'J ' ■;. . . • AJ

595.70673

A5I

S. I. LIBRARY.

I

MEMOIRS

OF THE

AMERICAN ENTOMOLOGICAL SOCIETY

NUMBER 3

A VENATIONAL STUDY
OF THE SUBORDER ZYGOPTERA

(ODONATA)

WITH KEYS FOR THE IDENTIFICATION
OF GENERA

BY
PHILIP A. MUNZ

/ WAR -3 1921 'l

PUBLISHED BY THE AMERICAN ENTOMOLOGICAL SOCIETY

AT THE ACADEMY OF NATURAL SCIENCES

PHILADELPHIA

1919

THE AMERICAN ENTOMOLOGICAL SOCIETY

PHILADELPHIA
Founded 1859 Incorporated 1862

Organized as "The Entomological Society of Philadelphia,"
the present title was adopted February 11, 1867.

President
Henry Skinner

Vice-President
James A. G. Rehn

Corresponding Secretary Recording Secretary

Morgan Hebard George M. Greene

Treasurer

Ezra T. Cresson

Publication Committee Finance Committee Property Committee

James A. G. Rrfm James A. G. Rehn Ezra T. Cresson, Jr.

Editor Chairman Librarian and Cvstodian

Ezra T. Cresson D. M. Castle Morgan Hebard

Philip P. Calvert Morgan Hebard Philip Laurent

Meetings held (alternately with those of the Entomological Section of the Academy
of Natural Sciences of Philadelphia) on the fourth Thursday of September, October,
November, January, February, March, April and May, and on the second Monday of
June and December (annual meeting). The offices, meeting room, library and collections
of the Society are in the Entomological Department of the Academy of ^fatu^al Sciences
of Philadelphia, Logan Square.

Serial Publications of the Society

Proceedings of the Entomological Society of Philadelphia, volumes I-VI, 1861-1867. Out
of print and cannot be supplied.

Transactions of the American Entomological Society, volumes I-IV, out of print and can-
not be supplied; V-X, $5.00 per volume; XI-XLIV, $4.00 per volume. Published in
annual volumes of quarterly parts.

The Practical Entomologist, volumes I-II, 1865-1867. Price, $1.00 for the two volumes.

Memoirs of the American Entomological Society, number 1, 1916, price $2.00; number 2,
1917, price $4.50. Published irregularly.

Entomological News, volumes I-XXIX. Price $2.00 per volume. Published in con-
junction with the Entomological Section of the Academy of Natural Sciences of
Philadelphia.

Various irregular, non-serial publications have been issued, which may be found in our
price lists under their respective authors.

To establish the date of issue, each paper, as soon as completed, is placed on sale and
a copy of it is mailed to each of the following institutions, with a request that it be placed
in their library for reference.

Boston Society of Natural History Field Museum of Natural History

American Museum of Natural History California Academy of Sciences
Smithsonian Institution Entomological Society of Ontario

Separates of papers published in the Proceedings and Transactions, if not out of print,
can be supplied at prices given in our price lists, which are sent upon application, stating
order or orders desired. The Society has also inaugurated a system of subscription to
papers published by it on individual orders, which arrangement is of material benefit to
the student who desires promptly the literature on any special group, but who does not
deaire to subscribe to entire volumes of the Transactions. A method of notification of the
appearance in its pirf>lication8 of papers relating to individual orders or subjects has been
developed by the Society. Information concerning these arrangements supplied upon
application.

In sending subscriptions or in making inquiries relative to publications, address:
THE AMERICAN ENTOMOLOGICAL SOCIETY,
Publication Department,

The Academy of Natural Sciences of Philadelphia,
Logan Square.

A VENATIONAL STUDY of the SUBORDER
ZYGOPTERA (ODONATA) WITH KEYS
for the IDENTIFICATION OF GENERA

A THESIS

Presented to the Faculty of the Graduate School

OF Cornell University for the degree of

doctor of philosophy

BY
PHILIP A. MUNZ

MEMOIRS OF THE AMERICAN ENTOMOLOGICAL SOCIETY

NO. 3. 1919

MEMOIRS

OF THE

AMERICAN ENTOMOLOGICAL SOCIETY

NUMBER 3

A VENATIONAL STUDY
OF THE SUBORDER ZYGOPTERA

(ODONATA)

WITH KEYS FOR THE IDENTIFICATION
OF GENERA

BY

PHILIP A. MUNZ

PUBLISHED BY THE AMERICAN ENTOMOLOGICAL SOCIETY

AT THE ACADEMY OF NATURAL SCIENCES

PHILADELPHIA

1919

{Issued March 13, 191 9.)

Memoirs

OF THE

American Entomological Society

Number 3

A VENATIONAL STUDY OF THE SUBORDER ZYGOPTERA
(ODONATA) WITH KEYS FOR THE IDENTIFICATION

OF GENERA 1

By Philip A. Munz
CONTENTS

PAGE

I. Introduction ^

Acknowledgements 3;

Nomenclature of wing-veins 4-

Explanation of figures 7-

II. A Discussion of the Phylogeny 7-

The Anisozygoptera 7-

The Zygoplera g-,

The A grionidae q

The Polythorinae 10

TheAgrioninae 1 1

The Epallaginae 12

The Disparocyphinae 13

The Coenagrionidae 13,

The Lestinae 15,

The Lestoidinae i j

The Megapodagrioninae 17

The Pseudostigmatinae 20^

The Coenagrioninae 21

The Protoneiirinae 24.

1 A Contribution from the Entomological Laboratory of Cornell University.

MEM. AM. ENT. SOC, 3.

2 A VENATIONAL STUDY OF THE ZYGOPTERA

III. Tendencies in Specialization 26

Vein Reduction ~1

1. Reduction in the number of supplementan' sectors 27

2. Reduction of the veins Cu^ and Cui . 28

3. Decrease in the number of antenodal cross-veins 28

4. Decrease in the number of postnodals 29

5. Loss of the cross-veins in the quadrangle and subquadrangle 29

6. Reduction in the number of those antenodal cells situated between M4
andCui 29

7. Reduction in the number of cross-veins behind the anal vein 29

Vein Shifting 3^

1 . Narrowing of the wing 30

2. Progressive petiolation 3^

3. Retraction of the nodus toward the base of the wing 31

4. Tendency Jor M2, Rg, and M3 to arise nearer the tip of the wing and
further from the subnodus 31

5. Migration outward of the base of Mia 33

6. The position of the second antenodals before the arculus 33

7. The tendency for the quadrangle to become acute distally 33

8. Change in the length of the quadrangle 35

9. Attachment of supplementary sectors 35

10. Straightening of the longitudinal veins 35

1 1 . Matching up of the cross-veins in trans\erse series 36

Miscellaneous Features 36

1. Differentiation of fore and hind wings 36

2. The bracing of the stigma 3^

3. Atrophy of the stigma 37

IV. Summary 37

V. Keys for the Identification of Genera 38

The Anisozygoptera 40

The Zygoptera 4^

"The Agrionidae 40

The Polythorinae 42

The Agrioninae 42

■ The EpaUaginae 45

The Disparocyphinae 47

The Coenagrionidae 4^

The Lestinae 4^

The Lestoidinae 49

The Pseudostigmatinae 49

The Megapodagrioninae 5t>

The Coenagrioninae 53

The Proloneurinae 63

VT . Bibliography 67

VII. Plates 72

PHILIP A. MUNZ 3

I— INTRODUCTION

At the present time it is not necessary to offer either explanation
or excuse for the use of venation as a guide to the phylogeny of a
group of insects, particularly of any of the Neuropteroids, the
exceedingly rich and varied veining of which has been the basis of
many excellent pieces of work. The purpose of this paper is to
make the fullest possible use of wing-veins as a means of interpre-
ting relationships and as a method of identifying the genera of one
of the suborders of the Odonata, the Zygoptera. It must be remem-
bered, however, that venation is not the only index to genealogy,
nor perhaps always the most reliable, but it is one of the most
evident and fundamental and therefore most valuable methods
which we have to enable us to determine relationships. In some
cases, without doubt, a study of genitalia and morphology will throw
more light on the situation and even change some of our concep-
tions which have been arrived at from the study of the wings only.

The nomenclature of the veins of the Zygoptera and their
homologies to those of other groups have been thoroughly worked
out, but, since the days of Baron de Selys, no attempt has been
made to consider the whole suborder in detail and to make keys
for identification. A word of explanation may be permissible in
regard to the keys herewith submitted; I have examined such
material as I have had access to in an endeavor to make them as
generally applicable as possible; but in some cases descriptions ,
only could be used, since during the past two years correspondences
with many European workers and hence the securing of many\
desirable figures ha\'e Ijeen imj^ossible. I feel, however, that the
paper is justifiable, if for no other reason than to ha\'e in one place
in the literature such wing-photographs as have been collected in
the course of the work, the \'enation of many genera not ha\-ing
been published before.

A cknowledgemeuts

To the many who have kindly contributed material and sugges-
tions for this study my gratitude is due and is thankfully given.
This problem was undertaken at the suggestion and under the
direction of Professor James G. Needham, whose unfailing interest

MEM. AM. ENT. SOC, 3.

'4 A VENATIONAL STUDY OF THE ZYGOPTERA

and valuable criticism have made possible what has been accom-
plished. He has been very generous, in the time and abundant
material which he has given me. Mr. C. H. Kennedy, also of Cor-
inell University, who has been working on the same phylogenetic
problem from another method of approach, has never hestitated in
giving me suggestions arrived at from his study, has made several
•drawings for me and brought material from museums which I was
not able to visit. Dr. P. P. Calvert and Dr. Henry Skinner of
Philadelphia very generously loaned specimens and permitted the
use of the extensive collections at the Academy of Natural Sciences
<of Philadelphia. Mr. Carl Ilg of Philadelphia kindly copied
tseveral descriptions which were not available at Cornell. Mr.

E. B. Williamson of Bluffton, Indiana, loaned valuable material;
Dr. Samuel Henshaw of Cambridge sent material for study and
permitted Mr. E. Avery Richmond of Cornell to photograph
several specimens at the Museum of Comparative Zoology. Mr.
Rolla P. Currie and Miss Currie were very kind and helpful in

/allowing me the use of the collections at the National Museum.
Dr. W. J. Holland and Mr. Hugo Kahl of the Carnegie Museum
permitted Mr. Kennedy to bring several specimens to Ithaca and
to study others at Pittsburg. Mr. Herbert Campion and Mr.

F. W. Campion sent several photographs of specimens at the
British Museum. Dr. F. Ris of Rheinau, Switzerland, kindly sent
proof-sheets of an article expressing his views on the venation of
the Zygoptera, and Mr. R. J. Tillyard of New South Wales, Aus-
tralia, sent several papers dealing with the same subject and with
new and rare Australian Zygoptera. To all of these contributors
is due much of the success in the completeness which it has been
possible to secure.

Nomenclature oj the Wing-Veins
The Comstock-Needham system of naming the veins has come
Into very general use among workers on the Odonata since the
Genealogic Study by Professor Ncedham in 1903; and it is to be
retained in all its essentials as presented in his paper. Tillyard,
however, (19 14) showed that the vein formerly known as A is in
reality partly a secondary structure built back toward the base of
the wing from one of the so-called "cubito-anal " cross-veins and

PHILIP A. MUNZ 5

that the vein A at the base of the wing is fused with Cu, but crosses
the space back of Cu and then proceeds as formerly thought
toward MA. He therefore proposes the name "anal-crossing"
(Ac) for this apparent cross-vein, and the abbreviations Ab for the
bridge toward the wing-base, and A' for the distal portion of the
anal vein. This system will be used in this paper when especial
attention is called to some part of this vein, otherwise A alone will
be used to designate the anal vein in general; the sign 1st A is not
necessary since the Zygoptera possess only one anal vein. I agree
with Mr, Tillyard that the sign A followed by an asterisk as used
by Dr. Ris is unfortunate; when I read one of his descriptions I find
myself involuntarily looking to the bottom of the page for the foot-
note which I always associate with the use of the asterisk.

Mr. Tillyard's contention (1915) and that of Dr. Ris (1916a)
that the Zygoptera present a difTerent condition than that of the
Anisoptera in the branching of the radius do not seem well founded.
To be sure none of the nymphs of the Zygoptera that have yet been
examined show the trachea Rg crossing over that of M1+2 as do the
Anisoptera, however it seems inconceivable that the vein Rg of the
Anisoptera can have in the Zygoptera a vein so completely analo-
gous in position and yet not be homologous. Mr. Tillyard's
name "median sector" is a matter of words only; it indicates
no homology to the veins of any other order, although in retaining
the same name for the different branches of media he admits that
the media has exactly the same position in both suborders, which
state cannot be true if the radius does not present the same con-
dition in both. Professor Needham (191 7) has shown how many
of 'Sir. Tillyard's arguments can be used to support the opposite
view of the case. I therefore have no hesitation in following the
usual system of nomenclature.

As an example of the Agrionidae the genus Agrion {Calopteryx)
may be used with the abbreviations given in fig. A. These abbre-
viations are given as follows:

C=costa, Sc=subcosta; Ri = radius-one, R,,= radical sector;
Mi = media-one, M2 =media-two, M3 = media-three, M4=media-
four, Mia = the longest sector between Mi and M2; Cu = cubitus,
Cui = cubitus-one, Cuo =cubitus-two, Cu2a=a branch of Cu2

MEM. AM. ENT. SOC, 3.

A VENATIONAL STUDY OF THE ZYGOPTERA

arising just beyond MA; A = first anal vein, Ac = anal crossing,
Ab=anal bridge, A' = anal vein from Ac to MA; N= nodus,
Sn=subnodus; St=stigma, B.S. =basal space, B.R.S. =basal
radial space; An=antenodal cross- veins, Pn=postnodal cross-
veins, Ar =arculus, MA =medio-anal link; Q = quadrangle, Sq =
subquadrangle.

a

AT,

Fig. A. Venation of Agrion (Calopteryx).

A/,

These terms are all familiar ones and need for the most part no
definition here. The subquadrangle is that space between Ac and
MA, and between Cu and A, or Cu and the hind margin of the
wing when A' is not present. Ac cannot always be located because
of the numerous cross-veins between Cu and A to be found in some
genera, but can generally be distinguished by its obliquity. The
term medio-anal link is one defined by Mr. Williamson (191 3) "as
the cross vein forming the distal end of the quadrangle and that
part of Cu2 opposed to Cu, and placed transversely to the long axis
of the wing." Supplementary sectors are those longitudinal veins
present between the principal veins. The quadrangle, subquad-
rangle, basal space, and basal radial space are said to be crossed
when they have cross \'eins, and Siva free when without cross-\eins.
The basal radial space is open when Mi+o docs not touch R, and
closed when M1+2 fuses with R. Of the antcnodals there are two
series: those of t\\Q first or costal series, i.e. those between C and Sc;
and those of the second or subcostal series, i.e. between Sc and R.

PHILIP A. MUNZ

Fig. B. N'enation of Argiocnemis.

As an example of a wing of the Coenagrionidae, Argiocnemis may
be used, for which see fig. B. The notations are largely as in fig. A,
except that there is no basal radial space present and that the ante-
nodel spaces are few enough to be numbered; the first antenodal
space is that between the base of the wing and the first antenodals,
the second antenodal space is that between the first and second
antenodals, and the third antenodal space is that between the second
antenodals and the nodus. Cu2a is not present.
Explanation of Figures

In referring to figures, A, B, C, etc. are the text figures, while
figures I, 2, 3, etc. are to be found in the plates at the end of the
paper.

n— A DISCUSSION OF THE PHYLOGEXY

Passing now from a review of the system of nomenclature used,
the next step is the consideration of the phylogeny of the Zygoptera
and their relationship to other groups.

The Anisozygoptera

In his treatise on fossil insects, Handlirsch (i 906-1 908) created
as a new suborder the Anisozygoptera, for many fossil forms and
for the living genus Palaeophlebia, to which he gave the name
Neopalaeophlebia, since the former name was preoccupied. Dr.
Calvert (1903), however, had already used the name Epiophlebia
for Palaeophlebia and it is his name that must be used. This
genus according to Handlirsch falls as readily into the Anisoptera
as the Zygoptera and to keep the two groui)s distinct, a third
intermediate one is necessary. Epiophlebia (fig. i) is the only
known survivor of a large suborder of Mesozoic insects and is

MEM. AM. ENT. SOC, 3. ' '

8 A VENATIONAL STUDY OF THE ZYGOPTERA

therefore of considerable importance phylogenetically. Hence,
while it is not now one of the Zygoptera, it deserves discussion in
this paper.

Aside from its venational primitiveness it has been shown by
Schmidt (1915) to be intermediate between the two modern sub-
orders in the structure of the male genitalia of the second and third
abdominal segments. Mr. Kennedy informs me that this is so
obviously the case that only a separate suborder will suffice to
show its points of difference. The broad area between M4 and Cui
just beyond MA; the shape of the quadrangle, especially that of
the hind wing where a cross-vein in the right position would give
the supertriangle and triangle of the Anisoptera the zigzag con-
dition of the supplementary sectors, — all these are venational fea-
tures which justify its position as something quite distinct from
the Zygoptera. In itself it does not need to be taken as a form from
which the Zygoptera and Anisoptera have descended, in fact it can-
not be so considered, but it does show affinities with both and, as a
survival of the ancestral group from which our two modern sub-
orders have sprung, it is worthy of note.

The Zygoptera Proper

The Zygoptera may therefore evidently be considered as deriva-
tives of the Anisozygoptera. They may themselves be di\ided
into two families, which were long made subfamilies by workers on
Odonata. It would be very desirable and lead to much less con-
fusion if these families could retain the old names; Calopterygidae
and Agrionidae, but since most writers on the group have recently
followed Kirby's Catalogue in using the genus names Agrion and
Coeiiagrion rather than Calopteryx and Agrion, and since Mutt-
kowski (19 10) has brought additional evidence to show that these
are the names to be used, I shall follow their example and speak of
the old group Calopteryginae as Agrionidae and of the former
Agrioninae as Coenagrionidae.

Of the generic names proposed, about one hundred and eighty
now seem valid and are recognized in the keys hereto appended.
Of these about forty come in the first family, the Agrionidae,
which is the more generalized of the families and ends in forms
very closely related to some of the Coenagrionidae. The two

PHILIP A. MUNZ 9

families, in fact, do not form a dichotomy, hut the one leads to the
other and is connected with it by so many genera that the dividing
line is in some cases difficult to find. In a general way the former
one is characterized by the presence of se\eral to many antenodals
and by cross-veins generally being present in the fjuadrangle and
subquadrangle. M3 separates from M1+2 nearer to the arculus
than to the subnodus, and the nodus is far out in the wing, at from
one-third to one-half of the length of the wing. These tendencies
in venation are all fairly constant and, as will be shown later, in-
dicate a rather primitive condition. The (\)enagrionidaje, on the
other hand, have two antenodals generalh' and no cross veins in
the quadrangle or subquadrangle, with but few exceptions. In
one line only, that of the Lestinae and Lestoidinae, does M3 retain
its original condition of branching off from M1+2 near the arculus,
in the other groups its point of origin migrates outward. The
nodus, however, is drawn toward the base of the wing and ranges
in position from about one-seventh to one-third of the length of
the wing. The more detailed account of the exact steps in phy-
logeny may better be discussed under the families and subfamilies.

The Agrionidae

As was said before, this famih- includes the old group Calop-
teryginae which was divided' l)y de Selys into five legions : Thore,
Calopteryx, Euphaea, Libellago, and Amphipteryx. Yenationally
these five groups are better arranged in three subfamilies, as was
done by Professor Needham (1903a), which I shall call : Polythorinae,
to include the legion Thore; Agrioninae, for the legion Calopteryx,
and the Epallaginae, for the last three legions. To these must now
be added a fourth subfamily, the Disparocyphinae for the genus
Disparocypha (fig. 37), recently described by Dr. Ris (1916b).
This very remarkable genus will be discussed later with the reasons
for its segregation from the others.

The affinities of the Agrionidae with Epiophlehia (fig. i) are
evident enough: the large number of antenodals, the origin of M3
far before the subnodus, and the location of the nodus far out in
the wing.

MEM. AM. ENT. SOC, 3.

10 A VENATION AL STUDY OF THE ZYGOPTERA

Proto/ieuri/nae

Lestoidi/sae

Lestimae

CoEMAGRIOniNAE,

[Aegapodagrionimae

PsEUDOSTIGMATI/iAE

Epallaginae

DlSPAROCYPHITSAE

Agrionimae

PoLYTIiORIMAE

tPIOPHLEBIMAE

Amisozygoptera

Fig. C. Proposed Phylogenetic Tree for the Zygoptera.

The Polythorlnae

The most shari)ly differentiated group of the Agrionidae, in that

it differs most from the others, is the Volythorinae. In it AI does

not descend the arcuhis to any extent, so that the quadrangle is

very much wider jiroximally than (list ally and is concave anteriorly.

PHILIP A. MUXZ II

This peculiar little group of four genera is a rather compact one, but
shows within itself the ever present tendencies of specialization to
be found in the other subfamilies. It retains a rich venation with
many cross-veins, but has these fewer in number in the higher
genera, has petiolation become more advanced, has some of the
sectors tend to become attached to the principal veins, and has M2
becoming branched from Mi beyond the subnodus. Polythore
(fig. 2) and Eiithore (fig. 3) are the two more generalized, Cora
(fig. 5) and Chalcopteryx (fig. 4) the two more specialized genera.
The fact that the Polythorinae are discussed first does not mean
that they are the most generalized Zygoptera; their quadrangle
and the continuing of the direction of Mi_3 by AI1+2 rather than by
Ms, as well as their petiolation, gainsay this. However, they are
the most distinct group and must have separated from the others
at rather a low level.

The Agrioninae

Venationally this group is quite distinct from the Epallaginae;
however, Schmidt (19 15) working on the male genitalia, finds
there no basis for such a division. The Agrioninae consists of the
Vestalinae as defined by Dr. Needham (1903a) and can be dis-
tinguished from the Epallaginae and Disparocyphinae by the
following venational tendencies: Mia has a course parallel to that
of Mo, the stigma becomes diffuse and finally atrophies altogether;
and the vein Cu2 in most forms has a branch Cu^a running toward
the hind margin of the wing. Vestalis(Jng. 18). Ileterina (fig. 19) and
Lais (fig. 20) do not possess this l)ranch, but the facts that the
stigma may be lost, that Mia is parallel to M2, that the quadrangle
becomes convex above, that some of the sectors tend to become
attached to the longitudinal veins, and that there is no petiolation-
and but slight reduction in the number of antenodals, indicate
affinities with the Agrioninae rather than the Epallaginae. The
subfamily Agrioninae has its own methods of specialization, but its
line ends in itself and shows no relationship to higher forms.

Aside from the Vestalis-IIeterina-Lais group, which is the most
distinct, the remaining genera can be divided into those which
have Cuoa running perpendicular to the hind margin of the wing or
toward the base of the wing, and those in which it is bent away

MEM. AM. ENT. SOC, 3.

12 A VENATIONAL STUDY OF THE ZYGOPTERA

from the base of the wing. The former group begins with such
genera as Archineura (fig. 8) and Echo (fig. 7) and continuing
through Umma (fig. 9), Sapho (fig. 10) and Agrion (fig. 11), ends
with Matroiia (fig. 13), Matronoides (fig. 12), and Neurobasis (fig.
14). The latter group consists of Mnais (fig. 15), Psolodesmus
(fig. 16), and Phaon (fig. 17), which approach Vestalis (fig. 18) in
some respects; for example in the attachment of sectors, in the
convexity of the quadrangle above, and in the general wing pro-
portion and number of cross-veins.

The Epallaginae

The discussion now leads to perhaps some of the most generalized'
of living Zygoptera: the genera Pseudophaea (fig. 21) and Dysphaea
(fig. 22). It is here that one must look for the beginning of the
line which leads to the higher family by reduction in the number
of cross-veins and of the supplementary sectors, by the location of
Mia halfway between Mi and M2, and especially by the bending of
Ms and Rg away from M1+2 so that it is the latter vein rather than
Ms which continues the direction of M 1+2+3. All the genera so far
discussed under the Zygoptera, except the Polythorinae, have M3
in line with M 1+2+3, but in the higher members of the Epallaginae
it is no longer the case. Petiolation too proceeds to a high degree
of advancement and the stigma persists. Beginning then, with
Pseudophaea (fig. 21) and Dysphaea (fig. 22) a series can be ar-
ranged (as has been done in the key to the Epallaginae) which
proceeds by reduction and petiolation through Bayadera (fig. 23),
Epallage (fig. 24), Anisopleura (fig. 25) and Rhinocypha (fig. 26)
to Micromerus (fig. 28) where the new condition is first evident as
regards the base of M3. Rhinocypha (fig. 26), Libellago (fig. 27),
and Micromerus (fig. 28) form rather a compact and distinct
group with M4 arching forward and Cu2 backward, more so than
in the related genera and with R^ and M2 parallel at their tips and
with two prominent sectors between. Cyanocharis (fig. 29),
Caliphaea, Ileliocharis (fig. 30), Dicterias (fig. 32) and Neocharis
(fig. 31) carry on the reduction process until in Devadatta (fig. 33)
and Amphipteryx (fig. 34) the continuing of Mi+2+3 by M1+2 is
again evident and firmly established. Diphlebia (fig. 34) is enough

PHILIP A. MUNZ 13

like a low form of the Lestinae to be taken for one but for the pres-
ence of too many antenodals; and in Philoganga (fig. 36) is to be
seen the beginning of the tendency for R^ and M3 to arise nearer
the subnodus. Thus the evolutionary steps in the Epallaginae
are in marked contrast to those in the Agrioninae.

