten, deren Ergebnisse bis 1999 in einem Report niedergelegt
wurden (Anonym 1999) und von den OECD Forschungs- und Technologieminis-
tern bei ihrem Treffen in Paris am 22.-23. Juni befürwortet wurden (OECD 1999).
Eine Global Biodiversity Information Facility (GBIF) soll geschaffen werden, die,
in enger Kooperation mit bestehenden Initiativen, sowohl die Sammlung, Bereit-
stellung und Vernetzung von Biodiversitätsdaten koordiniert als auch Aktivitäten
hinsichtlich der Entwicklung von speziellen Techniken und der akademischen
Entwicklung des neuen Fachgebiets Biodiversitätsinformatik anregen und interna-
tional aufeinander abstimmen soll. In Koordination mit der GBIF sollen dabei auf
nationaler Ebene folgende Aktivitäten im Bereich der Biodiversitätsinformatik
verfolgt werden (It. Anhang des Ministerbeschlusses):

— Interoperabilität von Biodiversitätsdatenbanken, einschließlich der Bereit-
stellung von Daten, Informationen und anderer Ressourcen innerhalb abge-
stimmter Rahmenbedingungen (Urheberrecht!)

— Entwicklung von neuen Benutzeroberflächen

— Standardisierung von Zugangs- und Verknüpfungsmechanismen, einschließlich
der Indizierung, Validierung, Dokumentation und Qualitätskontrolle

— Ermöglichung des Zugangs zu existierenden und neuen Datenbanken

— Entwicklung von Partnerschaften mit existierenden Organisationen und Projek-
ten

— Verbesserung informationstechnischer Infrastrukturen (Breitbandnetzwerke
Cl»)

— Arbeitsteilung bei der Speicherung von Massendaten

— Ausbildungsmaßnahmen im Bereich Biodiversitätsinformatik (Wissenschaftler,
Datenmanager und Techniker)

Seitens der Arbeitsgruppe wurden die Schaffung des taxonomischen Rahmens (vor
allem elektronische Register der Organismen aller Gruppen, Species 2000 In-
itiative) und die Digitalisierung der in biologischen Sammlungen vorhandenen
Information als Prioritäten für die Datenerfassung in der Anfangsphase der GBIF

° Von den Autoren dieses Gutachtens hat Häuser als deutscher Delegierter an der letzen
Sitzung der Arbeitsgruppe teilgenommen, Berendsohn nahm als Delegierter der
Europäischen Kommission an allen Sitzungen teil.

14

definiert. Es ist in politischer Hinsicht bemerkenswert, dass von allen 1996 einge-
richteten Arbeitsgruppen des Megascience Forums nur Biodiversitätsinformatik
und Radioastronomie es bis zu einem Ministerbeschluss gebracht haben.

G7 Environment and Natural Resources Management Project (ENRM)

Die G7-Staaten haben auf dem Gipfeltreffen von Halifax 1995 eine Reihe von
Projekten zur globalen Informationsgesellschaft beschlossen, u.a. auch das
Environment and Natural Resources Management Project (ENRM). Eine Exper-
tengruppe wurde eingerichtet, die in den Jahren 1995 bis 1998 besonders Fragen
internationaler Metadatenstandards erörterte. Ein konkretes Resultat ist der Global
Environmental Information Locator Service (GELOS), eine “virtuelle Bibliothek”
auf dem World Wide Web, von der aus man eine Vielzahl von Umweltinformation
bereitstellenden Datenbanken und Dokumenten erreicht. Der Prototyp des GELOS
ist z.Zt. nicht mehr zugänglich, dient aber als Modell für mehrere in Entwicklung
befindliche Systeme, die miteinander synchronisiert werden sollen (G8 1999).

Initiativen der Vereinten Nationen

Verschiedene Organisationen der Vereinten Nationen unterstützen Programme, die
direkt oder indirekt mit Biodiversitätsinformatik verbunden sind, allerdings fast
ausschließlich auf der Ökosystemebene bzw. im Bereich des Schutzes von Arten
und genetischen Ressourcen. Beispielhaft zu nennen sind hier:

United Nations Food and Agriculture Organization (FAO)

Im Bereich der genetischen Ressourcen sind das Domestic Animal Diversity Infor-
mation System, DAD-IS (FAO 1999a) und das World Information and Early
Warning System, WIEWS (FAO 1999b), im Internet verfügbar. Im Bereich der
globalen geographischen Informationssysteme (Bodenbedeckung, einschl. Vegeta-
tion) ist die FAO z.B. in Afrika mit dem AfriCover Programm (FAO 1998) aktiv,
in dessen Rahmen z.B. eine Software zur Erstellung von Standardklassifikationen
für Bodenbedeckung entwickelt wurde.

United Nations Environment Programme (UNEP)

Das UNEP koordiniert Aktivitäten im Bereich Umweltmonitoring und Global
Change, teilweise in enger Kooperation mit den anderen UN Agenturen. Im Inter-
net verfiigbar sind z.B. das Global Terrestrial Observing System GTOS (UNEP
1999a) und die Global Resource Database GRID (UNEP 1999b).

United Nations Educational, Scientific and Cultural Organization (UNESCO)

Die UNESCO ist vor allem im Zusammenhang mit dem Programm “Der Mensch
und die Biosphäre” (UNESCO 1999) zu nennen, in dessen Rahmen u.a. eine
Datenbank der Biosphärenreservate verfügbar ist und eine umfangreiche Vernet-
zung regionaler und nationaler Initiativen stattfindet. Die dem Bundesamt für
Naturschutz angegliederte Geschäftsstelle des Deutschen Nationalkomitees für
Man and Biosphere (MAB) betreut den deutschen Beitrag zum Programm.

15

Diversitas

Das internationale Programm für die Biodiversitätswissenschaften Diversitas ist
eine 1991 ins Leben gerufene, aus dem wissenschaftlichen Bereich stammende
globale Initiative, die heute allgemein als Dachorganisation für die weltweite
Forschung in diesem Gebiet anerkannt wird. Diversitas formuliert umfassend den
Forschungsbedarf für alle Ebenen und auch hinsichtlich des theoretischen Ver-
ständnisses von Biodiversität. Mehrere Programmpunkte nehmen Bezug auf die
Biodiversitätsinformatik, so wird z.B. betont, dass “die elektronische Erfassung
systematischen Wissens von hoher Priorität ist, da sie als Basis für die Organisation
und Koordination aller anderen Biodiversitatsinformation dient” (übersetzt aus
Diversitas 1996: Core Program Element 3, Inventorying and Classification of
Biodiversity). Der auf den systematischen Bereich bezogene Teil des Programms
beruht im wesentlichen auf der Systematics Agenda 2000, eines anfangs der 90er
Jahre von mehreren hundert Biologen erarbeiteten Grundsatzprogramms für die
Erschließung der Biosphäre. Eine der drei hier formulierten Hauptaufgaben ist die
“Aufbereitung der ... [Forschungs-] Ergebnisse in leicht zugänglicher und abruf-
barer Form ..” (Steininger 1996). Wie die OECD Megascience Forum Arbeits-
gruppe hat auch Diversitas (1999) das globale elektronische Organismenregister
(Species 2000 Initiative, Bisby 1997) als kritische Komponente der Informations-
infrastruktur identifiziert:

Core Program Element 3, Research Components. ... SP2000: There is a critical
need within the global scientific community for a register of all the world's 1.75
million known species. Species 2000, established as an IUBS/CODATA/IUMS
Scientific Programme, is comprised of an international federation of taxonomic
databases, working together to complete this task. The aim of Species 2000 is to
provide a world-wide service in the form of an index to all known animals, plants,
fungi and micro-organisms for use as a baseline reference system for communica-
tions about biodiversity. Data from an array of databases will provide both a
stabilised Annual Checklist, and a Dynamic Checklist accessed electronically.

EU Programme

Als Vertragspartei entwickelt auch die Europäische Union eine Strategie zur
Umsetzung der Biodiversitätskonvention. Gemäß dem Subsidiaritätsprinzip handelt
es sich aber im Ergebnis hauptsächlich um Maßnahmen zur Koordinierung und
Ergänzung von Programmen der einzelnen Mitgliedsstaaten. Unabhängig von der
Konvention spielt die EU jedoch bei der Einrichtung von Organisationen und
Institutionen im Bereich der Umwelt- und Biodiversitätsinformatik eine sehr
wichtige Rolle, allerdings handelte es sich dabei in der Vergangenheit vor allem
um die molekulare und die Ökosystemebene. Im ersten Bereich ist z.B. die Ein-
richtung des European Bioinformatics Institute und die Finanzierung anderer
Aktivitäten im Rahmen des EMBL (European Molecular Biology Laboratory) zu
nennen. Auf der Ökosystemebene ist vor allem die Organisation des sich aus der
Habitatdirektive (EU 1992) ergebenden Natura 2000 Programms und die Ein-

16

richtung und Finanzierung des CORINE (CoORdination of INformation on the
Environment) Programms sowie der Europäischen Umweltagentur (EUA, s. folgen-
der Abschnitt) zu nennen. Das inzwischen ausgelaufene, 1985 initiierte Programm
CORINE der Europäischen Gemeinschaften hatte zum Ziel, Erfordernisse und
angemessene Verfahren für die Zusammenstellung, Koordinierung und Abstim-
mung von Informationen über den Zustand der Umwelt in Europa und der natürli-
chen Ressourcen aufzuzeigen. Methoden und Sachverständigennetze, Datenbanken
und Informationssysteme wurden eingerichtet, spezielle Methoden für Landnut-
zung, Biotope und Emissionen in die Atmosphäre wurden zur Anwendungsreife
gebracht und Datenbanken auf EU-Ebene eingerichtet. Der entstandene Daten-
bestand bildet einerseits ein wesentliches Element des European Environmental
Information and Observation Network (EIONET) der Europäischen Umwelt-
agentur, andererseits steht er für nationale geographische Informationssysteme zur
Verfügung.

Hingegen ergab eine Durchsicht (August 1998) von über 1000 auf dem CORDIS
(Community Research and Development Information Service, EU 1999) Server der
Europäischen Kommission gespeicherten Projektprofile des 4. Rahmenprogramms,
dass in der direkten Forschungsförderung der Europäischen Kommission
Biodiversitätsinformatik auch im weitesten Sinne nur am Rande vertreten ist.
Projekte zur organismischen Ebene finden sich nur in Einzelfällen‘.

Europäische Umweltagentur (EUA)

Die EUA wurde 1990 u.a. als Reaktion auf Informations- und Koordinations-
defizite nach der Katastrophe von Tschernobyl ins Leben gerufen, um Umwelt-
informationen in geordneter, geprüfter und strategisch wirksamer Form zur Verfü-
gung zu stellen. Man konzentrierte sich hierbei vor allem auf die Bedürfnisse des
Umweltrisikomanagements, aber auch auf Informationen hinsichtlich der europäi-
schen Bioregionen und Ökosysteme. Die Finanzierung der EUA erfolgt durch die
EU, für die weiteren Vertragsstaaten durch die EFTA (European Free Trade
Association). Neuerdings sind auch Finanzierungen von Kooperationen mit ande-

* Environment and Climate Programme: Unter mehr als 400 Projekten nur The European Pollen
Database: A tool for paleoenvironment and palaeoclimate reconstructions at a continental scale und
European Diatom Database: an information system for palaeo-environmental reconstruction, daneben
wenige Ökosystem-Modellierungsprojekte. MAST III Programm (Marine Science and Technology):
Unter 139 Projekten nur A register of marine species in Europe to facilitate marine biodiversity research
and management. Biotechnologie-Programm: Dort nur Resource development for a biological collection
information service in Europe, neben einigen molekularbiologischen Bioinformatikprojekten. LIFE
Nature Programm: Unter 400 Projekten nur 2nd phase of the inventory taking and computer mapping
of habitat types and species according to directive 92/43/EEC in Spain. DG VI (Landwirtschaft), im
Bereich Genetische Ressourcen: A programme to conserve, characterise, evaluate and collect Allium
crops and wild species, die Fortsetzung eines Projekts, welches die European Allium Database (EADB)
enwickelt hatte. In den Programmen TERRA (15 Projekte) und ICZM (Integrated Management in
Coastal Zones, 35 Projekte) ist die Biodiversitätsinformatik nicht berücksichtigt.

7

ren Staaten durch andere Mechanismen (z.B. im Rahmen des Phare Programms des
Generaldirektoriats I (DG I) der Europäischen Kommission) vorgesehen.

In Bezug auf Biodiversitätsinformatik relevant sind die Ziele der EUA, den “Zu-
stand des Bodens, der Tier- und Pflanzenarten und der Biotope” zu dokumentieren
(EUA 1994), und die Bekämpfung des Rückgangs der Artenvielfalt (EUA 1998)
als eines der wichtigsten Umweltprobleme in Europa. Die EUA arbeitet mit zahl-
reichen nationalen und internationalen Organisationen zusammen, u.a. existiert
eine formale Kooperation mit dem UNEP.

Als Instrument der Informationserhebung, Bereitstellung und Verbreitung wurde
durch eine Eingangsfinanzierung im Programm Interchange of Data between
Administrations der DG XIII (Directorate General der EU) das Datenmanagement-
netz EIONET (European Environment Information and Observation Network)
geschaffen, welches heute etwa 600 Knoteninstitutionen umfaßt, davon etwa 200
mit aktiver Beteiligung (Saarenma 1998). Der Fokus liegt dabei auf den neun
European Topic Centres und den Beiträgen der 18 National Focal Points. Der
deutsche National Focal Point ıst das Umweltbundesamt in Berlin; als National
Reference Centre for Nature Conservation dient das Bundesamt für Naturschutz in
Bonn. Unter den European Topic Centres der EUA mit biodiversitätsinformati-
scher Relevanz sind drei besonders zu nennen:

1. Das European Topic Centre for Nature Conservation (ETC/NC) mit Sitz in
Paris. Es wird von 15 Mitgliedern aus 12 Staaten gebildet, das deutsche Mit-
glied des Konsortiums ist das Bundesamt fiir Naturschutz. Das ETC/NC ver-
folgt vor allem die Entwicklung der Naturschutzkomponente des EIONET und
unterstützt die EU bei der Entwicklung von Natura 2000 und ergänzenden
Initiativen. Es versucht z.Zt., eine Standardisierung im Bereich der geographi-
schen Information (einheitliches Koordinatensystem für Organismenkartierun-
gen), der roten Listen gefährdeter Arten und der Habitatklassifizierung durch-
zusetzen. Letzteres basiert auf der CORINE Habitat Klassifizierung (siehe
oben); aber nach eigenen Angaben wird auch z.B. mit dem European Vegetati-
on Survey zusammengearbeitet, der zur Zeit an einem European Overview of
Alliances arbeitet. Die Naturschutzinformation soll im European Information
System on Nature (EUNIS) bereitgestellt werden (ETC/NC 1999). Ein Report
aus dem Jahre 1996 (EEA-ETC/NC, 1997) liefert eine Übersicht von 200
Datenbanken mit Informationen zu europäischen Ökosystemen und Arten, die
sich fast alle auf nationale Ressourcen beziehen. Mit 40% der Quellen nehmen
zwar die Artdatenbanken einen großen Teil ein, es handelt sich aber fast aus-
schließlich um Beobachtungsdaten aus Kartierungsprojekten; auf Sammlungs-
belege wird nur ausnahmsweise Bezug genommen.

2. Das European Topic Centre on the Catalogue of Data Sources (ETC/CDS)
mit Sitz in Hannover und Partnern aus Italien, den Niederlanden, Österreich,
Spanien und Schweden. Das ETC/CDS steht unter Federführung des Nieder-
sächsischen Umweltministeriums; weitere deutsche Mitglieder sind das Um-
weltbundesamt (UBA) Berlin, das Forschungszentrum Informatik an der Uni-

versität Karlsruhe und die Lippke & Wagner GmbH, Berlin. In Zusammenar-
beit mit dem bereits erwähnten G7 Environment and Natural Resources Mana-
gement Programme wird ein Metainformationssystem für europäische Um-
weltinformation entwickelt, einschließlich der Software für die Datenerhebung
und Datenverbreitung, der Einrichtung eines multilingualen Thesaurus (GE-
MET) und der Definition von Qualitätskriterien für den Einschluß von Daten-
quellen (ETC/CDS 1999).

3. Das European Topic Centre on Landcover (ETC/LC) hat seinen Sitz am
Environmental Satellite Data Centre (MDC) in Kiruna, Schweden. Es ist ein
Konsortium aus 16 Organisationen aus 15 Staaten; der deutsche Partner ist das
Statistische Bundesamt ın Wiesbaden. Das ETC/LC führt die Arbeit des
CORINE-Landcover Projekts fort und soll dem Benutzer Daten zur europäi-
schen Bodenbedeckung in aktueller Form zur Verfügung stellen (ETC/LC
(S99)

Regierungsunabhängige Organisationen

Eine sehr große Anzahl von internationalen Nicht-Regierungsorganisationen ist im
Bereich des Naturschutzes tätig, und viele dieser Organisationen stellen Daten oder
Computerprogramme zur Verfügung. Hier sind besonders die am Biodiversity
Conservation Information System (BCIS 1999) beteiligten Organisationen zu
nennen, unter anderem BirdLife International, Botanic Gardens Conservation
International, Conservation International, International Species Information
System (Zoos), International Union for the Conservation of Nature (IUCN), The
Nature Conservancy, Wetlands International, World Conservation Monitoring
Centre, und World Wide Fund for Nature (WWF). Im europäischen Bereich exis-
tieren ebenfalls eine große Zahl derartiger Organisationen, das European Centre
for Nature Conservation hat dazu einen Katalog (ECNC 1999) veröffentlicht.

Zusammenfassung

Zusammenfassend läßt sich sagen, dass im Bereich der mit Ökosystemen befaßten
Biodiversitätsinformatik durch die Verzahnung mit der politisch unmittelbar
wichtigen Umweltinformatik und der Existenz internationaler Konventionen bereits
umfangreiche produktive Strukturen existieren, während auf der organismischen
Ebene im wesentlichen Rahmenbedingungen abgesteckt und Prioritäten gesetzt
wurden. Hier ist das konkrete Angebot, mit der Ausnahme von Informationen zu
geschützten und als genetische Ressource wichtigen Arten, noch sehr begrenzt. Es
existieren gegenwärtig zwar eine Reihe von Initiativen und Absichtserklärungen,
aber keine gezielten internationalen Förderungsmechanismen, um diesem Defizit
abzuhelfen.

19

3.3. Globale und europäische Initiativen auf der organismischen Ebene

Nachdem die Informationsverarbeitung auf der Ebene der Organismen als defizitär
herausgestellt wurde, soll hier detaillierter auf einzelne internationale Initiativen
eingegangen werden.

Organismenregister

Als gegenwärtig wichtigste internationale Initiative im Bereich der Bereitstellung
des globalen Organismenkatalogs ist Species 2000 (Sp2000 1999 und Bisby 1997)
zu nennen. Das Programm wurde von der IUBS (International Union of Biological
Sciences), in Kooperation mit CODATA (Committee on Data for Science and
Technology) und der IUMS (International Union of Microbiological Societies) im
September 1994 ins Leben gerufen. Nachträglich wurde es vom UNEP Biodiversity
Work Programme anerkannt, und es ist heute auch mit dem Clearing House Me-
chanism der Biodiversitätskonvention verknüpft. Species 2000 wird von einem
multinationalen Team mit Mitgliedern aus Australien, Brasilien, Großbritannien,
Japan, den Niederlanden, den Philippinen und den USA geführt. Organisatorisch
handelt es sich dabei um eine Föderation von Datenbanken, die jeweils eine be-
stimmte Gruppe von Organısmen abdecken und die über einen gemeinsamen
Abfragemechanismus dynamisch erreichbar sind. Eine aus den Beiträgen kompi-
lierte jährliche Checkliste soll als ein stabiler Artenindex dienen. Ein funktionie-
rendes, auf dem World-Wide-Web verfügbares Testsystem (Species Locator, siehe
unter Sp2000 1999) demonstriert die technische Machbarkeit dieses Konzepts. Der
Species Locator greift gegenwärtig auf folgende Gruppen zu: Bakterien (alle Arten;
Partner: Japan Collection of Microorganisms am Institute of Physical and Chemi-
cal Research [RIKEN] und die Deutsche Sammlung von Mikroorganismen und
Zellkulturen [DSMZ]), Pflanzen (bisher eine große Familie; Partner: ILDIS World
Database of Legumes) und Vertebraten (bisher nur Fische; Partner: Fishbase, eine
vom International Centre for Living Aquatic Resources Management [ICLARM]
in Zusammenarbeit mit mehreren anderen Partnern entwickelte Datenbank). Die
einzelnen Gruppen stellen, bis auf die Leguminosen, selbst wieder Föderationen
von Datenbanken dar.