The Disparocyphinae

The genus Disparocypha (fig. 37) is placed by Ris in the legion
Libellago and it is clearly related to that group. But its characters
of difference are so great that to be consistent with the arrange-
ment in the higher subfamilies, it must be placed in a separate one.
It differs from the members of the legion Libellago as much as the
Protoneura group does from the Coenagrion group or the genus
Lestoidea (fig. 45) from Lestes. Cui is short, not extending to the
level of the subnodus; Cu2 is reduced to a cross- vein; A ends in the
hind margin of the quadrangle; AI4 is almost perfectly straight in
its entire length; and the stigma is very much broader distallythan
proximally.

This reduction of Cui and Cuo, which seems to be a very highly
specialized tendency, will be discussed again under Lestoidea (fig.
45) and Protoneura. Its occurrence in three widely separated
groups is indeed an interesting parallelism.

The Coenagrionidae

As the two families now stand they can conveniently be sepa-
rated by placing in the Agrionidae those genera with fixe or more
antenodals and with M.3 arising nearer the arculus than the sub-
nodus. Of the Coenagrionidae those with more than two anteno-
dals have AI3 arising near the subnodus and often ha\-e the nodus
very near the base of the wing, while those with M3 arising near
the arculus have only two antenodals. Now from this state of
affairs it is evident that the division into two families is somewhat
artificial and yet, I think, it is a tenable one and it ct\rtainly is
convenient. The higher family includes only more sj)ecialized
forms than the lower; the only difiiculty in the whole situation is
that our series from one to the other is too complete to make the
divisions clearly marked. From the phylogenetic standpoint this
condition is of course fortunate. Diphlehia (fig. 34) and related
genera show very evident affinities to the Lestinae among the

MEM. AM. ENT. SOC, 3.

14 A VENATIONAL STUDY OF THE ZYGOPTERA

Coenagrionidae, especially in the position of the base of M3
and of Rs and in the petiolation and general arrangement of
veins. It connects the Lestinae to the Epallaginae through
the genera Pseudolestes and Ortholestes, and differs from these forms
chiefly in the number of antenodals. On the other hand, the
Lestinae are related very closely to the Megapodagrioninae by
Rhipidolestes (fig. 52). Yet in this same group of Megapodagrion-
inae it is necessary to include Thaumatoneura (fig. 51), which with
its number of antenodals and abundant sectors indicates a descent
from some Agrionid, possibly one like Philoganga (fig. 36) in which
the quadrangle is short, and Mz and Rg arise at some little distance
beyond the arculus. The retraction of the nodus in a form like
Philoganga would bring about a condition comparable to that in
Thaumatoneura, except for the numerous sectors and many cross-
veins of the latter. As will be shown later, Thaumatoneura in turn
leads to the Pseudostigmatinae and to the other Megapodagrion-
inae, which by reduction approach the Coenagrioninae and these in
their turn are connected with the Protoneurinae. After this addi-
tional evidence is presented, it will be seen that it is a very logical
thing to consider the Coenagrionidae merely a specialized offshoot
or number of offshoots from the Agrionidae, as arranged in the
tree in Figure C.

In this connection the following quotation is of interest. Till-
yard (1914b) in discussing the classification of the Zygoptera de-
clares, "It becomes now more than ever apparent that the Selysian
division into Calopterygidae and Agrionidae is quite untenable as a
natural dichotomy, more than this, it is pretty clear also that the
Zygoptera are not like the Anisoptera, derived from an original pure
line descent. Triangle-formation most certainly started once; and,
however far back new fossil discoveries may take us as regards the
first formation of the triangle, there can be no doubt about the
origin of all Anisoptera from that single line of descent, which
Palaeontology already places as far back as the Trias, and which
probably began in the Permian period. Most of the Zygoptera
(certainly all those with regular quadrilaterals) must have branched
away from the Anisoptera-Wnc before this. But who can tell how

PHILIP A. MUXZ 15

many separate branchings took place, even amongst this one por-
tion of the Suborder? The origin of the tribe Agrionini- is doubt-
ful, for their acute quadrilateral might yet be proved to be the
reduced remnant of an originally weak Anisopterid triangle-forma-
tion. Finally, standing out clear from all the rest as the most
recent offshoot from the Anisopterid line, we see the Epiophlebia-
Synlestes-Lestes line of descent, which branched off from that line
7iot very far from the beginnings of the Gomphinae, and whose
sharply angulated quadrilateral is almost certainly an Anisopterid
remnant."

Following out his theory, Mr. Tlllyard places this ''line of de-
scent" in a family Lestidae and divides it into the following sub-
families: Epiophlebinae, Lestinae, Synlestinae, and Heterophle-
binae, the last named containing fossils only. It is now known
that Epiophlebia (fig. i) is far from being near the Lestinae not
only venationally, but on the basis of the genitalia, it is of far too
generalized a nature to be placed near them, although an oblique
vein between ^SU and Rg is present in both subfamilies and many
of the sectors are very zigzag. The Zygoptera cannot be con-
sidered as derived from the xAnisoptera, but form a dichotomy
with them, both arising from the ancient Anisozygoptera. The
argument that all Zygoptera with the regular quadrangle are
primitive is also unfounded; the regularity of the quadrangle, that
is, its squareness distally, is not so fundamental a thing as it might
seem. Those forms which are perhaps the most highly specialized,
those in which Cu-. is reduced to a cross-vein, ha\'e the square
quadrangle; yet this condition arises independently three times, in
Disparocypha (fig. 37), in Lestoidea (fig. 45) and in the Protoneur-
inae. I cannot, therefore, agree with any of the \-iews expressed
in the above quotation as regards the phylogeny of the Zygoptera.
This subject can, however, be more fully discussed in treating each
subfamily independently.

The Lestinae

The Lestinae, as has been said before, include those Coenagrion-
idae in which the vein M3 arises nearer the arculus than the sub-
nodus, therefore it comprises the legion Lestes of de Selys. This
feature is an ancient and primitive one and specialization in this

2 Agrionini is Coenagrionidae of this paper.
MEM. AM. ENT. SOC, 3.

1 6 A VENATION AL STUDY OF THE ZYGOPTERA

group has been in a different direction from that in the other Coen-
agrionidae. The presence of the oblique vein between AI2 and Rg
is of course a character recalling the condition to be found in the
Anisoptera, but it seems to be a rather dangerous one on which to
depend too extensively if it means that Epiophlebia must be placed
in the same category as the Lestinae. Pseudolestes (fig. 38) and
Ortholestes (fig. 39) have the general venation of the Lestinae and
are intermediate between the higher Lestine genera and Diphlehia
(fig. 34). Yet they do not possess the oblique vein, at least it is
not distinct. Cannot this cropping out of a primitive feature be
a secondary matter, a reappearance? It is evidently so in the case
of the square quadrangle of Protoneura, Lestoidea, and Disparo-
cypha. A single character is dangerous to use too far to the exclu-
sion of others, although it may sometimes seem necessary to do so.
The oblique vein is present also in Chlorolestes (fig. 66) and Syn-
lestes (fig. 63) yet the rest of the venation does not suggest the
Lestinae. This oblique vein is of course present in those genera
in the nymphs of which the trachea for Rg is joined to the trachea
for M2. Whether this attachment of the trachea is to be taken as
more or less important than the adult venation is the question. It
seems reasonable to lay more stress on the latter since in it several
other characters unite in placing Chlorolestes and Synlestes in the
Megapodagrioninae.

Pseudolestes (fig. 38), Ortholestes (fig. 39) and Rhipidolestes (fig.
52) show a connection between the Lestinae and Megapodagrion-
inae, and these two subfamilies can be considered as offshoots from
the same general region of the Agrionidae. Their close affinity
warrants the retention of the subfamily name for the Lestinae
rather than the use of a family name. Of the forms included here
under Lestinae Pseudolestes (fig. 38) is undoubtedly the most primi-
tive, in having the quadrangle like that of Diphlehia (fig. 34) and
in its large number of supplementary sectors. Ortholestes (fig. 39)
comes next and is followed by the Lestine genera proper: Archi-
lestes (fig. 40), Megalestes (fig. 41), Orolestes, Platylestes, Lestes (fig.
44), Sympycna (fig. 42), etc. The general tendency is that of
reduction, combined with straightening of the veins at the tips.
Many sectors and veins are somewhat zigzag, and there develops

PHILIP A. MUNZ 17

also a narrowing and sharpening of the quadrangle distally. Peti-
olation does not proceed very far, ne\er to the level of the arculus.

The Lestoidinae

Air. Tillyard (1912a) described as a new genus a very peculiar
form which he called Lestoidea (fig. 45) and for which he proposed
a new legion of the same name. This remarkable insect has de-
cided Lestine affinities in the location of the base of Ms and R,, but
has the vein Cuo reduced to a cross-\ein and Cui but one cell long.
Yet in addition to the above characters it is distinguished from all
Protoneurinae by the presence of two sectors between M, and Mj^.

The decrease in length of Cui and Cu2 is one of the most extreme
tendencies in specialization occurring in the Zygoptera and its
origin in the various groups is an interesting parallelism. As al-
ways happens when Cuo is thus reduced MA is straightened and the
quadrangle ends squarely. The retention of the two sectors be-
tween Ml and M,,, is rather remarkable, in view of the fact that all
others are gone and that in none of the Lestinae are there any long
sectors between these two veins. Lestoidea must therefore have
arisen from some rather ancient form and is distinct enough to
merit a legion for itself or in the system here used, a subfamily.

The Megapodagrioninae
Here are included the genera ordinarily considered as belonging
to the legion Podagrion of de Selys; this is the group which leads
to the higher members of the Coenagrionidae. The genus most
distinct from the others is undoubtedly Thaiimatoneiira (fig. 51),
which has been given much attention by \-arious workers. Afc-
Lachlan (1897) and Foerster (1909) referred it to the Agrionidae;
Needham (1903a) spoke of its affinities to the Pseudostigmatinae
and Calvert (19 13) placed it in the Megapodagrioninae. All
agreed that it was an intermediate form. The safest course to
follow seems to be to leave it in the Megapodagrioninae, but to call
attention to the fact that it shows an undoubted approach to the
Pseudostigmatinae. It is unquestionably one of the Coeuagrion-
idae although it does have more than two antenodals. Neuro-
lestes (fig. 62), Trineuragrion {{\g. 54), and often Neuragrion have
three antenodals as the normal condition, so that the old idea of

MEM. AM. ENT. SOC, 3.

1 8 A VENATIONAL STUDY OF THE ZYGOPTERA

two as being an infallible means of determining Coenagrionidae
must gi\'e way, as has been stated by pre\'ious workers. The re-
traction of the nodus is a more fundamental difference than the
retention of four or five cross-veins and must place the genus in
the higher family. Now as to its subfamily ; it is convenient to
limit the Pseudostigmatinae to those genera in which the stigma
becomes diffused and which have Mia following Mo rather than
Ml. If this is done, Thaumatoneura may be placed near the base
of the Pseiidostigma line. I can readily see, however, that its gen-
eral vein-arrangement, its large postnodal area, its broad area
behind Cu and its retracted nodus may indicate a condition near
enough to that of Megaloprepus (fig. 46) to warrant placing the
two in the same subfamily.

To return to the general discussion of the Megapodagrioninae,
it may be said that Rhipidolestes (fig. 52) in the position of the base
of M3 shows a decided affinity to the Lestinae, to separate it from
which is perhaps as artificial an arrangement as to divide Thauma-
toneura from the Pseudostigmatinae. The extreme petiolation and
long narrow quadrangle, however, of Rhipidolestes throw it in this
subfamily, which may be said really to begin with Podopteryx
(fig- 53)- From there on, the adult venation at least, gives a well
connected series which has been carefully discussed by Dr. Cal-
vert (191 3). As nearly as they can be made to do so the keys
show my interpretation of its evolution. M3 and R^ at their points
of attachment move outward to beyond the subnodus, the quad-
rangle becomes elongate, the wing reaches the extreme of petiola-
tion, and the number of supplementary sectors is reduced to none
■other than M^.

Tillyard (1914), on the basis of the presence of the oblique vein
formed the Synlestinae as a group including such genera as Chloro-
lestes (fig. 66) and Synlestes (fig. 63), and placed them with the
Lestinae. But, as has been shown, this was a dangerous undertak-
ing. According to Schmidt (19 15) the genera Chlorolestes (fig. 66)
and Synlestes (fig. 63) belong to the Lestinae on the basis of the
structure of the male genitalia; his figures, however, fail to show
any very close connection. Venationally they are not to be placed
in the Lestinae; M3 arises very near to or at the subnodus although

PHILIP A. MUXZ 19

the general wing proportions are those of the Lestinae, in wliich it
arises far before the subnodus: petiolation has proceeded to the
level of half the length of the quadrangle; in the Lestinae it does
not reach the arculus. It is true that Cui arches forward just
beyond MA as in Lestes (fig. 44), but that is the case in some other
genera as well; for example Chorismagrion (fig. 69) and Trineura-
grion (fig. 54) and Dimeragrion (fig. 67) to a certain extent. Rhipid-
olestes (fig. 52) in its venation is nearer to Lestes (fig. 44). but
Schmidt considers it much nearer Megapodagrion (fig. 57) itself on
the basis of the penis.

I have spoken before of the progressive reduction of the sectors
in the Megapodagrioninae until all except Mia are gone; when this
state is arrived at, the condition to be found in the Coenagrioninae
is present. But there are two genera which agree with the Coen-
agrioninae in this respect, but which seem better retained in the
Megapodagrioninae. These are Chorismagrion (fig. 69) and P'eri-
lestes (fig. 70) ; their extreme petiolation, the remote origin of Rs
and M3, the arching forward of Cui beyond MA, and the heavy
stigma of Chorismagrion are indicative of affinity to Synlestes (fig.
63) and Heteragrion (fig. 68) and through its likeness to Chorismag-
rion, Perilestes is to be placed at the apex of the Megapodagrion
line.

Just the fact, then, of the presence or absence of supplementary
sectors is not a safe criterion by which to separate the Megapo-
dagrioninae from the higher forms. It is indeed an accompani-
ment of specialization and reduction and is a convenient character
to use. Generally mere presence or absence is not so reliable a
matter as might l)e supposed, since two forms without a gi\en char-
acter may ha\e arisen independently from two very distinct an-
cestors, one with the character ijresent in one wa>', another with it
in quite a different way.

It is probable, however, that the forms placed by de Selys in the
legions Agrion and Platycnemis have been derived from others,
closely related to Megapodagrion (fig. 57). in which petiolation has
not reached the lexel of Ac and in which Mn arises i^efore the sub-
nodus. A wing like that of Megapodagrion with the suj^plementary
sectors removed would be much like Platycnemis (fig. 74). I\t-
haps the distinction between the legion Podagriou on the one hand,

MEM. AM. EXT. SOC, .V

20 A VENATIONAL STUDY OF THE ZYGOPTERA

and Platycnemis and Coenagrion on the other is hardly great
enough to merit their being placed in separate subfamilies, but it
may be done, if it is kept in mind that the groups are close to each
other and that the second is an offshoot from the first.

So much may suffice for a discussion of the Megapodagrioninae;
those groups to which it has given origin may now be discussed.

The Pseudostigmatinae

As here used, this subfamily includes the Anormostigmatini of
Kirby's Catalogue, and is characterized by a diffusion of the stigma,
by having Mia parallel to Mo rather than to Mi, and by a very re-
tracted nodus with a long postnodal portion which has many cross-
veins and with the longitudinal veins running far out into the wing.
These insects are as specialized in a venational way as in their
elongate abdomens and peculiar habit of life, the nymphs living
in the small pools in epiphytic Bromeliads. If it were not for the
genus Thaiimatoneura (fig. 51) they would stand so far removed
from all other Zygoptera as perhaps to deserve a separate family,

Megaloprepus (fig. 46), however, the most generalized of the five
genera, has so many points of similarity to Thaiimatoneura that
there can be no doubt of their relationship. They agree in the
width and general proportions of the wings, in the presence of
abundant sectors and the position of the nodus. Megaloprepus
shows a more generalized condition than the remaining genera in
the short quadrangle, the broad area back of Cu2 with many at-
tached sectors, and in the fact that the stigma while diffuse is not
so much so as in the others.

\\'ithin the group the same tendencies of specialization appear
as in other groups: loss of sectors, narrowing of the wing, reduction
in number of cross-veins, straightening of veins, elongation of the
quadrangle, and petiolation of the wing. Next to Megaloprepus
comes Microstigma (fig. 47), with the quadrangle about five times as
long as wide and with M1+2 forking more than half the distance
from the subnodus to the tip of the wing. In these resi:)ects these
two genera differ from Anomisma (fig. 48), but agree with it in hav-
ing many secondary sectors and with Vuj, bearing many branches.
Anomisma is peculiar in being a genus of (\x'nagrionitlae having a
crossed quadrangle, Pseudostigma (fig. 49) and Mecistogasler (fig.

PHILIP A. MUXZ 21

50) have few sectors and no branches to Cu-.. Mecistogaster is the
most speciaHzed of all, with a single row of cells behind Cu2 and
between Mu and M2. It is interesting to note that in this group
the tendency is for the nodus to approach a more normal position
in the wing, so that in Mecistogaster (fig. 50) it is at one-fourth
the length of the wing, while in Megaloprepus (fig. 46) it is at
one-seventh.

The Coenagrioninae

Baron de Selys created two legions, Platycnemis and Agrion, for
the reception of genera which fall, it is true, into two groups, but
these have no definite nor sharply defined difl^erences. In general
it can be said that the quadrangle of the first of these groups has a
relatively long upper side, hence that the outer angle is not very
acute, but there are exceptions to this, as for example in Leptocne-
mis (fig. 80); in the second legion there areexceptions to the obliq-
uity of the quadrangle, since some of the more specialized forms,
as Hemiphlebia (fig. 106) and Agriocne?ms (fig. 104), have a fairly
regular one. For that reason a recent key to subfamilies of the
Zygoptera does not hold as applied in the following quotation:

"Quadrilateral regular Subfamily Platycneminae.

Quadrilateral irregular, usually with distal angle sharply acute.

Subfamily A grioninae.''^
Now these two groups have many points in common; they have no
supplementary sectors other than Mia, they have the stigma rather
short as compared with that of the IMegapodagrioninae and they
have Cu2 well developed. Since they do grade so from one into
the other, and are so fundamentally related, I cannot follow the
above example and place them in separate subfamilies; in fact, I
am not absolutely sure that as many such divisions ought to be
recognized as I ha^•e already done in this paper, that is, whether
these "subfamily" groups are really equivalent to subfamilies in
other orders.

This placing both in one group does not indicate, however, that
my feeling is that the legions Platycnemis and Coenagrion are not
rather distinct. Although the quadrangle alone does not suffice in

^ Agrioninae is here used for the legion Agrion of dc Selys, hence it is equivalent to
the Coenagrioninae of this paper minus the Platycnemis group.

MEM. AM. EXT. SOC, 3.

22 A VENATIONAL STUDY OF THE ZYGOPTERA

all cases to separate the two, the general wing proportion and ar-
rangement of the veins do. In Platycnemis (fig. 74) and its allies
the wing is long and with the longitudinal veins arranged in a very
regular and parallel fashion. The cells are largely square or rec-
tangular, while in the group containing Coenagrion (fig. 94) there is
a tendency for M1+2 to bend posteriorly at the subnodal region and
for M4 and Cu. to be zigzag so that many cells are pentagonal.
The methods of specialization in the two groups are also different,
in the Platycnemis s'eries are retained some of the features to be
found in the Megapodagrioninae; for example, (i) great petiola-
tion which may extend beyond the level of the arculus, and (2) an
outward migration of the base of Rs to a considerable distance be-
yond the subnodus. In the second group, on the other hand, petio-
lation is delayed for a long time, so that it never reaches the level
of the arculus, and Rs arises beyond the subnodus in Amphicncmis
(fig. 129) only. Many of the Coenagrion group are further special-
ized by a great reduction in the number of cross veins in the wing.
To come now to the discussion of the legion Platycnemis in de-
tail, it is perhaps to be noted first of all that this may not be a
natural group, it may represent a collection of forms each of which
is the result of the dropping out of sectors from some corresponding
Megapodagrionine. If Megapodagrion (fig. 57) itself, by way of
example, were to lose its supplementary sectors, the resulting vena-
tion-would be much like that of Platycnemis (fig. 74) or Calicnemis
(fig. 72). The same might be true'of Wahnesia (fig. 65) and Tatoc-
nemis (fig. 78), of Heteragrion perhaps, (fig. 68) and of Prionocne-
mis (fig. 79). However this may be, and although these genera
resembling Platycnemis may not have come from a common an-
cestor, they do form a remarkably close and well connected series
beginning with Metacnemis (fig. 71), Calicnemis (fig. 72), and
Platycnemis (fig. 74), in which the wing is not petioled to the le\el of
Ac, and R, begins at the subnodus. Proceeding through Idiocnemis
(fig. 75), Coeliccia (fig. 76), and Allocuemis (fig. 77), petiolation and
the outward movement of the base of R. achance until the two
tendencies reach their culmination in Talocucmis (fig. 78) and
Prionocnemis (fig. 79) respectiv'ely.

PHILIP A. MUNZ 23

Leptocnemis (fig. 79) and Paracnemis which are rather typical
forms of the Platycnemis group, except in the proportions of the
quadrangle, and Antiagrion (fig. 81) a segregate of the old genus
Coenagrion itself, offer themselves as transitional genera to the
remaining Coenagrioninae.

These other genera in turn, can be divided into two groups: the
Argia group and that of Coenagrion proper. In the former the
inner end of the quadrangle is generally longer than the upper part
of the arculus and as long as the upper side of the quadrangle, which
is widened distally. Another character, not venational, is the
presence of long tibial spines. This group has its own exolutionary
line which begins with Ilyponeiira (fig. S2) and Palaiargia (fig. 83)
and proceeds by greater petiolation, shortening of Cuo, and reduc-
tion of the number of postnodals through Argia (fig. 84) to Argial-
lagma (fig. 86) and Diargia (fig. 85).

Of the Coenagrion group proper, it may be said that venation
offers fewer marks for identification than in any other group of the
Zygoptera, and it is here that keys for determination must be
artificial. The genera are so numerous and so much alike that
other characters are necessary for their differentiation. Venation
does, however, indicate the general tendencies of phylogeny. Mr.
Tillyard (191 2) may be quoted as follows, "Those forms are the
least asthenogenetic in which the basal postcostal nervule^ is closest
to the base of the wing, the petiolation consequently least marked,
and the inferior sector of the triangle longest and best developed.
Such forms, while marking the 'high-water mark' of success-
ful coenogenetic specialisation in the legion, must nevertheless
be considered to be more In line with the ancestral stem of the
group than arc their asthenogenetic off-shoots." Following this
theory, one can readily arrange a scries which i)egins with Oreagrion
(fig. 87) and Pyrrhosoma (fig, 89) and, by a shortening of Cu-j and a
reduction In the number of postnodals, continues through Enallagma
(fig. 98), Ischnura (fig. 100) and Coenagrion (fig. 94) to another
dividing point, one branch leading off In the direction of reduction
In size and further decrease In the number of postnodals until
Ceratiira (fig. 102), Anomalagrion (fig. 97), Agriocnemis (fig. 104)

^ Basal postcostal nervule is the vein Ac of this paper. Inferior sector of the tri-
angle is Cu2.

MEM. ANr. ENT. SOC, 3.

24 A VENATIONAL STUDY OF THE ZYGOPTERA

and Ilemiphlebia (fig. io6) are arrived at; the other branch devel-
ops by a greater petiolation, which extends to or beyond Ac in
those genera above Pseudagrion (fig. 114). This line continues
through Oxyagrion (fig. io<^) , Acanthagrion (fig. 1 10) and others until
a new tendency is observable, namely the movement outward of
the base of Ms, until the origin of this vein very nearly coincides
with that of R^. This is true of such genera as Nesobasis, Aciagrion
(fig. 123), Leptobasis (fig. 127) and Teinobasis (fig. 128). In
Amphicnemis (fig. 129) the base of M3 reaches thesubnodus and Rs
arises a cell beyond, while petiolation extends at least to the level
of the arculus.

To return for a moment to the discussion of Ceratura (fig. 102),
Anomalagrion (fig. 97) and their allies; there are to be found here
several very interesting and peculiar conditions. The male of
Anomalagrion hastatum in the fore-wing has the stigma small and
three-sided, with the apex of this triangle not reaching the margin
of the wing, and with the base on Ri. Ceratura (fig. 102), Ans-
trocnemis (fig. 104) and Hemiphlebia (fig. 106) have the second
antenodals placed before the arculus, and have only from five to
seven postnodals, while Cu2 extends but a few cells beyond the
level of the subnodus. In Austrocnemis and Hemiphlebia petiola-
tion extends almost to Ac, since Ab and Ac are in line with each
each other. A' does not begin at the wing margin and the other
characters above mentioned place these two genera in this line
rather than in the Leptobasis one. The absence of the inner end
of the quadrangle in the fore-wing of the male of Ilemiphlebia (fig.
106) is an unusual state of affairs, and an interesting parallelism to
the Megapodagrionine Chorismagrion (fig. 69) recently described
by Morton (1914).

The Protoneurinae
This subfamily, the most specialized one among all the Zygop-
tera, includes the legion Protoneura of de Selys. It is distinguished
primarily by the reduction in length of Cuo which is generally
represented by a cross- vein only. In fact, this character was the
one first used by de Selys to determine whether a genus belonged in
the legion or not. On the discovery oiProueiira prolongata (fig. 131) ,
in 1889. with the vein Cuo one cell long, he modified the definition

PHILIP A. MUXZ 25

of the legion to include this form and the genus Chlorocnemis (fig.
130) which is likewise characterized by ha\ing Cu^ one cell in
length. It is of course evident that these two genera are interme-
diate between the old legions Agrion and Platycnemis proper. Yet
they are closer to the latter and may easily be considered as coming
in the same subfamily with it.

Chlorocnemis (fig. 130) with its general wing proportions and
arrangement of veins is a genus connecting Platycnemis (fig. 74)
with such genera as Palaemriema (fig. 136) and Platysticta, which
also have a long postnodal area with the longitudinal veins very
regularly and evenly placed. Proneura (fig. 131), on the other
hand, shows its afftnities with the Coenagrion group on the one side
and Xeoneura (fig. 133) and related genera on the other, in it are
found the pentagonal cells near M4, the bend of M 1+2+3 before the
subnodus, and the small number of the cells of the wing and a short
vein Mi.^.

Alter ('u2 has become reduced, the next step in the e\'olution of
the Protoneiirinae seems to have been the atrophy of that part of
the anal vein beyond Ac. Thus in Disparoneura (fig. 132), Peri-
sticta (fig. 134), Neoneura (fig. 133). and Idioneura (fig. 135) it
extends to the medio-anal link; in Palaemnema (fig. 136) and the
old genus Platysticta (fig. 137) it ends before MA on the posterior
edge of the quadrangle, and in Ilypostrophoneiira on the hind mar-
gin of the wing. In the higher forms it does not extend beyond Ac,
which then is generally moved nearer the base of the wing until in
Nososticta (fig. 151), Isosticta (fig. 150), and Selysioneura (fig. 153)
it lies before the level of the first antenodals. At the same time
Cui becomes shorter and the whole line culminates in Selysioneura
(fig. 153), which has only a cross-vein for both Cui and Cuo. In its
absence of sectors other than INIia, in the reduction of Cui and Cuo,
in the origin of both R, and M3 beyond the subnodus, in the
absence of that portion of A beyond Ac, in the retracted position
of Ac, and in the fact that the second antenodals lie before the
arculus, this genus presents a combination of characters of so high
a degree of specialization that it can hardly be equalled in its
reduction by any other member of the Zygoptera. It is therefore
fitting that it should l)e placed at the very apex of the genealogic

MEM. AM. ENT. SOC, 3.

26 A VENATIONAL STUDY OF THE ZYGOPTERA

tree of the suborder, and should bear the name of the man who has
done more than any other worker to further our knowledge of the
Odonata, namely Baron de Selys-Longchamps.

In contrast to the Coenagrioninae the Protoneurinae do not have
the reduction in the number of postnodals, they are as numerous in
Selysioneiira (fig. 153) as in Disparoneiira (fig. 132).

Ill— TENDENCIES IN SPECIALIZATION

These have been rather fully discussed in previous papers, par-
ticularly by Dr. Needham (1903a) for the order Odonata as a whole,
and by Dr. Calvert (19 13) for the Zygoptera with special reference
to the Megapodagrioninae. It is my purpose here to review their
opinions and to add a few observations of my own.

I think it must be apparent from the preceding discussion on the
phylogeny that the outstanding features in the development of the
Zygoptera are a constant reduction and simplification of venation.
At any rate this is true from the standpoint of comparative mor-
phology and to a certain degree it seems to be so paleontologically.
As might be expected this reduction is to some extent accompanied
by a decrease in the power of flight. Dr. Calvert says, "Many of
the ideas which have been expressed on the changes which have
taken place in the venation of the Odonate wings have been based
on the comparative morphology of living representatives of the
order or on mechanical advantage. It does not follow that these
foundations are trustworthy or that there has always been mechan-
ical improvement in the wings. Degeneration is just as probable
as progressive development. The actual course of phylogeny can-
not be deduced from these considerations or from the data of
morphology. It seems therefore desirable to attempt to ascertain
what paleontological evidence exists affording clues to the descent
of these insects, by tracing the modifications which the wings
exhibit from the Carboniferous period down to the present time."
This is what he has done and free use is here made of his
conclusions.

It should be noted that the tendencies here discussed are not all
present in the group, nor all to be found even in the highest groups,
but they are those which impress one, when studying o\'er all the
groups, as being indicative of advancement and departure from the

PHILIP A. MUXZ

27

generalized condition. They fall under two or three general heads
and may be so treated in their discussion. First come those which
may be classified under "X'^ein Reduction," secondly, those having
to do with "Vein Shifting," and as a last group "Miscellaneous
Features."

Vein Reduction
I. Reduction in the number of supplementary sectors. — This is one
of the most evident means of specialization within any group which
possesses such sectors, as well as in the whole suborder, of which
the two highest subfamilies have none. A comparison of the dif-
ferent subfamilies in this regard may be made in a general wa\' by
choosing for each one as its most typical genus the one after which
it is named. In Table I this has been done, the figures indicating
the number of sectors between the veins listed at the top of the
columns. The figures are for the front wings.

Table I
The Number of Supplementary Sectors present in the Sub-
families of Zygoptera

Subfamily Genus and

Species

Polythorinae Polythore

gigantea
Agrioninae Agrion

Virgo
Epallaginae Epallage

fatima
Disparocyphinae Disparocypha

biedermanni
Megapodagrioninae Megapodagrion
■ erin V5

Ml- Mia- M2- Rs- M3- M,- Cui- Cu2 Total
M,a M2 Rs Mj M4 Cui Cu2 Margin

040

O

7 8 8 3 S 45

3 4 4 4 4 28

.20000 8

o I o o o s

Lestinae

Lestoidinae

Pseudostigmatinae

Coenagrioninae

Protoncurinae

Lestes

reclangularis 0222
Lestoidea

000

conjiincta
Pseudostigma

aherrans
Coeti agrion

piiella
Protoneiira

capillaris

20000000

III I I o I

00000000

00000000

MEM. AM. ENT. SOC, 3,

28 A VENATIONAL STUDY OF THE ZYGOPTERA

It should be remembered of course that such a table is of value
in a general way only, since there is great variation in some sub-
families. Epiophlebia is a member of a lower suborder, yet its
supplementary sectors are not so numerous as those of many
Zygoptera. Thaumatoneura (fig. 51) is placed in the Megapoda-
grioninae, but its many sectors indicate a primitive condition, or
possibly they may represent a secondary acquirement. Neverthe-
less, this table shows that the highest subfamilies have the fewest
sectors.

Dr. Calvert has shown that in a general way the reticulation of
the oldest fossil order related to Odonata, the Protodonata, was
rich in supplementary sectors. In Anisozygoptera some wings
were richly veined, others less so. He concludes, that we have,
therefore, "only the general tendency to guide us when we assume
that reduction in density of venation means specialization."

2. Reduction of the veins Cu-i and Ciii. — This tendency appears
in three widely removed groups; by it are derived Disparocypha
(fig. 37) from the Epallaginae, Lestoidea (fig. 45) from the Lestinae,
and the Protoneurinae from the Coenagrioninae. As was shown
in the discussion of the Protoneurinae, Cui follows Cu2 in its reduc-
tion until nothing remains for both except the cross vein forming
the lower part of the medio-anal link. This condition is the one to
be found in Selysioneura (fig. 153) and is approached by a number
of forms, such as Nososticta (fig. 151) and Isosticta (fig. 150).

3. Decrease in the number of antenodal cross-veins. — This very
apparent tendency needs but little discussion, particularly after
the thorough treatment given it by Campion (191 3). The number
of cross-veins in the lower Agrionidae is very high; in the higher
Epallaginae it is reduced to seven, eight, and nine; in the Coenagri-
onidae, only Thaumatoneura (fig. 51) has over three. Neurolestes
(fig. 62), Trineuragrion (fig. 54), and Neuragrion ha\e three gener-
ally, in one wing at least; and Chlorolestes (fig. 66), Podopteryx (fig.
53), Paraphlebia (fig. 64), IIeteragrio?i (fig. 68), and Pyrrhosoma
(fig. 89) have occasionally been observed to have three; all others
commonly two. Dr. Calvert cites evidence to show that this
reduction is not in proportion wuth the retraction of the nodus, but
is apparently an independent and recent acquisition.

PHILIP A. MUNZ 29

4. Decrease i?i the number of poslnodals. — In the Zygoptera this
tendency comes in much later than does the reduction in the ante-
nodal series, nor does it ever reach so advanced a state. It ap-
pears very marked in the Coenagrioninae only, and there in the
Argia and Coenagrion series some of the highest genera such as
Diargia (fig. 85), Agriocnemis (fig. 104), Ilemiphlebia (fig. 106),
and others have very few postnodals.

5. Loss of the cross-veins in the qiiadrayigle and subquadrangle. —
The number of cross-veins in and before the subquadrangle seems
generally to correspond roughly to that of the antenodals. In
those Agrionidae with numerous antenodals the subquadrangle is
generally crossed and the number of "cubito-anal" cross-veins is
so great that Ac itself is distinguished only with difficulty. Neo-
charis (fig. 31) offers an interesting exception; in it the antenodals of
both series are numerous, but there are no true cross-veins back of
Cu. Diphlebia (fig. 34) and Amphipteryx (fig. 35) may also have
none. In the Coenagrionidae the lack of cross-veins in the sub-
quadrangle has become the normal state and it is only in such
forms as Para phi ebia (fig. 64) and Dimeragriori (fig. 67) that any
are found. The same tendency is true of the cross-veins of the
quadrangle; they are first lost in the Epallaginae and in the higher
forms are normally retained by one genus only, Anomisma (fig. 48).

6. Reduction m number of those antenodal cells situated between M4
and Cui. — As shown by Dr. Calvert "this character is in general
dependent on the degree of retraction of the nodus," and, "a reduc-
tion in the number of antenodal cells may occur independently of
retraction of the nodus, however, by the greater de\-elopment of
the quadrilateral distad." It is evident also that this reduction is
a part of the general tendency toward a smaller numl)er of cross-
veins in the wing as a whole; almost any part of which when ob-
served through a long ascending series, will show the cross-veins
becoming placed further apart. Other examples of the tendency
might be noted, for instance, the freeing of the basal radial space
where the latter exists.

7. Reduction in the number of cross-veins behind the anal vein.—
In part this is in line with the general tendency to drop out cross-
veins; in part due to progressive jjetiolation which reduces the

MEM. AM. ENT. SOC, 3.

30

A VENATIONAL STUDY OF THE ZYGOPTERA

amount of space back of the anal bridge and therefore the possi-
biHty of cross-veins occurring there.

Vein Shifting
I. Narrowing of the wing. — This tendency is coupled with the
first one mentioned in the previous group, namely the reduction in
the number of sectors. When the sectors drop out, the principal
veins come to lie closer together. The wing may become narrower
also by the veins Cui and Cuo becoming reduced. A comparison
of the typical genera of the subfamilies shows at a glance how preva-
lent this tendency is (see Table II).

Table II

The Ratio of the Width of the Wing to the Length in the Dif-
ferent Subfamilies.

Subfamily

Typical genus and species Ratio

as Decimal

Polythorinae

Polythore gigantea

380

Agrioninae

Agrion macula turn

345

Epallaginae

Epallage fatinia

228

Disparocyphinae

Diparocypha biedermanni

167

Megapodagrioninae

Mega podagr ion erinys

201

Lestinae

Lestes recta ngida ri s

203

Lestoidinae

Lestoidea conjuncta

252

Pseudostigmatinae

Psendostigma, sp ?

195

Coenagrioninae

Coenagrion resoliituni

210

Protoneurinae

Proton eur a capillaris

152

It is evident in this table, what effect the dropping out of Cuo in
Disparocypha and Protoneura has on the width of the wing. In
Lestoidea, however, the presence of two sectors between Mi and
Mia, has offset this tendency (see fig. 45).

The Pseudostigmatinae is a good example of the same tendency
(see Table III). The subfamilies Megapodagrioninae, Coena-
grioninae, and Epallaginae might be used for similar tables.

An examination of Handlirsch's figures in "Die Fossilen In-
secten" show some rather narrow wnngs in periods as remote as the
Mesozoic; evidently narrowing of the wing and loss of the sectors
are ancient tendencies.

PHILIP A. MUNZ 31

Table III
The Ratio of the Width of the Wing to the Length, in the
Pseudostigmatinae.

Genus and Typical Species Ratio, as Decimal

Megaloprepiis coeruleatus . 267

Mlcrostigma exustiim .237

Anoniisma ab)wrnie .204

Pseiidostigma aberrans .177

Mecistogaster uuirchali .148

2. Progressive petiolation. — In each subfamily except the Agrion-
inae, where speciaHzation is along a different direction, there is a
^'ery noticeable tendency toward greater petiolation; it has been
referred to constantly in the discussion on phylogeny and is used
often in the keys, offering as it does an easy index to the amount of
departure from the primitive. The Protoneurinae might be con-
sidered an exception, since in them, A becomes shorter and the
reduction process is a matter of retraction. But in both cases the
result is the narrowing of the wing near the base. Dr. Calvert
discusses petiolation from the paleontological standpoint and says,
"It seems justifiable to assume that the farther distad petiolation
has progressed, the more specialized the insect."

Petiolation reaches its highest development in the Megapoda-
grioninae where in Chorismagrion (fig. 69) and Perilestes (fig. 70)
it extends to beyond the level of MA.

3. Retraction of the nodus toward the wing base. — This is one of
the most common and evident occurrences in specialization and is
discussed in detail by both Professors Needham and Calvert. In
the Agrionidae the nodus is almost uni\'ersally at more than one-
third the distance from the base of the wing toward the tii), in the
Coenagrionidae at generally less.

4. Tendency for M->, i^s, and Mz to arise nearer the tip of the wing
and further from the subnodus. — As Dr. Calvert remarks this out-
ward "migration may be real or apparent. Retraction of the
nodus with no change in position of the separation point of Mo
would result in a greater interval between nodus and M- and hence
an apparent movement of the latter toward the stigma, but no real
movement in respect to the, wing as a whole." His table for

MEM. AM. ENT. SOC, 3.

32 A VENATIONAL STUDY OF THE ZYGOPTERA

various members of the Megapodagrioninae shows it is often ap-
parent only; but even if so a combination of the preceding ten-
dency and of this one indicates specialization.

A study of almost any group will show the same thing; one that
is a good example is the Platycnemis group of the Coenagrioninae
(see Table IV).

Table IV

The Ratio to the Length of the Wing of the Distance from the
Base of the Wing to the Subnodus and Points of Origin of M2, Rg,
and M3.

Genus and Species

Subnodus

M2

Rs

Ms

Metacnemis valida

.381

619

381

.362

Calicnemis eximia

•309

576

■309

.289

Psilocnemis marginipes

■354

631

354

•339

Platycnemis pennipes

•354

562

354

•323

Allocnemis leucosticta

•345

582

345

.322

Coeliccia membranipes

•311

613

311

.301

Tatocnemis malgassica

3"

561

327

■311

Prionocnemis haematopns

.191

591

435

.209

A first glance at the wings themselves might suggest to one that
M2, Rs, and Ms arise further out in the wing in the last genera than
in the first, but an inspection of the above table shows that in
general this is not true. It is not even equally true of the different
veins but may be more so of some than others; Rg for example in
Prionocnemis (fig. 79) does arise beyond what it does in the others,
M2 and Ms really arise nearer the base in the more specialized than
in the lower wings. The subnodus becomes more retracted, quite
consistently so, and the space between it and the points of origin of
M2 and Rs enlarges, while the relative positions of the subnodus
and the base of M3 change.

A comparison of the Agrionidae and of the Coenagrionidae shows
on the whole, that in the lower subfamilies of the latter family the
vein Mi+2 forks at about one-half the wing-length, as it does in the
former family, but in the higher subfamilies the movement out-
ward is evident. In the whole series Rg and Ms also migrate out-
ward. Dr. Calvert says that "the paleontological evidence is not
altogether satisfactory and at least admits of the possibility of the
separation — points of these three veins having fluctuated proximad
or distad from time to time."

PHILIP A. MUNZ 33

5. Migration outivard of the base of Mi^. — This tendency is cor-
related somewhat with the preceding, but seems to be independent
of it to a certain degree. Most genera among the Agrionidae,
Pseudostigmatinae, Lestinae, and Megapodagrioninae have Mia
arising about three or four cells beyond the fork of M1+2, or at least
rather near it. Perilestes, (fig. 70) however, which has M1+2 fork-
ing about two-thirds of the way out has Mia arising below the outer
end of the stigma, or about one-fourth of the wing-length beyond
M2. This is true to a lesser degree in some of the Coenagrioninae
such as Platycnemis (fig. 74), Antiagrion (fig. 81), Ilyponeura (fig.
82), Agriocnemis (fig. 104), Ceriagrion (fig. 117) and others. It is
not on the whole an important character, but does occasionally
seem worthy of notice.

6. The position of the second antenodals before the arciilus. — In
those Agrionidae with many antenodals, there is a tendency for two
of the costal antenodals to become more strongly developed than
the others, and to be in line with two corresponding ones of the sub-
costal series, while at the same time the second of these pairs is
directly above the arculus. In Epiophlebia (fig. i) this second pair
lies considerably beyond the level of the arculus; in Polythore (fig.
2) and Euthore (fig. 3) it is at that level, in Cora (fig. 5) and Chal-
copteryx (fig. 4) before it. In the Agrioninae, i.e. the legion Calop-
teryx, there is no evident hypertrophy of antenodals, but in the
Epallaginae the tendency is again e\-ident and the number of
antenodals is reduced, until the two pair typical of the Cocnagri-
onidae remain. Here, as in the Polythorinac and some Epallaginae,
in some of the most highly developed forms: Lestoidea (fig. 45),
Perilestes (fig. 70), Chorismagrion (fig. 69). Ceratura (fig. 102),
Agriocnemis (fig. 104), Ilemiphlebia (fig. 106), Caconeura (fig. 143),
Protoneura (fig. 146), Selysioneiira (fig. 153) and others, the second
antenodals are distinctly before the level of the arculus. Since
this tendency rarely is manifest except in the highest memi)ers of
the respective groups, it seems indicative of specialization.

7. The tendency for the quadrangle to become acute distally. — Dr.
Needham (1903a) has taken as the primitive type of quadrangle
one square distally and Mr. Williamson (1913) agrees that "in the
relations of MA," the Calopteryginae (Agrionidae of this paper)

MEM. AM. ENT. SOC, },.

34 A VENATIONAL STUDY OF THE ZYGOPTERA

shows a generalized structure and adds that for the Agrioninae
(Coenagrionidae here) he beHeves In two Hnes of evolutionary
activity, one in which MA is the center of activity and the quad-
rangle becomes pointed; the other in which MA is not changed,
but the posterior area of the wing is reduced ; here he refers of course
to the Protoneurinae. Dr. Calvert (1913) disagrees with the views
above expressed, partly because he finds no paleontological evi-
dence for their support. He argues that since "all the Jurassic
fossils, which are not clearly Anisoptera, show an oblique quad-
rilateral with its posterior distal angle acute," it seems "necessary
to regard the oblique quadrilateral as the more primitive and the
rectangular as a later appearance." It is true that the Anisozy-
goptera figured by Handlirsch have a quadrangle which is some-
what acute distally, and that the only living representative of. this
suborder, Epiophlebia, (fig. 10) has an acute quadrangle, so that
this condition seems the more ancient. When the living Zygoptera
are taken up, however, those that are in other respects the most
primitive have the quadrangle regular; in the Agrionidae, it is only
in the higher Epallaginae that the obliquity becomes noticeable;
and In the higher Megapodagrlonlnae, Lestlnae, Pseudostigma-
tinae, and those Coenagrlonlnae of the Argia and Coenagrion
groups it is carried to its highest development. Comparative
morphology, therefore, brings evidence to show that the obliquity
of the distal end of the quadrangle is not a direct heritage from the
Anisozygoptera, but rather by a parallelism or independent devel-
opment, and In general is a sign of specialization.

I think that enough evidence has now accumulated to show that
this obliquity may in turn, grow less and the quadrangle be re-
stored to distal squareness. In other words, the quadrangle is not
as fixed and unchangeable a part of the wing as might at first be
supposed. How else can the same shape of the quadrangle be
explained in three widely separate groups which agree only in the
reduction of Cu2? I refer to Disparocypha (fig. 37), Lestoidea (fig.
45), and the Protoneurinae. Evidently when this reduction takes
place new demands for bracing must be met and the quadrangle
again becomes square. The squareness of the quadrangle, then,
in the lower Agrionidae and in the Protoneurinae, is no proof of a

PHILIP A. MUNZ 35

primitive or of a specialized state in"botli groups, but of one in the
first, and of the other in the second.

8. Change in length of the quadrangle. — At first thought the
elongation of the quadrangle seems to be a tendency which may
point to departure from the ancient condition and which occurs
rather often. F'urther investigation, however, fails to permit of
the selection of any very well marked series in which this can be
traced. Apparent elongation may be due to narrowing of the
quadrangle itself or retraction of the nodus so that the quadrangle
almost reaches its level. -Beginning with Climacobasis (fig. 6)
and Echo (fig. 7) and proceeding to Psolodesmus (fig. 16), one
can observe some elongation, and likewise with Pseudophaea (fig.
21), Dysphaea (fig. 22), Anisopleura (fig. 25) and Rhinocypha
(fig. 26), and in some Megapodagrioninae the tendency is similar.
But for the whole suborder it has perhaps been rather dift'erent;
the highest Coenagrionidae have almost as short a quadrangle
as Pseudophaea (fig. 21), in fact often they may seem to have a
smaller one, partly because of distal obliquity and the resulting
shortness of the upper side. The length of the quadrangle, then,
does not greatly change, except in short series.

9. Attachment of supplementary sectors. — Here again is a tendency
not at all universal in its appearance, but occuring often enough to
deserve mention. Among the Polythorinae it is evident in Chal-
copteryx (fig. 4) ; among the Agrioninae it begins in Neurobasis
(fig. 14) and Phaon (fig. 17) and ends in Vestalis (fig 18); in the
Coenagrionidae, Thaumatoneura (fig. 51) exhibits it to a certain
degree, Megaloprepus (fig. 46),' Microstigma (fig. 47) and Anomisma
(fig. 48) to a higher one, but since in Pseudostigma (fig. 49) and
Mecistogaster (fig. 50) almost all secondary sectors drop out, the
tendency cannot longer manifest itself; nor can it in the (\)ena-
grioninae.

10. Straightening of the longitudinal veins. — Wherever there are
many supplementary sectors present, the princii)al veins tend to
curve, especially near the margin of the wing. But when these
sectors drop out the veins themsehes become straight and scarcely
curv^e. Among the Agrionidae, this is true of Micromerus (fig. 28)
and Philoganga; among the Coenagrionidae of Lestes (fig. 44),

MEM. AM. ENT. SOC, .V

36 A VENATIONAL STUDY OF THE ZYGOPTERA

Mecistogaster (fig. 50), the higher Megapodagrioninae, and the
Coenagrioninae and Protoneurinae generally. I think fossil evi-
dence bears out this tendency equally well ; it is, as was said above,
apparently one dependent on reduction in the number of sectors.

II. Matching up of the cross-veins in transverse series. — At the
same time that the preceding tendency becomes evident and,
largely in the same forms, it is very noticeable that the cross-veins
become matched up and form regular transverse series across the
wing. These are beautifully illustrated in the higher Pseudo-
stigmatinae, the Coenagrioninae and the Protoneurinae above all
others.

Miscellaneous Features

1. Differentiation of fore and hind wings. — This tendency so im-
portant among the Anisoptera has practically no place among the
Zygoptera. A marked difference in size and shape is noticeable in
only very few cases: Chalcopteryx (fig. 4) among the Polythorinae,
Pseudolestes (fig. 38) among the Lestinae, and sometimes Mecisto-
gaster (fig. 50) among the Pseudostigmatinae. These cases are
certainly evidence of a departure from the primitive condition.