Insgesamt wurden laut Programmsekretariat bisher weltweit über 150 Datenbanken
für den Einschluß in Species 2000 identifiziert. Bis auf den Beitrag der Deutschen
Sammlung von Mikroorganismen und Zellkulturen (DSMZ 1999a) und der in
Berlin installierten Global Plant Checklist (GPC, Berendsohn 1999b) der Interna-
tionalen Organisation für Pflanzeninformation (IOPI) sind dabei aus Deutschland
bisher nur 6 kleinere entomologische Datensammlungen und die EMBL Living
Reptile Datenbank (s.u.) vorgesehen (F. Bisby, pers. comm. 1998). Der deutsche
Beitrag im Bereich der Mikroorganismen erscheint durch den internationalen
Einfluß der DSMZ gesichert. Im ökonomisch wie ökologisch so bedeutenden
Bereich der Information zu Gefäßpflanzenarten entsteht zur Zeit hingegen eine
internationale Arbeitsteilung, die zur de-facto Belegung bestimmter Gruppen durch
bestimmte Institutionen führt. Ein gewisser deutscher Einfluß ist durch die in
Berlin geleistete Entwicklungsarbeit an der IOPI-GPC gegeben; es ist aber drin-

20

gend ein Beitrag zu fordern, der die Koordination der Information zu einer oder
mehreren bedeutenden Blütenpflanzenfamilien an Institute in Deutschland bindet.
Hier könnte eventuell auf das im Rahmen der Erstellung der Standardliste der Farn-
und Blütenpflanzen (ZfK 1993) geschaffene Expertennetzwerk zurückgegriffen
werden. Im zoologischen Bereich liegen andere Bedingungen vor; während bei den
Wirbeltieren auf Grund der relativ geringen Artenzahlen und verhältnismäßig
geringfügigen taxonomischen Schwierigkeiten ein fast vollständiges Inventar
existiert (Wilson & Reeder 1993), sind bei den Invertebraten die bekannten Arten
relativ schlecht katalogisiert, und in den meisten Gruppen ist (wie bei den Mikroor-
ganismen) der überwiegende Teil der Arten noch nicht einmal bekannt.

Ein vielversprechendes internationales Projekt mit deutscher Beteiligung zur
Schaffung eines Globalen Artenregisters der Tagfalter (ca. 16 000 Arten) wurde im
Oktober 1998 ım Rahmen eines Workshops in Washington initiiert. Die gegenwär-
tige Partnerschaft besteht aus den Museen in London (NHM), Leiden (RMNL),
Canberra (ANIC), Lima (MHNSM), Washington (Smithsonian) und Stuttgart.

Neben den globalen Vorhaben sind mehrere artbezogene biodiversitätsinformati-
sche Projekte auf europäischer Ebene zu nennen. Dabei bestehen oft Synergiebe-
ziehungen mit den globalen Projekten, z.B. sind sowohl die Flora Europaea Daten-
bank (Pankhurst 1999) als auch die Med-Checklist (Greuter et al. 1984, 1986,
1989) bereits in die IOPI-GPC und damit in Species 2000 integriert, und selbstver-
ständlich kann jedes globale Projekt einen Beitrag im europäischen Rahmen leis-
ten. Beispielhaft sollen hier zwei Initiativen genannt werden:

— Das Euro+Med PlantBase Projekt (Jury 1998) strebt eine Konsensustaxonomie
bei den europäischen und mediterranen höheren Pflanzen an, ähnlich der im
Rahmen der Flora Europaea in den 60er und 70er Jahren erreichten Überein-
kunft (die aber selbst Anstoß für intensive Forschungstätigkeit war und daher
aufgearbeitet werden muß). Als Grundlage dienen die existierenden Daten-
banken der Flora Europaea (verwaltet am Royal Botanic Garden, Edinburgh)
und der Med-Checklist (Botanischer Garten und Botanisches Museum Berlin-
Dahlem), die bereits als Basısdatensätze in die IOPI Global Plant Checklist
integriert wurden. Ein von der EU im 5. Forschungsrahmenprogramm ge-
fördertes Projekt soll in den Jahren 2000 bis 2002 die Grundlage für das Projekt
legen. Das Datenbankmodell von IOPI wird auch weiterhin als Grundlage der
Datenbankausgabe dienen, die Software PANDORA (Pankhurst et Pullan 1999)
als Eingabeinstrument benutzt werden.

— Das European Register of Marine Species (ERMS 1998) ist eine von der EU
(DG XII) geförderte konzertierte Aktion, die bis März 2000 eine Checkliste der
Meeresorganismen Europas, verbunden mit bibliographischen Verweisen auf
Bestimmunssliteratur und einer Expertenliste erstellen und im WWW verfüg-
bar machen soll. In einem vom Forschungsinstitut Senckenberg koordinierten
Nachfolgeprojekt sollen die Daten in ein relationales System überführt und
Informationen zu Verbreitung, Typusbelegen und Originalbeschreibungen mit
dem Register verknüpft werden (Türkay, pers. comm. 1999).

2

Fauna Europaea ist eine weitere von der EU im 5. Rahmenprogramm geförderte
Initiative, die das zoologische Gegenstück zur Flora Europaea darstellen soll
(Los, pers. comm.).

Einen Übergang zwischen artbezogenen und belegbezogenen Informationen stellen
die Datenbanken mit chorologischer Information (Daten zur Verbreitungskartie-
rung) dar. Die Projekte weisen einen sehr unterschiedlichen Grad an Vollständig-
keit und Organisation auf. Im europäischen Rahmen sind hier beispielhaft zu
nennen:

Vögel: Die European Union for bird ringing (EURING 1999). In der EURING-
Zentrale in den Niederlanden werden alle europäischen Beringungsdaten (von
insgesamt 33 Beringungszentralen) zusammengeführt und zentral verwaltet.
Alle Beringungszentralen verwenden einen einheitlichen Code für die Art-
namen und erheben die Daten nach identischen Kriterien (Zeitraum zwischen
Beringung und Wiederfund mindestens | Jahr, Wiederfundlokalität mindestens
100 km vom Beringungsort entfernt). EURING beinhaltet 1,2 Millionen Daten-
sätze; bei der weiteren elektronischen Erfassung werden aus osteuropäischen
Staaten grob geschätzt noch zusätzliche 30% an Daten eingebracht werden
(Fiedler, pers. comm. 1998). Außerdem wurde vom European Bird Census
Council EBCC ein Atlas of European Breeding Birds veröffentlicht, in dem die
Präsenzdaten zu 495 Arten in über 4400 50x50 km Quadranten dokumentiert
werden (Hagemeijer & Blair 1997). Dazu liegt auch eine Atlas Datenbank bei
der Organisation SOVON, Niederlande, und eine Bird Database bei Birdlife
International, England, vor (Rheinwald, pers. comm.).

Säuger: Als Resultat des EMMA Projekts (European Mammal Mapping Atlas)
der Societas Europaea Mammalogica wurde der Atlas of European Mammals
(Mitchell-Jones et al. 1999) publiziert, der basierend auf mehr als 93 000
Registrierungen 194 Arten in 50x50 km UTM Quadranten dokumentiert.

Amphibien: Auch die Societas Europaea Herpetologica veröffentlichte als
Ergebnis internationaler Kooperation in Rahmen ihres Mapping Committees ein
Atlantenwerk mit Verbreitungsdaten, den Atlas of Amphibians and Reptiles in
Europe (Gasc et al. 1997).

Wirbellose: Als deutscher Beitrag im European Invertebrate Survey sei hier die
seit Anfang der 70 Jahre erfolgende Molluskenkartierung Deutschlands unter J.
H. Jungbluth (pers. comm.) genannt; Datenhaltung am Universitats-Rechenzen-
trum Heidelberg.

Pflanzen: Ein gedruckter Atlas Flora Europaea (Jalas et al. 1972-1999) wurde
in bisher 12 Bänden veröffentlicht und umfaßt die Resultate einer bis zu 150
Jahre alten Registrierungsstradition. Seit 1992 arbeitet man an der Einrichtung
einer Atlas Flora Europaea database; die Karten der veröffentlichten Bände
wurden inzwischen erfaßt (Lampinen 1999). Ein umfangreicher deutscher
Beitrag wurde im Rahmen der Florenkartierung Deutschlands (Datenbank
Gefäßpflanzen, s.u.) geleistet.

DD

Biologische Sammlungen

Unter dem Begriff biologische Sammlungen werden hier sowohl die naturkundli-
chen Forschungssammlungen in Museen und Universitäten als auch Lebendsamm-
lungen, also Kultursammlungen von Mikroorganismen sowie botanische und
zoologische Garten, verstanden.

Im globalen Rahmen gibt es eine biodiversitätsinformatische Grundstruktur nur bei
den zoologischen Gärten, die mit ihrem International Species Information System
(ISIS) etwa die Hälfte aller anerkannten Zoos und Aquarien weltweit abdecken
(annähernd 500 Institutionen aus 54 Ländern). ISIS stellt Software für die Ver-
waltung der Bestände und artenschutzorientiertes Sammlungsmanagement zur
Verfügung und faßt die so gewonnene Information im Netzwerk zusammen. Von
etwa 250 000 lebenden Exemplaren (6000 Arten) und etwa 750 000 ihrer Vorfah-
ren sind Informationen vorhanden. Alle gespeicherten Wirbeltiere der beteiligten
Zoos der Welt lassen sich im World Wide Web abfragen; geboten werden wissen-
schaftliche Namen (Gattung/Art), englische Trivialnamen und die Anzahl Männ-
chen/Weibchen pro Zoo (ISIS 1999). Noch umfangreichere Daten stehen als [S/S
Specimen Reference CD-ROM zur Verfügung (Information zu 1 200 000 Akzessio-
nen von 7500 Arten in über 500 Zoologischen Gärten, einschließlich historischen
und Stammbaumdaten).

Eine vergleichbare Zusammenarbeit gibt es bei anderen Sammlungsarten nicht,
allerdings sind vielfach positive Ansätze zu vermerken. Die großen kommerziellen
mikrobiologischen Sammlungen Europas haben im Rahmen des CABRI (Common
Access to Biotechnological Resources and Information) Projekts ein gemeinsames
Katalogsystem aufgebaut (CABRI 1999). Auch die großen naturkundlichen Mu-
seen Europas haben ein Konsortium gebildet, um die Erschließung der Sammlungs-
information gemeinsam zu verfolgen (CETAF, Consortium of European large-
scale TAxonomic Facilities). Aus Deutschland sind hier vertreten: Forschungs-
institut und Naturmuseum Senckenberg in Frankfurt, Naturhistorisches Museum
der Humboldt-Universität Berlin und Botanischer Garten und Botanisches Museum
Berlin-Dahlem. BioCISE (Resource Identification for a Collection Information
Service in Europe), ein derzeit (bis Ende 1999) unter deutscher Federführung
durchgeführtes EU-Projekt, hat sich zum Ziel gesetzt, die existierenden
biodiversitätsinformatischen Ressourcen in europäischen Sammlungen und Kartie-
rungsprojekten zu identifizieren. Darauf aufbauend sollen Projektanträge formuliert
werden, mit dem Ziel, eine gemeinsame WWW-Schnittstelle zu allen diesen
Ressourcen bereitzustellen (BioCISE 1999).

Nomenklatur

Eine der wichtigsten internationalen Grundstrukturen der biologischen Wissen-
schaften ist die Nomenklatur, also die Regeln, nach denen Organismen wissen-
schaftlich benannt werden. Diese sind in den sogenannten Nomenklaturcodes
festgelegt, und zwar getrennt für die Tiere (Ride et al. 1999), Pflanzen (einschließ-

23

lich der Pilze, Greuter et al. 1993), Bakterien (Sneath 1992), Viren (Francki et al.
1990) und Kulturpflanzen (Trehane et al. 1995). Voraussetzungen für die Anwend-
barkeit eines Namens sind z.B. die korrekte Form der Veröffentlichung, die Ein-
deutigkeit sowie die Hinterlegung sogenannter Typus-Exemplare im Rahmen der
Neubeschreibung. Nomenklatorische Listen mit den Namen und Veröffentli-
chungsdaten sind von großer Bedeutung, um Fehlbenennungen oder Fehlanwen-
dungen von Namen zu vermeiden. In der Zoologie veröffentlicht Biosis den seit
1865 geführten Zoological Record, der neuerdings (für Abonnenten) auch in
elektronischer Form zur Verfügung steht (Biosis 1999). Von Interesse in diesem
Zusammenhang ist übrigens, dass die Einnahmen nur etwa 20% der Kosten decken,
finanziert wird Zoological Record 1m wesentlichen aus den Gewinnen, die Biosis
aus dem Verkauf der Biological Abstracts erzielt (M. Dadd, pers. comm., 1998).

In der Botanik hat bei den Gefäßpflanzen traditionell der vom Botanischen Garten
Kew bei London herausgegeben /ndex Kewensis diese Rolle gespielt, der inzwi-
schen als Datenbank-CD verfügbar ist. Eine über das World Wide Web frei zu-
gängliche nomenklatorische Datenbank für Pflanzen (PNP 1999) entsteht gegen-
wärtig im Rahmen des /nternational Plant Name Index Projekts (IPNI), das die
vorhandenen Daten aus dem Index Kewensis, dem Gray Card Index der Harvard
University (nordamerikanische Pflanzennamen) und des Australian Plant Name
Index zusammenfaßt und in einem Kontributionssystem fortlaufend aktualisiert
werden soll (Croft et al. 1999). Das Projekt arbeitet mit dem IOPI Global Plant
Checklist Project zusammen, das neben den in IPNI behandelten nomenklatori-
schen Daten auch die Umschreibung der mit den Namen belegten Pflanzengruppen
behandelt (aufgrund der Nomenklaturregeln ist es durchaus möglich, dass in ihrer
Umschreibung sehr unterschiedliche Gruppen mit demselben Namen belegt wer-
den). Die IOPI Liste berücksichtigt außerdem auch falsch angewandte Namen, die
in Standardwerken wie Roten Listen, landwirtschaftlichen Publikationen etc.
verwendet werden.

Für die Pilze existiert der Index of Fungi, der von CAB International (CABI) in
Großbritannien vermarktet wird.

In der Botanik gibt es auch Bestrebungen, die Registrierung von Pflanzennamen
obligatorisch zu machen; dies wurde aber vom letzten Internationalen Botanischen
Kongreß (1999 in St. Louis) nach einer sehr kontrovers geführten Debatte abge-
lehnt. Aus dem von der International Association of Plant Taxonomists in Berlin
erstellten Datenbankprototyp werden nun komplementär zum IPNI Projekt die
Indizierung der Algen und der fossilen Pflanzen ausgegliedert und als Indizierungs-
projekte weitergeführt.

In Berlin wurde auch ein Projekt ins Leben gerufen, die gebräuchlichen Namen
(Names in current use) zu dokumentieren und ihre gültige Publikation nach den
Regeln des Code zu überprüfen (z.B. Greuter et al. 1993). Eine Sanktionierung
solcher Listen durch den Botanischen Kongress, die bislang ebenfalls keine aus-
reichende Unterstützung gefunden hat, könnte Namensänderungen aufgrund
entdeckter älterer Quellen überflüssig machen. Für die Bakterien ist eine solche
Liste (Approved List of Bacterial Names) im International Code of Nomenclature

24

of Bacteria bereits festgelegt worden; diese wird von der DSMZ in Braunschweig
als Bacterial Nomenclature up-to-date auf dem WWW zur Verfügung gestellt
(DSMZ 1999b).

3.4. Standardisierung

Eine globale Informationsstruktur setzt eine gewisse Standardisierung voraus.
Bestes Beispiel dafür ist das Internet und seine verschiedenen Protokolle (tcp/ip,
ftp, http etc.), das auch ein Vorbild für die Möglichkeit der weitgehenden Selbst-
organisation solcher Standards setzt.

Biodiversitätsinformatische Strukturen setzen auf diesen technischen Standards der
Informationsübermittlung auf. Aber auch im Bereich der Datenformate und Daten
selbst sind Standards und Normen Voraussetzung für ein reibungsloses Zusammen-
spiel verschiedener Informationsquellen. Hierbei muß man drei Bereiche unter-
scheiden: (1) Datenstrukturinformation, d.h. Felddefinitionen und Informations-
modelle, (2) Standarddatenkataloge und Thesauri und (3) Metadaten (Daten über
Daten).

Pionierarbeit, besonders für botanische Datenbanken, wurde von der 1985 ins
Leben gerufenen Taxonomic Databases Working Group (TDWG) geleistet. Dieser
Arbeitsgruppe gehört ein großer Teil der wichtigen naturkundlichen Institutionen
weltweit an und sıe findet sich jährlich zu einem Arbeitstreffen zusammen. In
mehreren Untergruppen wurden Datenaustauschstandards, z.B. in Form von
Datenbank-Felderlisten für Herbarien, botanische Namen, allgemein taxonomische
Daten und Akzessionen in Botanischen Gärten entwickelt und von der TDWG
anerkannt. Diese Datenstrukturstandards finden heute weite Anwendung bei der
Konzeption von biologischen Datenbanken. Auch mehrere Standarddatenkataloge
wurden verabschiedet, so Standardabkürzungen für Zeitschriften, Autorenzitate
wissenschaftlicher Namen, geobotanische Regionen und Präsenzdaten. Die TDWG
hat inzwischen ihren Wirkungskreis auf den Bereich der gesamten Biologie ausge-
dehnt (TDWG 1999).

Neben zahlreichen veröffentlichten Modellen implementierter Systeme existieren
relativ wenige theoretische Informationsmodelle. Die US-amerikanische Associati-
on of Systematics Collections veranstaltete 1992 einen Workshop, aus dem ein
Kernmodell (das ASC Model, ASC 1993) hervorging, welches in seinen Grundzü-
gen heute von vielen Datenbanksystemen in der Systematik eingesetzt wird. Als
Resultat zweier von der DG-XII zwischen 1993 und 1999 finanzierten Projekte
wurde ein allgemeines Informationsmodell für Biologische Sammlungen (Berend-
sohn et al. 1999) publiziert. Wie bereits erwähnt, existiert im Zusammenhang mit
dem IOPI Projekt ein allgemeines Informationsmodell für botanische Taxa, wel-
ches unschwer in Richtung eines auch die Zoologie abdeckenden Modells erweitert
werden kann (das /OPI Model, Berendsohn 1997).

Im Zuge der zunehmenden Datenverfügbarkeit und der Vernetzung verschiedener
Datenbanken werden standardisierte Metadaten, also Daten zur Qualitäts-,
Herkunfts- und Aktualitätsbezeichnung vorhandener Daten, immer wichtiger. Dies

25

ist ein weit über den Bereich der Biodiversitätsinformatik hinausgehendes Problem,
und es können durchaus bestimmte allgemeine Standards wie z.B. der Dublin Core
(Weibel et al. 1998) ganz oder teilweise übernommen werden.

Zu Standards, Modellen und Metadaten, die im Rahmen von
Sammlungsinformationssystemen wichtig sind, existiert eine Referenzliste der
TDWG Subgroup on Accession Data (in Zusammenarbeit mit dem BioCISE
Projekt) auf dem WWW (Berendsohn 1999a).