Many Zygoptera ha^T a fairly constant difference in the length
of the quadrangles of the fore and hind wings, enough so that one
can tell which is the front and which the hind wing. The Poly-
thorinae have this difference well developed in that the quadrangle
of the fore wing is much the shorter and with fewer cross-veins.
Among the Coenagrionidae several subfamilies have the upper side
of the quadrangle of the fore wing shorter than that of the hind
wing. But, generally speaking, the differentiation of fore and
hind wings among the Zygoptera is not of enough importance for
it to be considert^d a sign of specialization.

2. The bracing of the stigma. — Dr. Needham has shown two ways
by which this may proceed, one by the cross-vein between Ri and
Ml below the inner end of the stigma becoming oblique and often
hypertrophied. This sort of bracing is not met with in the Agrioni-
dae, nor in the lower Lestinae nor Megapodagrioninae. In most
higher Lestinae, Megapodagrioninae, Coenagrioninae, and Pro-
toneurinae it is the usual thing. The plates in Handlirsch's
"Die fossilen Insecten" show bracing in Tarsoph/ehid, a genus

PHILIP A. MUNZ 37

of the Anisozygoptera and among the possible ancestors of the
Zygoptera.

The other method of bracing is for the inner end of the stigma
to be pointed and form a \\ the base of which is made l)y that por-
tion of Ri situated just before it. In this way Ri holds the stigma
firmly. Both methods may occur in the same wing. The Poly-
thorinae have this Y well developed.

3. Atrophy of the stigma. — The diffusion and even total disap-
pearance of the stigma in genera of two subfamilies, the Agrioninae
and Pseudostigmatinae, may not be considered by some as a sign
of advancement, but since the Odonata as far back as the Mesozoic
do have the stigma, and since the modern genera which stand at the
base of the Agrionine and Pseudostigmatine lines have it, and its
gradual disappearance is traceable in a series, its atrophy must
be taken as a matter of specialization, if not necessarily one of
advantage.

I\'— SUMMARY

F'rom the preceding discussion it is evident that venation gives
in most cases an easy and fairly clear means of determining rela-
tionships, but in certain ones it may be capable of more than one
method of interpretation, and must therefore be checked up by
other characters if a final decision is to be reached. Sometimes
this doubtful condition is due to parallelism, sometimes to contra-
dictory evidence from several sources; one character in the wing
may indicate one tendency, another quite a different one.

The keynote to the phylogeny of the Zygoptera as based on
venation is reduction. With this in mind one can di\ide the sub-
order into two families, the lower of which, the Agrionidae, has as
lateral offshoots: (i) the Polythorinae (figs. 2-5) in which M4 does
not descend the arculus and the quadrangle is therefore broad at
the base and concave above; (2) the Agrioninae (figs. 6-20), where
the wing fails to become petiolated, but the stigma atrophies and
the quadrangle becomes somewhat con\-ex abo\'e and widened
distally; and (3) at a much higher le\el. the Disparocyphinae (fig.
37) as a direct offshoot of the main line of descent, by a shortening
of Cuo, until it is represented by a cross-vein only. This main line
of descent is to be found in the Epallaginae (figs. 21-36) which

MEM. AM. ENT. SOC, ,^.

38 A VENATIONAL STUDY OF THE ZYGOPTERA

advance by petiolation, by reduction of the number of cross-veins,
especially of the antenodals, and by a retraction of the nodus
toward the wing-base. ^ The Epallaginae lead to the second family,
the Coenagrionidae, with which they are connected by various inter-
mediate forms: Diphlebia (fig. 34), Philoganga (fig. 36), Thaiima-
toneiira (fig. 51), Ortholestes (fig. 39) and Pentaphlehia. Early in
the Coenagrionidae the nodus becomes more retracted, the number
of antenodals becomes established as two, and the number of sup-
pleiTientary sectors shows a decided tendency to be decreased.
There are two general lines of descent as regards the points of
origin of Rg and M3; in the first one these veins retain the condition
of the Agrionidae and arise nearer the arculus than the subnodus;
this is true of the Lestinae (fig. 38-44) and Lestoidinae (fig. 45).
In the Lestinae is developed a quadrangle which is acute distally,
and in the other subfamily the vein Cu2 is again reduced so that it
does not extend beyond MA. The other line in the Coenagrionidae
has Rs and Ms migrate outward so that they arise nearer the sub-
nodus than the arculus. The Pseudostigmatinae (figs. 46-50) are
an immediate offshoot, with greatly elongated wings in which the
nodus is retracted and the stigma becomes diffuse. The Megapo-
dagrioninae (figs. 51-70) form the central branch which leads to
the Coenagrioninae (figs. 71-129) by the loss of all sectors except
Mia, and also continues in its own direction by greater petiolation
and outward progression of the base of M3. The Coenagrioninae
are also characterized by a slight movement outward of Mg and by a
great reduction in the number of postnodals; petiolation on the
whole does not proceed very far. This family in turn gives rise to
the Protoneurinae (figs. 130-153) by the reduction of Cu2, a change
which has thus taken place in three groups of Zygoptera. In the
Protoneurinae Cui also becomes short, and in the highest form,
Selysioneura (fig. 153), Cu does not branch at all.

V— KEYS FOR THE DETERMINATION OF GENERA

For the most part these keys indicate as nearly as they can be
made to do so, the same views regarding the phylogeny of the
Zygoptera as have been expressed in the preceding pages. The
Coenagrioninae have been the stumbling block lor such a system,

PHILIP A. MUXZ 39

and although an arrangement can be made which seems to follow
the general line of descent in this subfamily, and to indicate what
tendencies have been followed out venationally, it has been, so far
impossible to base any key on venation, since the genera as now
recognized have considerable variation within themselves. It has
been necessary, therefore, to make use of structural characters for
this subfamily, and to make a key for which no claims of natural
relationship can be made, but which is merely an attempt to aid in
identification. Before this difftcult group can be finally worked
out, however, much more material must be studied than it has
been possible to study for this paper.

The nomenclature of veins used in the keys is the same as that
shown in the wings of Agrion and Argiocnemis in the first part of
the paper. The number of the figure appearing in connection
with each genus refers to the plates at the end of the paper. For
the following genera neither figures nor specimens have been
available: Paraphaea., Caliphaea, and Pentaphlebia among the
Agrionidae; and Orolestes, Platylestes, Protolestes, Mesopodagrion,
Metagrion, Allolcstes, Argiagriou, Ileteropodagrion, Mesagrion,
Mesocnemis, Mornbagrioiu Myagrion, Nesocnemis, Nesobasis, Neur-
agrion, Onychargia, Paracnemis, Pericneniis, Phasmoneum, Sten-
ocnemis, and Thermagrion among the Coenagrionidae.

The material used for study of the other genera was that in the
collections at Cornell University, at the Academy of Natural
Sciences at Philadelphia, at the National Museum, In the private
collection of Mr. C H. Kennedy, and some borrowed from Mr.
E. B. Williamson, the Carnegie Museum, and the Museum of
Comparative Zoology. In addition to this free use was made of
those figures and keys which appear in the literature, and of some
photographs sent b\' Mr. H. (^^mp^()n of the British Museum.

Key to the Suborders of Odoxata

1 (4). Fore and hind wings essentialh' similar; in repose, held vertically or half-
opened; eyes far apart. A quadrangle present. Males with two superior and
two inferior abdominal a])pendages 2

2 (3). Area between M4 and Cui just beyond MA as broad as that between Cui
and posterior margin of the wing. Quadrangle of hind wing and that of fore
wing quite different in shape, the former twice as wide distalK- as i)r()\imally.
Sectors of the wing (except M,a) zigzag ANISOZ YGOPTERA (p. 40)

MEM. AM. ENT. SOC, 3.

40 A VENATIONAL STUDY OF THE ZYGOPTERA

3 (2). Area between M4 and Cui just beyond MA generally narrower than that be-
tween Cui and the posterior margin of the wing. The quadrangle of the fore
wing of the same general shape as that of the hind wing; one may be longer
than the other, but if one is widened distally, both are. Sectors of the wing
tend usually to be continuous veins, not broken nor zigzag.

ZYGOPTERA (p. 40)

4 (1). Fore and hind wings dissimilar, held horizontally in repose. Eyes close to-
gether. Supertriangle and triangle present. Males with two superior and one
inferior abdominal appendage ANISOPTERA^

Key to the Suborder Anisozygoptera
Family Epiophlebiidae. Subfamily Epiophlebiinae.
I. But one living genus. Characters those of the suborder. Japan and Eastern

Asia. (Fig. i) Epiophlebia Calvert, 1903

{Palaeophlebia Selys, 1899)

Key to the Families and Subfamilies of the Suborder Zygoptera

1 (8). Antenodal cross-veins 5 or more. Usually cross-veins present in the sub-
quadrangle beyond the vein Ac. M3 separating from M1+2 nearer to the arculus
than to the subnodus. Nodus usually more than one-third of the distance from
the base of the wing to the tip. The quadrangle often crossed.

(Family Agrionidae) ■ • ■ 2

2 (3). M not descending the arculus; quadrangle therefore concave anteriorly and
at least twice as wide proximally as distally. Basal space crossed. Wings
petioled at leasfone-half way to the arculus. Mia closely parallel to IMi.

Polythorinae (p. 42)

3 (2). M descending the arculus for some distance; the quadrangle not markedly
concave anteriorly, and not twice as wide proximally as distally, at least in the
hind wing. Basal space crossed or not. Wings petioled or not 4

4 (5). Mia parallel to M2. Stigma often atrophied. Quadrangle elongate, often
as long as the basal space, and convex anteriorly. Cuo beyond the quadrangle,
especially in the hind wing, sends off a branch (Cu2a) to support the area posterior
to it, or Cu2 itself bends back to support this area Agrioninae (p. 42)

5 (4). Mia not parallel to M2, but more or less intermediate between Mi and M2
(not in Micromerus) and with rather prominent sectors on either side. Stigma
well developed, at least in the hind wing. Quadrangle not so long as the basal
space, or if so, not generally convex anteriorly. Cu2a not present 6

6 (7). Cui and Cu2 well developed veins, A extending as far as the distal end of the
quadrangle. M4 generally somewhat curved beyond MA. Stigma usually not
much broader distally than proximally Epallaginae (p. 45)

7 (6). Cui short, not reaching the level of the subnodus. Cu2 reduced to a cross-
vein. A following the hind margin of the wing beyond Ac for some distance,
then bending forward and ending at the posterior side of the quadrangle. M4
not curved beyond MA. Stigma much broader distally than proximally.

Disparocyphinae (p. 47)

' Not further discussed in this paper.

PHILIP A. MUXZ 41

8(1). Antenodals generalh' two, rarely three to five. Usuall\- no cr()ss-\eiiis in the
pubciuadrangle beyond Ac. M3 separating from AIi^2 nearer the siihnodus than
the arcLilus, except in the Lestinae, none of which have more than two antenodals.
Xodus usually from one-fifth to one-third of the distance from the base of the
wing to the tip. The quadrangle crossed in Anomisma only.
(Family Coenagrionidae) 9

9 (12). M3 arising nearer the arculus than the subnodus. The basex)f R, (properly,
the base of the bridge to Rs) arising before the subnodus and generally as near the
arculus as the subnodus 10

10 (11). Supplementary sectors present between R^ and M3, and between .M3 and
M4. Cui and Cu. well developed, extending considerably beyond the subnodus.

Lestinae (p. 48)

11 (10). No supplementary sectors between R3 and M3, and between M3 and M4;
two between Mi and Mia. Cui one cell long and not reaching the subnodus.
Cu-> reduced to a cross-vein Lestoidinae (p. 49)

\2 (9). M3 arising nearer the subnodus than the arculus (half-way in Rhipidolestes).
The bridge to Rs short, arising at or beyond the subnodus (slighth' before in
Rhipidolestes) i ,>

13 (18). Cuonot reduced to a vein one cell long nor lacking 14

14 (15). Stigma difi"used or lost, not made up of one cell, but if present, usually of
several cells, and irregular in shape, not braced. The longest vein between Mi
and M2 rather more parallel to M2 than to Mi. Nodus at from one-fourth to
one-seventh the wing length. Many postnodal cross-veins.

Pseudostigmatinae (p. 49)

15 (14). A true stigma present, usually braced at the inner end. Mi,, not usually
parallel to Mo, when any other sectors are present. Nodus generally at more
than one-fourth of the wing length. The number of ])Ostnodal cross-\eins often
reduced 16

16 (17). Definite sui^iilementary sectors present other than Mh, and arising at least
as far in the wing as under the stigma. (No such sectors in Chorismagrion or
Perilestes, but these genera can be recognized b\- the j^etiolation of the wing which
extends almost to the level of MA, by the quadrangle coming near to the posterior
wing-margin, by the arching of Cui beyond MA, and b\- R,, arising sexeral cells
be\-ond the subnodus.) The stigma generally rather hea\>-.

Megapodagrioninae (p. 50)

17 (16). NO definite sui)i)lementary sectors other than Mi, are present, except oc-
casional very irregular ones which are only a few cells long and do not arise more
than a few cells from the wing-margin. Stigma generally rather small.

Coertagrioninae (p. 53)

18 (13). Cu2 reduced to a cross-vein, ending in the margin of the wing, or in two
genera to a vein one cell long Protoneurinae (p. 63)

MEM. AM. ENT. SOC. 3.

42 A VENATIONAL STUDY OF THE ZYGOPTERA

Key to the Genera of the Subfamily Polythorinae

1 (8). Cuo branched, more than one row of cells between Cu, and Cuoand between
Cuo and the wing-margin -

2 (7). Quadrangle of hind-wing one and one-half to two times as long as that of the
fore-wing. Wings petioled about one-half of the way to the arculus. Fore-wing
about three and one-half times as long as broad. Cuj forking, with two rows of
cells between the branches. Arculus at slightly more than one-third of the dis-
tance from the base of the wing to the subnodus 3

3 (6). Fore and hind wings of same general shape and size. Sectors arising between
the main veins, which do not appear forked. M2 arising within a cell or two of
the subnodus 4

4 (5). Venation very complex, many sectors and cross-veins; for example, 40-50
antenodals; often 12 or more cross-veins under the stigma; rather prominent
sectors between M4 and Cui, at least in the hind wing. In the hind wing, several
sectors between Mo and Rs arise as far proximad as the inner end of the stigma.
Northern South America from Colombia to Peru and Brazil. (Fig. 2.)

Polythore'5 Calvert. 191 7
{Thore Selys, 1853)

5 (4). Venation not so complex, antenodals 20 to 35; cross-veins under the stigma
6 to 10; usually no sectors between M4 and Cui, more than three or four cells
long. Between M2 and Rs no sectors arise proximad of stigma. South America,
Venezuela to Peru and Brazil. (Fig. 3.) Eu thore Selys, 1869

6 (3). Fore-wing evidently longer and narrower than the hind wing. Main veins
appear forked because of the attached sectors. Mo arising two to four cells be-
yond the subnodus, especially in the fore-wing. Brazil. (Fig. 4.)

Chalcopteryx Selys, 1853

7 (2). Quadrangle of hind wing scarcely longer than that of the fore wing. Wings
petioled about two-thirds of the distance from the base of the wing to the arculus.
Fore wing about four and one-half times as long as broad. Arculus often at almost
one-half the distance from the wing-base to the subnodus. Mexico to Bolivia.
(Fig. 5.) Cora Selys, 1853

8(1). Cua unbranched, one row of cells between Cui, and Cuo, one row of cells back
of Cu2 Miocora Calvert, 1917

Key to the Genera of the Sithfaniily Agrioninae

1 (24). Cu2a present. R, in hind wing usually strongly waved 2

2 (19). Cu2a running perpendicular to the posterior margin of the wing or toward the
base of the wing, and usually with one or more sectors toward the anal angle
which are longer than the quadrangle 3

3 (12). In the hind wing the sectors of that portion of the wing posterior to Cu>
and distad to Cui,, are generally arched from near the wing margin and not di-
rectly attached to Cu^a- True stigma present 4

" Calvert, 1917, proposes the name Polyllwre for this genus, since the name Thore
is preoccupied.

PHILIP A. MUNZ 43

4 (9). Basal space crossed. Arculus usually distinctly angled and not markedly
oblique 5

5 (6). Eight or more cells across the anal area along Cu-ia- More than two sectors
from Cu-2a run toward the anal angle, but they are irregular. Anal margin of
hind wing convex. Space between Cui and Cuo in hind wing reticulate just be-
yond MA. China. (Fig. 8.) Archineura Kirby, 1894

6 (5). Not more than four to five cells along Cu2a between Cuo and the wing-margin.
In the fore wing there is one sector from Cuoa toward the anal angle; anal margin
not convex. One row of cells between Cui and Cu2 at their origin 7

7 (8). Nodus less than one-half of the distance from the base of the wing to the
stigma. Southern and south-eastern Asia. (Fig. 7.) Echo Selys. 1853

8 (7). Nodus more than one-half the distance from the base of the wing to the stigma.
Southeastern Asia. (Fig. 6.) Climacobasis" Laidlaw, 1902

9 (4). Basal space free. Tendenc}' for arculus to be less angled and more
oblique 10

10 (11). Cu-2,i in the hind wing running almost toward the margin, so that it does
not reach the length of the quadrangle before sending off a branch toward the
anal angle. Western and Central Africa. (Fig. 9.) Umma Kirb}-, 1890

{Cleis Selys, 1853)

11 (10). Cuja in the hind wing running toward the base of the wing, almost the
length of the quadrangle before sending a branch toward the anal angle. West
Africa. (Fig. 10) Sapho Selys, 1853

12 (3). Sectors in that area of the hind wing posterior to Cua and distad to Cu2a,
usually (sometimes not in Agrion) attached directly to Cu2a. Stigma false or
lacking 13

13 (14). Basal space free. Almost cosmopolitan, except for most of Africa and
Australia. (Fig. 11.) Agrion Fabricius, 1775

{Calopteryx Leach, 18 15)

14 (13). Basal space crossed 15

15 (16). Broad area in hind wing posterior to Cuo, with five rows of cells between end
of Cu2a and the wing margin which is ver>' convex; in the hind wing the length is
less than three and one-half times the breadth. R^ very strongly waved in the
hind wing, cutting ofT a sector anterior to it. All four wings of the male opaque;
only one or two cells of the basal space divided. Borneo. (Fig. 12.)

Matronoides Foerster, 1897

" I am not sure that Climacobasis is generically distinct from Echo. Laidlaw (1903)
says of Climacobasis lugens and Echo modesta, "There can be no doubt that the two spe-
cies which I founded on a single specimen in each case, are merely male and female of
the same species. The generic name may be retained as that of a section of the genus
Echo, distinguished by the great relative length of the pterostigma in the male, in which
it is about three times as long as in the female." Williamson (1904) says, "Echo is
known to me only from descriptions and figures, but such differences between it and
Climacobasis appear to exist in the anal region and in the stigmas of the males that, in
so far as the two names have been published, their separation seems to me desirable, at
least till more critical study can be made."

Since I have had no access to specimens of Echo, I shall here follow tiie precedent made
by Williamson. The stigmatic differences in figures six and seven are due to sex rather
than genus.

MEM. AM. ENT. SOC, 3.

44 A VENATIONAL STUDY OF THE ZYGOPTERA

16 (15). Area of hind wing posterior to Cuo narrower, with two to four rows of cells
between end of Ciha and wing margin which is not ver>' convex; hind wing about
three and one-half times as long as broad. Rs not usually so strongly waved as
to cut off an anterior sector 17

17 (18). M2and Ms not appearing forked. Most cells of basal space divided into
two by median nervules. All four wings of male opaque. Cu2a alike in fore and
hind wings. Japan and southeastern Asia to India. (Fig. 13.)

Matrona Selys, 1883

18 (17). M2 and M3 appearing forked because of attached sectors. Only one or two
cells of basal space divided. Front wings of male hyaline. Cu.^ in fore wing
tends to parallel A. Malaysia. (Fig. 14.) Neurobasis Selys, 1863

19 (2). Cu2a turning away from the wing-base; the sector from Cu2a toward the anal
angle shorter than the quadrangle 20

20 (23). True stigma present. M1+2 fusing with R at least in the hind wing. Mi-i
and M4 arise separately from arculus. Asian genera 21

21 (22). Quadrangle approximately as long as the basal space and with four to seven
cross-veins; posterior wing-margin near the anal angle not concave. Stigma
about three times as long as broad; the distance from the wing-base to the nodus
more than two-fifths of the wing length. Posterior end of the medio-anal link
more basal than the anterior end. China and Japan. (Fig. 15)

Mnais Selys, 1853

22 (21). Quadrangle one and one-half times as long as the basal space and with ten
or more cross-veins; posterior wing-margin near the anal angle very concave.
Stigma four times as long as broad; the distance from the wing-base to the nodus
less than two-fifths of the wing length. Posterior end of MA not so basal as
anterior end. China and Formosa. (Fig. 16.) . .Psolodesmus McLachlan,i870

23 (20). Stigma present or not. M1+2 not fusing with R. Mi-sand M4 usually arise at
a common point, especially in the hind-wing. R^ in the hind wing waved so as
to approach M2 very closely. Africa and Madagascar. (Fig. 17)

Phaon Selys, 1853

24 (i). Cu2a not present. Area posterior to Cu2 supported by a wide bend ot Cu2.
Rs in the hind wing not strongly waved. Quadrangle often very convex anter-
iorly and widened distally. Arculus not angled, but oblique 25

25 (26). Basal space free. M2, M3, Cui, and Cu2 with attached sectors. Veins and
sectors turn sharply toward wing margin. Quadrangle almost as long as the
basal space. India to Malaysia. (Fig. 18.) Vestalis Selys, 1853

26 (25). Basal space crossed, distinctly longer than the quadrangle. Sectors inter-
mediate, generally not attaching to the main veins. Veins and sectors not bend-
ing sharply toward the wing-margin. R^ and M., parallel at the end, with two
intervening sectors. Genera of the Western Hemisphere, ranging from (\inada
to South America. (Figs. 19 and 20.)

Heterina^ Selys, 1853. and Lais Selys, 1853

PHILIP A. MUNZ 45

Key to the Genera of the Subfamily Epallaginae

1 (lo). Wings petioled not more than one-half way from the wing base to the
arculus. Quadrangle not more than four times as long as broad; with not more
than two or three cross-veins. Area posterior to Cuo. at its widest point, at least
one-third the width of the wing at that point with several sectors beyond MA. . .2

2 (5)- Quadrangle with one or two cross-^■eins; subquadrangle usually with more
than two. In the hind wing the basal space generally not more than twice as
long as the quadrangle 3

3 (4). Area back of anal vein with only one row of cells to beyond MA, at least in
the fore wing. Stigma with anterior proximal angle away from the wing-base.
In the fore wing the antenodal cross-veins are less numerous than the postnodal.
India to China and Malaysia. (Fig. 21.) Pseudophaea' Kirby, 1890

{Euphaea Selys, 1840)

4 (3). Area back of anal vein with two rows of cells to beyond MA. Anterior prox-
imal angle of stigma toward the wing-base. Antenodal cross-veins more numer-
ous than the postnodal. Indo-China to Malaysia. (Fig. 22.)

Dysphaea Selys, 1853

vS (2)- Quadrangle without cross-veins; subquadrangle usually with two or fewer.

Basal space generally more than twice as long as the quadrangle 6

6 (7). Basal radial space closed and free. India. (Fig. 23.)

Bayadera Selys, 1853

7 (6). Basal radial space not closed in both wings 8

^ The genera Heterina and Lais now contain so many species with so many inlergrad-
ing characters, that it is impossible to fix upon any clearly defined characters for the sep-
aration of the two genera, without a more detailed study than I have been able to give
them. Many species of Heterina have several rows of cells back of the anal bridge in the
forewings, while Lais is supposed to have but two such rows. However, in specimens of
H. donna, vtilnerata, criienlata, tricolor, titia, caja, macropus, etc. some have only two.
In many species of Hetaerina the males have the base of the wing red. Beyond these
characters further ones are unknown to me.

^ The genus Paraphaea, Martin, 1903, would seem, from the description, in the Bulle-
tin of the Museum of Paris 1903, to come in this part of the key, but the description is
too incomplete to use for the key. The transUition reads as follows; "This new genus is
created from a single specimen from Manilla and belonging to the Museum of Paris.
It is of the Legion Euphaea of M. de Selys. The basal space is free, the median sector
(Ms) branches from the principal (Mi) toward the end of the quadrangle; the feet are
quite long, the body slender, the fundamental color of the body red, but the ventral
surface of the thorax entirely black. The nodus is placed exactly at the middle of the
wing. The superior appendages are slightly bent not dilated, the inferior small; the
posterior margin of the tenth segment bears a papilla."

Under the specific description of P. barbata: "Abdomen 55 mm.; lower wing 41
mm.; 24 antenodals and 24 postnodals to the upper wings; 24 anienodals and 24 post-
nodals to the lower wings. Wings hyaline, slightly tinged with saffranin, the distal tip
of the stigma bordered with brown; the stigma brown, covering se\cn cells, broad at the
middle, narrowed at the two ends which are pointed, especially proximally. Quadrangle
small, short, crossed."

MEM. AM. ENT. SOC, 3.

46 A VENATIONAL STUDY OF THE ZYGOPTERA

8 (9). Basal radial space crossed. M2 usually arising not more than one cell from
sub-nodus. The longest sector between Cui and Cu2 arising below a point sev-
eral cells before the level of the subnodus. Greece to Persia. (Fig. 24.)

Epallage Charpentier, 1840

9 (8). Basal radial space free; M2 tends to arise two cells beyond the subnodus.
The longest sector between Cui and Cuo, especially in the hind-wing, arises almost
below the subnodus. In the hind-wing of the male, C has a hump before the
arculus. India to Burma. . (Fig. 25.) Anisopleura Selys, 1853

10 (i). Wing usually petioled two-thirds or more of the way to the arculus. Length
of quadrangle and the number of cross-veins vary. The area posterior to Cui
narrow, at its widest point usually less than one-fourth the width of the wing at
that point, and with n^t more than two distinct longitudinal sectors 1 1

11 (16). Area posterior to Cu2 a single row of cells (occasionally two cells wide for a
short distance; but if so, Cuo is not arched). Quadrangle in the fore wing usually
at least six times as long as wide. Beyond MA, M4 arches forward, Cuo back-
ward. M2 and Rs near the wing-margin generally are parallel and with two rather
prominent parallel sectors between 12

12 (13). Quadrangle generally with two or more cross-veins, at least in one wing.
The sector anterior to Mia arising before the inner end of the stigma, at least in
the hind wing. India to Malaysia. (Fig. 26.) Rhinocypha Selys, 1853

13 (12). Quadrangle usually with less than two cross-veins. The sector anterior to
Mia, if present, not arising before the inner end of the stigma 14

14(15). Short sectors between Ml and Mia. Third antenodal over arculus. Quad-
rangle uniform in width or widened distally. West Africa to Zanzibar and in
Philippines. (Fig. 27.) Libellago Selys, 1853

15 (14). No sectors between Mi and Mia. Second antenodal over arculus. Quad-
rangle wider proximally. India to Malaysia. (Fig. 28.)

Micromerus Rambur, 1S42

16 (11). Area posterior to Cuo with more than one row of cells beyond MA, except
in Dicterias and Heliocharis. (Of these two, Heliocharis often has basal space
crossed; Dicterias has M2 and Rs ending very near to each other.) M4 generally
not prominently arched beyond MA 17

17 (34)- Quadrangle at least four times as long as wide. Wings generally widest
beyond the level of the subnodus 18

18 (27). Antenodal cross-veins usually more than twelve. Area posterior to Cu2
of one row of cells only, sometimes two rows in Neocharis: or Cu2 forks and has
two rows of cells between the forks. Quadrangle crossed 19

19 (20). Cu2 forking; two rows of cells between the forks. First costal antenodal
almost over the arculus. Brazil. (Fig. 29.). . . .Cyanocharis Needham, 1903

20 (19). Cu2 not forking. First costal antenodal generally some distance before the
arculus 21

21 (22). Stigma covering only two to three cells. Supplementary sectors in apex
of wing numerous and curved; some attached to the principal veins. Spines on
the legs very long. India and China Caliphaea^o Selys, 1859

1" Same as Notholesles, McLachlan, 1.SH7. See McLachlan, 1894, Annals and Magazine
of Natural History, (6) 13: 435. I have seen neither figures nor specimens of this genus,
but descriptions seem to place it in the Epallaginae.

PHILIP A. MUXZ

47

22 (21). Stigma covering four to se\en cells. Supplementar\- sectors not attached.
S[jines on legs medium in length. Genera of South America 23

23 (24). Basal space often crossed. Sectors between Mia and M. do not arise before
the level of the inner end of the stigma. Mi-^ forks alxjut two cells beyond the
subnodus. Quadrangle with one cross-\ein. Brazil. (Fig. 30. )

Heliocharis Selys. 1853

24 (23). Basal space free. M2 arises within a cell of the subnodus. Quadrangle
with one or two cross-veins 25

25 (26). Some sectors between Mia and Mj arise far before the level of the stigma.
Quadrangle with one cross-\ein. Brazil. (Fig. 32.). . . .Dicterias Selys, 1S53

26 (25). Xo sectors between Mi^and M.arise before the level of the stigma. Quad-
rangle with two cross- veins. British Guiana. (Fig. 31.)

Neocharis Foerster. 1906

27 (18). Antenodal cross-veins usually less than nine. Area posterior to Cu* of
more than one row of cells, but Cuj not s\mmetrically forked 28

28 (29). Two cross-veins in the quadrangle; two or three often in the subquad-
rangle. Stigma about three times as long as wide. Fourth antenodal over the
arculus. M2 arising four to five cells beyond the subnodus. Straits Settle-
ments. (Fig. 33.) Devadatta Kirby, 1890

(Tetraneura Selys, 1859)

29 (28). Xo cross-veins in the quadrangle: usually none in the subquadrangle.
Stigma about five times as long as wide, fusiform in outline. Second or third
antenodal over the arculus • ■ .30

30 (33). Genera of the Old World. M2*not arising more than one cell beyond the
subnodus. Sectors present between Mi and M^. Second antenodal over the
arculus 31

31 (32). M2 arising at the subnodus. or within one-half cell of it: M^ arising three
to five cells beyond M2. Arculus less than one-half way from the base of the
wing to the nodus. Antenodals fi\'e to seven. Australia and Malaysia. (Fig.
34.) Diphlebia Selys, 1869

32 (31). M2 arising one cell beyond the subnodus. Mi^ five cells beyond Mj.
Arculus at one-half the distance from the base of the wing to the nodus. An-
tenodals generally five. West Africa Pentaphlebia" Foerster, 1909

33 (30). A genus of the Xew World. .M; arising two to four cells beyond the sub-
nodus. Often no sectors between Mi and M^. The third antenodal over the
arculus. Colombia. (Fig. 35.) Amphipteryx Selys. 1853

34 (17)- Quadrangle about twice as long as wide. Wings widest before the level of
the nodus. Antenodals about fifteen. India to China. (Fig. 36.)

Philoganga Selys, 1879

Key to the Disparocyphinae

I. One genus onl\-. characters those of the subfamily. .AntenodaU ten to twelve
in each series. South Celebes. (Fig. 37.) Disparocypha Ris, 1916

" Only the description was available for this genus.

MEM. XM. EST. SOC. 3.

48 A VENATIONAL STUDY OF THE ZYGOPTERA

Key to the Genera of the Subfamily Lestinae

1 (2). The distal end of the quadrangle cut off almost squarely. The front wing
over one-third longer than the hind wing. A short sector present in the fore
wing between Mi and M,a. In the hind wing a sector present between M4 and
Cui. The distance between the points of origin of Ms and the bridge to R, is
almost as long as that from the origin of M3 to the arculus. Hainan. (Fig. 38.)

Pseudolestes Kirby, 1900

2 (l). The distal end of the quadrangle is not cut off squarely. The fore and hind
wings of about the same length. No sector between M, and Mia (except in
Ortholesles). The distance between the origin of M,, and that of the bridge to Rs
is less than the distance from the origin of Ms to the arculus 3

3 (4). The quadrangle widened distally and with the superior side over one-half the
length of the inferior. Xo sectors between Ma and M4. Two rows of cells be-
tween M4 and Cui at the end of Cui, at least in many specimens. West Indies.
(Fie 3Q.) Ortholestes Calvert, 1891

4 (3). The quadrangle not widened distalh and with the superior side scarcely
one-half the length of the inferior. One or more sectors. between M, and M4,
except in Orolestes. One row of cells between M4 and Cui.

5 (6). The inner end of the quadrangle in the fore wing almost one-half the length
of the lower side. M1+2 forking less than two cells beyond the subnodus. The
third antenodal space almost twice the length of the first. California to X'ene-
zuela. (Fig. 40.) ■ Archilestes Selys, 1862

6 (5). The inner end of the quadrangle one-third or less of the length of the lower
side except in Orolestes. M,+2 forking more than two cells beyond the subnodus,
at least in one wing. The third antenodal simce not more than one and one-
half times the length of the first 7

7 (8). Wing petioled to within Ac's length of itself. No sectors between M,, and R^.
The longest sector between Mo and Rs arises about one-half of the way between
the stigma and the origin of Mia. Mia not zigzag. India. (Fig. 41.)

Megalestes Selys, 1862

8 (7). Wing petioled to Ac or nearly so. Sectors present between M3 and R^. The
longest sector between M2 and R., not arising more than three or four cells before
the level of the stigma 9

9 (10). Wings, at least those of the male, with large opaque areas. Cu-- is the only
zigzag vein. Stigma \ery large, surmounting four to six cells. M1+2 forking
five to se\en cells beyond the subnodus. India Orolestes^-^ McLachlan. 1895

10 (9). Wings hyaline. Many of the veins strongly zigzag; i. e. the cells are angled.
Stigma shorter, surmounting two to three cells. Mi^2 forks three to seven cells
beyond the subnodus ■ . . . 1 1

1 1 (12). Stigma broad, scarcely two and one-half limes as K)ng as wide. R. zigzag.
The anal appendages of the female o\oid, \ery much depressed, as long as the
tenth segment. A single species, from Fast Indies.. .Platylestes'- Selys, 1862
'2 Descriptions only have been available for these genera.

PHILIP A. MIXZ 49

12 ( 1 1 ) . Siignia narrower, usualh- three or more times as long as wide 13

13 (14). Wings raised, when at rest. Inner end of quadrangle only one-tilth as
long as the inferior side. The prothorax with three scallops postericjrK , the mid-
dle one of which is the longest. Euroix> to Madagascar, New duinea, and New
Caledonia. (Fig. 42.) Sympycna Chari)eniier, 1840

14 (13). Wings held horizontalK, when at rest . The i)rothora\ not as aho\e 15

15 (16). The sectors of the arculus arising toward the middle, M.i about one-half of
the wa\' down; hence the ciuadrangle is narrow with the inner end between one-
fourth and one-fifth as wide as the inferior side is long. The ([uadrangle in the
hind wing has a distinctly longer upper side than does that of the fore wing.
The upper and lower sides are almost parallel. A genus from .Australia and sur-
rounding islands. (Fig. 43.) Austrolestes TilKard, 1913

16 (15). The sectors of the arculus arising nearer its u[)per end, Mi.ihdut one-third
of the way down, hence the inner end of the quadrangle is from one-third to
one-fourth of the length of the inferior side. The cjuadrangle of the front and
hind wings essentiall\- alike and triangular in shape l)ecause ol the narrow-
ing distalh-. Cosmopolitan, excejit not in Australia. (I'ig. 44.)

Lestes Leach. 1815

Key to the Subfamily Lestoidinae

I. One genus only; the characters those of the subfamiK'. .Australia. (F"ig. 45.)

Lestoidea TilKard, 11J13

Key to the Genera of the Subfamily Fseudostigmatiiiae

1 (6). Secondary- sectors numerous; two or more rows of cells between M,.. and M2
for some distance. Mr^i forking at more than one-hall of the distance trom the
subnodus to the tiyj of the wing. Cuj short, ending at the level ot less than one-
half the distance from the fork of Mi-j to the tip of the wing. Cui with many
forks; several to many cells between Cu2and the i)osterior wing margin 2

2 (5). Quadrangle not crossed; about five times as long as wide. Xo cross-\ein in
the subquadrangle. M.A before the le\el of the subnodus. M1-2 forking be-
yond the half-way point between the subncKius and the tij) of the wing 3

3 (4). The nodus at one-sixth of the wing length; M2 arched forward under the
stigma so as to make a sharp bend. M forks under the inner end of the stig-
matic area, at about fifteen cells from the margin of the wing; Cui forks many
cells from the wing-margin. Mexico to C.uatemala. (Fig. 46.)

Megaloprepus Rambur, 1842

4 (3). The nodus at one-fifth of the wing length; M^ not so strongl\- arched. M.3
forks about five cells from the wing margin or not at all. Cui forks about five
cells from the wing margin. Colombia to Brazil and Fcuador. (Fig. 47.)

Microstignia l\,imi)in-. 1S42

5 (2). The (luadrangle with two or three cross-veins, sexeral in the sub(iuadrangle,
some of which line up with tho.se in the ([uadrangie. M.A in the hind-wing es-
pecially, is at the level of the subnodus. .M1+2 forking at about half wa\- between
the subnodus and the wing tip. Peru and Ecuador. (Fig. 48.)

Anomisma .Mil.aclilan, 1877

MEM. AM. ENT. SOC, 3.

50 A VENATION AL STUDY OF THE ZYCiOPTERA

6 (i). Secondarx- sectors few; a tendency to only one row of cells between Mu and
Mo, if two rows, it is for a short distance only, M1+2 forking at less than one-half
the distance from the subnodiis to the tip of the wing. Cua longer, ending at
one-half or more than one-half of the distance from the fork of Mi+-. to the tip
of the wing, at least in the fore wing. Cu-> iinbranched. One or two rows of cells
between Cu2 and the posterior wing margin. One row of cells between Ri and
Mi+2 before the fork of the latter 7

7 (8). The nodus at one-fifth of the wing length. The hind wing, at least, tends to
ha\-e two rows of cells between Mia and M2 for a short distance. Mi has several
short sectors at its end; there are two rows of cells posterior to.Cu2. Mexico to
Colombia. (Fig. 49.) Pseudostigma Selys, i860

8 (7). The nodus at one-fourth to one-fifth of the wing length. One row of cells
between M^ and M2. No sectors more than a few cells long, except between Mi
and Mia. M4 usually unbranched. One row of cells between Cu2 and the pos-
terior wing-margin. West Indies and Mexico to Ecuador. (Fig. 50).

Mecis togas ter Rambur, 1S42

Key to the Genera of the Subfamily Megapodagrioniuae

1 (2). Antenodals three to five, postnodals over fifty; venation very complex, many
supplementary sectors from Mi to the posterior margin of the wing. About ten
cross-veins under the stigma. Quadrangle almost rectangular at both ends.
Wings not petioled to the level of the arculus. Panama to Costa Rica. (Fig. 51 .)

Thaumatoneura McLachlan, 1897

2 (i). Antenodals two to three, postnodals generally not over thirty; venation
simpler; supplementary sectors not usually present between Mi and Mia, and
Cui and Cui. Cross-veins under stigma six or fewer. Quadrangle generally
acute di'^tally. Wings usually petioled beyond the level of the arculus 3

3 (4). Rs arising one cell before the subnodus, Ms about one-half the distance be-
tween the arculus and the subnodus. Supplementary sectors Mia to Cu2. \\'ing
petioled to the level of one-third to one-half of the distance between the arculus
and MA. Formosa. (Fig. 52.) Rhipidolestes Ris, 191 2

4 (3) . R„ not arising before the subnodus. M3 arising nearer the subnodus than the
arculus '"^

5 (30). R. arising at or ven.' near the subnodus, Ms at least one cell before the sub-
nodus

6 (7). .'\rea posterior to Cwi with four or fixe rows of cells. Wing petioled to or
beyond the le\el of the arculus. Supplementary sectors from Mia to Cui or Cu2.
Malaysia and Australia. (Fig. 53-) Podoptery.x Selys, 1871

7 (6). Area posterior to Cu2 with one or two, rarely three rows of cells. 8

8 (19). Wing not petioled to the level of the arculus, it may be almost so in Tri-
neuragriou, a genus with a third antenudal 9

9 (14). Short sectors present between R, and M3 as well as some of the other longi-
tudinal veins ^^^

PHILIP A. MUX/ 51

ic (11). One or two sectors Ix-twcen Mi and Mia; others fmm Mh. to M., and M4

and Cui. Malaysia. (Fig. 53.) Rhinagrion Calvert, 1913

(Aviphilestes Selys, 1862)

11 (10). No sectors between Mi and Mi^ 12

12 (13). Three antenodals, the third one being costal and beyond the arculus. One
row of cells back of Cuj. New Caledonia. (Fig. 54.) . . Trineuragrion Ris, 1915

13 (12). Two antenodals. Generalh^ at least two more or less complete rows of
cells back of Cih. Australian region. (Fig. 56.) Argiolestes Selys. 1862

14 (9). Xo sectors between Rsand M3 15

15 (16). A genus of the Old World. Three cells below the stigma, postnodals about
twenty. Thibet? Mesopodagrioni^ McLachlan, 1896

16 (15). Genera of the New World. One cell below the stigma: postnodals about
fifteen 17

17 (18). Three rows of cells in the area between Mu, and M2, with two unbroken-
sectors between M.. and Rs. In the male the stigma is two and one-half times as
long as wide. Brazil. (Fig. 57.) Megapodagrion Selys, 1885, sens, strict.

18 (17). Two rows of cells in the area between M^, and M2 and one sector between
M2 and Rs- Stigma about twice as long as wide. Xorthi'rn South' America.
(Fig "^8.) Allopodagrion Foerster, 1910

19 (8). -Vt least one wing petioled to be\()nd the level of the arculus 20

20 (21). At least one sector between M4 and Cui. Sectors present Mia to Cui.
Generally two or more rows of cells back of Cu2, at least part way. Australian
region. (Fig. ^56.) Argiolestes Selys, 1862, sens, strict.

21 (20). Xo sectors between M4and Cui. One row of cells back of Cu-j 22

22 (29). Supplementary sectors between R, and Ms as well as Mh. to Rs and be-
tween M.j and M4- Cui not strongly arched beyond MA 23

23 (24). The anterior of the sectors between Rsand M3 shorter than the posterior
one. Fi\e cross-veins under the stigma. The longest sector between Mh^ and
M2 arising four to six cells before the inner end of the stigma. Between M3
and M4 only one sector is more than three or four cells long. Costa Rica to
Brazil. (Fig. 59.) Philogenia Selys. 1862

24 (23). The anterior of the sectors between Rsand M3 is longer ihan the posterior
one. Two or three cross-\eins under the stigma. The longest sector between
Mir, and M2 arises not more than one or two cells before the inner end of the
stigma. Between M3 and M4 art' often two sectors more than three cells long,
especially in the hind wing 25

25 (26). Wing not petioled beyond ihe arculus. At least three cells Ijetween MA
and the level of the subnodus in the area between M4 and Cui. Sui)i)lementary
sectors (one or two each) from Mu, to M4. Labuan and Sumatra. (Fig. 60.)

Podolestes Selys, 1862

26 (25). Wing is petioled be> ond the arculus. Two cells between M.\ and the level
of the subnodus in the area between M4 and Cui- Supplementary sectors (two
each) from M^, to -M4 27

'^ Description only was available for ihis genus.

MEM. AM. EXT. SOC, 3.

52 A VENATIONAL STUDY OF THE ZYGOPTERA

27 (28). Wing petiuled to the level of about one-third the length of the quadrangle.
Three antenodals. Tendency to a cubito-anal cross-vein before the subquad-
rangle in addition to the vein Ac. Arculus not beyond the last antenodals.
West Africa. (Fig. 62.) Neurolestes Selys, 18S2

28 (27). Wing petioled to the level of one-half to two-thirds of the length of the quad-
rangle. Two antenodals. No accessory cross-\eins before the subquadrangle.
Arculus distinctly beyond the second antenodals. Madagascar. (Fig. 61.)

Nesolestes Selys, 1891

29 (22). No sectors present between Rs and M^. Cui strongly arched beyond IMA.
One sector between Mia and M. and one between M2 and Rs. Australia. (Fig.
5^.) Synlestes Selys, i86q

30 (5). Rs arising at least a cell beyond the subnodus. M3 usually not before the
subnodus, at least not a cell's length before it. The wing petioled to beyond the
level of the arculus 3 ^

31 (32). Sectors Mi to Cu2. Many sectors between Cui and Cu2 so that the dis-
tance on the hind margin between these two veins is as great as that from the
level of MA to the end of Cu2. Quadrangle very long, almost or quite reaching
the subnodus. Wing greatly petioled. Mexico to Central America. (Fig. 64.)

Paraphlebia Hagen, 1861

32 (31). No sectors between M.and M^. Cui and Cuj usually ending near together
at the wing-margin 33

33 (36). The area behind Cu2 of two or three rows of cells. Supplementary sectors
at least from M la to Cui 34

34 (35). Ms arising at the subnodus. Ac is slightly beyond the le\el of the first
antenodal. New Guinea. (Fig. 65.) Wahnesia Foerster, 1900

35 (34). Ms arising slightly beyond the subnodus. New Guinea.

Metagrion'^ Calvert, 1913

36 (33). The area posterior to Cu2 of one row of cells only 37

37 (48). Supplementary sectors rather numerous. Some between Rs and M3. . . .38

38 (41) . Supplementary sectors from Mia to M4. Old World genera 39

39 (40). Ms arising from M1+2 at the subnodus. Often more than two sectors be-
tween Mia and M2. Quadrangle not reaching the level of the nodus. South
Africa. (Fig. 66.) Chlorolestes Selys, 1862

40 (39). Ms arising before the subnodus. Not more than two sectors between Mia
and M2. Quadrangle long, reaching the level of the nodus. Seychelles.

Allolestes^^ Selys, 1869

41 (38). No sectors between Ms and M4. Genera of South America 42

42 (43). Antenodals more than two, at least in the hind wings. Fcuador.

Neuragrion'^ Karsch, 1891

43 (42) . Antenodals two only 44

44 (47). Ms sei)arating from M1+2 before the subnodus 45

45 (46). Su])i)lementary sectors from M^, to Ms, often none between M2 and R3.
Usually an accessory cross-vein between Cu and A. Guiana. (Fig. 67.)

Dimeragrion Cahert, 1913

" Ucscriptions only were available for these genera.

PHILIP A. MIX/ 53

46 (45). Sui)i)lemL-iuar\ sectors helwecMi Mjaiid R, and two lon^ ones Ijetween R3
and M:i. Xo cross-veins in the subquadrangle or before it. Ecuador.

Heteropodagrion'* Selys. 1885

47 (44). Ms separating from Mr- beyond the subnodus. Supplementary sectors
between Mia and M2, and Rs and M3. Colombia Mesagrioni^ Selys, 1885

48 (37). Xo supplementary sectors between R^ and M., 49

49 (50). Rs arising one cell beyond the subnodus. Two sectors between M,,. and
Mo, some also between M- and Rs. and M., and M4. Cui arched beyond MA.
South Africa. (Fig. 66.) Chlorolestes Selys, 1862

50 (49). Rs arising at least two or three cells beyond the subnodus. Xo sectors be-
tween Mo and Re nor between M.3 and M4 5^

51 (54). At least some supplementary sectors present. Quadrangle not approach-
ing very closeh' the posterior wing-margin. Cui not strongly arched beyond
MA 52

52 (53)- One sector between M^, and Mj. Africa and Madagascar.

Protolestes'^ Foerster, 1899

53 (52). Two sectors between M,,, and -Mj. The Xew World, from Mexico to
Argentina. (Fig. 68.) Heteragrion Selys 1862

54 (51). Xo supplementary sectors excei^t Mia- The quadrangle at its distal end
approaches the hind margin of the wing. Cui strongly arched beyond MA .... 55

55 (56)- The fore wing of the male with the inner end of the quadrangle lacking.
Ac about one-third of the way from the first to the second antenodals. R,
arising tour cells beyond the subnodus in the fore wing, three in the hind wing.
The upper side of the quadrangle distinctly shorter than the lower side. Aus-
tralia. (Fig. 69.) Chorismagrion Morton, 1914

56(05). The quadrangle touching the jmsterior wing-margin. The wing petioled
to Mia. Ms arising one to one and one-half cells beyond the subnodus, R, at
five to se\en cells beyond the subnodus. Costa Rica to Brazil. (Fig. 70.)

Perilestes Selys, 1862

Key to the Genera of the Subfamily Coenagrioninae

1 (22). Quadrangle relatixeh' regui^ir, the upper side two-thirds to three-fourths
or more of the length of the lower in both wings. M^ and Cu-. scarcely zigzag, so
that most of the cells of the wing are four-sided rather than five-sided; the longi-
tudinal veins arranged in a parallel and \ery regular fashion. Tibial sjjines and
quadrangle relati\ely long. (Legion Platyciiemis of Sehs for the most jxirt) .... 2

2 (13). Wings generalh- plainly not petioled to Ac. R^arisingat the subnodus. . . .3

3 (6). The upper side of the (|uadrangle more than three-fourths the length of the
lower side. Antenodal cells between M4 and Cui are three to five in number.
usualK- at least four. .\c near the level of the first antenodal cross-veins. The
first and second antennal segments short, the third long and slender 4

4 (5). Three to five antenodal cells between Mi and Cu,. Arculus under the second
antenodals. Africa. (Fig. 71.) Metacnemis Selys, 1863

'5 Descriptions only were available for these genera.

MEM. .\M. ENT. SOC, 3.

54 A VENATIONAL STUDY OF THE ZYGOPTERA

5 (4). Four antenodal cells between M4 and Cui. The arculus between the first
and second antenodals, much before the second. West Africa.