4. STRUKTUREN IN DEUTSCHLAND
4.1. Umsetzung internationaler Übereinkommen

Die in Abschnitt 3.1 genannten Konventionen und EU-Richtlinien sind vor allem
auf nationaler Ebene umzusetzen. Zur Erfüllung der damit verbundenen Informa-
tionspflichten werden im Bereich des Arten- und Biotopschutzes generell die
Umweltinformationssysteme (s. Abschnitt 4.2) benutzt oder benutzt werden. Im
Rahmen der Biodiversitätskonvention wurde vom Bundesministerium für Umwelt
ein Projekt zur Einrichtung des Deutschen Clearinghouse Mechanismus (CHM) auf
dem WWW finanziert, welches zur Zeit an der Zentralstelle für Agrardokumentati-
on und -information (ZADI) des BML eingerichtet ist. Hier wird, neben der Dar-
stellung der Konvention und ihrer politischen Umsetzung, Zugang zu Information
über genetische Ressourcen geschaffen (s. Abschnitt 4.3). Abgesehen davon gibt es
auf der Ebene der Organismen erst seit 1999 einen Ansatz zu einem Förderpro-
gramm, welches im Sinne der Biodiversitätskonvention die Erschließung der in
Deutschland vorhandenen Information zur globalen Biodiversität über den Einsatz
von Informationssystemen zum Ziel hat (Biolog-Programm des Bundesministeri-
ums für Bildung und Forschung).

4.2. Umweltinformationssysteme
Organisation von Umweltinformationssystemen

Wie auf der europäischen Ebene im Rahmen der EUA, so sind auch in Deutschland
die öffentlichen Einrichtungen auf Bundes-, Länder- und kommunaler Ebene vor
allem mit Umweltinformationssystemen befaßt. In diesem Bereich werden sub-
stanzielle Förder- und Unterhaltungsmittel aufgebracht. Die meisten (teilweise
schon seit den 70er Jahren laufenden) Projekte befassen sich mit Umweltinformatik
im weiten Sinne (Gewässerschutz, Umweltverträglichkeitsprüfung, Abfallver-
wertung, Toxine, radioaktive Belastung usw.). Biodiversitätsbezogene Daten
spielen insgesamt meist eine eher untergeordnete Rolle, in der Regel handelt es
sich um Daten auf der Ökosystemebene (Biotopaufnahmen) oder um Daten im
Rahmen des Artenschutzes. So ergab z.B. eine Nachfrage im Umweltministerium
des Landes Nordrhein-Westfalen im August 1998, dass das dortige Umweltinfor-
mationssystem (UIS) im Hinblick auf Biodiversität zwar Standorttypen, Biotop-
ausdehnungen und Arten der roten Listen erfaßt, ansonsten aber keine biodiversi-
tätsbezogenen Daten berücksichtigt. Nach Durchsicht der auf dem WWW publik

26

gemachten Unterlagen des Bund-Länder Arbeitskreises Umweltinformations-
systeme ist davon auszugehen, dass dies auch auf die Systeme in den anderen
Bundesländern zutrifft.

Innerhalb der zur Zeit laufenden Koordinierungsbestrebungen sind aber durchaus
Ansätze vorhanden, im Rahmen der UIS eine verstärkte Einbeziehung von auf
Deutschland beschränkter allgemeiner Information zur Biodiversität zu bewerkstel-
ligen, sei es durch deren direkte Integration oder durch eine Einbeziehung als
beigeordnetes Fachinformationssystem.

Die durch die föderalen Strukturen bedingte starke Zersplitterung der Zuständig-
keiten hat zweifellos zu erheblichen Parallelentwicklungen geführt. So gibt es z.B.
sowohl ın den Bundesländern als auch auf Bundesebene zentrale Umweltinforma-
tionssysteme, die bei oft gleichen Ansprüchen zum Teil erhebliche konzeptionelle
Unterschiede aufweisen (siehe Projekt “Dokumentation der Umweltinformations-
systeme des Bundes und der Länder”, früher “Fortlaufende Bestandsaufnahme von
UIS-Konzepten in Bund und Ländern”, im Auftrag des Bund-Länder-Arbeits-
kreises Umweltinformationssysteme [Page et al. 1996]). Zusätzlich werden vielfach
kommunale Umweltinformationssysteme installiert, teilweise mit recht guter
finanzieller und personeller Ausstattung. So ist zum Beispiel aus dem Kommunalen
Umweltbeobachtungs- und Informationssystem (KUBIS) der Stadt Bonn heute
(Oktober 1999) das Umweltinformationssystem UIS (auf WINCAD-SD-Basis)
geworden. Mit dem Programm sollen alle umweltbezogenen Daten (mit geographi-
scher Lage) zusammengeschlossen, verwaltet und, wenn möglich, visualisiert
werden. Dabei werden auch Altstandorte mit möglicher Umweltbelastung berück-
sichtigt (Zeitachse). Die Primärdaten für die Biotopkartierung wurden teils durch
Mitarbeiter der Universität Bonn, zum größten Teil aber von Mitarbeitern der
Landesanstalt für Ökologie und Landschaftsplanung (LÖLF, heute LÖBF) erhoben.
Der Stellenwert des UIS in der Stadtverwaltung ist offenbar relativ hoch; vier
permanente Mitarbeiter betreuen das UIS (Datenbankpflege, Programmanpassung,
Berichte etc.), während Abfragen und auch Dateneingaben durch die in der Stadt-
verwaltung verteilten Sachbearbeiter erfolgen. Zunehmend wird auch die
Intranet/Internet-Technologie für die verwaltungsweite Verbreitung der Umwelt-
informationen genutzt.

Kommunale Umweltinformationssysteme sollen künftig verstärkt in lokale Agenda
21 Prozesse mit einbezogen werden. Das BMU/UBA ist hier unterstützend tätıg
und bietet diverse Informations- und Beratungsprodukte an: 1998 erschien ein
Handbuch mit Praxisanleitungen zur Durchführung lokaler Agenda 21 Prozesse
(BMU, UBA 1998) und 1999 ein europaweiter Vergleich (BMU, UBA 1999).
Schließlich wird ein Literatur- und Adressen-Wegweiser (BMU 1997) fortgeschrie-
ben, der in Kürze mit aktuellen und nützlichen Internet-Adressen neu herausgege-
ben werden soll.

Es gibt aber auch zahlreiche Bestrebungen, durch gemeinsame Planungen und
Entwicklungen in diesem Bereich Synergieeffekte auszunutzen und Doppelent-
wicklungen zu vermeiden. In diesem Sinne ist z.B. der Bund-Länder-Arbeitskreis
Umweltinformationssysteme zu nennen, ein von der Umweltministerkonferenz

2A

Anfang der 80iger Jahre eingerichtetes Gremium der für Umwelt zuständigen
Ministerien der Länder und des BMU (Leitung BMU, beratende Mitglieder aus
UBA und BfN). Auf Bundesebene (BMU/UBA) liefert das als Auskunftssystem für
die Belange des UBA und des BMU konzipierte UmweltPLanungs- und Informa-
tionsSystem des Umweltbundesamtes (UMPLIS) umfangreiche Hintergrunddaten
auch für die Ländersysteme. Hierzu gehören auch die Umweltforschungsdatenbank
(UFORDAT), Umweltliteraturdatenbank (ULIDAT) und die Umweltrechtsdaten-
banken (URDB) am UBA.

Die LANA (Länderarbeitsgemeinschaft für Naturschutz, Landschaftspflege und
Erholung) besteht aus den obersten Naturschutzbehörden aller Bundesländer und
dem BfN als Vertretung des Bundes. Die LANA dient der Zusammenarbeit von
Bund (BMU) und Ländern in Fragen des Naturschutzes und der Landschaftspflege
(z.B. 1m Rahmen des Bundesnaturschutzgesetzes, der Entwicklung von Biodi-
versitätsstrategien, etc.). Das BfN als deutsche Kontaktstelle Naturschutz zum
europäischen Themenzentrum Naturschutz in Paris (ETC/NC, siehe unter Ab-
schnitt 3.2, EUA) soll in und mit der LANA den Informationsfluß zwischen Bun-
desländern und der europäischen Ebene sichern.

Mit dem LANIS-Bund (Landschafts- und Naturschutzinformationssystem) des BEN
soll ein internes Fachinformationssystem entstehen, das den Zugriff auf relevante
Informationen aus den Fachgebieten und Arbeitsbereichen des BfN realisiert
(geographische Daten, Naturschutzgebiete, Biotoptypen, in Zukunft auch Zugriff
auf einzelne Datenbanken wie z.B. der Datenbank Gefäßpflanzen und diverser
Standard- und Referenzlisten, also Rote Listen, Master-Listen bestimmter Gruppen
etc.). Ein externer Zugriff auf Datenbestände des BfN soll über drei Schienen
realisiert werden: (1) statische Webseiten unter der Homepage des BfN, (11) Beiträ-
ge zum German Environmental Information Network (GEIN), u.a. durch Zugriff
auf Datenbanken und thematische Karteninhalte mit einem Web-fähigen GIS-Tool,
und (iii) Beiträge des BfN zum Bundesinformationssystem Genetische Ressourcen
(BIG), so z.B. Zugriffe auf relevante Datenbankinhalte und thematische Karten für
das Thema einheimische Gefäßpflanzen (R. May, pers. comm.).

Im Sinne der Biodiversitätsinformatik bedeutsam sind im Rahmen der genannten
Aktivitäten vor allem die Daten zum Natur- und Artenschutz sowie die Floren- und
Faunenkartierung (s.u.). Die federführende Rolle spielt hier zumeist das BMU und
vor allem das BfN, aber viele der Zuständigkeiten, insbesondere ım Vollzug, liegen
bei den Ländern. So liegen z.B. im ALBIS (Arten-, Landschafts- Biotop- Informa-
tionssystem) bei der Landesanstalt für Umweltschutz Karlsruhe die Daten aus ca.
35 000 Biotopkartierungen in Baden-Württemberg vor. In der Landesanstalt für
Ökologie, Bodenordnung und Forsten/Landesamt für Agrarordnung des Landes
Nordrhein-Westfalen werden im Rahmen der Biotopkartierung artbezogene Daten
erhoben; eine Datenbank ist im Aufbau. Das im Aufbau begriffene sog. FOGIS
(Forstliches Geographisches Informationssystem) der Landesforstverwaltung
Baden-Württemberg (auch in Thüringen und Niedersachsen) soll ebenfalls Biotop-
daten enthalten.

28

Eine interessante Mischung aus UIS-Komponenten und umweltinformatischer
Forschung stellt das am Fraunhofer Institut für Informations- und Datenverarbei-
tung in Karlsruhe realisierte Elbe-Ökologie-Informationssystem ELISE (IITB
1998) dar. Im Rahmen des Forschungsverbunds Elbe-Ökologie, einer Fördermaß-
nahme des BMBF, wird ELISE an der Bundesanstalt für Gewässerkunde aufge-
baut, um den Informations- und Datenaustausch der beteiligten Projekte zu unter-
stützen. Auch das LOTSE (Land Ocean Thematic Search Engine) Projekt am
GKSS Forschungszentrum Geesthacht fällt in diese Kategorie (GKSS 1999). Über
LOTSE können die Daten verschiedener Einzelprojekte zum Bereich küstennaher
Gewässer eingesehen werden; z.B. Sensitivitäts- und Schadstoffkartierung des
gesamten deutschen Wattenmeeres oder Ökosystemforschung Wattenmeer in
Schleswig-Holstein bzw. in Niedersachsen - jeweils unterteilt in strukturelle und
funktionale Ökosystemforschung.

Eine besonders vielversprechende Koordinierungsmaßnahme stellt der
Umweltdatenkatalog UDK dar. Der UDK wurde 1991-1995 im Niedersächsischen
Umweltministerium im Rahmen eines vom Umweltbundesamt geförderten For-
schungsvorhabens entwickelt. Seit Januar 1996 ist eine von 13 Bundesländern
unterzeichnete Bund-Länder-Verwaltungsvereinbarung in Kraft, die die Weiter-
entwicklung, Pflege und Einführung des UDK zum Ziel hat. Auf der Grundlage
dieser Vereinbarung wurde 1m Niedersächsischen Umweltministerium die Koordi-
nierungsstelle UDK eingerichtet. Seit 1993 besteht eine enge Kooperation mit
Österreich, wo seit 1.1.1995 der Einsatz eines Umweltdatenkataloges gesetzlich
vorgeschrieben ist.

Das 1994 begonnene Projekt WWW-UDK stellt die Daten des UDK im World
Wide Web zur Verfiigung (Technik: Nikolai et al. 1999, Geschichte: Swoboda et
al. 1999). Wie das vom ETC/CDS (s. Abschnitt 3.2) entwickelte System (mit dem
zusammengearbeitet wird) handelt es sich um ein objektorientiertes Metainforma-
tionssystem, welches Information über die Struktur und Beschaffenheit (Gültigkeit,
Fehlertoleranz etc.) von Datenbeständen bereitstellt. Gleichzeitig soll es eine
benutzergerechte Navigation innerhalb dieser Informationen und ggf. auch einen
direkten Zugriff auf gefundene Datenbestände ermöglichen.

Floren- und Faunenkartierung und Listen, Beringung’

Die wohl umfangreichsten Einzeldatensammlungen zur Biodiversitätsinformatik in
Deutschland liegen im Bereich der nationalen Floren- und Faunenkartierung und
Beringung vor, und zwar sowohl auf Bundes- als auch auf Länderebene. Diese
Datenbestände illustrieren bereits den Nutzen einer breiten Informationsbasis auf
organismischer Ebene, sowohl für Naturschutzbelange als auch besonders für die
wissenschaftliche Auswertung. Hier ist zu nennen:

° Das primäre Ziel von Beringungsdatenbanken ist zwar nicht die Faunenkartierung, die auf
Zeitpunkt und Fundort bezogenen Daten können aber wie Kartierungsdatensätze benutzt
werden.

29

Gefäßpflanzen: Am BfN wird, gefördert durch Mittel aus dem Umweltforschungs-
plan des Bundesministeriums für Umwelt, die “Datenbank Pflanzen” aufgebaut.
Neben Charakterisierungsdaten zur Biologie und Ökologie sowie zur Gefährdung
(Rote Listen) von Gefäßpflanzenarten enthält diese Datenbank als Kernbestand
(derzeit ca. 14 Millionen Datensätze) die Basisinformation zur Verbreitung und
Bestandssituation der wildwachsenden Gefäßpflanzenarten in Deutschland (May
1994). Die Daten stammen aus regionalen Projekten (teilweise auf Länderebene,
teilweise aus privaten Erhebungen), die von der Zentralstelle für die floristische
Kartierung Deutschlands organisiert und betreut werden. So wird den regionalen
Projekten das PC-Programm FlorEin kostenlos für den Aufbau und die Auswertung
eigener Kartierungs-Datenbanken zur Verfügung gestellt (May & Subal 1992). Die
Zentralstelle gliedert sich entsprechend der Geographie des Bundesgebietes und der
Aufgabenverteilung in die vier Zentralstellenbereiche Nord (H. Haeupler, Bo-
chum), Ost (E. Jäger, Halle), Süd (P. Schönfelder, Regensburg, projektfederfüh-
rend) und Zentrale Datenbank (R. May, BfN). Leider ist diese Datenbank bislang
nicht im Internet zugänglich, im Rahmen der Beiträge des BfN zum GEIN und zu
BIG (s.o.) soll aber im Jahre 2000 auf die Verbreitungsinformationen und weitere
Charakterisierungsdaten von Gefäßpflanzen in der Datenbank per Internet zu-
gegriffen werden können (R. May, pers. comm.). Darüber hinaus existieren mehre-
re Florenkartierungsprojekte auf Länderebene, z.B. eine Datenbank sämtlicher
bereits elektronisch erfasster Daten der floristischen Kartierung Baden-Württem-
bergs im Rahmen des Arten-, Landschafts-, Biotop- Informationssystems (ALBIS)
an der Landesanstalt für Umweltschutz Karlsruhe (100 000 Datensätze Moose,
80 000 Flechten; 800 000 Phanerogamen). Die Daten sollen künftig über den
Umweltdatenkatalog im Netz verfügbar gemacht werden. Als Resultat der Floren-
kartierung wurde auch eine Standardliste der Farn- und Blütenpflanzen Deutsch-
lands (ZfK 1993) veröffentlicht.

Mollusken: Eine umfangreiche Datenbank zur Molluskenkartierung Deutschlands,
die bereits als Basis mehrerer Publikationen diente (gedruckte Schnecken-Faunen
oder -Atlanten u.a. für Baden-Württemberg, Rheinland-Pfalz und Hessen). Leiter
der Projektgruppe: H. Jungbluth; Datenverarbeitung am Universitätsrechenzentrum
Heidelberg.

Fledermäuse: Eine zentrale Datenbank für Fledermausberingungen in Deutschland
(BatRing) mit über 60 000 Datensätzen, gefördert vom BfN. Die Leitung hat R.
Hutterer (ZFMK Bonn). Die finanziellen Mittel werden Ende 1999 verbraucht sein;
voraussichtlich sind dann mehr als 60% von insgesamt 100 000 Datensätzen
eingegeben. Es werden alle deutschen Fledermausarten berücksichtigt.

Vögel: Mehrere Vogelberingungsdatenbanken (Wiederfunde beringter Vögel), in
den Vogelwarten Helgoland (114 000 Datensätze), Radolfzell (42 000) und Hid-
densee (240 000). Die Datenstruktur ist einheitlich und richtet sich nach EURING
(s. Abschnitt 3.3). Der Dachverband Deutscher Avifaunisten unterhält eine Daten-
bank der Verbreitung und Häufigkeit der Brutvögel von Deutschland (Rheinwald
1993) und eine Datenbank mit den Daten des Bestandsmonitorings häufiger Brut-
vögel Deutschlands für den Zeitraum 1989-1999 (G. Rheinwald, pers. comm.).

30

Wandernde Tierarten: Ein Datenbankprojekt zur Erstellung eines Global Register
of Migratory Species (GROMS) im Sinne der Konvention von Bonn (Riede 1999).
Es handelt sich um ein Kooperationsvorhaben zwischen dem Zentrum für Entwick-
lungsforschung der Universität Bonn und dem Zoologischen Forschungsinstitut
und Museum Alexander Koenig (ZFMK), Bonn. Dieses vom BMU finanzierte
Projekt erscheint aufgrund der Anwendung eines Populationskonzepts und dem
Datenbank-gekoppelten GIS Einsatz besonders interessant.

Die Kartierungsprojekte gehen zwangsläufig mit der Erstellung von Standardlisten
der betroffenen Organismengruppe für die kartierte Region einher. Solche Listen
werden aber auch ohne Verbindung zu Kartierungsprojekten erstellt, teilweise mit
interessantem biodiversitätsinformatischen Hintergrund. So fördert das BMU z.B.
ein Kollaborationsprojekt zwischen der Universität Göttingen und dem BfN zur
Erstellung einer nomenklatorischen und taxonomische Referenzliste der Moose
Deutschlands (Koperski et Sauer 1999). Als Werkzeug wurde das Datenbank-
programm TAXLINK - entwickelt, eine Datenbank zur Verwaltung unterschiedli-
cher taxonomischer Auffassungen, mit dem die Behandlung der verschiedenen
existierenden taxonomischen Konzepte auf Art- und Gattungsebene im Sinne des
IOPI Datenmodells (Berendsohn 1995, 1997) praktisch angewandt und evaluiert
werden.

Die meisten sogenannten Roten Listen gefährdeter bzw. geschützter Pflanzenarten
liegen als Datenbanken vor. Am BfN wurde die Datenbank WISIA (Wissenschaft-
liches Informationssystem im Internationalen Artenschutz) entwickelt, in der
wissenschaftliche Angaben über die gesetzlich in nationalem und internationalem
Rahmen (Convention on International Trade on Endangered Species |CITES],
Bundesnaturschutzgesetz, Berner Konvention, Flora-Fauna-Habitat-Richtlinie)
geschützten Arten gespeichert sind. Die Datenbank soll vor allem den Vollzug der
gesetzlichen Vorgaben erleichtern.

4.3. Informationssysteme zu genetischen Ressourcen in Deutschland

Die in Land- und Forstwirtschaft und in der Tierzucht zum Einsatz kommenden
genetischen Ressourcen sind von erheblicher wirtschaftlicher Bedeutung. Die
direkte Förderung von Biodiversitätsprojekten durch das Bundesministerium für
Landwirtschaft (BML) konzentriert sich auf diesen Bereich, oft in Zusammenarbeit
mit dem BMU bzw. der BfN. So fördert das BML z.B. OEKOGEN - ein von der
Bundesforschungsanstalt für Forst- und Holzwirtschaft (BFH) in Großhandorf
entwickeltes Simulationsprogramm zur Modellierung des Einflusses verschiedener
Maßnahmen auf die Biodiversität (BMU 1998). Ein weiteres Projekt befaßt sich
mit der Abschätzung der genetischen Vielfalt der europäischen Rotbuche Fagus
sylvatica. Hier sollen vom Institut für Forstgenetik und Pflanzenzucht rund 35 000
Individuen aus über 100 Beständen untersucht werden. Daneben existieren zahlrei-
che Datenbanken zu wichtigen Kulturpflanzen bzw. Sorten.