Mesocnemis''^ Karsch, 1891

6 (3). The upper side of the quadrangle not more than three-fourths the length of
the lower side; the antenodal cells three or less between M4 and Cui. Ac gen-
erally beyond the level of the first antenodals. Antennae not as above except in
Platycnemis and Stenocnemis 7

7 (10). Cuo long, extending beyond the level of the base of Mia which is six to seven
cells before the level of the stigma. Postnodals fifteen or more. Posterior
tibiae of males not dilated, superior anal appendages at least as long as the in-
ferior 8

8 (9). Three antenodal cells between M4 and Cuj. About fifteen postnodal cells in
the front wing. The third segment of the antenna shorter than the first plus the
second. Cuo long, extending beyond the level of the base of ^NIi^, which vein arises
six to se\-en cells before the stigma. India to Burma. (Fig. 72.)

Calicnemis Selys, 1S63

9 (8). Two antenodal cells between JVI4 and Cui. Front wing with eighteen to
twenty postnodals. The third segment of the antenna longer than the first plus
the second. West Africa Stenocnemisi^ Karsch,_ 1899

10 (7). Cu2 not extending beyond the level of the base of AIu, the latter arising two
to four cells before the stigma. About thirteen postnodals. Two antenodal
cells between M4 and Cui. Posterior tibiae of males more or less dilated, su-
perior anal appendages shorter than the inferior 11

11 (12). Cu2 extending almost to the level of the base of Mia. The second antennal
segment as long as the third. Syria to Japan and Sumatra; Madagascar..
(Fig. 73-) Copera Kirb>-, 1 890

{Psilocnemis Selys, 1863)

12 (11). Cu2 not extending to the level of the base of Mia by several cells. The third
antennal segment as long as the first plus the second. Mediterranean Region and
China. (Fig. 74.) Platycnemis Charpentier, 1840

13 (2). Wings petioled close to or beyond Ac; R^ usualh' arising beyond the sub-
nodus (sometimes at the subnodus in AUocneiftis) 14

14 (21). R3 arising within one or two cells of the subnodus 15

15 (20). Wings not petioled to the level of the arculus. Ac almost as near the level
of the second antenodal as of the first 16

16 (17). The end of the wing with a denticulate margin. Superior appendages of
the male longer than the inferior, which are straight. Malaysia. (Fig. 75.)

Idiocnemis Selys, 1878

17 (16). The end of the wing not denticulate at the margin. Superior appendages
not longer than the inferior 18

18 (19). The upper side of the quadrangle not more than three-fourths the length of
the lower. Cu2 extending beyond the subnodus to a distance of thirteen to six-
teen cells. Usually six to nine cells between the base of R, and the point where
M,-|-2 forks. India to Malaysia. (Fig. 76.) Coeliccia Kirby, 1894

{Trichflcucniis Selys, 1857)
" Description only was available for this genus.
" Studied from descriptions only.

PHILIP A. MUXZ

00

19 (iS). The upjXM- side (if ihe (luadrangle scarceh' shorter than tlie lower. Cu2
extending six to eight cells be>ond the level of the siibnodiis. Four to h\e cells
between the base of Rs and the forking of Mi-o. West Africa to South Africa.
(Fig. 77.) Allocnemis SeKs, 1863

20 (15). Wings petioled beyond the level of the arculus. Ac under or near the level
ot the first antenodal. Rs arising one or two cells beyond the subnodus. M3 at the
subnodus. Madagascar. (Fig. 78.) J Tatocnemis Kirby, 1889

Nesocnemisi^ Sehs, 1891

21 (14). Rs arising fourorniore cells be\ond the subnodus, at least in the frontwing.
Ac about one-half way between the antenodals. Philii)i)ines. (Fig. 79.)

Prionocnemis Kirby, 1890
{Ilypocnemis SeKs, 1863)

22 (i). The quadrangle generally with a very acute distal lower angle, the upper side
of the quadrangle usually less than two-thirds the length of the upper side.
M4 and Cui usually zigzag, so that many cells are pentagonal. Wings generally
not long nor with the straight appearance of those under rubric (i). (If the
wings have the general appearance of those under (i), the quadrangle is acute at
the end. If the quadrangle is regular, the postnodals are few in number and the
second antenodals are before the arculus.) 23

23 (28). M4 and Cuinot strongly zigzag, so that the princii)al wins are long, straight
and give the wing the straight, uniform ai)pearance of the Platycnemis group.
Tibial spines relatively long 24

24 (27). In the front wing the inner end of the quadrangle is about as long as the
upper side. Mi+o forks much nearer the origin of Mia than the subnodus 25

25 (26). The upper side of the quadrangle about one-third as long as the lower side.
The superior appendages of the male longer than the inferior, and compressed,
and bent at the end. Seychelles. (Fig. 80.) Leptocneniis Kirby, 1890

{Ilemicneniis Selys, 1882)

26 (25). The upi)er side of the quadrangle about one-half as long as the lower side.
The superior appendages of the male thick, short, compressed, and conical in
form. Madagascar Paracnemis'' Martin, 1903

27 (24). The inner end of the ([uadrangle of the fore wing about one-third the length
of the upper side; the quadrangle long and narrow. Mij forking about one-
half way between the subnodus and the base of Mia. Chili. (Fig. 81.)

Antiagrion Ris. 11)04
2>^ (23). M4 and Cui zigzag; Mi+2 usually curved somewhat at or be\-ond the sub-
nodal region, instead of running straight out 29

29 (40). Tibial si)incs long, generally about twice as long as the intervening sjjaces.
In the front wing the inner end of the quadrangle generally longer than the upper
l)art of the arculus and as long as the upper side of the quadrangle. Quadrangle
widened distalK. Wing not i)etioIed to the le\el of Ac 30

•* De Selys' description of Ncsocnemis tits that of Tatocnemis almost exactly; the two
are evidently the same genus.

1^ Studied from the description only.

MEM. AM. ENT. SOC, 3.

56 A VENATIONAL STUDY OF THE ZYGOPTERA

30 (31). Tarsal claws symmetricalh- cleft, so that the superior and inferior teeth are
about equal. The ninth segment of the abdomen almost equals the eighth. The
first segment of the antenna scarcely visible; superior male appendages as long
as the inferior. Malaysia Onychargia Selys, 1865

31 (30). Tarsal claws not symmetricalh' cleft, the superior and inferior teeth not
equal 3-

32 (33). The area posterior to Cu. tends to be two cells wide and the second anteno-
dal is often slightly beyond the level of the arculus. IVh becomes zigzag at the
level of the fork of ]\Ii+2. Mexico to Colombia. (Fig. 82.)

Hyponeura Selys, 1854
;'^;i, (32). The area back of Cua of one row of cells only 34

34 (35). The prolongation of the m)dal cross-\ein through the vein M1+2 is generally
at first perpendicular to Mi+.>. Mi+o in the front wing forks six cells beyond the
subnodus; in the hind wings four cells beyond. Tibial spines three times as long
as the intervening spaces. New Guinea. (Fig. 83.) Palaiargia Foerster. 1903

35 (34). The prolongation of the nodal cross-vein through the \-ein M1+2 is not per-
pendicular to Mi+2 • 3^^

36 (37). Cui long, extending to the level of the origin of M^, especially in the hind
wing, and at least ten or twelve cells beyond the subnodus. Generally fourteen
or more postnodal cells; three to seven antenodal cells between M4 and Cui.
The New World, Moluccas, Cape of Good Hope, Kurile Islands. (Fig. 84.)

Argia Rambur, 1842

37 (36). Cu2 short, not extending more than four or five cells beyond the level of the
subnodus. About ten cells present between the nodus and the stigma 38

38 (39). An apical spine on segment eight of the female. The arculus slightly be-
yond the second antenodal. Three antenodal cells between M4 and Cui. West
Indies and Central America. (Fig. 86.) Argiallagma Selys, 1876

39 (38). No apical spine on segment eight of the female. The arculus at the second
antenodal. Two antenodal cells between M4 and Cui. Brazil. (Fig. 85.)

Diargia Calvert, 1909

40 (29). Tibial spines shorter, scarcely k)nger than the inter\ening spaces; it long,
the wings are petioled to or beyond Ac. The inner end of the quadrangle usually
not longer than the upper part of the arculus in the fore wing, nor longer than the
upper side of the quadrangle. Quadrangle generally not widened distally 41

41 (88). No pale post-ocular sjiots present on a dark background 42

42 (53); Female with apical spine on segment eight'-" 43

(.■\) Wing petioled to the vein Ac. The quadrangle in the front wing with a very i)ointed

distal lower angle, so that the upper side is shorter than or only as long as the inner
end. Tenth segment of the abdomen with a deeply cut posterior margin. Con-
sideraiiii- blue in the coloration. South .America. (Fig. 121.)

Skiallagma Foerster, 1906

'" The following two genera cannot lie fitted into the key, because the females are un-
known, iiul a brief description ma>' aid in their determination.

PHILIP A. Ml.NZ 57

(B) "Arciilus at, or vcr\- near to, the second anlenodal; M> oa the hind wings arising
nearest the fourth postnodal; costal edge of the stigma of the hing wings shorter
than the proximal or distal edges; Ciii and Cu2 not reaching as far distad as the

level of origin of Mia and M2 ; A separating from the hind

margin proximal to, or less frequently at, the cubito-anal cross-vein ,

which is situated nearer the level of the first than of the second antenodal; venation
of the tips of front and hind wings similar, abdomen chieBy red, tenth segment of
the male not elevated doVsally, vulvar spine?" Considerable orange and red in
the coloration, witli black markings. Brazil. (Fig. 96.)

Tigriagrion Calvert, 1909

43 (4(S). Wing not i)eti()led all the \va\- to .\c 44

44 (47). ( ienera of the Xew World 45

45 (46). .-\i)donien chieH\- red. The costal side of tlie stigma not shorter than the
proximal or distal sides. Ciii extending at least to or beyond the level of the base
of Mia. Cii-.. to or bexond that of M .. In the hind wing the inner end of the quad-
rangle not as long as the upjter side. North and South America. (Fig. 88.)

Amphiagrion Selys, 1876

46 (45). Abdomen generalK- with some red. The costal edge of the stigma is shorter
in the hind wing than the proximal or distal sides. \'enation otherwise much like
the preceding. Mexico and United States. (Fig. c)^.)

Hesperagrion CaKert, 1902

47 (44). Cienus of the Old World. Front margin of the stigma longer than the pos-
terior one. Quadrangle exceedingly narrow. South Ab\s.sinia.

Thermagrion-' Foerster. 1906

48 (43). W ing petioled to or beyond Ac 49

49 (50). Agenus of the Old World. Coloration largeh- black and blue, 'i'ibiaewith
long spines. Lower tooth of tarsal claw almost as large as theui)i)er. Oceanica.
(Fig. 1 16.) Xiphiagrion SeKs, 1876

50 (49) . (ienera of the Xew World 51

51 (52). M:i not arising very close to the subnodus, but at least one-half cell Ix'tore it.
Coloraticjn light, much red. Color ])attern of dorsum of thorax with no markings
darker than brown. Some blue often present. South .America and Western
North .America. (Fig. ux).) Oxyagrion Si'lxs. 1876

52 (51). Ms arising \cr\ near the subnodus, not more tiian oiie-hfth ol a cell before
it. .Abdomen chieflv reddish and xcllowish. (Fig. 127.)

Leptobasis SeKs, 1.S77
54 (42). F'emale without apical si)ine on segment eight 54

54 (5rt)- Two comi^lete rows of cells be\ ond the stigma between C and Ri. this area
much broader than the corresjionding one between R, and M,. The wing-margin
slightK- indented at tlu' stigma. Considerable red in the bod\ coloration. Fast
Indies. (Fig. 87.) Oreagrion Ris, i9i,>

55 (54). Rareh two comiilete rows of cells be\ond the stigma between C and Ri-
Wing-margin not indented at the stigma 5^'

-' Studied from (l{'scri|)lion oiil\'.

MEM. .\M. E.\T. SOC. ."v

58 A VEXATIONAL STUDY OF THE ZYGOPTERA

56 (59). Wing ver\" evidently not petioled to the level of Ac; ])etiolation ceasing as
far before it as Ac is long 57

57 (58). Colors red and dark bronze. Europe and Mediterranean Region to China.
(Fig. 89.) Pyrrhosoma Charpentier. 1840

58 (57)- Coloration without red, but is bronze \aried with blue, green, or yellow.
Europe, Asia Minor, and Chili? (Fig. 90.) . . . .Erythromma Charpentier, 1840

59 (v56) . W ing petioled very close to the level of Ac . . 60

60 (71). Genera of the Xew World 61

61 (68). Wing not petioled beyond Ac. Tarsal claw with rather strong inferior
tooth 62

62 (63). Postnodals eighteen to twenty, Cu-j extending to the level of the base of
Mia, which arises four or more cells before the stigma. Colors of the dorsum blue
and black; yellow beneath the thorax. United States and Canada. (Fig. 108.)

Chroniagrion Xeedham, 1903

63 (62). Postnodals fourteen or fewer, Cu-. not extending to the le\el of the base of
Mia. the latter arising one to two cells before the stigma 64

64 (67). Coloration not reddish. Abdomen usiialK' thirt\' or mure millimeters
long 65

65 (66). The descending cross-vein from the subnodus continuous to the wing-
margin. Rs and M3 narrowly separated at the ])roximal cross-vein between them.
Tooth on tarsal claw well developed but distinctly shorter than the claw. Male
inferior appendages well developed. South America. (Fig. 120.)

Aeolagrion-2 Williamson, 191 7

66 (65). The descending cross-vein from the subnodus not continued directly to the
wing-margin. Tooth on the tarsal claw large, almost as large as the claw; male
inferior appendages rudimentary. Rs and Ms widely separated at the proximal
cross-vein between them. South America. (Fig. 119.)

Leptagrion Selys, 1876, sens, strict.

67 (64). Coloration reddish. Tarsal claw with a small inferior tooth. Abdomen
generally less than thirty millimeters long. United States and West Indies to
Brazil. (Fig. 122.) Telebasis Selys, 1865

{Erythragriou Sel>-s, 1876)

68 (61). Wing petioled be>'ond Ac. Tarsal claw with a very small or no inferior
tooth 69

69 (70). M2 arising nearest the fifth or sixth postnodals in the hind wings; at the
sixth or beyond in the front wings. Tarsal claw toothless; superior appendages
ot male not bent down almost at a right angle in their apical half. Genital valves
of the female not reaching far beyond the level of the tips of the abdominal ap-
pendages. Central America to Brazil. (Fig. 126.)

Metaleptobasis Calvert, 1907

" This genus includes Leptagrion dor sale, hence Hylaeagrion of Foerster, 1906. Wil-
liamson (Entom. News, 191 7, p. 241) arrives at the conclusion that VocrstGv's Hylaeagrion
argentolinealum is in reality L. dorsale, and his H. croceiim Selys is Leptobasis, probably
vacillans. Therefore the generic name Hylaeagrion falls, L. dorsale becomes the type
of Aeolagrion and L. macriirum is that of Leptagrion.

PHILIP A. .MUX7 59

70 (69). M2 arising nearest the fourth postnodal in the hind wings, the fifth in the
front wings. Tarsal claws toothed. The superior appendages of the male bent
down almost at a right angle in their apical half, (ienital \al\es of the female
reach far bej'ond the level of the tips of the abdominal api)endages. West Indies
and Brazil. (Fig. 127.) Leptobasis Sehs. 1.S77

71 (60) . Genera of the Old World 72

72 (83). M3 arising very dislincth- before the subnodus 73

73 (80). Inferior tooth of tarsal claws well developed, when comi)ared with the su-
perior tooth 74

74 (75). Wing ten to tweKe mm. long; arculus distinctly bexond the second an-
tenodal, inner end of (juadrangle lacking in the male. Postnodals six to eight.
Australia. (F~ig. 106.) Hemiphlebia Sehs, 1869

75 (74). Wing usualK' twent\" or more mm. long: arculus near the le\'el of the sec-
ond antenodals; inner end of quadrangle present 76

76 (77). Coloration yellowish or reddish. India to Australia and Africa. (Fig.
117.) Ceriagrion-3 Sehs. 1876

77 (76). Coloration not light 78

78(79). Largeh" black. Tibial si)ines long. Fiji Islands and Xew Cuinea.

Nesobasis-^ Selys. iSgi

79 (78). Bronze and dark. Tibial spines \er>' short. Aru and Bi\ak Islands.
(Fig. 118.) Papuagrion Ris. 1913

80 (73). Inferior tooth of the tarsal claw lacking or very small 81

81 (82). Largely black. Tibial spines long. Malaysia and Pohnesia.

Nesobasis-^ Selys, 1891

82 (81). Coloration bronze, \aried with green and black, wings broad. Black
bands on head, thorax, and abdomen. Africa Argiagrion-^ Seh's. 1876

83 (72). M3 arising almost or cjuite at the subnodus 84

84 (87). Rs arising at the subnodus 85

85 (86). Stigma square or oblong, followed by a single row of cells. Coloration not
ver>- metallic. Malaysia. (Fig. 128.) Teinobasis Kirby, 1890

(Telebasis Seh's, 1877)

86 (85). Stigma five-sided and followed 1)\- two rows of cells. Malaysia and Poly-
nesia Pericnemis'-^ Selys. 1863

87 (84). Rs arising beyond the subnodus; M 3 also slighth' l)e\ond or at the sub-
nodus. Wing petioled to or be\ond the level of the arculus. Mala\sia. (Fig.
129.) Amphicnemis Sehs. 1863

88 (41). Pale postocular spots present on a dark background, although sometimes
confluent with the light coloring of the rear of the head; apparentK' so in those
genera with dichromatic females 89

-^ Ceriagrion is here used to include Brachybasis; in the opinion of Dr. Ris the latter
cannot stand as a genus. See \). 317 of his "Odonatanach den Sammkingen L.
Schultzes unci Katalog dcr Odonalen von Sucdafrika," 1908. Denkschr Meii.-natwiss.
C.es. Jena, 13: 303-346.

-■' Studied from description only.

MEM. .\M. ENT. SOC, 3.

6o A VENATIONAL STUDY OF THE ZYGOPTERA

89 (94). Stigma in fore and hind wings of males dissimilar in color, size or both
Tenth segment of abdomen of male generally with a bifid dorsal process. Fe-
male often with apical spine to segment eight of abdomen. Postocular spots, in
female at least, tend to disappear. Females dichromatic or polychromatic. . . .90

90 (91). In the male the stigma of the front wing does not touch the costa. Fe-
male with no external black stripe on second or third tibiae. Coloration black,
orange, and yellowish. Postocular spots, in females at least, tend to disappear.
United States to Venezuela and West Indies. (Fig. 97.)

Anomalagrion Sel\s, iS^j

91 (90). Stigma in male touching C. Female generally with an external black
stripe on second and third tibiae 92

92 (93). Penis with pair of erect spines on the penultimate segment. In American
species at least, the anterior surface of thorax not solid black. Cosmopolitan.
(Fig. 100.) Ischnura Charpentier, 1840. {Micron ympha Kirby, 1890)

93 (92). Penis with pair of erect spines on penultimate segment. Anterior surface
of thorax solid black, except in green females, middle prothoracic lobe with small
tooth on each side. Arizona to Baja California. (Fig. loi.)

Celaenura Kenned\-, 191 7

94 (89). Stigma in fore and hind wings of male similar 95

95 (102). The second antenodals distinctly before the level of the arculus. Both
wings not petioled to Ac, but the vein from Ac back to the wing-margin, that is
A', does not continue the direction of the petioled part of the margin, instead it
passes up toward Ac and Ab 96

96(97)- Female with apical spine on segment eight. Postnodals six to seven.
Coloration dark bronze, with yellow and blue. Dichromatic females. A genus
of the New World — tropical America and the West Indies. (Fig. 102.)

Ceratura Sehs, 1876

97 (96). Female without apical spine. Old World genera 98

98 (loi). Postnodals five to nine; females dichromatic; very small insects, usually
less than 24 mm. long 99

99 (100). Quadrangle of fore wing with upper side less than half as long as the lower
side; hence distal angle very acute. Legs very long in comi>arison with rest of
body. Australia. (P^ig. 103.) Austrocneniis Tilhard, 1913

100 (99). Quadrangle of fore wing with upper side one-half or more than one-half
of length of lower. Legs short. Australia to India. (Fig. 104.)

.\griocneniis Sehs, 1809

loi (98). Postnodals ten to twelve; females not dichromatic; chiefl\' insects of more

moderate size i.e. abdomen of twenty-five or more nun. India to Australia,

and in Zanzibar and Madagascar. (Fig. 105.) . . . . Argiocnemis Selys, 1877

102 (95). The second antenodals not distinctly before the arculus, or if so, either one
wing iH'tioK'd ti) .\c or the win A' conlinncs direction of ilic hind margin of wing
\\\) to point of petiolation 103

103 (i iC)). Both wings distincth' not jietioled to the le\el of .\c 104

104 (in). I'emale without ajiical spine to segment eight 105

I

PHILIP A. MUXZ 6 1

105 (106). Stature very slender, abdomen especially so. Coloration metallic and
brilliant, similar in both sexes. Europe, China, Xorth and South America.
(Fig. 91.) Nehalennia Selys. i<S5()

106 (105). Stature more robust, abdomen not extremeK' slender; coloration not
ver>- brilliant and dissimilar in sexes 107

107 (108). Postocular spots elongated, almost linear, with other transverse lines be-
tween them. Two small clear spots between the ocelli. Superior appendages of
male longer than last abdominal segment and arcuate. Europe, Asia Minor.
Japan, South Africa. (Fig. 92.) Cercion" Xa\as, ii;o7

108 (107). Postocular spots cuneate or rounded, no spots between the ocelli. Su-
perior appendages often short and straight 109

109 (no). Species of Hawaiian Islands. Characters variable. (Fig. 9,"^.)

Megalagrion-*^ McLachlan, 1S83

1 10 (109). Xot Hawaiian, nor Australian, otherwise quite cosmoi)olitan. (P'ig. 94.)

Coenagrion Kirb>-, 1890
{Agrion Fabricius, 1775)

1 1 1 (104). Female with apical spine on segment eight 112

112 (113). Costal edge of stigma shorter than the proximal or distal edges. Cui
reaching the level of base of iVlia, Cuo that of origin of Mo. Postocular spots
sometimes disappearing. United States and Mexico. (Fig. t)^.)

Hesperagrion Calvert, 1902

113 (112). Costal edge of stigma as long as proximal or distal edges 114

114 (115)- Mj of hind wing generally arising at or beyond the fourth i)ostnodal,
tenth segment of male not elevated. Coloration of male generalh- blue, of fe-
male green or yellow, sometimes blue. Cosmopolitan, not in Australia. (Fig.
98.) Enallagma Charpentier, 1840

115 (114). Tenth segment of male elevated and with a bifid dorsal process. Penis
with transverse row of short spines at base of distal or spatulate lobe. M-. of
hind wing arises at about the fifth postnodal. California. (Fig. 99.)

Zoniagrion Kennedy. 1917

116 (103). One or both wings petioled jiractically to or even be\ond the le\el of
- Ac 117

1 17 (108). Pale postocular spots widely confluent with the i)ale color on the rear of
the head. M2 arising nearest the fourth jjostnodal in the hind wings and the
filth in the front wings. Tarsal claws toothed. Superior appendages of the male
bent down almost at a right angle in their apical half, (ienital valves of female
reaching far beyond the level of the tips of the abdominal api)endages. West

Indies to Brazil. (Fig. 127.) Leptobasis-" SeKs. 1877

-^ Cercion has been shown l)y Kennedy, 1917, Knt. News, to include C. lindenii, C.
([Kadrigerum, and C. piinctum, all formerly considered species of Coenagrion.

-^ Megalagrion was said by Perkins (Fauna Hawaiiensis \'ol. I, p. clxxv and \"ol.
II, pp. 63 to 77 and 693 to 696) to have been separated by McLachlan from the genus
Coenagrion on inconstant characters. Kennedy (Ent. News, 191 7), in a study of the penes
of many of the Hawaiian species, considers it probable that all the species of those islands
may be considered as comprising the genus Megalagrion.

" Pale postocular spots not always present. Females sometimes with, someiiincs
without, vulvar spine. See (52) and (70).

MEM. .\M. ENT. SOC, 3.

62 A VENATIONAL STUDY OF T«E ZYGOPTERA

Il8 (117). Pale postocLilar spots not widely confluent with the pale color on the

rear of the head 1 19

IK) (124). Female with apical spine to segment eight^^ 120

120 (121). Genus of the New World. Colors bronze and blue or greenish. The
eighth or ninth or tenth segments blue. End of tenth more or less prolonged or
turned back. Mexico to South x'\merica. (Fig. no.)

Acanthagrion Selys, 1876

121 (120). Genera of the Old World 122

122 (123). Ms arising before the subnodus; front wing with twelve to fifteen post-
nodals. Coloration yellowish, greenish and bluish with black markings. Kili-
manjaro and Meru, Africa Mombagrion^s Sjostedt, 1909

123 (122). Ms arising almost at the subnodus; front wing with nine to twelve post-
nodals. Coloration black, blue, and dark green. Malaysia and West Africa.
(Fig. 123.) Aciagrion Selys, 1892

124 (119)- Female without apical spine on segment eight 125

125 (128). Genera of the New World ' 126

126 (127). Tips of the wings similar, abdomen very long and slender. Tenth seg-
ment of male not forked. Coloration greenish or yellowish. Mia arising one or
two cells before the stigma. Eastern United States and Brazil. (Fig. in.)