Das vom Informationszentrum für Genetische Ressourcen (IGR) in der Zentral-
stelle für Agrardokumentation und -information (ZADI) des BML entwickelte

31

Informationssystem Genetische Ressourcen (GENRES) ist ein beispielhaftes
WWW Informationssystem (IGR 1999). GENRES soll eine Verbindung zwischen
den beim IGR zentral gespeicherten Meta-, Fakten- und Auswertungsdaten zu
genetischen Ressourcen in Deutschland und den dezentralen Datenbeständen, die
in den an GENRES mitwirkenden Einrichtungen vorhanden sind, herstellen. Im
Zugang getrennt nach den vier Bereichen pflanzengenetische, tiergenetische,
forstgenetische und mikrobielle Ressourcen bietet es zahlreiche Verweise auf
Datenbanken, Projektinformationen, Institutionen und Dienstleistungen. Es liefert
auch Informationen über deutsche, europäische und internationale Maßnahmen zur
Erhaltung und nachhaltigen Nutzung genetischer Ressourcen für die Ernährung,
Land- und Forstwirtschaft. Hier finden sich auch Verweise auf Bund-Länder-
Kooperationsgremien, wie die in den 80er Jahren gebildete Bund-Länder-Arbeits-
gemeinschaft “Erhaltung forstlicher Genressourcen”.

Ein wichtiges, vom IGR koordiniertes Kooperationsprojekt hat sich zum Ziel
gesetzt, ein Bundesinformationssystem Genetische Ressourcen (BIG) als Online-
System 1m Internet zur Verfügung zu stellen. Das BMBF-finanzierte Projekt soll
Datenbanken und Fachwissen aus vier wichtigen Organisationen zusammenfassen:
(1) Teile der Datenbanken am Bundesamt für Naturschutz, so das Arteninventar,
Verbreitung, Bestandssituation und Ökologie der einheimischen Wildpflanzen,
Artenschutz; (2) das unter SysTax (s.u.) in Kooperation mit dem Verband Bota-
nischer Gärten implementierte Informationssystem Botanischer Gärten mit den
Pflanzenbeständen vieler deutscher Botanischer Gärten; (3) die Datenbank des
Institut für Pflanzengenetik und Kulturpflanzenforschung (IPK) in Gatersleben mit
fast 100 000 Kulturpflanzen-Akzessionen der Genbank und einer projektierten
Datenbank zu Mansfeld's World Manual of Agricultural and Horticultural Crops;
(4) fruchtartenspezifische Datenbanken des IGR. Die zentrale Datenbank soll vom
IGR vorgehalten werden.

Unabhängig von BIG existiert ein BMU/BfN gefördertes Forschungs- und Ent-
wicklungsprojekt “Beitrag der Deutschen Botanischen Gärten zur Erhaltung Biolo-
gischer Vielfalt und Genetischer Ressourcen. - Bestandsaufnahme und Entwick-
lungskonzept” im Auftrag des Verbandes Botanischer Gärten und in Kooperation
mit Botanic Gardens Conservation International (BGCI). Als Ziel gilt die Ent-
wicklung eines Konzepts zur Erhöhung der Wirksamkeit der Botanischen Gärten
ım nationalen und internationalen Arten- und Naturschutz, auch durch Koordinie-
rung von Datenverarbeitung (Barthlott et al. 1999).

4.4. Deutsche Informationssysteme zur globalen Biodiversität
auf der organismischen Ebene

Es soll hier nicht weiter auf die Ökosystemebene eingegangen werden, da existie-
rende Informationssysteme entweder bereits in Abschnitt 4.1 abgedeckt wurden,
oder es sich um Komponenten von Forschungsprojekten handelt, für die For-
derungsmechanismen existieren. Zu dieser Kategorie sind z.B. zu rechnen: Projekt-
bezogene, institutsinterne Datenbanken des Zentrums für marine Tropenökologie,

SV)

Bremen, z.B. das BMBF finanzierte MADAM-Projekt (MAngrove Dynamics And
Management in Brazil); Teile der PANGAEA Datenbank des Alfred-Wegener-
Instituts für Polar- und Meeresforschung (A WI) in Bremerhaven; die Datenbanken
der im Programm Man and Biosphere anerkannten Projekte zur “Okosystemfor-
schung Bornhöveder Seenkette” im Ökosystemzentrum in Kiel (gefördert vom
BMBF) und zum Ökosystemforschungsprogramm Wattenmeer (interdisziplinäres
Verbundprojekt von BMU, BMBF und den Ländern Bremen, Niedersachsen und
Schleswig-Holstein); auch das bereits erwähnte ELISE gehört hierher.

Auf der Ebene der Organısmen werden im Folgenden zunächst Datenbanken
dargestellt, die potentiell vollständige Beiträge zu globalen oder zumindest euro-
päischen Organısmenregistern darstellen, d.h. die eine Organismengruppe voll-
ständig abdecken sollen. Dabei ist anzumerken, dass die (ungeordnete) Liste
keineswegs als vollständig anzusehen ist; ein mit Aussicht auf Förderung verbun-
dener öffentlicher Aufruf würde vermutlich eine ganze Reihe weiterer, teilweise
von Wissenschaftlern individuell gehaltener, aber sehr wertvoller Datenbestände zu
Tage fördern. Dies gilt im Prinzip auch für die weiter unten aufgelisteten Samm-
lungsinformationssysteme, deren Bedeutung erst kürzlich in einem Papier der
Direktorenkonferenz Naturwissenschaftlicher Forschungssammlungen Deutsch-
lands herausgestellt wurde (Naumann & Greuter 1997). Schließlich wird anhand
von Beispielen ein Bereich angeschnitten, der im vorgehenden Text nur im Zu-
sammenhang mit ökologischen und molekularen Fragestellungen implizit an-
gesprochen wurde, auf der Ebene der Organismen aber weitgehend vernachlässigt
wurde: Die Erschließung von an Organismenregister oder Belegdatenbanken
gebundener Information.

Taxonbezogene Datenbanken und Informationssysteme in Deutschland
Wirbeltiere

Die EMBL Reptile Database (Uetz 1999) am European Molecular Biology Labora-
tory in Heidelberg ist eine von Freiwilligen im Rahmen der AG Systematik der
Deutschen Gesellschaft für Herpetologie (DGHT) unterstützte, nicht-kommerzielle
Datenbank aller lebenden Reptilienarten (Datenbank der Fossilien in gleicher
Struktur vorhanden, aber zur Zeit nicht gepflegt). Es handelt sich um knapp 8000
Taxa mit zahlreichen Synonymen (diese allerdings noch nicht vollständig). Zur
Realisierung des Systems wurde ein im molekularbiologischen Bereich erprobtes
Datenbanksystem verwandt.

Insekten

Am Zoologischen Forschungsinstitut und Museum Alexander Koenig in Bonn ist
das Projekt Faunistik der Blatt- und Samenkäfer Mitteleuropas (Coleoptera: Chry-
somelidae, Bruchidae) angesiedelt. Das Projekt geht auf eine Privatinitiative des
Koordinators M.T. Schmitt zurück und wurde im Rahmen einer bereits abge-
laufenen 2-jährigen AB-Maßnahme begonnen. Die systematischen Einträge be-

53

rücksichtigen nur die Arten Mitteleuropas; jede Art wurde einer sorgfältigen
Validitätsprüfung unterzogen. Die Liste der Synonyme ist unvollständig.

An der Universität Bielefeld arbeitet M. von Tschirnhaus mit einem englischen
Kollegen (Henshaw) an einem Weltkatalog für die Dipteren-Familien Agromyzidae
und Chloropidae mit vollständiger Synonymisierung, der als Buch erscheinen soll.
Es handelt sich um 2736 Arten (insgesamt mehr als 5000 Taxa), die auch als PC-
Datenbank (dBASE) vorliegen. Ein Verzeichnis der Käfer Deutschlands (Köhler &
Klausnitzer 1998) ist in der Serie Entomofauna Germanica erschienen. Der Katalog
wird auch als dBASE-Datei angeboten und umfaßt fast 9000 Datensätze. Ein
zweiter Band dieser Serie ist die Checkliste der Dipteren Deutschlands mit fast
9200 Arten aus 117 Familien (Schumann et al. 1999); der Vertrieb einer digitali-
sierten Form ist vom Herausgeber bislang nicht vorgesehen. Schließlich soll ein
dritter Band, das Verzeichnis der Schmetterlinge Deutschlands (Gaedicke &
Heinicke), noch 1999 als Beiheft 5 zu den Entomologischen Nachrichten und
Berichten (Dresden) erscheinen.

Marine Organismen

Auch am Forschungsinstitut und Naturmuseum Senckenberg liegen in Form von
Arbeitslisten einzelner Bereiche Datenbestände aus der Zoologie vor. Hier ist z.B.
eine Liste der europäischen Crustacea Decapoda zu nennen (eine solche für das
Rote Meer ist im Aufbau), die langfristig zu einer globalen Liste ausgebaut werden
soll (M. Türkay, pers. comm. 1999). Die Taxonomische ArbeitsGruppe (TAG) ist
eine zentrale deutsche Einrichtung für die Forschung zur Taxonomie mariner
Organismen, die vormals an der Biologischen Anstalt Helgoland angesiedelt war
und jetzt im Deutschen Zentrum für Marine Biodiversitätsforschung (DZMB)
aufgehen soll. Die TAG arbeitet zur Zeit an mehreren globalen Artenlisten; so ist
eine EDV-gestützte globale Artenliste der Dinoflagellaten in Vorbereitung, in die
die vorliegenden handgeschriebenen Karteien überführt werden, außerdem liegt
eine von K. Riemann-Zürneck erstellte globale Artenliste der Tiefsee-Actinaria
handschriftlich auf Karteikarten vor (M. Elbrächter, pers. comm.).

Pflanzen

Im Systax System, einem Datenbanksystem für Systematik und Taxonomie unter
dem Datenbankverwaltungssystem Oracle, das in langjähriger Zusammenarbeit
zwischen Botanikern und Informatikern an der Universitat Ulm unter der Leitung
von T. Stiitzel entwickelt wurde, werden neben den Namen aus dem im Abschnitt
4.3 erwähnten Informationssystem Botanischer Gärten auch die Daten mehrerer
botanischer Checklistenprojekte (Annonaceae, Asclepiadaceae pro parte, Boragina-
ceae, Eryocaulaceae, Gesneriaceae und epiphylle Moose und Flechten) geführt
(J.R. Hoppe, pers. comm. 1998).

Die seit 1994 am Botanischen Garten und Botanischen Museum Berlin-Dahlem
(BGBM) aufgebaute Global Plant Checklist Database (Vascular Plants) der
Internationalen Organisation für Pflanzeninformation (IOPI) umfaßt inzwischen

34

über 240 000 Datensätze, die von zahlreichen Mitgliedern der Organisation einge-
bracht wurden. IOPI ist als Kommission der International Union of Biological
Sciences (IUBS) anerkannt. Der IOPI gehören 77 botanische Institutionen in 41
Ländern an, darunter fast alle bedeutenden Herbarien. Unterstützung des Global
Plant Checklist Projekts erfolgte (in kleinem Rahmen) u.a. durch CODATA, IAPT
(International Association for Plant Taxonomy) und das USDA (United States
Department of Agriculture). Auf Grund des Fehlens internationaler Fördermittel
konnte der 1993 ım Projektplan festgeschriebene Aufbau eines globalen Systems
bislang nicht erfolgen, obwohl eine fast alle Gruppen der Gefäßpflanzen abdecken-
de Liste von zur Mitarbeit bereiten Taxonomen vorliegt. Das Berliner System
(unter dem Datenbankverwaltungsprogramm Microsoft SQL-Server) erlaubt
aufgrund seines Datenmodells eine Parallelhaltung von verschiedenen taxono-
mischen Auffassungen zu einem Namen, so dass eine solche Aufarbeitung auch
nachträglich erfolgen kann.

Für das Pilotprojekt der IAPT zur globalen Erfassung neu veröffentlichter Pflan-
zennamen wurde am BGBM ein Datenerfassungsprogramm in MS-Access 97
erstellt, das neben dem Namen selbst die Eingabe aller nomenklatorisch relevanten
Informationen erlaubt. Eher im Sinne einer Stabilisierung der Nomenklatur ge-
dacht, aber dennoch ein globales Register darstellend, ist hier auch die Liste der
verwendbaren Pflanzengattungsnamen (Names in current use of extant plant
genera) als Datenbank mit 28 041 Datensätzen verfügbar.

Die Med-Checklist Database, eine bislang zu 60% fertiggestellte Liste (Greuter et
al. 1984, 1986, 1989) der in den an das Mittelmeer angrenzenden Ländern vor-
kommenden Pflanzen (20 500 Datensätze), wird ebenfalls vom BGBM verwaltet.
Neben nomenklatorischen Details, Synonymen und zahlreichen Verweisen von
falsch angewandten Namen wird der Vorkommensstatus in jeder betroffenen
Region angegeben. Diese Datenbank soll zusammen mit der Flora Europaea
Database den Grunddatenbestand für die projektierte Euro+Med Plantbase (Jury
1998) bilden.

Sammlungsdatenbankprojekte in Deutschland

Im Einklang mit weltweiten Initiativen (vergl. Butler 1998) wie Systematics Agen-
da 2000 und Diversitas (s. Abschnitt 3.1), der europäischen CETAF Initiative und
vor dem Hintergrund der immer drängenderen finanziellen Restriktionen haben
sich 1996 die Direktoren der größten Forschungssammlungen in Deutschland zu
einer ständigen “Direktorenkonferenz naturwissenschaftlicher Forschungssamm-
lungen Deutschlands” (DNFS) zusammengeschlossen. Als erstes Ergebnis wurde
Ende 1997 ein Papier zu Funktion, Situation und Perspektiven der biologischen
Sammlungen vorgelegt (Naumann & Greuter 1997), in dem auf das völlige Fehlen
eines Förderungsinstruments hingewiesen wurde, welches den Sammlungen die
Erfüllung der neu an sie herangetragenen Aufgaben ermöglichen könnte. Hierzu
zählt vor allem die Erschließung der enormen, mit den Belegen verbundenen

Oo
Nn

Biodiversitätsinformation mittels der elektronischen Datenverarbeitung. Solche
Collection Management Grants existieren seit mehreren Jahren z.B. in den USA,
Australien, Kanada und Mexiko, aber auch in einigen europäischen Staaten (Däne-
mark, Niederlande, Schweiz, Großbritannien).

Wie bereits erwähnt, liegt bisher keine vollständige Übersicht zum Stand der
Datenerfassung in den biologischen Sammlungen Deutschlands vor, aber auf
Grund der bereits eingegangenen Resultate der BioCISE Projektumfrage, der
Ergebnisse der EDV-Arbeitsgruppe der DNFS und eigener Erfahrungen der Auto-
ren lassen sich mehrere Aussagen treffen:

Die Verantwortlichen in den Sammlungsinstitutionen sind sich der Dringlich-
keit des Einstiegs in die Informationsgesellschaft vollauf bewußt und tun ihr
Mösglichstes, um die Bestände elektronisch zu erfassen.

Es findet nur in wenigen Teilbereichen eine echte Koordination statt, dadurch
kommt es zu erheblichen Doppelarbeiten.

Die mit der Erstellung eines Sammlungsdatenerfassungsprogramms verbunde-
nen Schwierigkeiten, von der Planung des Systems über die Erstellung einer
Datenbank bis hin zur Datenerfassung selbst, wurden von den Leitungen oft
weit unterschätzt. So werden informationstechnisch unzulänglich qualifizierte
Mitarbeiter, oft Kustoden, mit dem Problem der Erstellung einer Datenbank
konfrontiert, wodurch es zu einer erheblichen Beeinträchtigung der kustodialen
Aufgaben kommt. Oder es wird versucht, die Datenbankerstellung einem mit
der komplexen Sammlungsinformation nicht vertrauten Programmierer zu
überlassen.

International bereits gemachte Erfahrungen sowie Standards werden selten
berücksichtigt.

Die Lebendsammlungen weisen generell einen besseren elektronischen Erfassungs-
stand auf, wohl aufgrund der Tatsache, dass hier im Gegensatz zu vielen Präparate-
sammlungen schon traditionell Erfassungen der Belege in Karteien erfolgten.
Neben den bereits erwähnten Zoologischen Gärten (Abschnitt 3.3) und Samm-
lungen genetischer Ressourcen (4.3.) weisen auch die mikrobiologischen Samm-
lungen generell einen sehr guten Erfassungsstand auf (z.B. die an der DSMZ oder
am Institut für Pflanzenvirologie, Mikrobiologie und biologische Sicherheit der
Biologischen Bundesanstalt für Land- und Forstwirtschaft gehaltenen). Die Bota-
nischen Gärten Deutschlands weisen ebenfalls zentrale Organisationsansätze auf.
So werden z.B. die Botanischen Gärten Ulm und Bochum gemeinsam über das
Netzwerk im SysTax System in Ulm verwaltet. Weitere sechs übergeben regel-
mäßig Bestandslisten an SysTax zwecks gemeinsamer Abfrage über das System
und weitere Gärten stellen ihre Samentauschlisten im System zur Verfügung.
Mehrere Botanische Gärten verwalten ihre Bestände im “DIDEA-FR” Programm,
einem Ende der 80er Jahre entwickelten PC-Programm, welches auf einem damals
von BGCI (Botanic Garden Conservation International) erstellten und von der
TDWG anerkannten einfachen Datenaustauschformat beruht.

36

Im Gegensatz zum ersten Eindruck gilt, dass auch in den deutschen naturkundli-
chen Präparatesammlungen überraschend viele Datenbanken existieren. Hier sollen
nur einige exemplarisch genannt werden, wobei wiederum die strukturellen Details,
von denen letztendlich die Qualität der gewonnenen Information und ihre Inter-
operabilität in Netzwerken abhängt, nicht bewertet werden.

Am Zoologischen Forschungsinstitut und Museum Alexander Koenig, Bonn
(ZFMK), wurde mittels des seit 1990 in Kooperation mit dem Zoologischen Mu-
seum des Naturhistorischen Museums der Humboldt-Universität Berlin (ZMB),
dem Museum für Tierkunde in Dresden (MTD) und dem Deutschen Entomologi-
schen Institut in Eberswalde (DEI) entwickelten Programms BIODAT
(Sammlungsinformations-Managementsystem) eine Erfassung von Sammlungs-
daten vorgenommen. BIODAT ist ein unter DOS lauffahiges, auf Paradox (runti-
me) aufsetzendes Einzelplatzprogramm. Derzeitiger Stand der Sammlungserfas-
sung am ZFMK: 6000 Wirbeltiere, 21 000 Vögel und 7500 Insekten (Zikaden), am
MTD ca. 20 000 Datensätze Insekten, 10 000 Datensätze Mollusca, 15 000 Daten-
sätze Wirbeltiere, und mehrere 1000 Fische; am DEI liegen u.a. für Coleoptera:
Curculionidae etwa 50 000, für Hymenoptera 25 000 Datensätze vor. Die Erfassun-
gen am MTD und am DEI wurden mit auf dem BIODAT-Stammprogramm auf-
bauender Erfassungssoftware vorgenommen.

Am Forschungsinstitut und Naturmuseum Senckenberg in Frankfurt/M. wurde
schon Mitte der 80er Jahre mit der computergestützten Sammlungserfassung
begonnen. Das Großrechnersystem wurde Mitte der 90er Jahre auf ein Client-
Server System mit Sybase als Backend umgestellt. Das Frontend, SeSam (Sencken-
bergisches Sammlungsverwaltungsprogramm) wurde in MS-Access programmiert
und greift auf Serverdaten zu. Bisher wurden die Daten zu 70 000 zoologischen
Belegen erfaßt.

Am Phyletischen Museum der Friedrich-Schiller-Universität Jena liegen im Kata-
log der zoologisch-paläontologischen Sammlungen bereits über 100 000 Datensät-
ze (von insgesamt ca. 450 000) vor. Bei der Datenbank handelt es sich um ein
bereits 1986 entwickeltes dBASE III+ System unter MS-DOS, welches sich nach
Aussage der Benutzer gut bewährt hat.