Telagrion Selys, 1876

127 (126). Tips of front and hind wings of male unlike; Ri and Mi in the hind wing
almost unite at the tip. Segment ten of the male with a tubercle above and two
spines. Coloration bluish or reddish, or brownish. Mexico to Central America.
(Fig. 125.) Anisagrion Selys, 1876

128 (125). Genera of the Old World 129

129 (138). M3 arising cjuite distinctly before the subnodus. Tarsal claw generally
with an inferior tooth 130

130 (133) . Prothorax of female provided with two hooks which project forward . .131

131 (132). Ac at or but slightly beyond the level of the first antenodals, not more
than one-fourth the distance to the second ones. Coloration blue. Australia.
(Fig. 107.) Caliagrion Tillyard, 191 2

132 (131). In Australian species, at least, Ac is about one-half way between the first
and second antenodals. Superior appendages of male usually forked or hollowed
at tips. Africa and Asia to Australia. (Fig. 114.) Pseudagrion Selys, 1876

133 (130)- Prothorax of female simple, no hooks present 134

-* Skiallagma sometimes has traces of postocular spots; female unknown. See descrip-
tion in footnote to rubric forty-two.

Myagrion, female unknown. Male described as follows: Color carmine-red. Quad-
rangle irregular, upper side contained in the lower two and one-half times, in the front
wing. Cu2 thirteen cells long. M2 arises six cells beyond the subnodus in the front
wing, five in the hind wing. Abdomen very slender, legs short, tibial spines weak.
Tergum of tenth abdominal segment produced into two spines, but without a tubercle
such as occurs in Anisagrion. Appendages scarcely as long as these spines. South
America, Foerster, 1914.

^^ Studied from description only.

PHILIP A. MUXZ 63

134 (135). Ac at the le\el of the first antenodals. Postiiodals ele\"en to fifteen.
Coloration red. AustraHa and Xew Zealand. (Fig. 1 15.)

Xanthagrion Sehs, 1S76

135 (134)- Ac about half-way between the first and second antenodals 136

136 (137)- Front wing not petioled quite to Ac, the hind wing so petioled. Inferior
appendages of male much longer than superiors. Xew Zealand. (Fig. 112.)

Xanthocnemis Tillyard, 1913

137 (136). Both wings petioled to Ac. Superior apjjendages of male longer than in-
ferior; small species of blue and bronze color. (Fig. 113.)

Austroagrion Tillyard, 1913

138 (129). Ms arising \er\- near to subnodus, almost touching base of Rs. Tarsal
claw without an inferior tooth. Body slender. Malaysia to Australia. (Fig.
124.) Archibasis Kirby, 1890

{Stetwbasis Selys, 1877)

Key to the Genera of the Subfamily Protoneiirinae

1 (4). Cuo one cell long 2

2 (3). Usually more than 14 postnodal cells. Ac near the level of the first antenodals.
Mia arising six to eight cells before the stigma. Cui is six to se\'en cells long.
West Africa. (Fig. 130.) Chlorocnemis Selys. 1863

3 (2). Less than 10 postnodal cells. Ac about midwa\- i)etween the antenodals.
Mia arising one cell before the stigma. Cui is lour cells long. Brazil. (Fig. 131.)

Proneura Selys, 1889

4 (i). Cu2 reduced to a cross-\ein or lacking, not projecting be\()nd MA 5

5 (18). A not reduced to Ac, but extending be>ond 6

6 (13). A ending in MA, and not in Cu on the posterior side of the quadrangle, nor
in the hind margin of the wing before MA. Xo "cubito-anal" cross-veins before
Ac. That edge of the stigma along the margin of the wing as great as that along
Ri. The quadrangle practically uniform in width - 7

7 (8). Labium with short lobes. The wings often not petioled to Ac. Cui ending
at or be\ond the level of the subnodus. India, Africa, Malaysia. (Fig. 132.)

Disparoneura Selys, i860 sens, strict.

8 (7). Labium with elongate lobes 9

9 (10). Cui ending three to five cells beyond the level of the subnodus. Wing
petioled almost to Ac. Front wings with three cells between the subnodus and
fork of Mi+2. Brazil. (Fig. 134.) Peristicta Selys, i860

10 (9). Cui ending within a cell of the level of the subnodus. Wing petioled to or
slightly beyond Ac. Often four or more cells between the subnodus and fork of
Ml+2 II

11 (12). Ac not nearer the level of the second antenodal than of the first. The first
antenodal space generally more than one and one-half times as long as the second.
Five or more cells in the hind wing beyond the stigma. Tibial spines short.
West Indies. Central America to Brazil. (Fig. 133.) Neoneura Selys, i860

{Caenonenra Kirby, 1890)

MEM. AM. EST. SOC, 3.

64 A VENATIONAL STUDY OF THE ZYGOPTERA

12 (ii). Ac nearer the level of the second antenodal than of the first. The first
antenodal space less than one and one-half times the length of the second. Three
or four cells beyond the stigma. Brazil. (Fig. 135.). . .Idioneura Selys, i860

13 (6). A ending in Cu under the quadrangle or in the wing margin, but not reach-
ing MA ; ^ +

14 (17). A ending in Cu on the posterior side of the quadrangle. A "cubito-anal"
cross-vein generally present before Ac. That edge of the stigma along the wing
margin shorter than that along Ri . . 1 5

15 (16). Cui ending generally some distance beyond the level of the fork of INI 1+2,
about two-thirds of the distance from the nodus to the stigma. Mi+o forking
between one-third and one-half the distance from the subnodus to the stigma.
Mexico to Venezuela. (Fig. 136.) Palaemnema Selys, i860

16 (15). Cuiending not more than a cell or so beyond the fork of M1+2, usually before
it, Mi+2 often forking less than one-third the distance from the subnodus to the
stigma. Ceylon to Malaysia. (Fig. 137.) Platysticta Selys, i860

17 (14). A ending in the hind margin of the wing between Ac and MA. No cubito-
anal cross-vein before Ac. The costal edge of the stigma as long as that edge
along Ri. Southeastern Asia and Malaysia. (Fig. 138.)

Hypostrophoneura^" gen. nov.

18 (5). The vein A reduced to Ac, not extending beyond 19

19 (43). Ac distinctly beyond the level of the first antenodal; Cui usually extending
beyond the level of the subnodus -'-^

20 (33). The subquadrangle shorter than the portion of the wing behind Cu-f A and
before Ac. Genera of the New World and two of the East Indies and Southern
Asia 21

21 (24). Genera of the Old World. Labium with short lobes. About sixteen post-
nodals --

22 (23). Accessory cross- vein before Ac. The costal edge of the stigma shorter than
that along Ri. India to Malaysia. (Fig. 144.) Protosticta Selys. 1885

23 (22). No accessory cross-\ein before Ac. The costal edge of the stigma not
shorter than that along Ri. Malaysia. [ Disparoneura Selys, i860, sens. lat.

\ Caconeiira^^ Kirby, 1890 sens, strict.
■ [ AUoneura, Selys, i860 sens, strict.

24 (21). Genera of the New World. Labium with elongated lobes. About twelve
postnodals 25

3" This genus is created for those species of Disparoneura placed by Selys (1886)
in the third group, and has as its type gomphoides Rambur. It is distinct from Dis-
paroneura, sens, strict., by having A end before MA and not in MA, and by Cui usually
being much shorter than in Disparoneura. It receives its name from the fact that A
turns downward: ItroaT po4>-n = a turning backward, and vevpov= nerve.

3» As here treated this genus comprises only the first group of AUoneura as given by
Selys (1886). It, therefore, contains the following species: C. dorsalis Selys, C. lansbergei
Selys, C. graciUima Selys, C. hyperylhra Selys, and C. dohrni Krueger. See rubric 49
for Risioneura. Dr. Ris and Mr. Laidlaw consider this section of Caconeura as belong-
ing to Disparoneura; the genus Caconeura thus passes out of existence, since C. dorsalis
is the genotype. See Ris (1915) and Laidlaw (1913)-

PHILIP A. Ml XZ 65

25 (30). M:iand R9 arising separately and the latter arising at the suhnodus 26

26 (27). Legs quite long; femora strongly compressed. Tibial sj)ines long, numer-
ous. \odus not very oblique, so that Sc ends beyond the level of the base of M3.
Cuba. (Fig. 145.) Microneura Selys, 1886

27 (26). Legs short; femora not compressed. Nodus oblique, so that Sc ends before
the level of the base of M3 28

28 (29). Wings narrow, about one-sixth or one-seventh as wide as long. The first
antenodal space twice as long as the second and also longer than the tliird. Ac
otten v'er>' near the level of the first antenodals. Cuba and Central America to
Brazil. (Fig. 146.) Protoneura Selys, 1857

29 (28). Wings broader, about one-fifth as wide as long. The first antenodal space
twice as long as the second, and equal to the third. .\c usually at a level inter-
mediate between the antenodals. British C.uiana. (I'ig. 147.)

Epipleoneura Williamson, 1915

30 {2~,). M3 and Rs joined at their origin into a single vein arising beyond the sub-
nodus; or they may arise as separate veins, but if so, Rs still arises beyond the
subnodus 31

31 (32). Cui about two and one-half cells long, so that it passes through the cross-
vein coming down from the subnodus. The first and third antenodal costal
spaces about equal; the second much shorter, more than one-half of the third.
Stigma long and narrow, twice as long as wide. British Guiana. (Fig. 148.)

Epipotoneura Williamson, 191 5

32 (31). Cui ending at the cross-vein descending from the subnodus. The three
antenodal costal spaces about equal. Wings rather broad, almost one-fourth as
wide as long. Stigma shorter 33

33 (34). iM'in the front wing arises at the se\enth postnodal ; in the hind wing at the
fifth. British Guiana Phasmoneura'^ Williamson, 1916

34 (33)- ^^i iri the front wing arises before the seventh postnodal, usually at or be-
fore the sixth; in the hind wing at the fourth or before. British Guiana. (Fig.
149.) Psaironeura Williamson, 191 5

35 (20). The subquadrangle ot about the same length as the space before Ac.
Australian genera 36

3^ (39)- Cui more than one cell long and traversing the cross-vein coming down
from the subnodus. The space between the fork of Mi^o and the base of
Alia is not o\er two-thirds as great as that between the fork of Mr 2 and the
subnodus 37

37 (38). Mia arising about five cells before the stigma in the hind wing. The quad-
rangle almost as long as the second antenodal space. Cui extending six to seven
cells beyond the quadrangle. New South Wales. (Fig. 139.)

Neosticta Tilhard, 191 3

38 (37)- Mia arising about three cells before the stigma in the hind wing. The quad-
rangle shorter than the second antenodal space. Cui extending three to four
cells beyond the quadrangle. Australia. (Fig. 140.)

Austrosticta Tillyard. 1908
'^ Description only was available for this genus.
MEM. AM. ENT. SOC, 3.

66 A VEXATIOXAL STUDY OF THE ZYGOPTERA

39 (36). Cui not more than one cell long and not reaching the cross-\-ein which
comes down from the subnodus (this true in the hind hind wing at least). The
space between the base of INIu and I\Ii+2 at least four-fifths of that between the
fork of Mi+2and the subnodus, in the hind wing 40

40 (41). Cui less than one cell long. Ms arising before the subnodus. The stigma
longer than wide. The second antenodal space about one-half as long as the third.
The space between the forking of AI1+2 and the base of Mia is longer than that be-
tween the subnodus and the forking of Mi+2, in the hind wing. Australia.
(Fig. 141.) Notoneura Tilhard, 1913

41 (40). Cui about one cell long. Ms arising beyond the subnodus. The stigma
about as wide distally as long. The second antenodal space as long as the third ;
the space between the base of Mia and the forking of M1+2 not as long as that be-
tween the subnodus and the forking of M1+2 in the hind wing. Australia. (Fig.
142.) Oristicta Tillyard, 1913

42 (19). Ac at or before the level of the first antenodals; Cui usually not extending
to the level of the subnodus 43

43 (50)- Cui generally not reduced to a cross-vein, but extending beyond MA. Cu2
may be lacking. Ri on lower side of stigma not zigzag, but straight 44

44 (47) . Quadrangle widened distallj-. Area back of M4 beyond MA broad, at least
one-third as great as the width of the wing at that point 45

45 (46). Head very small, thorax narrow, legs short, abdomen exceedingly attenuate.
About nine postnodals. Australia. (Fig. 141.) Notoneura Tillyard, 1913

46 (45). Head larger, thorax not so narrow, but form robust, tibiae with many spines.
About fourteen postnodals. Australia. (Fig. 151.) . .Nososticta Selys, 1S60

47 (44). Quadrangle not widened distally. x^rea beyond MA not so broad as
above 4^

48 (49). Ms in some species arising just beyond subnodus. Cuiin species from East
Indies is at least two cells long, in Australian species less. In Australian species
at least, "both superior and inferior appendages of males somewhat forcipate, the
inferior pair prolonged beyond the superior." New Caledonia, Bismarck Arch-
ipelago and Torres Straits. (Fig. 150.) Isosticta Selys. 1885

49 (48). Ms always arising before the subnodus. Cui in the front wing almost
always less than two cells long, generally ending in the first cell beyond MA.
Malaysia to Bismarck Archipelago. (Fig. 152.) Risioneura^s gen. nov.

50 (43). Cui reduced to a cross-vein, not extending beyond MA. Ri on lower side
of stigma zigzag. Ms arising beyond the subnodus. New'Guinea. (Fig. 153.)

Selysioneura Foerster, 1900

33 This genus is made for those species of Caconeura (Alloneura) included by Selys
(1886) in his second group and differs from Caconeura, sens. .strict., in having Ac situated
under or before the first antenodals and Cui rarely long enough to reach the cross-vein
coming down from the subnodus. R. plagiata Selys is the genotype; the other species
included are R. insignis Selys, R. erythroprocta Selys, R. wallacii Selys, R. exul Selys,
R. moluccensis Selys, R. flavipennis Selys, R. salomonis Selys, R. circumscripta Selys. R.

PHILIP A. .MUXZ 67

VI_A PARTIAL LIST OF THE MOST IMPORTAXT REF-
ERENCES COXSULTED IX THE PREPARATIOX
OF THIS PAPER

Calvert. P. P. 1891. A new genus and species of Odonata from Jamaica.
Entom. News, 2: 199.

1893 (a). Catalogue of the Odonata (dragonflies) of the vicinity of Phila-
delphia, with an introduction to the study of this group of insects.
Trans. Amer. Entom. Soc, 20: 152-172. pis. 2-3.

1893 (b). The Odonate genus Ortholestes.
Proc. Acad. Xat. Sci. Philadelphia, 45: 377-3S2, fig.

1898. Odonata. A synonym and the bearing of its discoxen,- on the
classification ol Agrion.
Entom. Xews, 9: 72.

1 901-1908. Odonata. (in)
Biologia Centrali-Americana, Xeuroptera: 17-420. 10 i:»l5.
London, R. H. Porter and Dulau and Co.

1902. On the systematic position of Thaumatoneiira inopinata McLachlan
(Order Odonata), with some remarks on the classification of the suborder
Zygoptera.
Entom. Mo. Mag., 38: 29-32.

1904. Thaumatoneiira again.
Entom. Xews, 15:216.

1909. Contributions to a knowledge of the Odonata of the Xeotropical
region, exclusive of Mexico and Central America.
Ann. Carnegie Mus., 6: 73-280. pis. 1-9.

1913. The fossil Odonate Phenacolesles. with a discussion of the venation
of the legion Podagrion Selys.

Proc. Acad. Xat. Sci. Philadelphia, 65: 225-272. pi. 14.
Calvert, P. P. 191 7. Studies in Costa Rican Odonata, 8. A new genus allied to
Cora.

Entom. Xews, 2S: 259-263, figs.
Campion. H. 1913. The antenodal reticulation of the wing? of Agrioninc dragon-
flies.

Proc. Acad. Xat. Sci. Philadelphia, 65: 220-224.
Comstock, J. H. and Xeedham. J. G. 1898. The wings of insects. Chaj). 4. III.
The \enation of the wings of the Odonata.
Amer. Xatural.. 32:903-<;ii.

5^/3-5/ Foerster, R. finisterrae Foerster, R. eburnea Foerster, R. jriihstorferi Krueger, R.
pseiidexul Ris, R. nigrijrons Ris, R. marina Ris, R. rosea Ris, R. dorsonigra Martin, and R.
torrenticola Foerster. It is dedicated to Dr. Ris, who has given us the most careful re-
cent treatment of this group. See F. Ris, 1913, Odonata von den Aru-und Kei-Inseln,
pages 508 to 514.

MEM. AM. ENT. SOC, 3.

68 A VEXATIOXAL STUDY OF THE ZYGOPTERA

Foerster, F. 1897 (a) Beitrage zur kenntniss der indo-australischen Odonaten-
fauna, IV.
Wien Entom. Zeitung, 16: 1 01-103.

1897 (b) Beitrage zur kenntniss der indo-australischen Odonatenfauna,
VI.
Entom. Nachrichten, 23:333-338.

1899. Zur Odonatenfauna von Madagascar.
Entom. Nachrichten, 25: 1 86-191.

1900. Odonaten aus Neu-Guinea.
Termes. Fuzetek, 23:81-108, figs.

1903. Odonaten aus Neu-Guinea III.
Ann. Mus. Hungar., I: 509-554.

1906(a). Neotropische Libellen.
Insektenborse, 23: 15-16.

1906 (b). Libellen. (Forschungsreise durch Siidschoa, Galla und die
Somalilander v. C. Frhr. \'. Erlanger.)
Jahrb. Ver. Naturk. Wiesbaden, 59: 299-344, i pi.

1909. Beitrage zu den Gattungen und Arten der Libellen.
Jahrb. Ver. Natk. Wiesbaden, 62: 211-235.

1 910. Beitrage zu den Gattungen und Arten der Libellen.
Entom. Zeitung, Wien, 29: 51-56.

Hagen, H. 1861. Synopsis of the Neuroptera of North America.

Smithsonian Mis. Coll., 345 pp.
Handlirsch, A. 1906-1908. Die fossilen Insecten und die Phylogenie der rezenten

Formen. Ein Handbuch fuer Palaeontologen und Zoologen. 2 vols., 51 pis.

Leipzig, Verlag von Wilhelm Engelmann.
Karsch, F. 1891 (a). Neue Odonaten von Ecuador.

Societas Entom., 6: 105-121.

1891 (b). tjbersicht der von Herrn Dr. Paul Preuss in DeutschAVest-

Afrika 1890 gesammelten Odonaten.

Entom. Nachrichten, 17:65-80.

1895. Archineiira basilactea W. F. Kirhy = Echo incarnata Karsch.

Entom. Nachrichten, 20: 84.

1899. Odonaten von Johann-Albrecht-Hohe (Nord-Kamerun), gesammelt

von Herrn Leopold Conrad t.

Entom. Nachrichten, 25, 161-176.
Kennedy, C. H. 191 7. Notes on the life-histor\' and ecology of the dragonflies

(Odonata) of central California and Nevada.

Proc. U. S. Nat. Mus., 52:483-635, 404 figs.
Kirby, W. F. 1889. Descriptions of new genera and species of Odonata in the

collection of the British Museum, chieHy from Africa.

Proc. Zool. Soc. London, 299-303.

1890. A synonymic catalogue of Neuroptera Odonata, or dragonflies, with

an appendix of fossil species.

Ciurne\- and Jackson, London.

PHILIP A. MUXZ 69

Kirby, W. F. 1894 (a). Catalogue of the described Neuroptera Odonata (dragon-
flies) of Ceylon, with descriptions of new species.
Jour. Linn. Soc, 24: 545-566, pis. 41-42.

1894 (b). On a new genus and species of Agrionidae from Foo Chow.
Ann. Mag. Nat. Hist., (6) 13: 84-86.

1894 (c). Note on Ardmieura hasiladea Kirby.
Ann. Mag. Nat. Hist., (6) 13:450-451.

1900. On a small collection of Odonata (dragonfiies) from Hainan collected
by the late John Whitehead.
Ann. Mag. Niit. Hist., (7) 5: 530-539. p'- 12.
Laidlaw, F. F. 1902. On a collection of dragonfiies made b>- members of the Skeat
expedition in the Mala\- peninsula in 1 899-1 900.
Proc. Zool. Soc. London: 63-92, 3''^i-389. pls- 5-6-

1913. Contributions to a study of the dragonfiy fauna of Bornt-o. Part L
Proc. Zool. Soc. London: 63-80, pi. 4.

1915. Notes on Oriental dragonfiies in the Indian Museum, No. 3. Indian
species of the "legion" Protoneura.

Rec. Ind. Mus. Calcutta, ii:3S7-3()i.

1916. Notes on Oriental dragonfiies in the Indian Museum, No. 4. The
genus Pseudagrion.

Rec. Ind. Mus. Calcutta. 12: 21-25.
Martin, R. 1903. Odonates indo-oceaniens des collections du museum.

Bull. Mus. Paris, 8: 506-512.
McLachlan, R. 1870. Description of a new genus and four new species of Calop-

terygidae, and of a new genus and species of Gomphidae.

Trans. Ent. Soc. London, 165-170.

1895. Some new species of Odonata of the legion Lestes, with notes.
Ann. Mag. Nat. Hist., (6) 16: 19-28.

1896. On Odonata from the province of Szechuen, in western China, and
from Moupin in Eastern Thibet.

Ann. Mag. Nat. Hist., (6) 17:364-374.

1897. Thaumatoneiim inopiiiuta. a new genus and species of Calopteryginae.
Entom. Mo. Mag., (2) 8: 130-131.

Morton, K. J. 1914. A remarkable new genus and species of Odonata, of the

legion Podagrion Sehs, from North Queensland.

Trans. Flntom. Soc. London: 169-172, pi. 9.
Muttkowski, R. .\. 1910 (a). The applicability of certain generic names of

Odonata. i. Agrion versus Calopteryx.

Bull. Wis. Nat. Soc, Milwaukee, 8: 158-160.

1910 (b). Catalogue of the Odonata of North America.
■ Bull. Public Museum City of Milwaukee, i: 1-207.
Na\as, L. 1907. Neuropteros de Espana y Portugal.

Broteria, S. F"iel. 6: 43-100.

MEM. AM. F.NT. SOC, 3.

70 A VENATIOXAL STUDY OF THE ZYGOPTERA

Needham, J. G. 1903 (a). A genealogic study of dragonfly wing venation.
Proc. U. S. Nat. Mus., 26: 703-764, pis. 31-54-

1903 (b). A new genus and species of dragonfly from Brazil.
Proc. Ent. Soc. Washington, 16: 55-58.

1903 (c). Aquatic insects in New York State. Part 3, Life histories of
Odonata, suborder Zygoptera.
Bull. New York Mus., 68: 218-279, pis. 11-19.

1917. Notes on some recent studies of dragonfly wing tracheation. (Odon.)
Entom. News, 28: 169-173.
Perkins, R. C. L. 1899. Neuroptera. (in)

Fauna Hawaiiensis, 2: 31-89, pis. 3-5.
Rambur, J. P. 1842. Histoire naturelle des insectes Neuropteres.

Paris.
Ris. F. 1898. Neue Libellen vom Bismarck-Archipel.
Entom. Nachrichten, 24:321-327.

1900. Libellen vom Bismarck-Archipel gesammelt durch Prof. Friedr. Dahl.
Arch. Naturg., 66: 175-204, pis. 9-10.

1904. Odonaten.
Hamburg. Magalhaen. Reise, 7: 44 pp.

191 1. Libellen von Sintang, Borneo gesammelt von Dr. L. Martin.
Ann. Soc. Entom. Belgique, 55: 231-255.

1912. Neue Libellen von Formosa, Siidchina, Tonking, und den Philip-
pinen.

Suppl. Entom., Berlin, 1:44-85, 3 pis.

1913 (a). Die Odonaten von Dr. H. A. Lorentz's Expedition nach SUd-
west New-Guinea 1909 und einige Odonaten von Waigeu. Nova Guinea.
Resultats de I'expedition scientifique neerlandaise a la Nou^■elle Guinee, 9
Leiden: 471-512.

1913 (b). Odonata von den Aru- und Kei- Liseln gesammelt durch Dr. H.
Merton 1908, nebst Ubersicht liber die von den Aru-Inseln bekannten Odon-
aten.
Abh. Senckenb. Ges. Frankfurt a. Main, 34: 501-536, i pi.

1915. Libellen (Odonata) von Neu-Caledonien und den Loyalty-Inseln.
F. Sarasin und J. Roux, Nova Caledonia, Zoologie. 2: 57-72, 11 fig.

1916 (a), tjber Ontogenese der Flugeladerung bei den Libellen.
Mitth. Schweiz. Entom. Ges., 28: jip.

1916 (b). Zwei Notizen liber Calopterygiden (Odonata) \om Malaiischen
Archipel.