Am Staatlichen Museum für Naturkunde Görlitz existiert eine Sammlungsdaten-
bank unter dBASE IV (Einzelplatz), die 1994 entwickelt wurde und seither unver-
ändert in Benutzung ist. Bisher wurden etwa 100 000 Datensätze (von 900 000)
eingegeben.

Ein Beispiel für eine außerhalb der üblichen naturhistorischen Sammlungen stehen-
de Institution mit projektbezogenem Fokus ist die Forschungsstelle für Ökosystem-
forschung und Ökotechnik (FSÖ) der Christian-Albrechts-Universität Kiel. In der
“Faunistisch ökologischen Datenbank” der FSÖ, gehalten in einer relationalen
Datenbank unter Ingres auf AIX sind z.Zt. bereits 650 000 Datensätze zu Inverte-
bratenaufsammlungen registriert.

Als ein Beispiel aus einer kleineren Institution mag hier das Museum für Ur- und
Frühgeschichte in Bottrop dienen; mittels des Programms ”Museumsmanager”

Sn

(unter Superbase Professional und Windows) wurden hier seit 1988 18 000 der
insgesamt 30 000 zoologischen und paläontologischen Sammlungsobjekte erfaßt.

Auch im privaten Bereich existieren umfangreiche, teilweise institutionsüber-
greifende zoologische Sammlungsdatenbanken (meist in direkter Verbindung mit
Artenregistern). Als Beispiel sei hier die von H. Wolf in Plettenberg aufgebaute
dBASE-Datenbank europäischer und palaearktischer Pompilidae genannt, welche,
basierend auf dem von ihm seit 1950 gesichteten Material, etwa 50 000 Samm-
lungsdatensätze mit Fundort, Sammlername, Sammlungsverbleib und Sammeljahr,
und damit verbunden ca. 5000 validitätsüberprüfte Taxa enthält.

Am Naturhistorischen Museum der Humboldt-Universität Berlin hält D. Lazarus
größere Datenbestände zu den Radiolarien (rezent und fossil), die auf Grund ihrer
Bedeutung in der Paläoozeanographie, der Global Change und der Evolutionsfor-
schung besondere Beachtung verdienen. Ein internationales Netzwerk von Spezia-
listen besteht, zumal das Naturhistorische Museum letztlich als eines von acht
globalen Radiolarian Marine Micropaleontology Reference Centers anerkannt
wurde.

Die großen Herbarien haben bisher, wohl aufgrund der generell schwierigen
kuratoriellen Lage, kaum Sammlungserfassung betrieben. In Berlin wurde Anfang
der 90er Jahre im Rahmen einer Arbeitsbeschaffungs-Maßnahme die getrennt vom
Generalherbar aufbewahrte Frucht- und Samensammlung erfaßt. Das damals
erstellte dBASE Programm wurde später in ein MS-Access System umgewandelt.
Daneben existieren Erfassungen von Teilen des Herbars, die in laufende For-
schungsprojekte eingebunden sind; so z.B. die aus Griechenland stammenden
Belege in der Flora Hellenica Database in Kopenhagen, sowie die El-Salvador-
Belege in der gemeinsam mit dem dortigen Botanischen Garten seit Ende der 80er
Jahre aufgebauten Datenbank.

Das Systax System wurde um ein Modul zur Herbarverwaltung erweitert, welches
z.Zt. im Herbarium der Universität Ulm und der Universität Gießen sowie im
Herbarium der Staatssammlung München eingesetzt wird.

Verknüpfte Information

Organismenregister und Belegdatenbanken bilden die primäre Informationsgrund-
lage für die globale Biodiversitätsinfrastruktur. Information auf der molekulargene-
tischen wie auf der Ökosystemebene muß letztendlich an Organismen festgemacht
werden, aus oder von denen die Gene (ab)stammen, oder aus denen die Ökosyste-
me zusammengesetzt sind. Die Belege sichern dabei die Reproduzierbarkeit des
wissenschaftlichen Ergebnisses und bilden gleichzeitig die Ausgangsbasis für neue
Analysen.

Aber auch auf der organismischen Ebene selbst bestehen Verknüpfungen zu ande-
ren Themen. Als Beispiel soll hier die Verbindung zu deskriptiver Information
(Wie sieht der Organismus aus, wie kann man ihn von anderen unterscheiden), die
zu chemischer Substanzinformation (Was enthält der Organismus’) und die in der
Biodiversitätsforschung wichtige Verbindung zu geographischer (Wo kommt der
Organismus vor, wieviele Organismen kommen wo vor?) angeschnitten werden.

38

Für das computergerechte Kodieren taxonomischer Beschreibungen wurde bereits
vor über einem Jahrzehnt die Descriptive Language for Taxonomy (DELTA)
entwickelt und auch als TDWG Standard anerkannt. Ursprünglich vor allem als ein
Hilfsmittel für die computergestützte Erstellung von Bestimmungsschlüsseln und
Beschreibungstexten gedacht, ist DELTA inzwischen vor allem auch durch deut-
sche Initiativen in den Bereich der deskriptiven Datenbanken und damit der Analy-
se großer Merkmalsdatensammlungen vorgedrungen. Hier sollen genannt werden:

— Das LIAS System (DELTA-based determination and data storage system for
LIchenized and lichenicolous AScomycetes, Rambold & Triebel 1999) an der
Botanischen Staatssammlung Miinchen, mit Forderung durch die Deutsche
Forschungsgemeinschaft. Hier entsteht eine DELTA Datensammlung fiir alle
Flechtengattungen weltweit, die bereits weitgehend komplett ist. Ein generi-
scher Bestimmungsschlüssel kann über das WWW benutzt werden. LIAS ist
ein gutes Beispiel für eine erfolgreiche internationale Kooperation vieler Wis-
senschaftler, die ohne die Bereitstellung einer technischen Infrastruktur nicht
zustande gekommen wire.

— DeltaAccess ist ein public domain Programm, welches von G. Hagedorn (1999,
Biologische Bundesanstalt für Land- und Forstwirtschaft) entwickelt wurde,
und welches die eigentlich an ein Textformat gebundene DELTA Struktur in
eine relationale Datenbank umsetzt. Damit wird es möglich, die deskriptive
Information mit zahlreichen statistischen und numerisch-taxonomischen Funk-
tionen zu analysieren, neben der traditionellen on-line Bestimmung und Be-
schreibungstextausgabe, die ebenfalls im Funktionsumfang des Programms
enthalten ist. Dem open source software Konzept entsprechend wurden in
internationalen Kooperationen u.a. bereits zwei World-Wide-Web Schnittstellen
entwickelt, die z.B. interaktive Identifikation über das WWW erlauben.

Zwei ganz andere Projekte sollen das Potential für interdisziplinäre Integration von
Organismendaten aufzeigen, die sowohl eine ökonomische als auch eine taxono-
mische Bedeutung haben:

— Im Verlauf ihrer in den 60er Jahren begonnenen Forschungen über die Chemie
der Compositen hat die Arbeitsgruppe F. Bohlmann an der TU-Berlin eine
Kartei zu Inhaltsstoffen und Arten der Compositen aufgebaut, die nach Bohl-
manns Tod von C. Zdero und J. Jakupovic weitergeführt wurde und als zu 98%
komplett gelten kann; d.h. sie enthält fast alle jemals für Compositenarten
publizierten oder in eigenen Arbeiten gefundenen Inhaltsstoffe. Seit 1994
wurden die Daten in eine ISIS/PC Datenbank überführt (die Bohlmann Files),
die auf dem WWW zugänglich gemacht werden kann. Ein Entwurf für ein
relationales Datenbanksystem für die mit den Strukturen verknüpfte taxono-
mische und bibliographische Information liegt vor. In Kombination mit einem
aktualisierten und von Experten revidierten Register der Compositen könnten
die hier vorhandenen Angaben (über 5000 Arten, über 20 000 Substanzen) als
Basis für weitergehende systematische und anwendungsorientierte (vor allem
pharmazeutische) Forschungen dienen.

39

Phytopathogene Pilze gehören größtenteils klar abgegrenzten taxonomischen
Gruppen an und stellen gleichzeitig eine wirtschaftlich und ökologisch be-
sonders wichtige Gruppe dar. Die friihzeitige Bestimmung von Schadorganis-
men kann den Einsatz umweltschädlicher Pestizide reduzieren oder sogar
vermeiden. Ein Wirtspflanzenindex phytopathogener Pilze (Arbeitstitel: Ho-
stIndex), verbunden mit der Möglichkeit, Bestimmungen interaktiv über das
WWW durchzuführen, bietet die besonders interessante Möglichkeit, mehrere
Organismenregister (Wirte und Parasiten) zusammen mit deskriptiver Informa-
tion in einer gemeinsamen Anwendung zu vereinen. Von G. Deml in Berlin
werden in der MINOS Datenbank (G. Hagedorn) des Instituts für Pflanzenviro-
logie, Mikrobiologie und biologische Sicherheit der BBA ca. 3200 nomen-
klatorische Datensätze und 3700 Sammlungsbelege von Brandpilzen (Ustilagi-
nales s.l., Basidiomycetes) gehalten. H. Bauch, Auenwald, verfügt über eine
unveröffentlichte Datenbank mit 9000 erfaBten Wirt-Parasit-Interaktionen von
Rostpilzen (Uredinales, Basidiomycetes). Beide Datenbanken würden sowohl
für ein globales Register der pflanzenparasitischen Pilze als auch für ein
Wirt/Parasit Indexprojekt zur Verfügung stehen.

Schließlich soll anhand eines Projektbeispiels noch auf den geographischen The-
menkomplex eingegangen werden. Globale Verbreitungsinformation zu Arten
beruht im allgemeinen auf Beleginformation, und die Herbarien und zoologischen
Sammlungen enthalten enorme Mengen derartiger Informationen. Der Einsatz von
Geographischen Informationssystemen zur Darstellung und Analyse derartiger
Daten steht im Mittelpunkt des folgenden Projekts.

Das Projekt Kartierung der globalen Phytodiversität - BIOMAPS (Biodiversity
Mapping for Protection and Sustainable Use of Natural Resources) an der
Universität Bonn (Arbeitsgruppe Barthlott) soll Erkenntnisse zur Verteilung der
globalen Biodiversität bereitstellen. Dazu werden auf der Basis vorliegender
Einzelstudien detaillierte Karten der Phytodiversität erstellt, die funktionalen
Zusammenhänge zwischen Geodiversität und Phytodiversität erarbeitet und
skizziert, und operationalisierbare Indikatoren für ein kontinuierliches Monito-
ring der Phytodiversität untersucht und identifiziert. Als Basis dienen floristi-
sche und vegetationskundliche Daten, räumlich differenzierte Geobasisdaten,
Vegetationskarten und aus Fernerkundungsdaten abgeleitete phänologische
Indikatoren, die integrativ in einem Geographischen Informationssystem (GIS)
verarbeitet werden. Das Projekt wird in enger Kooperation mit dem Deutschen
Fernerkundungsdatenzentrum (DLR/DFD), Abteilung Umweltsysteme (G.
Braun) und dem Geographischen Institut der Universität Bonn durchgeführt.

4.5. Zusammenfassung

Eine bundesweite Koordinierung von Projekten zur Biodiversitätsinformatik findet
bislang nur im Umfeld der Umweltinformationssysteme und des Umweltdatenkata-
loges sowie der Organismenkartierung statt und bezieht sich (unter Ausnahme der

40

genetischen Ressourcen) nur auf die Biodiversität in Deutschland selbst. Aufgrund
der durch die föderalen Strukturen bedingten Zersplitterung sind hier noch erhebli-
che Synergiereserven zu erschließen. Auf der Ökosystemebene ist hier durch die
notwendige Anwendung der Richtlinien der EU (siehe unter Abschnitt 3.2) ein
erheblicher Informationszuwachs zu verzeichnen, der sich teilweise durch Erfas-
sungsmaßnahmen und vor allem durch die langjährigen Kartierungsmaßnahmen
auf Bundes- und Länderebene auch auf die Artebene bezieht. Hier ist durch Projek-
te wie die Kartierungsmaßnahmen auf europäischer Ebene (s. unter Abschnitt 4.2),
die Zusammenarbeit im marinen Bereich und durch die Einbindung in UN Pro-
gramme wie Man and Biosphere eine gewisse Koordination zu verzeichnen.

Hingegen ist ein koordiniertes Vorgehen in Bezug auf die enormen Informations-
reserven zur globalen Biodiversität, die in der deutschen systematischen Forschung
und in den biologischen Sammlungen zu finden sind, nur in ersten Ansätzen zu
verzeichnen. Aufgrund des auf nationaler Ebene fehlenden zusammenfassenden
Rahmens sind die hier vorhandenen Informationen nur schwer zugänglich, ihre
Einbindung in internationale Vorhaben wird erschwert, und es ist anzunehmen,
dass vielfach auf Grund persönlicher wissenschaftlicher Initiative geschaffene
Datensammlungen nicht erschlossen werden. Den naturkundlichen Forschungs-
sammlungen Deutschlands, die überwiegend gleichzeitig systematische
Forschungsinstitute sind, kommt auf der Ebene der organismen-bezogenen Primär-
information der Biodiversitätsforschung eine Schlüsselrolle zu.

5. STRATEGIE UND PRIORITÄTEN IM BEREICH
BIODIVERSITÄTSINFORMATIK

5.1. Strategie für eine national koordinierte Forschungsförderung

Aus der vorangehenden Analyse ergibt sich, dass im Bereich der Biodiversitäts-
informatik auf der organismischen Ebene das größte Förderungsdefizit besteht; hier
handelt es sich um eine besonders im Hinblick auf die internationale Kooperations-
und Konkurrenzfähigkeit der deutschen Biodiversitätsforschung zu füllende Lücke.
Hierbei ist die nationale Seite (also die Verarbeitung von Informationen zur Biodi-
versität der in Deutschland vorhandenen Organismen) durch vom BMU (BfN) und
vom BML unterstützte Projekte in Teilbereichen bereits relativ weit fortgeschritten.
Ein Ausbau der Informationsbereitstellung im international ausgerichteten organis-
mischen Bereich leistet im Sinne der Biodiversitätskonvention einen direkten
Beitrag zur geforderten Verbesserung der Erhaltung, nachhaltigen Nutzung und
Entwicklung der globalen Biodiversität und kommt zusätzlich der Informationslage
im Ökosystembereich (Umweltinformation) sowohl unmittelbar als auch langfristig
zugute.

Im Sinne einer effizienten Förderung ist hierbei zunächst ein gezieltes Verbessern
und Homogenisieren der biodiversitätsinformatischen Infrastruktur ins Auge zu
fassen. Hierzu gehören Maßnahmen, die auf eine verbesserte Koordination der

41

verschiedenen Initiativen und Projekte abzielen, ebenso wie die Förderung der
Nutzung von vorhandenen Standards und die Entwicklung von Verfahren (Hard-
und Software-Empfehlungen zur Datenerfassung, Abfragen auf verteilten inhomo-
genen Datenbanken, etc.).

Diese Maßnahmen müssen parallel durch die vermehrte Bereitstellung bzw. digita-
le Erschließung der - besonders in Deutschland umfangreich vorhandenen - primä-
ren Datenbestände unterstützt und erprobt werden. Die Erschließung zusätzlicher
Daten kann zunächst nur fokussiert erfolgen, muß aber mittel- bis langfristig
umfassend ausgebaut werden. Es handelt sich zumeist um anfänglich personal-
intensive Maßnahmen, die nicht unter die gegenwärtig zur Verfügung stehenden
Förderungsinstrumente fallen. Im Rahmen der internationalen Forschungskoopera-
tion (und auch der durchaus vorhandenen internationalen wissenschaftlichen
Konkurrenz) ist dabei auch dem Besetzen von Themen durch die deutsche Wissen-
schaft eine Rolle zuzuschreiben.

Für alle genannten Maßnahmen gilt, dass sie unter verstärktem Einsatz innovativer
Informationstechnologien (z.B. verteilte Objekttechnologie im Datenzugriff,
Replikationstechniken für Datenbanken, Bildverarbeitung, Verbindung mit Geo-
graphischen Informationssystemen), soweit möglich im interdisziplinären Ansatz,
unbedingt aber in direkter Anbindung an internationale, aktuell durchgeführte
Vorhaben und Projekte realisiert werden sollten.

Die Förderung in diesen Bereichen hat einen deutlichen Anschubcharakter; viele
der hier angestrebten Lösungen sind wegbereitend für eine Integration der Erfas-
sungsverfahren in den normalen Betrieb der mit Biodiversitätsforschung befassten
Institutionen. Mittelfristig wird damit eine allgemeine Koordination der Daten-
erfassung bei Biodiversitätsforschungsprojekten in Deutschland angestrebt, wobei
das US-amerikanische Beispiel zeigt, dass die verschiedenen Förderungsträger hier
einen entscheidenden Einfluss ausüben können. So findet dort eine ausdrückliche
Förderung u.a. in den NSF Programmen Database Activities in the Biological
Sciences, Biological Infrastructure, Biological Research Collections und Biotic
Surveys and Inventories statt. Das vom BMBF aufgestellte Förderprogramm
BIOLOG macht einen Schritt in diese Richtung, obwohl erneut eine gezielte
Unterstützung sowohl von groß angelegten Erfassungsmaßnahmen als auch von der
Erstellung globaler Organismenregister in Deutschland unterbleibt.

5.2. Verbesserung der Infrastruktur

Koordination

Das Ziel von Projekten in diesem Bereich ist, die bereits vorhandenen und die
entstehenden biodiversitätsinformatischen Ressourcen miteinander zu verknüpfen,
und zwar sowohl in sachlicher als auch in organisatorischer und persönlicher
Hinsicht:

42

— In sachlicher Hinsicht ist hier vor allem die Entwicklung und Veröffentlichung
von Datenstruktur- und Austauschstandards anzustreben, die auf geeigneten
Datenmodellen aufsetzen. Es handelt sich dabei teilweise um Anpassungen und
Übersetzungen aus dem Englischen von bereits veröffentlichten internationalen
Standards.

— In organisatorischer und persönlicher Hinsicht geht es um den Erfahrungs-
austausch unter den Institutionen und den jeweils für Biodiversitätsinformatik
zuständigen Fachleuten und ggf. die gemeinsame Entwicklung von Projekten.
Mittelfristig ist hier in Form einer Selbstorganisation das Ziel einer Schaffung
von nationalen Koordinierungsstellen für verschiedene Informationssysteme ins
Auge zu fassen, wobei es vermutlich sinnvoll wäre, die Bereiche Mikrobiolo-
gie, terrestrische Zoologie, Botanik, marine Biologie und Paläontologie zu
trennen.

Eine wichtige Grundlage hierfür wäre die Erfassung von deutschen Projekten und
Strukturen im Sinne eines ständig gepflegten Projektregisters. Dies könnte u.U. als
eine Komponente in den deutschen Clearinghouse Mechanismus integriert werden.
Im Bereich biologische Sammlungen kann dabei auf den Datenbestand und die
Datenbank des BioCISE Projekts zurückgegriffen werden. Weiterhin sollte hier
auch eine zentrale Erfassung (und damit Koordinationsmöglichkeit) für Biodi-
versitätsdaten, die 1m Rahmen deutscher Entwicklungshilfeprojekte im Ausland
erhoben werden, geschaffen werden.

Die Bedeutung einer institutions- und länderübergreifenden Koordination in
Deutschland wird in der Übersicht der in Deutschland gegenwärtig durchgeführten
Projekte im Bereich organismischer Biodiversitätsinformatik (vgl. Abschnitt 4.4)
deutlich. Es mangelt durchaus nicht an substanziellen Ansätzen und teilweise
beachtlichen, bereits digitalisiert verfügbaren Datenbeständen. Als nachteilig
erweist sich jedoch das weitgehende Fehlen einer überregionalen Koordination, die
einen effektiven Datenaustausch sowie die Zusammenführung der zahlreichen
unterschiedlichen Projekte und Vorhaben, die sich oft mit denselben Organismen-
gruppen befassen, erleichtern würde. Erschwerend kommt hinzu, dass viele In-
stitute immer noch nicht über eine effiziente Anbindung an internationale Daten-
netze verfügen. Der Mangel an Integration auf nationaler Ebene erzeugt zusätzliche
Probleme bei der internationalen Einbindung deutscher Vorhaben, da hier in
Deutschland oft zu viele potenzielle Ansprechpartner vorhanden sind. Weiterhin
wirkt sich in Deutschland das Fehlen zentraler Förderinstrumente für die Informa-
tionserschließung als nachteilig aus (Naumann et Greuter 1997, vgl. NSF 1998),
die beispielsweise die Möglichkeiten zur Schaffung und Einhaltung allgemeiner
technischer Standards und einheitlicher Informationsstrukturen deutlich verbessern
würden.

Schaffung von - oder Anbindung an - Standarddatenkataloge

Hiermit sollte das Ziel verfolgt werden, übergreifende Standarddatenkataloge zu
schaffen oder eine Mitbenutzung solcher Kataloge zu sichern. Dabei handelt es sich

43

um Datensammlungen (von einfachen Tabellen bis zu komplexen Systemen), die
in andere Biodiversitätsdatenbanken integriert, aber an der Ursprungsstelle gepflegt
werden. Es handelt sich also um eine den Organismenkatalogen analoge Vernet-
zungsaufgabe; die Priorität sollte bei den Themenbereichen geographische An-
gaben, Literaturdatenbanken und taxonomische Informationssysteme oberhalb der
Artebene liegen.

Die Begründung für eine getrennte Förderung solcher Projekte ist, dass es eine
ganze Reihe von Datenbereichen gibt, die in verschiedenen Organismenregistern
und Sammlungsdatenbanken Verwendung finden und deren Parallelentwicklung in
jedem einzelnen Projekt verhindert werden sollte. So ist z.B. für das vorrangige
Ziel einer vollständigen Inventarisierung der Biosphäre die Ermittlung der räumli-
chen Verbreitung der Organismenarten von zentraler Bedeutung, d.h. die Erfassung
von geographischer Information ist dringend geboten. Konkrete Verbreitungsdaten
bei Sammlungsbelegen bestehen in der Regel in der Angabe von Ortsnamen bzw.
sind an Ortsnamen ausgerichtet. Eine Angabe geographischer Koordinaten, wie sie
für eine informationstechnische Umsetzung solcher Verbreitungsangaben am
günstigsten ist, ist in der Mehrzahl der Fälle nicht vorhanden. Wichtig aus der Sicht
der Biologie sind daher die Verfügbarkeit von globalen Ortsnamenregistern (gazet-
teers) bei der Datenerfassung bzw. die Nutzung von Informationssystemen, die
eine Zuweisung von Koordinaten zu Ortsnamensangaben (unter Angabe des da-
durch entstehenden Fehlers) ermöglichen. So besteht dann umgekehrt auch die
Möglichkeit, Daten zum Vorkommen bestimmter Organismenarten in bereits
bestehende, z.B. umweltorientierte geographische Informationssysteme einzuspei-
sen, wie dies auf lokaler Ebene bereits erfolgreich praktiziert wird (vgl. Abschnitt
4.1, z.B. ALBIS). Eine einheitliche Strukturierung geographischer Angaben zu
Organismenarten erleichtert zudem für externe Nutzer den Zugang zur jeweils
relevanten Information.

5.3. Informationserschließung

Gegenüber den offenkundigen Defiziten bei der biodiversitätsinformatischen
Infrastruktur scheint die Situation hinsichtlich der Verfügbarkeit von erschließ-
baren Datenbeständen und der Fachkompetenz auf der organismischen Ebene in
Deutschland besser, wenn sie auch nicht als generell gut bezeichnet werden kann
(vgl. Abschnitte 4.2, 4.4). Durch die zwar zwischen zahlreichen naturkundlichen
Sammlungen und Forschungsinstituten aufgeteilten, in ihrer Summe aber weltweit
bedeutenden Sammlungsbestände bestehen in Deutschland für die Erschließung
und Bereitstellung primärer Daten wie systematisch-taxonomischer Kenntnisse
beachtliche Potenziale (Cracraft 1995, Hawksworth 1995, Naumann & Greuter
1997), die bisher jedoch noch wenig oder nur unzureichend genutzt wurden. So
beherbergen z.B. allein die neun großen deutschen zoologischen Forschungssamm-

44

lungen zusammen über 2,4 Mio. Wirbeltierpräparate (und liegen hiermit nach den
USA und Großbritannien weltweit an 3. Stelle), in den Herbarien deutscher Uni-
versitäten und Forschungssammlungen lagern über 17 Mio. Pflanzen-Belege
(Naumann & Greuter 1997) und 10 deutsche Institutionen besitzen Insektensamm-
lungen von weltweiter Bedeutung mit jeweils mehr als | Mio. Exemplaren. Trotz
fortschreitendem Abbau von Stellen im Bereich organismischer Biologie an deut-
schen Hochschulen (insbesondere in der Zoologie, Naumann, pers. comm.), exis-
tiert aufgrund alter Traditionen in der biosystematischen Forschung hierzulande
immer noch ein hohes Maß an taxonomischer Fachkompetenz (Cracraft 1996), das
allerdings in zunehmendem Maße auch von Privatpersonen abgedeckt wird (z.B.
Schminke 1996). Weiterhin sind auch aus für die Biodiversitätsinformatik relevan-
ten anderen Fachbereichen, insbesondere aus der Geographie, weltweit bedeutende
Datenbestände und Kompetenzen in Deutschland vorhanden, die allerdings von der
biologischen Seite bisher wenig genutzt werden. Insgesamt findet sich für die
Biodiversitätsforschung auf nationaler Ebene im Bereich der Informationserschlie-
Bung das größte kurzfristig verfügbare Potenzial (vgl. Linsenmair 1998), das daher
gezielt mobilisiert werden sollte. Hierzu bieten sich folgende Themenbereiche an:

Organismenregister

Zur Erreichung des Ziels der Schaffung globaler Organismenregister der bekannten
Arten aus bestimmten systematischen Gruppen müssen die für begrenzte Regionen
existierenden elektronischen Artenregister zusammengeführt und zusätzlich in der
Literatur vorhandene Informationen digitalisiert werden. Dies erfordert sowohl
Expertenwissen (ggf. mehrerer Wissenschaftler in verschiedenen Ländern) für die
taxonomische Koordination als auch für die Abstimmung der Datenstrukturen
zusammenzuführender Bestände, außerdem wird interdisziplinäres Fachwissen für
den Einsatz und die Entwicklung von Werkzeugen zur Digitalisierung von Litera-
turdaten, der Bereitstellung der Information in Netzwerken und der Integration mit
den globalen Strukturen benötigt. Die Komponente Verbreitungsinformation kann
u.U. in Zusammenarbeit mit Geographen mittels Geographischer Informations-
systeme abgedeckt werden.

Auf die Bedeutung der globalen Organismenregister wurde bereits unter Abschnitt
1.2 und 3.3 eingegangen. Sie fehlen bislang für die überwiegende Mehrzahl aller
taxonomischen Gruppen. Ihre Erstellung ist sowohl im Sinne der Erfüllung der
Biodiversitätskonvention als auch im Sinne der im Rahmen der internationalen
Wissenschaftsagenda geforderten Inventarisierung der Biosphäre eine aktuell

° Hier sei der Hinweis erlaubt, dass die überwiegende Zahl der Arten auf der Erde bisher
nicht bekannt ist (vergl. Steininger 1996 und Environment Australia 1998). Dem wird ein
biodiversitätsinformatisches Förderprogramm nicht abhelfen, es kann aber unserer Ansicht
nach Synergieeffekte hervorrufen, die letztendlich eine Freisetzung von Reserven für die
Biodiversitätsforschung selbst erlaubt.

45

vordringliche Aufgabe (UNEP 1995, Diversitas, Agenda Systematik 2000; s.
Abschnitt 3.2). Durch die Bereitstellung von wichtiger Rahmeninformation für
interdisziplinäre Informationssysteme (z.B. Schadinsekten, phytopathogene Pilze,
Pflanzeninhaltsstoffe, Indikatororganismen für Luftverschmutzung und vieles
andere mehr) wird durch solche globalen Organismenregister zudem für eine
Bewältigung der Probleme des Globalen Wandels notwendiges Grundlagenwissen
bereitgestellt (Linsenmair 1998). Im diesem Bereich bieten sich derzeit noch gute
Chancen zur “Besetzung” von Themen durch die deutsche Forschung im interna-
tionalen Rahmen, wie die Übersicht der momentan noch fragmentarischen Bemü-
hungen zeigt (vgl. Abschnitt 3.3; Species 2000, IOPI), obwohl sich in manchen
Bereichen (z.B. Blütenpflanzen) bereits eine gewisse globale Verteilung der The-
men andeutet.

Als Kriterium für die Auswahl von Projekten muß zunächst die Auswahl der
Organismengruppe herangezogen werden. Für eine konzentrierte Förderung kom-
men Organismengruppen in Frage, für die in Deutschland wissenschaftliche Kom-
petenz und möglichst auch eine Datengrundlage vorhanden sind, die eine aus-
reichende Größe und eine weite (möglichst auch tropische) Verbreitung aufweisen
und für die eine entsprechende Lücke im Konzert der bereits vorhandenen interna-
tionalen Bemühungen besteht. Weiterhin ist eine Kombination mit Projekten zur
Erschließung von Sammlungsinformation und/oder mit verknüpften Informations-
systemen wünschenswert. Mit einer Bildung von Arbeitsgruppen, die unter Be-
teiligung ausländischer Wissenschaftler eine gleichzeitige Sicherung der Datenhal-
tung in Deutschland gewährleisten, wäre eine internationale Anbindung solcher
Vorhaben zu erreichen, die gleichzeitig Synergieeffekte für die Förderung biosyste-
matischer Forschung in Deutschland aufweisen würde.

Erschließung der Sammlungsinformation

Um das Ziel zu erreichen, die in den biologischen Sammlungen liegenden
Informationsreserven zu mobilisieren, müssen einerseits die bereits ın Einzel-
institutionen und aus zahlreichen Projekten digitalisiert vorliegenden Daten (vgl.
Abschnitt 4.4) allgemein verfügbar gemacht werden (was oft eine Konvertierung
der Datenbestände mit dem Ziel, sie den heutigen Standards anzupassen, ein-
schließt). Andererseits muß die ganz überwiegende Anzahl der vorliegenden
Belege neu erfaßt werden. Aus der Natur der Objekte heraus bieten sich dabei
durchaus je nach Sammlungstyp verschiedene Ansätze. So ist z.B. bei den höheren
Pflanzen normalerweise die Sammlungsinformation (Name, Herkunft der Belegs,
Sammler etc.) auf dem Etikett des Belegs selbst konzentriert, während bei vielen
anderen Sammlungen sich die Information im wesentlichen in Akzessions- und
Notizbüchern oder Karteien findet. Etiketten und Notizbücher können mit der heute
zur Verfügung stehenden digitalen Phototechnik preisgünstig aufgenommen
werden und damit weltweit zur Erschließung zugänglich gemacht werden. Der
gegenwärtige Stand der Technik erlaubt aber auch bereits ohne großen finanziellen

46

Aufwand die Digitalisierung bestimmter ganzer, im wesentlichen 2-dimensionaler
Belege (z.B. mikroskopischer Algen), der Herbaretiketten, oder auch der Notizbü-
cher. Insbesondere bei den besonders wertvollen Typusexemplaren’ ist eine solche
Volldigitalisierung anzustreben, wobei der rapide Fortschritt der Technik durchaus
auch dreidimensionale und hochauflösende Darstellungen mittelfristig realistisch
erscheinen läßt. Für geographisch/ökologische Fragestellungen, taxonomischen
Zugang und statistische Analysen sind aber wohlstrukturierte Datenbanken unbe-
dingt notwendig, die in der Datenerfassung erheblich arbeitsintensiver sind. Kurz-
fristig und bei den “normalen” Belegen ist vermutlich ein Arbeitsprozess am
vielversprechendsten, der eine Kombination von Bild- und Textdaten vorsieht,
wobei der Text sich in der Ersterfassung auf Kerndaten (wissenschaftlicher Name,
ggf. Lagerort des Belegs, eventuell geographische Angaben und/oder Sammler-
angaben) beschränkt, um dann im Bedarfsfall mittels der in Bildform digitalisierten
Information weiter detailliert zu werden.

Die Bedeutung der biologischen Sammlungen wurde bereits in Abschnitt 1.3
angesprochen. Die in deutschen naturkundlichen Forschungssammlungen und
Universitätsinstituten enthaltenen naturkundlichen Sammlungsbelege sind in ihrer
Summe von grundlegender Bedeutung für die globale Biodiversitätsforschung
(Naumann & Greuter 1997, Steininger 1997). Eine vollständige Erschließung und
die allgemeine Verfügbarmachung dieser Datenbestände erfüllt nicht nur eine der
zentralen Verpflichtungen aus der Ratifizierung der Biodiversitätskonvention
(vergl. 3.1.; Repatriierung von Daten‘), sondern beinhaltet einen substanziellen,
den Leistungen anderer Länder entsprechenden Beitrag zur internationalen Biodi-
versitätsforschung. Entsprechende Förderungsmaßnahmen zur Erschließung biolo-
gischer Forschungssammlungen existieren bereits seit längerem z.B. in den USA
(vgl. NSF 1998) und haben wesentlich zur heutigen Führungsposition der Ver-
einigten Staaten im Bereich organismischer Biodiversitätsforschung beigetragen.
Eine unter Einsatz moderner Informationstechnologie durchgeführte Datenerschlie-

’ Typus-Exemplare sind präparierte Belegexemplare, die (normalerweise) vom Autor des
Tier- oder Pflanzennamens festgelegt wurden. Sie sind die unersetzliche international
verbindliche Referenz für die Identität der wissenschaftlichen Namen einzelner Taxa (also
bestimmter Arten, Gattungen, Familien, etc.) und besitzen daher einen besonderen
wissenschaftlichen Wert (vgl. Naumann & Greuter 1997 und die Nomenklaturcodes: Ride
& al. 1999 und Greuter & al. 1994).

* Die Abgabe von so gewonnenen Teildatenbeständen an das Ursprungsland stellt eine sehr
effektive Maßnahme des Informations- und Technologietransfers im Sinne des Clearing
House Mechanisms der Biodiversitätskonvention dar. Dies wurde z.B. bereits in großem
Umfang von der mexikanischen Biodiversitätskommission CONABIO initiiert und
durchgeführt und hat maßgeblich zum Aufbau der beispielhaften biodiversitäts-
informatischen Infrastruktur in Mexiko beigetragen.

47

Bung hilft weiterhin der in vielen Fällen in Deutschland aktuell bedrohten, lang-
fristigen Sicherung dieser Bestände (vgl. Schminke 1996).

Obwohl eine konzentrierte Förderung der großen Institutionen kurzfristig den
größten Effekt zeigen dürfte, bietet die Erschließung kleinerer, oft stark vernachläs-
sigter Sammlungen mit überregionaler Bedeutung die Chance, diese oft selbst der
Fachwissenschaft kaum bekannten oder unzugänglichen Bestände zu erschließen
(Schminke 1996). Derartige, vom Gesamtumfang her kleinere Spezialsammlungen
sind aufgrund der Kulturhoheit der Länder und des föderativen Systems in
Deutschland zahlreich vorhanden und oft aufgrund mangelhafter konservatorischer
Pflege direkt in ihrem Bestand bedroht. Allerdings stellen die relativ hohen Mit-
gliedsbeiträge zum deutschen Forschungsnetz (Internet) gerade für kleinere In-
stitutionen ein starkes Integrationshindernis dar.

Schaffung von verknüpften biologischen Informationssystemen

Ziel ist hier die Schaffung oder Einbindung von Informationssystemen, die in
direkter Beziehung zu Organismenregistern oder Sammlungsdatenbanken stehen
und damit langfristig zu deren Erhaltung und Ausbau beitragen können. Hierzu
zählen sowohl Themen innerhalb der Biowissenschaften als auch interdisziplinäre
Projekte. Es ist hier durchaus auch an eine Verbindung mit Verlagen oder anderen
kommerziellen Informationsbereitstellern zu denken. Ein Beispiel hierfür ist die
Zusammenarbeit zwischen ILDIS, der Weltdatenbank der Blütenpflanzenfamilie
der Leguminosen (Hülsenfrüchtler) und dem Verlagshaus Chapman & Hall, die zur
Produktion der CD Phytochemical Dictionary of Leguminosae führte.

Aufgrund der zentralen Position der Organismenregister in der Biodiversitäts-
informatik ist das Feld der möglichen verknüpften Information sehr breit. Dies
beginnt im taxonomisch-systematischen Feld selbst, z.B. durch die Schaffung und
Bereitstellung von Informationssystemen, die, auf vorhandenen und zu schaffenden
Standards aufbauend, Merkmalsinformation zu den Organismen liefern, aufgrund
derer interaktive Bestimmungen und Analysen durchgeführt werden können, über
die Verknüpfung mit geographischen Informationssystemen zur Beantwortung von
Fragen der globalen Biodiversitätsforschung, bis hin zu interdisziplinären Anwen-
dungen. Hinzu kommt der weite Bereich der multimedialen Information (v.a. Bild-
und Tonarchive). Die anzustrebende Interdisziplinarität ist naturgemäß nicht in
gleichem Maße auf alle Wissenschaftsbereiche außerhalb der Biologie sinnvoll
anwendbar. Die direkte Kooperation mit entsprechenden Bereichen der Informatik
bzw. Informationstechnologie ist vorgegeben, und die enge Verzahnung mit der
Geographie wurde bereits erwähnt. Sinnvollerweise sollte aber auch eine Anbin-
dung an die Wissenschaftsbereiche erfolgen, die unmittelbar mit der Nutzung bzw.
Erhaltung der internationalen Biodiversität befaßt sind, wie z.B. der Entwicklungs-
zusammenarbeit im Bereich der Landwirtschafts- und Forstwissenschaften und des
Naturschutzes, in der Medizin (besonders in der Epidemiologie und Pharmazie)
und der (Naturstoff-)Chemie. Daneben existieren durchaus auch Beziehungen mit
einigen geistes- bzw. sozialwissenschaftlichen Fächern wie der Ethnologie (Kultur-
anthropologie) und Archäologie.

48

Bei der verknüpften Information zu Belegdatenbanken kann es sich, neben den
bereits erwähnten Bilddaten der Belege selbst und dem in Abschnitt 5.2. dargestell-
ten Geographiebezug, vor allem um Literaturdaten handeln, insbesondere die
Verknüpfung der Typusexemplare mit den entsprechenden Literaturstellen. Beim
Aufbau solcher Datenbanken ist eine Überprüfung der Typen vorzunehmen. Ein
anderes Gebiet ist die Integration von beschreibenden Belegdaten (z.B. Größen-
messungen etc.) in auf Standards wie DELTA beruhende Informationssysteme in
der taxonomischen Forschung.

Die Bedeutung der verknüpften Datenbanken in der Anfangsphase der Biodi-
versitätsinformatik liegt vor allem in ihrer unterstützenden Funktion für die infra-
strukturbildenden Organismenregister und Belegdatenbanken. Wie jede Infra-
struktur benötigt auch diese ständige, langfristig gesicherte Pflege. Im Fall von
Informationsinfrastrukturen wird diese Pflege am besten durch intensive Nutzung
(und damit ständige Ergänzung und Verbesserung) erreicht. Dies muss aber zu-
mindest in der Anfangsphase direkt mitgefördert werden, um mittelfristig einen
Selbstorganisationsprozess in Gang zu setzen, der die biodiversitätsinformatische
Infrastruktur langfristig voll in die bestehende wissenschaftliche Infrastruktur
integriert.

DANKSAGUNG

Wir danken Gregor Hagedorn, Biologische Bundesanstalt in Berlin, und Rudolf
May, Bundesamt für Naturschutz in Bonn, für die kritische Durchsicht des Manu-
skripts und für wertvolle Anregungen und Ergänzungen.

ZITIERTE LITERATUR

Anmerkung: Aufgrund der häufig nicht stabilen Adressen von Dokumenten auf
dem World Wide Web haben die Autoren versucht, solche Zitate auf gedruckte
Werke zurückzuführen. Wo dies nicht möglich war, empfiehlt sich eine Suche des
Titels mittels einer Suchmaschine, sollte die URL nicht mehr aktuell sein.