Entom. Mitth., Berlin, 5: 303-318, 9 figs.
Schmidt, E. 1915. Vergleichende Morphologic des zweiten und dritten Abdom-
inalsegments bei mannlichen Libellen.
Zool. Jahrb., 39: 87-196, pis. 9-1 1, figs. A-Z.

PHILIP A. MUXZ 71

Selys-Longchamps, E. de. 1853. S\nopsis des Caloptengines.
Bull. Acad. Belg., (2) i : 73 pp.

1854.. Monographie des Caloptengines avec collaboration H. Hagen.
Mem. See. R. Science de Liege, 9.

1857. Neuropteres de I'isle de Cuba.
De la Sagra Hist. Cuba, 7: 183-201.

1859. Additions au synopsis des Calopterygines.
Bull. Acad. Belg., (2) 7:437-451.

i860. Synopsis des Agrionines. Premiere legion: Pseudostigma.
Bull. Acad. Belg., (2) 10: no. 6.

i860. Derniere legion : Protoneura.
L. c. (2) 10: 431-462.

1862. Seconde legion: Lestes.
L. c. (2) 13: 288-338.

1862. Troisieme legion: Podagrion.
L. c, (2) 13 :,5-44.

1863. Quatrieme legion: Platycnemis.
L. c, (2) 16: 147-176.

1865. Cinciuieme legion: Agrion.
Genre Argia, 1. c. (2) 20: 375-414.

1S75. Genre Agrion, 1. c, (2) 41 : 247-322. 496-539, 1 233-1309.
(2) 42:490-531,953-999.

1877. Genres Telebasis, Argiociietnis, et Hemiphlebia.
L. c, (2) 43:97-159-

1882. Odonates des Phillippines.

Anal, de la Soc. Esp. de Hist. Xat., 11 : 1-5, 19-31-

1883. Les Odonates du Japon.

Ann. Soc. Entom. Belg. 27: 82-87, 128-138.

1886. Revision du synopsis des Agrionines.
Mem. Cour. Acad. Belg., 2^: 1-233.

1889. Palaeophlebia.
Ann. Soc. Entom. Belg., 33.

1 891. Causeries odonatologiques, no. 5.
Conipt. Rendu. Soc. Entom. Belgique, 35: cccxcxiii-cccciii.
Sjostedt, Y. 1909. Pseudoneuroptera. 1. Odonata.

Sjostedt's Kilimandjaro-Mcru Expedition, 14: 1-52, pis. 1-2. Stockholm.
Tillyard, R. J. 191 1. On the genus Diphlebia with descriptions of new species
and life-histories.
Proc. Linn. Soc. New So. Wales, 36: 584-604, pis. 19-20.

\()12. On seme new and rare Australian Agrionidae. (Odonata).
L. c, 37:404-479.

1913. Some descriptions of new forms of Australian Odonata.
L. c, 38: 229-241, pi. 15.

1914. On some problems concerning the develoi^ment of the wing \ena-
tion of the Odonata.

L. c, 39: 163-216, pis. 11-13.

MEM. .\y\. ENT. SOC, 3.

72 A VEXATIOXAL STUDY OF THE ZYGOPTERA

TilKard, R. J. 1915. On the dexelopment of the wing venation in Zygopterous
dragonflies, with special reference to the Calopter\gidae.
L. c, 40: 212-230, tigs.

1916. Further researches upon the problems of the radial and zygopterid
sectors in the wings of Odonata, and upon the formation of bridges.
L. c, 41:871-887, 6 figs.
Williamson, E. B. 1901. The dragonflies of Indiana.
Report Dept. Geol. Indiana, 24: 229-333.

1904(a). The dragonflies (Odonata) of Burma and Siam. I. Subfamily
Calopteryginae.
Proc. U. S. Nat. Mus., 28: 165-187, figs.

1904 (b). A new species of Psolodesnins (order Odonata) from Formosa.
Entom. News, 15:247-250, i pi.

1913. The medio-anal link in the Agrioninae.
L. c, 24: 258-261.

1915. Notes on neotropical dragonflies or Odonata.
Proc. U. S. Nat. Mus., 48: 601-638, pis. 36-40.

1916. A new dragonfly genus of the legion Protoneura, (Odonata).
Entom. News, 27:30-33, figs.

191 7. Some species of Leptagrion with descriptions of a new genus and a new
species (Odonata).

L. c, 28: 241-255, pis. and figs.

Explanation of Plates
Plate I

Fig. I. — Wings of Epiophlehia super stes Selys. (Female, Cornell Collection.)

Fig. 2. — Hind wing of Polythore gigantea Selys. (IVIale, Cornell Collection.)

Fig. 3. — Hind wing of Eiithore fasciata Selys. (Female, Cornell Collection.)

Fig. 4.— Hind wing of Chakopteryx rutilans Rambur. (Male, Cornell Collection.)

Plate II

5. — Hind wing of Cora marina Selys. (Male, Cornell Collection.)

Fig
Fig
Fig
Fig
Fig
Fig

6. — Hind wing of CUmacohasis modesta Laidlaw. (Male, after Williamson.)
7. — Front wing of Echo margarita Selys. (Female, after Selys.)
8. — Hind wing of Archineura basilactea Kirby. (Male, after Kirb>-.)
9. — Hind wing of Umnia longhtigma Selys. (Male, Cornell Collection.)
10.— Hind wing of Sapho cilkita Fabricius. (Female, Acad. Nat. Sci. Phila-
delphia.)

Plate III

Fig. II. — Hind wing of /4 ^rzo)? I'/rgo Linnaeus. (Female, Cornell Collection.)
Fig. 12. — Hind wing of Matronoides cyaneipeunis Foerster. (Male, after Foerster.)
Fig. 13. — Hind wing of Matrona japonica Foerster. (Female, Cornell Collection.)
Fig. 14. — Front wing of Neurobasis chinensis Linnaeus. (Female, Cornell Col-
lection.)
Fig. 15. — Front wing of Mnais pruinosa Selys. (Male, Cornell Collection.)

Fig.

I7--

Fig.

i8.-

Fig.

19.-

Fig.

20.-

Fig.

21.-

Fig.

22.-

Fig.

23--

Fig.

24.

Fig.

25--

Fig.

26.

Fig.

27-

Fig.

28.-

PHILIP A. MUXZ 73

. Plate IV

Fig. 16. — Front wing of PsoIodesmns dorothea Williamson. (Male, alter William-
son.)
-Front wing of Phaon iridipennis Burmeister. (Male. Cornell Collection. j
-Front wing of Vestal is gracilis Rambur. (Female, Cornell Collection.)
-Front wing of Iletaerina americana Fabricius. (Female, Cornell Col-
lection.)
-Front wing of Lais globifer Hagen. (Male, after Selys.)
-Hind wing of Pseudophaea ochracea Selys. (Female, Cornell Collection.)

Plate \'

-Front wing of Dysphaea diniidiata Selys. (Male, after Seh's.)
-Hind wing of Bayadera indica Selys. (Female, Cornell Collection.)
-Hind wing of Epallagefatima Charpentier. (Female, Cornell Collection.)
-Front wing of Anisopleiira fiircata Selys. (Female, after Williamson.)
-Front wing of Rhinocypha biforata Selys. (Male, after Williamson.)
-Front wing of Libellago caligata Selys. (Female, Cornell Collection.)
-Front wing of Micromeras lineatiis Burmeister. (Female, after W il-
liamson.)

Plate \\

-Front wing of Cyanocharis valga Xeedham. (Male, Cornell Collection.)
-Front wing of Heliocharis amazona Selys. (Male, after Cah-ert.)
-Front wing of Neocharis cothiirnata Foerster. (Male, Williamson Col-
lection.)
-Front wing of Diiierias alrosan guinea Selys.- (Male, after Selys.)
-Hind wing of Devadatta argyoides Selys. (Female, after Williamson.)
-Front wing of Dipklebia lestoides Selys. (Cornell Collection.)
-Front wing of Amphipteryx agrioides Selys. (Male, Cornell Collection.)

Plate VII

-Hind wing of Pliilogangu molilalia Selys? (.Male. Cornell Collection.)
-PVont wing of Disparocypha biedermanni Ris. (Male, after Ris.)
-Front wing of Pseiidolestes mirabilis Kirb\-. (Male, in British Museum.

From photograph by F. W. Campion.)
-Hind wing of Orlholestes clara Cahert. (Male, Cornell Collection.)
-Front wing of Archilestes grandis Rambur. (Male, Cornell Collection.)
-Front wing of Megalestes tnajor Selys. (Male, Cornell Collection.)
-Front wing of Sympycna fiisca \'an der Linden. (Female, Cornell

Collection.)
Fig. 43. — Front wing of Aiistrolestes ciiigitlatiis Burmeister. (Male, Cornell

Collection.)

MEM. AM. ENT. SOC, -\.

Fig.

29.

Fig.

30

Fig.

31-

Fig.

32

Fig.

vi3

Fig.

34

•Fig.

35

Fig.

3^^

Fig.

37

Fig.

38

Fig.

39

Fig.

40

Fig.

41

Fig.

42

74 A VENATIONAL STUDY OF THE ZYGOPTERA

Plate VIII

pi^ _m^_ — Front wing of Lestcs sponsa Hansem. (Male, Cornell Collection.)
Fig. 45. — Hind wing of Lestoidea conjuncta Tillyard. (Male, after Tillyard.)
Fig. 46. — Hind wing of Megaloprepus coeruleatus Drury. (Female, Cornell

Collection.)
Fig. 47. — Hind wing of Microstigma anomahtm Rambur. (Specimen in Acad. Xat.

Sci. Philadelphia. Photograph by C. H. Kennedy.)
Fig. 48. — Hind wing of Anomisma terminutum McLachlan. (Male, Acad. Nat.

Sci. Philadelphia.)
Fig. 49. — Hind wing of Pseiidostigma abermns Selys. (Male, Acad. Nat. Sci.

Philadelphia. From photograph by C. H. Kennedy.)
Fig. 50. — Hind wing of Mecistogaster linearis Fabricius. (Female, Acad. Nat.

Sci. Philadelphia. From photograph by C. H. Kennedy.)

Plate IX

Fig. 51. — Front wing of Thaiimatoneuni pcUitcida Calvert. (Male, Acad. Nat.

Sci. Philadelphia.)
Fig. 52. — Hind wing oi Rhipidolcstcs aciileata Ris. (Male, after Ris.)
Fig- 53-^Fi'ont wing of Podopteryx roseonotata Selys. (Male, after Ris.)
Fig. 54. — Front wing of Trincuragrion percostak Ris. (Male, after Ris.)
Fig. 55. — Front wing of Rhinagrion macrocephalum Selys. (Male, in Museum of

Comparative Zoology. From photograph by E. Avery Richmond.)
Fig. 56. — Hind wing of Argiolestes icteromelas Selys. (From photograph by J. G.

Need ham.)
Fig. 37. — Front wing of Megapodagrion erinys Ris. (Male, after Ris.)
Fig. 58. — Front wing of Allopodagrion sp.? (Male, Cornell Collection.)

Plate X

Fig- 59- — Front wing of Pkilogcnia terruha Calvert. (Male, in Acad. Nat. Sci..
Philadelphia.)

Fig. 60. — F"ront wing of Podoleslcs orieutalis Sehs. (Male, in Acad. Nat. Sci.
Philadelphia.)

Fig. 61. — Hind wing of Nesolestes alboterminata Selys. (Female, Cornell Col-
lection.)

Fig. 62. — Hind wing of Neurolesks trinervis Selys. (Male, in Acad. Nat. Sci.
Philadeljihia.)

Fig. 63. — Hind wing of Synkstcs weyersi Selys. (Male, Cornell Collection.)

Fig. 64. — Hind wing of Paraphlchia duodecima Calvert. (Male, Cornell Collec-
tion.)

Fig. 65. — Front wing of Walnicsia iiwntivagans Foerster. (Male, in Acad. Xat.
Sci. Philadeli^hia.)

Fig. 66. — Hind wing of Chlorolestes fasciata Biu'meister. (Female, in British
Museum. From photograph by F. W. Campion.)

PHILIP A. MUXZ

/O

Fig.

67

Fig.

68

Fig.

69

Fig.

70

Fig.

71

Fig.

72

Fig.

73

Fig. 74-

Fig. 75

Fig. 76

Fig. 77

Fig. 78

Fig. 79

Fig. 80

Fig. 81

Fig. 82

Fig. 83.—

Fig. 84

Fig. 85

Fig. 86

Fig. 87

Fig. 88

Fig. 89

Fig. 90

Fig. 91.
Fig. 92.

Plate XI

Hind wing of Dhneragrion petruhitale CaKert. (Male, after Calvert.)
Hind wing of Heteragrion flavovittahim Selys. (Cornell Collection.)
Front wing of Chorismagrion risi Morton. (Male, after Morton.)
Front wing of Perilestes fragilis? Selys. (Male?, after Calvert.)
Front wing of Metacneniis valida Selys. (Male, Museum of Comijara-

tive Zoology. From photograph b\- E. Avery Richmond.)
Hind wing of Calicnemis exiviia Selys. (Female, Cornell Collection.)
Hind wing of Copem marginipes Rambur. (Male, in Acad. Xat. Sci.

Philadelphia.)
Front wing of Phitycnemix pcnnipes Pallas. (From i)hotograph by

J. G. Xeedham.)

Plate XII
Hind wing of Idiocnemis nicrtoni Ris. (Female, after Ris.)
Front wing of Coeliccia memhmnipes Rambur. (Female, after Ris.)
Front wing of AUocuemis leiicosiicta Selys. (Male, in collection of E. B.

Williamson.)
Hind wing of Tatocueruis malgassica Kirby. (Cornell Collection. From

photograph by J. G. Xeedham.)
Front wing of Prionocnemis haematopus Selys. (After Selys.)
Hind wing of Leptocnemis hilineata Selys. (Male, Cornell Collection.

Photograph by J. G. Xeedham.)
Hind wing of Antiagrion gayi Selys. (Male, in Xational Museum.)
Hind wing of Ilypoueiira lugens Hagen. (Female, Cornell Collection.)

Plate XIII

Hind wing of Palaiargia flavovittata Selys. (Male, in collection of E. B.

Williamson.)
Vront wing oi Argia fiimipeuuis Burmcister. (Cornel! Collection.)
Front wing of Diargia bicellulata Calvert. (After Cahert.)
Hind wing of ArgiaUugnia miuiita Selys. (Male, in Acad. Xat. Sci.

Philadelphia.)
Front wing of Oreagrion lorenlzi Ris. (After Ris.)
FVont wing of Amphiagrion sauciiim Burmeister. (Female, Cornell

Collection. From photograph b\- J. G. Xeedham.)
Front wing of Pyrrhosovia nymphida Sulzer. (Male, Cornell Collection.)
Hind wing of Erythromma najas Hansem. (Male, Cornell Collection.

From jihotograph by J. G. Xeedham.)

Plate XI\'

Front wing of Xehalennia irene Hagen? (Photograph, b\' J. Ct. Xeed-
ham.)

Hind wing of Cercion lindenii Selys. (Male, in Acad. Xat. Sci. Phila-
deljihia.)

MEM. AM. ENT. SOC, 3.

76

A VENATIONAL STUDY OF THE ZYGOPTERA

Fig. 93

Fig. 94-

Fig. 95-

Fig. 96.

Fig. 97-

Fig. 98.—

Fig. 99- —

Fig. 1

00

Fig.

01

Fig. 1

02

Fig.

03

Fig.

04

Fig.

05

Fig. ]

06

Fig. 1

07

Fig.

Fig.

09.

Fig.

10.

Fig. ]

II.

Fig.

12.

Fig.

13-

Fig. 1

14.

Fig.

15-

Fig.

16.

Fig.

H7.

Fig.

118.

08.—

Hind wing of Me^ala<!_rioii occauiciini McLachlan. (Male.)

Hind wing of Coenagriov piiella Linnaeus. (Female, Cornell Collection.)

Hind wing of Hesperagriou heterodoxum Selys. (From photograph by

J. G. Needham.)
Hind wing of Tigriagrion aiirantimgrum Calvert. (After Calvert.)
Front wing of Anomalagrion hastatiim Say. (Male, Cornell Collection.

From photograph b\- J. G. Xeedham.)
Front wing of EnaUagnui cyathigerum Charpentier. (Cornell Collec-
tion.)

Plate XV

Front wing of Zoniagrion exdamationis Selys. (Male, in collection of

C. H. Kennedy.)
Hind wing of Ischniira verticalis Say. (Male, after Needham.)
Hind wing of Cclaemira dentkollis Burmeister. (Male, in collection of

C. H. Kennedy.)
Hind wing of Ceratiira capreola Hagen. (Cornell Collection. From

photograph by J. G. Needham.)
Hind wing of Austrocnemis splendida Martin. (Male, in collection of

Carnegie Museum. From drawing by C. H. Kennedy.)
Front wing of Agriooicniis puhenihius Selys. (After Needham.)
Front wing of Argiociiemis nibesceiis Selys. (Female, in Carnegie

Museum. From drawing by C. H. Kennedy.)
Front wing of Hemiphlebia mirahilis Selys. (Male^ Cornell Collection.)
Part of hind wing of CaUagrion hillinghursti Martin. (Male, after

Tillyard.)
Hind wing of Chromagrion condituni Selys. (After Needham.)

Plate XVI

Front wing of Oxyagrion rufulum Hagen. (Male, Cornell Collection.)
Front wing of Acanthagrion gracile Rambur. (From photograph by

J. G. Needham.)
Front wing of Telagrioii loiigum Selys. (After Cah-ert.)
PVont wing of Xanthocncniis zdandicum McLachlan. (Female, Cornell

Collection.)
Part of hind wing of Austroagrion cyane Selys. (Male, after Tillyard.)
Front w^ing of Pseiidagrion fiircigentm Rambur. (Male, from photo-
graph by J. G. Needham.)
Front wing of Xanthagrion erylhrouciinim Selys. (From photograph by

J. G. Needham.)
Front wing of Xiphiagrion cyanomclas Selys. (Male, after Ris.)

Plate XVII

Hind wing of Ceriagrion corotuandcliaiiuiii Fabricius. (Male, alter

Needham.)
Hind wing of Papiiagrion occipitale Selys. (After Ris.)

PHILIP A. Ml XZ

77

Fig.

119.

Fig.

120.

Fig.

121.

Fig.

122.

• Fig.

123.

Fig.

124.

Fig.

1^5-

Fig.

126.

Fig.

127.

Fig.

128.

Fig.

129.

Fig.

130.

Fig.

131-

Fig.

132.-

Fig.

I33--

Fig.

Fig.

135-

Fig.

136.-

Fig.

137-

Fig.

138.

Fig.

139-

Fig.

140.

Fig.

141.

Fig.

142.-

Fig.

143.

-Hind wing of Lepta'^rion macninmi Burnieister. (Male, afler Xecd-
ham.)
. — Hind wing oi Aeolanrioii flamnieiitu Selys. (Collection of E. B. William-
son.)

-Hind wing of SkialUigma simidacnim Calvert. (Male, in Arad. Xat.
Sci. Philadelphia.)

-Front wing of Tekbnsis salva Hagen. (From photograph l)y J. G.
Xeedham.)

-Hind wing of Aciagriou pallidum Selys. (Cornell Collection.)
. — Hind wing of Archihasis mimetcs Tillj-ard. (Male, after (\il\crt.)

Plate XVIII

-Hind wing of Aiiisagrio)i intncatipcnne Calvert. (Male, Cornell Col-
lection.)
-P>ont wing of Mclulcplobasis manicaria Williamson. (Male, Cornell

Collection.)
-Front wing of Leptohasis vacillaiis Selys. (Male, after Williamson.)
-Hind wing of Teiiiobasis meUdlica Foerster. (After Ris.)
-Front wing of Amphicnemis wallacei Selvs. (Female, after Ris.)
-Front wing of CJdorocnemis clonoatn Selys. (Male, in Acad. Nat. Sci.

Philadelphia.)
-Front wing of Proneiira prolongata Selys. (Male, after Selys.)
-Front wing of Disparoueura sp.? (After J. G. Xeedham.)
-Hind wing of Neoneurn amelia Calvert. (Male, in collection of K. B.
Williamson.)

Plate XIX

34-—

Hind wing of Perislicta forceps Selys. (Female, in Museum of Compar-
ative Zoolog>^ From photograph by E. Avery Richmond.)
Hind wing of Idioneura muilla Selys. (Cornell Collection.)
Front wing of Palaemnema paidina Drury. (Male, Cornell Collection.)
Front wing of Platysticta maadata Selys. (Female, Cornell Collection.)
Hind wing o{ IlypostropJwneitra sp.? (Male, Acad. Nat. Sci. Philadel-
phia.)
■Hind wing of Neosticta cnnescens Tillyard. (Male, after Tilhard.)
Hind wing of Aiistrosticta fieldi Tillyard. (Male, after Tillyard.)
•Hind wing of Notoneiira soUtaria Tillyard. (Male, after Tillyard.)
•Hind wing of Orisiicta filicicola Tillyard. (Male, after Tillyard.)
•Front wing of Disparonciim dorsalis Selys. (Male, in Acad. Nat. Sci.
Philadelphia.)

MEM. .\M. EXT. SOC, 3.

78

A VENATIOXAL STUDY OF THE ZYGOPTERA

Plate XX

Fig. 144. — Hind wing of Protostida sp.? (Female, in British Museum. From
photograph by F. W. Campion.)

Fig. 145. — Hind wing of Microneiira caligata Selys. (Male, Museum of Com-
parative Zoology. From photograph by E. A. Richmond.)

Fig. 146. — Hind wing of Protoneiira capillans Rambur. (Male, Cornell Collec-
tion.)

Fig. 147. — Front wing of Epipleoneura inciisa Williamson. (Male, after William-
son.)

Fig. 148. — Front wing of Epipotoneura nehalennia Williamson. (Female, after
Williamson.)

Fig. 149. — Hind wing of Psaironeiira remissa Williamson. (Male, after William-
son.)

Fig. 150. — Front wing of Isosticta spinipes Selys. (Male, after Ris.)

Fig. 151. — Hind wing of Nososticta solida Selys. (Female, Cornell Collection.)

Fig. 152. — Hind wing of Risioneura ehiirnea Foerster. (Male, Cornell Collection.)

Fig. 153. — ^Wing of Selysioneura cervicornit Foerster. (Male, after Foerster.)

Mem. Am. Ent. Sue, Xo. 3.

Plate 1.

Trnniil

MUNZ— ZVGOPTERA VEXAl lOX

Mem. Am. Ext. Soc, No. 3.

Pl.ATK II.

Ml XZ ZVGOPTERA XEXAIIOX

Mem. Am. Ent. Soc, No. 3.

Plate III.

''''^^ita^^^ai^aste*""-'""

IMUNZ— z^'('.()I•^l•:K.\ \ i:.\.\i lox

Mem. Am. Ent. Soc, No.

Platk IV

MITXZ— ZV(x()lTERA XEXATION

Mem. Am. Ent. Soc, No. 3.

Plate \'

^n^NZ-ZVGOPTERA VENA IK ).\

i

MiLM. Am. Ext. Soc, Xo.

Plate \1,

All XZ -ZVCiOlTERA XKXATIOX

IVli'M. Am. Ent. Soc, No. 3.

rLAih Ml,

iMUNZ— ZYGOP^ERA VExXATION

Mem. Am. Ent. Soc, No. 3.

Plate VI II,

MUNZ— ZVGOPTERA \ EX ATlOX

Mem. Am. Ent. Soc, No. 3.

Plate IX.

MUNZ— ZYGOPTERA \'EXATiOX

Mem. Am. Ent. Soc, No. 3.

Plate X.

]\IUNZ— ZYGOI'TICRA \EX ATIOX

Mem. Am. Est. Soc, No. 3.

Plate XI

l^IUNZ— ZVGOPTERA XKXA'riON

AIem. Am. Ent. Soc, No. 3.

Plate XII.

MUNZ— ZYGOPTERA VEXATION

Mem. Am. Kxt. Soc, Xo. :-,.

Tlatf. XIII,

Muxz— zv(;()1ti:r.\ \ iaaiiox

Mem. Am. Ent. Soc, Xu.

Plate XIV.

MIXZ— ZVGOPTKRA \ i:\Ar I()\

Mem. Am. Ent. Soc, Xo. 3.

Plate XV.

MUNZ— ZYGOPTERA VENATION

Mem. Am. Ent. Soc, No. 3.

Plate XVI.

MUNZ— z^'(;()^TI:R.\ xkxatiox

Mem. Am. Ext. Soc, Xo. 3

— H

Plate X\'II

MUNZ-ZYGOPTERA VENATION

Mem. Am. Ent. Soc, No. 3.

Plate XX'III.

AiLxz zv(;()P'I"i:ra xexaiiox

I\Ii:m. Am. Kxt. Soc, Xo. 3-

Platk XIX,

AllXZ— ZYGOPTERA \ i:XAl lOX

Mem. Am. Ext. Soc, Xo. 3.

Plate XX.

MUNZ— ZVGOITERA VEXATIOX

\

>

UA

Ul

-O

M

'--TV

•

O

H

^

"-<

1 1 1 1 1 i 1 1

C!