Anonym (1990): Understanding our genetic inheritance — the U.S. human genome
project: the first five years. U.S. Dept of Health and Human Services, U.S. Dept
of Energy. [http://www.nheri.nih.gov/HGP/HGP_goals/Syrplan.html]

Anonym (1991): Ubereinkommen zum Schutz der Alpen (Alpenkonvention).
Salzburg 1991 [http://deutsch.cipra.org/texte/alpenkonvention/alpenkonven-
tion_hauptseite.htm]

Anonym (1996): Dublin Core Metadata Element Set: Reference Description.
OCLC Online Computer Library Center, Inc. [http://purl.org/metadata/dublin
_ core _elements]

49

Anonym (1999): Final report of the OECD Megascience Forum Working Group
on Biological Informatics, January, 1999. [http://www.bgbm.fu-berlin.de/bio-
divinf/docs/OECDMSWGEBI.pdf]

ASC (1993): An information model for Biological Collections (Draft). — Report
of the Biological Collections Data Standards Workshop, August 18-24, 1992.
Association of Systematic Collections, Committee on Computerization and
Networking. [gopher://kaw.keil.ukans.edu:70/1 1/stan-dards/asc |

Eeamenlotinwe NE vonrden Diriesch, iP Latbisch.W. Lebin &G.
Rauer (1999): Botanische Gärten und Biodiversität. — Bundesamt für Natur-
schutz, Bonn, 70 pp.

BCIS (1999): Biodiversity Conservation Information System — better data for
better decisions. — IUCN, Chicago. [http://www.biodiversity.org/]

Berendsohn, W.G. (1995): The concept of "potential taxa" in databases. —
Taxon 44: 207-212.

Berendsohn, W.G. (1997): A taxonomic information model for botanical
databases: The IOPI Model. — Taxon 46: 283-309.

Berendsohn, W.G. (1998): Datenstrukturforschung und international vernetz-
te Umweltinformation auf der Ebene der Organismen. pp 33-45 in: Hoppe, J.,
Helle, S. & Krasemann, H. L. (ed.): Vernetzte Umweltinformation. — Praxis der
Umweltinformatik 7. Metropolis Verlag.

Berendsohn, W.G. (ed.) (1999a): Standards, Information Models, and Data
Dictionaries for Biological Collections. TDWG Subgroup on Accession Data,
Berlin. [http://www.bgbm.fu-berlin.de/TDWG/acc/Referenc.htm]

Berendsohn, W.G. (ed.) (1999b): Provisional global plant checklist. Informa-
tion systems committee, International Organization for Plant Information,
Berlin. [http://www.bgbm.fu-berlin.de/iopi/gpc |

Berendsohn, W.G., A. Anagnostopoulos, G. Hagedorn, J. Jaku-
DONC yale Nimis, By. Valdés, A.Guntseh, R.J. Pankhurst &
R.J. White (1999): A comprehensive reference model for biological collec-
tions and surveys. Taxon 48: 511-562. [http://www.bgbm.fu-
berlin.de/biodivinf/docs/CollectionModel/]

BioCISE (1999): BioCISE - Resource Identification for a Collection Informati-
on Service in Europe. Botanischer Garten und Botanisches Museum Berlin-
Dahlem. [http://www.bgbm.fu-berlin.de/biocise/]

Biosis (1999): Zoological Record. Biosis and the Zoological Society of London.
[http://www.york.biosis.org/zrdocs/zrprod/zoorec.htm]

Bisby (1997): Putting names to things and keeping track: the Species 2000
programme for a coordinated catalogue of life. — pp 59-68 in: Bridge, P., Jef-
fries, P., Morse, D. R. & Scott, P. R. (ed.): Information technology, plant
pathology and biodiversity. - CAB International, Wallingford.

BMU (Hrsg.) (1997): Wegweiser lokale Agenda 21 — Literaturansprechpartner;
erarbeitet vom Umweltbundesamt (UBA), Dezember 1997.

50

BMU (1998): Bericht der Bundesregierung nach dem Ubereinkommen tiber die
biologische Vielfalt. - Bundesministerium fiir Umwelt, Naturschutz und Reak-
torsicherheit, Bonn, März 1998.

BMU, UBA (Hrsg.) (1998): Handbuch Lokale Agenda 21 — Wege zur nachhalti-
gen Entwicklung in den Kommunen; vorgelegt vom Internationalen Rat für
kommunale Umweltinitiativen (ICLEI), Juni 1998.

BMU, UBA (Hrsg.) (1999): Lokale Agenda 21 im europäischen Vergleich;
vorgelegt vom Internationalen Rat für kommunale Umweltinitiativen (ICLEI),
Mai 1999.

Butler, D. (1998): Briefing Museums. — Nature 394: 105, 115-119.

CABRI (1999): Common Access to Biotechnological Resources and Informati-
on. — The CABRI Consortium. [http://www.cabri.org/]

CITES (1973/1979): Convention on International Trade in Endangered Species
of Wild Fauna and Flora. Signed at Washington, D.C., on 3 March 1973,
amended at Bonn, on 22 June 1979. — CITES Secretariat, Geneve. [http://
www.wemc.org.uk/ CITES/english/text.htm]

Collins, F. & D. Galas (1993): A new five-year plan for the U.S. human
genome program. — Science 262:43-46.

Collins, E.S., A. Patrinos,E. Jordan; Ar ChakravartipekenGice
steland & L. Walters (1998): New goals for the U.S. human genome
project: 1998-2003. — Science 282:682-689.

COP (1998): Convention on Biological Diversity. Texts of the decisions adopted
by the fourth meeting of the Conference of the Parties. [http://www.biodiv.
org/cop4/FinalRep-/1.html]

Council of Europe (1979): Convention on the Conservation of European
Wildlife and Natural Habitats (Bern Convention). — Council of Europe, Europe-
an Treaties ETS 104. [http://www.nature.coe.int/english/cadres/berne.htm]

Cracraft, J. (1995): The urgency of building global capacity for biodiversity
science. — Biodiversity and Conservation 4: 463-475.

Cracraft, J. (1996): Systematics, biodiversity science, and the conservation of
the Earth's biota. — Verhandlungen der Deutschen Zoologischen Gesellschaft
89: 41-47.

Croft, J., N. Cross, S. Hinchcliffe, E. Nic Luschacdhasstzsrr
Stevens, J.G. West & G. Whitbread (1999): Plant names for the 21st
century: the International Plant Names Index, a distributed data source of
general accessibility. — Taxon 48:317-324.

Diversitas (1996): Diversitas, an international programme of biodiversity
science. Operational Plan. Diversitas, Paris.

DSMZ (1999a): DSMZ - Deutsche Sammlung von Mikroorganismen und Zell-
kulturen GmbH. [http://www.dsmz.de/]

DSMZ (1999b): Bacterial Nomenclature up-to-date. [http://www.dsmz.
de/bactnom/bactname.htm]

5]

Duden Informatik (1993): Duden Informatik — Ein Sachlexikon fiir Studium
und Praxis. 2. Aufl., Dudenverlag.

ECNC (1999): Directory of Nature Conservation Organizations. European Centre
for Nature Conservation, Tilburg. [http://www.ecnc.nl/doc/europe/ organi-
za/director.html]

EEA-ETC/NC (1997): Databases on Species, Habitats and Sites, Survey and
Analysis 1995-96. European Environmental Agency, Kopenhagen.

Ehlers, E. & T. Krafft (ed.) (1998): German Global Change Research.
German National Committee on Global Change Research, Bonn. 128 pp.

Environment Australia (1998): The Darwin Declaration. Australian Biolo-
gical Resources Study, Environment Australia, Canberra.

ERMS (1998): European Register of Marine Species. U. of Southampton. [http://
www.sobs.soton.ac.uk/erms/]

ETC/CDS (1999): European Topic Centre on Catalogue of Data Sources. — Nieder-
sächsisches Umweltministerium, Hannover. [http://www.mu.niedersachsen.
de/cds/etc-cds_neu/etc-cds_home.html]

ETC/LC (1999): European Topic Centre on Land Cover. — Satellus AB, Kiruna.
[http://etc.satellus.se/]

ETC/NC (1999): European Topic Centre for Nature Conservation. — European
Environment Agency, Paris. [http://www.mnhn.fr/ctn/]

EU (1979): Council Directive of 2 April 1979 on the conservation of wild birds
(79/409/EEC). — Off. J. EC 25 April 1979: L 10371.

EU (1992): Richtlinie 92/43/EWG des Rates vom 21. Mai 1992 zur Erhaltung der
natürlichen Lebensräume sowie der wildlebenden Tiere und Pflanzen. — ABI.
22, Jet GOZ AN 206: 7:

EU (1994): Verordnung (EG) Nr. 1467/94 des Rates vom 20. Juni 1994 tiber die Er-
haltung, Beschreibung, Sammlung und Nutzung der genetischen Ressourcen der
Landwirtschaft. — Off. J. EC L 159:1-10. [http://europa.eu.int/comm/dg06/
res/gen/]

EU (1999): Community Research and Development Information Service COR-
DIS. [http://www .cordis.lu/]

EUA (1994): Europäische Umweltagentur — Informationen sinnvoll nutzen. —
Europäische Umweltagentur, Kopenhagen.

EUA (1998): Europas Umwelt — der zweite Lagebericht. — Europäische Umwelt-
agentur, Kopenhagen.

EURING (1999): Bird Ringing in Science and Environmental Management. —
Netherlands Institute of Ecology, Heteren. [http://www.nioo.knaw.nl/euring2.
htm]

FAO (1998): The FAO AFRICOVER Programme. [http://www.fao.org/ WAI-
CENT/FAOINFO/SUSTDEV/Eldirect/EIre0053.htm]

FAO (1999a): Domestic Animal Information System DAIS. [http://www.
fao.org/dad-is/entry.htm]

32

FAO (1999b): World Information and Early Warning System on plant genetic
resources WIEWS. [http://apps2.fao.org/wiews/]

Francki, R.1l.B., CoM Fauquer, DIE? Knudson dc Fy bimowan
(1990): Classification and nomenclature of viruses. — Archives of Virology
Suppl. 2:1-445.

G8 (1999): Final Report of the G8 Information Society Pilot Project - Theme 6
Environment and Natural Resources Management. U.S. Geological Survey,
Washington. [http://www.g7.fed.us/enrm/final.html |

Gase, I. P., Ar Cabela, I. Ernobrnja’lsarlovie, DzDolmenrarg
Grossenbacher, P. Haffner; J. Lescure, H. Martens 22
Martinez Rica, H. Maurin, M. E.Oliveira, T. S-Soflantdonz
M. Veith & A. Zuiderwijk (1997): Atlas of Amphibians and Reptiles in
Europe. — SEH & SPN-MNHN, Paris, 494 pp.

Gilbert, W. (1991): Towards a paradigm shift in biology. — Nature 349: 99.

GKSS (1999): LOTSE - Land Ocean Thematic Search Engine. [http://w3g.gkss.
de/lotse/]

Greuter, W.,H.M. Burdet & G. Long (1984): Med-Checklist. A critical
inventory of vascular plants of the circum-mediterranean countries, vol. 1.
Geneve & Berlin.

Greuter, W.,H.M. Burdet & G. Long (1986): Med-Checklist. A critical
inventory of vascular plants of the circum-mediterranean countries, 3. Dicotyle-
dones (Convolvulaceae-Labiatae). Geneve & Berlin.

Greuter, W.,H.M. Burdet & G. Long (1989): Med-Checklist. A critical
inventory of vascular plants of the circum-mediterranean countries, 4. Dicotyle-
dones (Lauraceae-Rhamnaceae). Geneve & Berlin.

Greuter, W., F. Barrie, H. M. Burdet, W. G. Chaloner; V. De-
moulin, D. L. Hawksworth, P. M. Jorgensen, D2 BeaNiicol-
son, P. C. Silva, P. Trehane & J. McNeill (ed.) (1994): Interna-
tional Code of Botanical Nomenclature (Tokyo Code). — Regnum Vegetabile
eile

Greuter, W.,R.K. Brummit, E. Farr, N. Kilian, P.M. Kirk & P.
C. Silva (eds.) (1993): NCU-3. Names in Current Use for Extant Plant
Genera. — Regnum Vegetabile 129. [database: http://www.bgbm.fu-
berlin.de/iapt/ncu/genera/]

Hagedorn, G. (1999): DeltaAccess Version 1.6 — an SQL interface to DELTA
(Description Language for Taxonomy), implemented in Microsoft Access.
[www.bgbm.fu-berlin.de /Projects/DeltaAccess/ DeltaAccess.html]

Hagemeijer, W.J.M. & MJ. Blair (ed.) (1997); the eB C@yalawon
European breeding birds. — Poyser, London.

Hawksworth, D.L. (1995): The Resource Base for Biodiversity Assessments.
— pp 545-605 in: UNEP: Global Biodiversity Assessment.

I

HELCOM (1992): Convention on the Protection of the Marine Environment of
the Baltic Sea Area, 1992. — Helsinki Commission - Baltic Marine Environment
Protection Commission, Helsinki. [http://www.helcom.fi/]

IGR (1999): GENRES - Informationssystem Genetische Ressourcen. Zentralstelle
für Agrardokumentation und -information, Bonn. [http://www.dainet.
de/genres/]

ISIS (1999): International Species Information System. [http://www .isis.org/]

ITTB (1998): ELISE, das Informationssystem für den Forschungsverbund “Elbe-
Ökologie” des BMBF. — Fraunhofer-Institut für Informations- und Datenver-
arbeitung IITB, Karlsruhe. [http://elise.bafg.server.de/]

iid 7 Suominen, R7 Tampinen (voli I1ff.), & A. Kurtto (vol.
12) (ed.) (1972-1999): Atlas Florae Europaeae Vol. 1-12. — Akateeminen
Kirjakauppa, Helsinki, 2039 pp.

Jury, S. (1998): Euro+Med PlantBase — a new initiative in plant systematics. U.
of Reading. [http://www.herbarium.reading.ac.uk/EuroMed/]

Kohler, F. & B. Klausnitzer (Hrsg.) (1998): Entomofauna Germanica 1.
Verzeichnis der Käfer Deutschlands. — Entomologische Nachrichten und Be-
richte. Dresden. Beiheft 4: 1-185.

Koperski, M. & M. Sauer (1999): Das Projekt “Referenzliste der Moose
Deutschlands” — Dokumentation unterschiedlicher taxonomischer Auffassun-
gen mit Hilfe des Datenbankprogrammes TAXLINK. — Stuttgarter Beitr. Na-
turk. Ser. A. 590: 1-10.

Lampinen, R. (1999): Atlas Florae Europaeae Database. -Finnish Museum of
Natural History, Helsinki. [http://www.helsinki.fi/kmus/afe/database.html]

Lane, M. (1998): Biological informatics, weaving a web of wealth. — Australian
Academy of Sciences, Canberra, 26 pp.

Linsenmair, K.E. (ed.) (1998): Biodiversity Research: General aspects and
state of the art in Germany. — pp 12-35 in: Ehlers, E. & Krafft, T. (eds): German
Global Change Research. Bonn.

May,R. (1994): Die Datenbank der Floristischen Kartierung. Ein Beispiel für die
dezentrale Erhebung und zentrale Zusammenführung von raumbezogenen,
naturschutzrelevanten Informationen. — pp 155-17 in: Kremers, H. (Hrsg.)
(1994 ): Praxis der Umweltinformatik. Band 5 Umweltdatenbanken. Marburg.

May,R.& W. Subal (1992): Einsatz von EDV für Kartierungsprojekte. — pp
87-93 in: Bergmeier, E. (Hrsg.) (1992): Grundlagen und Methoden Floristischer
Kartierungen in Deutschland. — Floristische Rundbriefe Beiheft 2, Göttingen.

Mitueimeliewones, Are Pr Reijnders & J. M. Ziman (ed) (1999): The
Atlas of European Mammals. Academic Press, 433 pp.

Naumann, C. & W. Greuter (1997): Die Biologischen Sammlungen —
Funktion, Situation, Perspektiven. Im Auftrag der Direktorenkonferenz natur-
wissenschaftlicher Forschungssammlungen Deutschlands (DNFS). — Zoo-
logisches Forschungsinstitut und Museum Alexander Koenig, Bonn.

54

NCBI (1999a): A decade of data at NCBI. - NCBI News Summer 1999: 1-4.

NCBI (1999b): Genetic Sequence Data Bank 15 August 1999, NCBI-GenBank
flat file release 113.0, distribution release notes. [ftp://ncbi.nlm.nih.gov/ gen-
bank/gbre I.txt]

Nikolai, R., W. Kazakos, R. Kramer, S. Behrens,W. Swoboda
& F. Kruse (1999): WWW-UDK 4.0: Die neue Generation eines Web-
Portals zu deutschen und österreichischen Umweltdaten. Tagungsband des 13.
Internationalen Symposiums "Informatik für den Umweltschutz”, Magdeburg.

NSF (1998): Biological Research Collections (BRC), Program Announcement
(NSF 98-126). — National Science Foundation, Directorate for Biological
Sciences, Division of Biological Infrastructure.

OECD 1999: Meeting of the OECD Committee for Scientific and Technological
Policy at ministerial level, Paris, 22-23 June 1999, conclusions. - OECD News
Release 23 June 1999, Paris. [http://www.oecd.org/news_and_events/release/
nw99-68a.htm]

Page, B.,E.Schikore & J. Mack (1996): Dokumentation der Umwelt-
informationssysteme des Bundes und der Länder. Erstellt im Auftrag des Bund-
Länder-Arbeitskreises Umweltinformationssysteme (BLAK UIS). — Hamburg
1996.

Pankhurst, R. (ed.) (1999): Database of the Flora Europaea. [http://www.rbge.
org.uk/forms/fe.html]

Pankhurst, R. & M. Pullan (1999): The Pandora taxonomic database
system. — Royal Botanic Gardens, Edinburgh. [http://www.rbge.org.uk/re-
search/pandora.home]

PNP (1999): The Plant Names Project. — The Royal Botanic Gardens, Kew, The
Harvard University Herbaria, and the Australian National Herbarium, Harvard.
[http://www.ipni.org/]

Rambold, G. & D. Triebel (1999): An online interactive key to the genera
of lichenized and lichenicolous Ascomycetes. — In: LIAS — a DELTA-based
determination and data storage system for lichenized and lichenicolous Asco-
mycetes. [www.botanik.biologie.uni-muenchen.de/botsamml/lias/liasonline.
html]

Rheinwald, G. (1993): Verbreitung und Häufigkeit der Brutvögel von
Deutschland. — Dachverband Deutscher Avifaunisten Nr.12, 264 Seiten.

Ride, W.D.L., H.G. Cogger, €. Dupuis,:O. Kraus a. Vitor
F.C. Thompson & P.K. Tubbs (eds.) (1999): International Code of
Zoological Nomenclature 4" Edition. — International Trust for Zoological
Nomenclature, British Museum (Natural History), London.

Riede, K. (1999): Das “Weltregister wandernder Tierarten” am Museum Koe-
nig. Bonn. [http://www .biologie.uni-freiburg.de/data/riede/gromsger.html]
Robbins,R. J. (1998): What is Biological Informatics? - Conference of Biolo-
gical Informatics, 6-8 July 1998, Canberra. [http://www.esp.org/rjr/canberra.

pdf]

3

Saarenmaa, H. (1998): Development of the EIONET, the European Environ-
ment Information and Observation Network. — In: Conference Agenda and
Papers, Intern. Conf. on the environment and related transport telematics re-
sults, June 4-5, 1998.

SBSTTA (1996): Practical approaches for capacity building for taxonomy. —
Note by the Secretariat. Subsidiary Body on Scientific, Technical and Techno-
logical Advice, Second Meeting, Montreal, 2 to 6 September 1996.
UNEP/CBD/ SBSTTA/2/5 19 July 1996.

Schminke, H.K. (1996): Naturkundliche Sammlungen — das vernachlässigte
Erbe? — Spektrum der Wissenschaft 1996:116-119.

Schumann, H., Bährmann, R. & A. Stark (Hrsg.) (1999): Entomofauna
Germanica 2. Checkliste der Dipteren Deutschlands. — Studia dipterologica.
Halle (Saale). Suppl. 2: 1-354.

Sneath, P.H.A. (ed.) (1992): International Code of Nomenclature of Bacteria,
1980 Revision. Washington.

Sp2000 (1999): Species 2000: Indexing the World’s known species.
[http://www.sp2000.org/]

Steininger, F.F. (ed.) (1996): Agenda Systematik 2000 — Erschließung der
Biosphäre. — Kleine Senckenberg-Reihe 22. Kramer, Frankfurt/M.

Swoboda, W.,F. Kruse, R. Nikolai, W. Kazakos, D. Nyhuis &
H. Rousselle (1999): The UDK Approach: the 4th Generation of an Envi-
ronmental Data Catalogue Introduced in Austria and Germany. Proceedings of
the Third IEEE Meta-Data Conference, Bethesda, Maryland. [http://computer.
org/conferen/proceed/meta/1999/papers/45/wswoboda.html |

Szentendre, Hungary. [http://www.rec.org/REC/Programs/Telematics/Deter-
mine/EuroInfoSession/HSaarenmaa.html |

TDWG (1999): International Union for Biological Sciences, Taxonomic Data-
base Working Group (TDWG). [http://www.tdwg.org/]

hammer Br Ze Da Brickell, Ba Ro Baum! W: Ls A. Hetter-
Seattle aA ec. Leslie, J. McNeill, S. A.:Spongbere’ & FE.
Vrugtman (1995): International Code of Nomenclature for Cultivated Plants.
[ICNCP or Cultivated Plant Code.] — Quarterjack Publishing, Wimborne.

Uetz, P. (1999): The EMBL reptile database, an online information resource on
reptile taxonomy. [http://www.embl-heidelberg.de/~uetz/LivingReptiles.html]

UN (1992): Text of the Convention on Biological Diversity. [http://www.biodiv.
org /]

UNEP (1976): Convention for the Protection of the Marine Environment and the
Coastal Region of the Mediterranean (Barcelona Convention). [http://www.
unepmap.org/barce.zip]

UNEP (1979): Convention on the Conservation of Migratory Species of Wild
Animals (Bonn Convention). [http://www.wcmc.org.uk/cms/cms_conv.htm]

UNEP (1995): Global Biodiversity Assessment. (ed. by V. H. Heywood). —
Cambridge University Press, Cambridge; xi, 1140 pp.

56

UNEP (1999a): Global Terrestrial Observing System. [http://www.fao.org/
GTOS/]

UNEP (1999b): Global Resource Information Database GRID. [http://www. grid.
unep.ch/gridhome.html |

UNESCO (1994): Übereinkommen über Feuchtgebiete, insbesondere als Le-
bensraum für Wasser- und Watvögel, von internationaler Bedeutung; geändert
durch das Pariser Protokoll vom 3.12.1982 und die Regina-Änderungen vom
28.5.1987. — Büro für internationale Normen und rechtliche Angelegenheiten
Organisation der Vereinten Nationen für Erziehung, Wissenschaft und Kultur,
Paris. [http://iucn.org/themes/ramsar/ key_conv_g.htm]

UNESCO (1999): Man and Biosphere Programme. [http://www.unesco.org/
mab/]

Weibel, S., J. Kunze, C. Lagoze & M. W 01174999) ZDubina@ore
Metadata for Resource Discovery. IETF #2413. — The Internet Society, Septem-
ber 1998

Wilson, D.E. & D.M. Reeder (ed.) (1993): Mammal Species of the World.
Smithsonian Institution Press, Washington. 1206 pp. On-line version unter
[http://www.nmnh.si.edu/msw/]

ZADI (1999): Deutscher Clearing-House-Mechanismus (CHM), nationale Um-
setzung des Übereinkommens über die biologische Vielfalt. — Zentralstelle für
Agrardokumentation und -information, BML. [http://www.dainet.de/bmu-cbd/]

ZfK (1993): Standardliste der Farn- und Blütenpflanzen der Bundesrepublik
Deutschland. Zentralstelle für die floristische Kartierung der Bundesrepublik
Deutschland (Nord), Bochum. — Flor. Rundbr. Beih. 3: 1-480.

Adressen der Autoren:

Walter G. Berendsohn, ZE Botanischer Garten und Botanisches Museum Berlin-
Dahlem, Freie Universität Berlin, Königin-Luise-Str. 6-8, D-14191 Berlin
Christoph L. Häuser, Staatliches Museum für Naturkunde, Rosenstein I, D-70191
Stuttgart

Karl-Heinz Lampe, Zoologisches Forschungsinstitut und Museum Alexander
Koenig, Adenauerallee 160, D-53113 Bonn

ANIC
ASC
BBA
BENE
BCIS
BDTWG
BIN
BGBM

BGCI
BIG
BIN21
BioCISE

BioFoS

BIOSIS/ZR

BMBF

BML
BMU
CABI

CABIKEY

CABRI

CBD

CDEFD

CETAF
CHM

CIESIN
CiES

ANHANG
Verzeichnis verwendeter Abkiirzungen

Australian National Insect Collection, Canberra

Association of Systematics Collections

Biologische Bundesanstalt fiir Land- und Forstwirtschaft
Biodiversity and Ecosystems International

Biodiversity Conservation Information System >BMU, BfN
Biodiversity Topic Working Group

Bundesamt fiir Naturschutz

Botanischer Garten und Botanisches Museum Berlin-Dahlem (Zen-
traleinrichtung der Freien Universitat Berlin)

Botanic Gardens Conservation International
Bundesinformationssystem Genetische Ressourcen > BMU, BfN
Biodiversity Information Network

Biological Collection Information Service for Europe - Resource
Identification (von der EU, DG-XII geförderte Konzertierte Aktion,
1997-1999)

Biologische Forschungssammlungen

Zoological Record (index to world zoology literature produced since
1864)

Bundesministerium für Bildung, Wissenschaft, Forschung und Tech-
nologie

Bundesministerium für Landwirtschaft

Bundesministerium für Umwelt, Naturschutz und Reaktorsicherheit
"Commonwealth" Agricultural Bureaux International

Computer Aided Biological Identification System, >CABl
Common Access to Biotechnological Resources and Information
(von der EU, DG-XII gefördertes Demonstrationsprojekt)
Convention on Biological Diversity (Übereinkommen über die biolo-
gische Vielfalt)

A Common Datastructure for European Floristic Databases (von der
EU, DG-XII geförderte Konzertierte Aktion, 1993-1995)
Consortium of European Large-Scale Taxonomic Facilities

Clearing House Mechanism of the Convention on Biological Di-
versity

Consortium for International Earth Science Information System

Convention on International Trade on Endangered Species; >UNEP

58

CMS
CODATA

CONABIO

COR

CORINE
DEIN
DELTA
DEG

DG
DNFS

DPD
DSMZ

EBI
ECNC
BEA
EFTA
EIONET

EMBL
ENRM
ERIN
ERMS
ETC/CDS
ERC/EE
ETC/NC
ETI

EU

EUA

EUNIS

Bonn Convention on Migratory Species; UNEP

Committee on Data for Science and Technology of the International
Council of Scientific Unions

Comision Nacional para el Conocimiento y el Uso de la Biodiversi-
dad (Mexiko)

Conference of Parties (Vertragsstaatenkonferenz) of the Convention
on Biological Diversity

Coordination of Information on the Environment
Deutsches Entomologisches Institut, Eberswalde
Descriptive Language for Taxonomy

Deutsche Forschungsgemeinschaft

Directorate General (EU Kommission Generaldirektoriat)

Direktorenkonferenz Naturwissenschaftlicher Forschungssamm-
lungen Deutschlands

Database of Plant Databases >IOPI

Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH,
Braunschweig

European Bioinformatics Institute
European Center for Nature Conservation
European Environmental Agency
European Free Trade Association

European Environmental Information and Observation Network >
EEA

European Molecular Biology Laboratory

G7 Environment and Natural Resources Management
Environmental Resources Information Network

European Register of Marine Species

European Topic Center on the Catalogue of Data Sources ~EUA
European Topic Center on LandCover ~EUA

European Topic Center on Nature Conservation ~EUA

Expert Center for Taxonomic Identification, Amsterdam
Europäische Union

Europäische Umweltagentur

European Information System on Nature

EURING
FAO
GBIF

GEF
GEIN
GELOS
GENRES
GIS
GMSD's
GPC
GROMS
GUS
IAPT
ICLARM
IGR
ILDIS
IMI
INBIO
IOP
IOPI
IPK
IPNI
ISIS
IUBS
IUCN
IUMS
LIAS

MAB
MHNSM
MTD
NCBI
NCU

European Union for Bird Ringing
Food and Agriculture Organization of the United Nations

Global Biodiversity Information Facility; OECD Megascience
Working Group on Biological Informatics

Global Environmental Facility

German Environmental Information Network

Global Environmental Information Locator Service
Informationssysteme Genetische Ressourcen ~BML, ZADVIGR
Geographisches-Informations-System

Global Master Species Databases; >SP 2000 (Species 2000)
Global Plant Checklist; >IOPI

Global Register of Migratory Species

Gemeinschaft unabhängiger Staaten (ehemalige Sowjetunion)
International Association for Plant Taxonomy

International Centre for Living Aquatic Resources Management
Informationszentrum für genetische Ressourcen (?ZADI, > BML)
International Legume Database & Information Service
International Mycological Institute (Index of Fungi); ~CABI
Instituto Nacional de Biodiversidad (Costa Rica)

International Organization for Palaeobotany

International Organization for Plant Information »*GPC, SP2000
Institut fiir Pflanzengenetik und Kulturpflanzenforschung Gatersleben
International Plant Name Index

International Species Information System

International Union of Biological Sciences

International Union for Conservation of Nature

International Union of Microbiological Societies

DELTA-based Determination and Data Storage System for Licheni-
zed and lichenicolous Ascomycetes (DFG-Projekt)

Man and Biosphere »~UNESCO

Museo de Historia Natural, “San Marcos”, Lima, Peru

Museum fiir Tierkunde Dresden

National Center for Biotechnology Information der USA( >GenBank)
Names in Current Use >IAPT

60

NHM
NSF
OECD
OPTIMA

BERS
PNP
RMNL
SA 2000
SBSTTA
SCBD
SP2000
SRE
TDWG

TITAN
TRITON
UBA
UDK

UIs

UN
UNDP
UN/ECE
UNEP
UNESCO
USDA
US FGDC
US-OBI
WBD
WCMC
WDCM
WFCC
WRI
WWE

The Natural History Museum, London, UK
National Science Foundation der USA
Organization for Economic Cooperation and Development

Organization for the Phytotaxonomic Investigation of the Mediterra-
nean Area

Plant Fossil Record Information System >IOP

The Plant Names Project

Rijksmuseum van Natuurlijke Historie, Leiden, Niederlande
Systematics Agenda 2000

Subsidiary Body on Scientific, Technical and Technological Advice
Secretariat of the Convention on Biological Diversity

Species 2000 project

Species Plantarum Project >IOPI

Taxonomic Databases Working Group (IUBS Commission on Taxo-
nomic Databases)

Taxonomic Index to Animal Names; >BIOSIS/ZR

Taxonomy Resource & Index To Organism Names >BIOSIS/ZR
Umweltbundesamt

Umwelt-Daten- Katalog

Umwelt-Informations-System

United Nations

United Nations Development Program

UN Economic Commission for Europe

United Nations Environment Program

United Nations Educational, Scientific and Cultural Organization
United States Department of Agriculture

US Federal Geographic Data Commission

US Organization for Biodiversity Information

World Biodiversity Database >ETI

World Conservation Monitoring Center

World Data Centre for Microorganisms

World Federation for Culture Collections

World Resource Institute

Worldwide Fund for Nature

WWW
ZADI
ZADVIGR
EK

ZFMK

ZMB

61

World Wide Web (Internet)
Zentralstelle für Agrardokumentation und -information; >BML
Information Centre for Genetic Resources "BML

Zentralstelle für die floristische Kartierung der Bundesrepublik
Deutschland

Zoologisches Forschungsinstitut und Museum Alexander Koenig,
Bonn

Zoologisches Museum des Naturhistorischen Museums der Humboldt
Universitat Berlin

=

PRU

x

BR P
’ ri EG
re Aa

In der Serie BONNER ZOOLOGISCHE MONOGRAPHIEN sind erschienen:

13.

14.

13:

16.

ik

18.

9:

20.

2.

22.

23.

Naumann, C.M.: Untersuchungen zur Systematik und Phylogenese der holark-
tischen Sesiiden (Insecta, Lepidoptera), 1971, 190 S., DM 48,—

Ziswiler, V., H.R. Güttinger & H. Bregulla: Monographie der Gattung
Erythrura Swainson, 1837 (Aves, Passeres, Estrildidae). 1972, 158 S., 2 Tafeln,
DM 40,—

Eisentraut, M.: Die Wirbeltierfauna von Fernando Poo und Westkamerun.
Unter besonderer Berücksichtigung der Bedeutung der pleistozänen Klimaschwan-
kungen für die heutige Faunenverteilung. 1973, 428 S., 5 Tafeln, DM 106,—
Herrlinger, E.: Die Wiedereinbürgerung des Uhus Bubo bubo in der Bundes-
republik Deutschland. 1973, 151 S., DM 38,—

Ulrich, H.: Das Hypopygium der Dolichopodiden (Diptera): Homologie und
Grundplanmerkmale. 1974, 60 S., DM 15,—

Jost, O.: Zur Okologie der Wasseramsel (Cinclus cinclus) mit besonderer Beriick-
sichtigung ihrer Ernährung. 1975, 183 S., DM 46,—

Haffer, J.: Avifauna of northwestern Colombia, South America. 1975, 182 S.,
DM 46,—

Eisentraut, M.: Das Gaumenfaltenmuster der Säugetiere und seine Bedeutung
für stammesgeschichtliche und taxonomische Untersuchungen. 1976, 214 S.,
DM 54,—

Raths, P., & E. Kulzer: Physiology of hibernation and related lethargic states
in mammals and birds. 1976, 93 S., 1 Tafel, DM 23,—

. Haffer, J.: Secondary contact zones of birds in northern Iran. 1977, 64 S., 1 Falt-

tafel, DM 16,—

. Guibé, J.: Les batraciens de Madagascar. 1978, 144 S., 82 Tafeln, DM 36,—
. Thaler, E.: Das Aktionssystem von Winter- und Sommergoldhähnchen (Regulus

regulus, R. ignicapillus) und deren ethologische Differenzierung. 1979, 151 S.,
DM 38,—

Homberger, D.G.: Funktionell-morphologische Untersuchungen zur Radiation
der Ernährungs- und Trinkmethoden der Papageien (Psittaci). 1980, 192 S.,
DM 48,—

Kullander, S.O.: A taxonomical study of the genus Apistogramma Regan, with
a revision of Brazilian and Peruvian species (Teleostei: Percoidei: Cichlidae). 1980,
152 S., DM 38,—

Scherzinger, W.: Zur Ethologie der Fortpflanzung und Jugendentwicklung des
Habichtskauzes (Strix uralensis) mit Vergleichen zum Waldkauz (Strix aluco). 1980,
66 S., DM 17,—

Salvador, A.: A revision of the lizards of the genus Acanthodactylus (Sauria:
Lacertidae). 1982, 167 S., DM 42,—

Marsch, E.: Experimentelle Analyse des Verhaltens von Scarabaeus sacer L. beim
Nahrungserwerb. 1982, 79 S., DM 20,—

Hutterer, R., & DC.D. Happold: The shrews of Nigeria (Mammalia: Sorici-
dae). 1983, 79 S., DM 20,—

Rheinwald, G. (Hrsg.): Die Wirbeltiersammlungen des Museums Alexander
Koenig. 1984, 239 S., DM 60,—

Nilson, G, & C. Andrén: The Mountain Vipers of the Middle East — the
Vipera xanthina complex (Reptilia, Viperidae). 1986, 90 S., DM 23,—
Kumerloeve, H.: Bibliographie der Säugetiere und Vögel der Türkei. 1986,
132 S., DM 33,—

Klaver, C., & W. Böhme: Phylogeny and Classification of the Chamaeleonidae
(Sauria) with Special Reference to Hemipenis Morphology. 1986, 64 S., DM 16,—
Bublitz, J.: Untersuchungen zur Systematik der rezenten Caenolestidae Trouess-
art, 1898 — unter Verwendung craniometrischer Methoden. 1987, 96 S., DM 24,—

24.

Zar

26.

2

28.

29;

30.

31.

32:

33

34.
35.

36.

aT.

38.
39,

40.

41.

2.
43.
44.

45.

-

Arratia, G.: Description of the primitive family nalen (sin
Teleostei, Pisces): Morphology, taxonomy zur a implicati Ss.
120 S., DM 30,— 2

Nikolaus, G: Distribution atlas of Sudan’s birds with notes ( on ha vite
status. 1987, 322 S., DM 81,—

Löhrl, H.: Etho-ökologische en an ne Klei erat
tidae) — eine vergleichende Zusammenstellung. 1988, 208 S., DM 52, 2

Böhme, W.: Zur Genitalmorphologie der Sauria: Funktionelle unk
geschichtliche Aspekte. 1988, 175 S., DM 44,— | 2

Lang, M : Phylogenetic and biogeographic patterns of Basiliscine
(Reptilia: Squamata: “Iguanidae”). 1989, 172 S., DM 43, — | 2

Hoi-Leitner, M.: Zur Veranderung der Säugetierfauna des Neusi
Gebietes im Verlauf der letzten drei Jahrzehnte. 1989, 104 S., DM 26,

Bauer, A. M.: Phylogenetic systematics and Biogeography of the ‚Car
lini (Reptilia: Gekkonidae). 1990, 220 S., DM 55,— oe

Fiedler, K.: Systematic, evolutionary, and ecological inplicione
phily within the Lycaenidae a we N
DM 53,—

Arratia, G.: Development and variation of the use ae 0 7
fishes (Teleostei: Ostariophysi) and their phylogenetic ae
DM 37,— .

Kotrba, M.: Das Reproduktionssystem von Cyrtodiopsis a u
sidae, Diptera) unter besonderer Berücksichtigung der inneren ae
schlechtsorgane. 1993, 115 S., DM 32,—

Blaschke- Berthold, U.: Anatomie und Phylogenie de Bibione Ä
secta, Diptera). 1993, 206 S., DM 52,— =
Hallermann, J.: Zur Morphologie der Ethmoidalregion der fen. «
— eine vergleichend-anatomische Untersuchung. 1994, 133 S., DM 3 ;
Arratia, G., & L. Huaquin: Morphology of the lateral line syste
skin of Diplomystid and certain primitive Loricarioid Catfishes and
and ecological considerations. 1995, 110 S., DM 28,
Hille, A.: Enzymelektrophoretische Untersuchung zur see Pope
struktur und geographischen Variation im Zygaena-transalpina-Supe
Komplex (Insecta, Lepidoptera, Zygaenidae). 1995, 224 5, DM 36 =
Martens, J, &S. Eck: Towards an Ornithology of the Himalayas: S
ecology and vocalizations of Nepal birds. 1995, 448 S., 3 Farbtafeln, D
Chen, X: Morphology, phylogeny, biogeography and Bie cn F
(Pisces: Cyprinidae). 1996, 227 S., DM 57,— 5
Browne, DIL &C.H Scholtz The morphology of the Wind o g
and wing base of the Scarabaeoidea (Coleoptera) with some PD
cations. 1996, 200 S., DM 50,—
Bininda-Emonds, O.R.P,&A.P. Russell: A morphologies a
the phylogenetic relationships of the extant phocid seals Me
Phocidae). 1996, 256 S., DM 64,—

Klass, K.-D.: The external male genitalia = the phylogeny of Bl
Mantodea. 1997, 341 S., DM 85, — |

Hörnschemeyer, T.: Morphologie und Evolution des Fligelgelenk
optera und Neuropterida. 1998, 126 S., DM 32,— eee
Solmsen, E-H.: New World nectar-feeding bats: biology,

craniometric approach to systematics, 1998, 118 S., DM 30,—
Berendsohn, W.G., C.L. Hauser & K.-H. Lampe: Biodiversitäts
Deutschland: Bestandsaufnahme und Perspektiven, 1999, 64 S., DM 16

HECKMAN

BEE NED EO RY, TaNEG;
Bound-To-Please®

MAY 01

N. MANCHESTER, INDIANA 46962

SMITHSONIAN INSTITUTION LIBRARIES

AMON

3 01206 9902

ua
BS ehnger

Ahr

Mo tee ne

st mag