Back to the Trigonalidae Homepage
Phylogenetic revision of the parasitoid wasp family Trigonalidae
(Hymenoptera)
David Carmean (carmean@sfu.ca)
Biological Sciences
Simon Fraser University
Burnaby, BC CANADA V5A 1S6
Lynn Kimsey (lskimsey@ucdavis.edu)
Bohart Museum
Department of Entomology
University of California
Davis, CA USA 95616-8584
Running title: Revision of the Trigonalidae
Abstract
A phylogenetic analysis of generic relationships and revised generic
concepts for the Trigonalidae are presented. The Trigonalidae
is divided into two subfamilies, the Orthogonalinae and the Trigonalinae.
Orthogonalinae consists of a single primitive genus, Orthogonalys,
sharing many generalised apocritan characters, but lacking advanced
trigonalid characters including antennal tyloids and female metasomal
armature. No synapomorphies have been detected for the Orthogonalinae.
Trigonalinae are characterised by the presence of tyloids. The
Trigonalinae were originally defined by the absence of female
armature, and were therefore polyphyletic because armature has
been lost several times. Within the revised concept of Trigonalinae,
the tribe Trigonalini is characterised by the presence of elongate
parameres and an inter-torulus distance subequal to the distance
between the torulus and the eye. A second tribe, Nomadinini, is
based on the secondary loss of tyloids, and comprises the previous
subfamilies Seminotinae, Nomadininae, Bareogonalinae, and Lycogastrinae.
The two tribes Trigonalini and Nomadinini share the presence of
female armature, though in some Trigonalini it has been secondarily
lost. The genera Labidogonalos Schulz, Lycogastroides
Strand, Lycogonalos Bischoff, Nanogonalos
Schulz, Poecilogonalos Schulz and Taiwanogonalos
Tsuneki are synonymised under Taeniogonalos Schulz.
The species Lycogaster zimmeri Bischoff,
Tapinogonalos maschuna Schulz, and
Trigonalys pictifrons Smith (placed
in Lycogaster by Schulz 1906) are transferred to
Taeniogonalos. The genera Discenea
Enderlein and Stygnogonalos Schulz are synonymised
under Trigonalys Westwood. The species Labidogonalos
flavescens Bischoff, Labidogonalos
sanctaecatharinae Schulz, Trigonalys
lachrymosa Westwood (placed in Lycogaster
by Bischoff 1938), Trigonalys maculifrons
Cameron (placed in Labidogonalos by Bischoff 1938),
and Trigonalys rufiventris Magretti
(placed in Lycogaster by Schulz 1907) are transferred
to Trigonalys. Trigonalys costalis
Cresson is sunk under Taeniogonalos gundlachii
(Cresson). Xanthogonalos fasciatus
Bertoni and X. severini Schulz are
synonymised under Trigonalys sanctaecatharinae
(Schulz). Mimelogonalos partiglabra
Riek and Mimelogonalos punctulata
Riek are synonymised with Mimelogonalos bouvieri
Schulz. Lectotypes are designated for Trigonalys
melanoleuca Westwood and Taeniogonalos
fasciatipennis (Cameron). The author of Trigonalys
maculifrons is Sharp 1895, not Cameron 1897, and
the author of Taeniogonalos enderleini
is De Santis 1980, not Schulz 1906. Viereck in 1914 designated
Trigonalys pulchella Cresson as type
of the genus Tapinogonalos Schulz, preceding Bischoff's
1938 designation, making Tapinogonalos a synonym
of Orthogonalys. A new genus, Afrigonalys,
is proposed for the three species that were described in "Tapinogonalos"
sensu Bischoff, nec Viereck.
KEYWORDS: Trigonalidae, Trigonalyidae, phylogeny, Hymenoptera,
parasitoid, hyperparasitoid, biogeography, homoplasy, convergence.
Introduction
Trigonalids have a biology unique within the Hymenoptera. Thousands of eggs are laid on foliage (Figs 1-5) and these eggs will not hatch unless consumed by a herbivorous caterpillar or sawfly larva. The ultimate hosts of trigonalids are the parasitoids or predators which attack the herbivore, although some species will directly parasitise sawflies. Some species of trigonalids may parasitise hosts of different orders and sizes, resulting in a great deal of intraspecific variation and confusion. Weinstein & Austin (1991) recognised about 100 valid species and here we recognise 16 genera and 88 species, and have seen approximately 12 additional undescribed species. This group is rarely collected; and the hosts for most species, and many other aspects of the biology remains unknown. The biology of trigonalids was reviewed in detail by Clausen (1940), Carmean (1991), and Weinstein & Austin (1991, 1995b). Chen (1949), Riek (1954, 1962a), Teranishi (1929), Townes (1956), Tsuneki (1991), and Smith (1996) are important regional references with keys, though they cover a limited number of taxa.
Most Trigonalidae are relatively large (5-15 mm long), though some are as small as 3 mm. Some species are thin and elongate, mimicking ichneumonids (Fig. 6), while others are relatively stout-bodied, mimicking vespids (Figs 4-5). Superficial convergence among phylogenetically distant trigonalids is common and appears to be due to unrelated species mimicking a common wasp model.
Schulz (1907a) undertook the only global study of trigonalid generic relationships, and regional studies consistently mention the need for generic revision. Schulz probably had about 250 specimens available to him, about half of which were Pseudogonalos hahnii (Spinola), and most of his specific and generic descriptions were based on one or two specimens. As a result, his genera tended to be narrowly defined and based on variable characters. Bischoff (1933) remarked that under Schulz's system a new genus would be necessary for each new species. Benoit (1951) stressed the need for broadening generic concepts. Strand (1912) stated that Schulz had gone too far in his classification of genera and that his subfamilies were built on a weak foundation of variable characters. Weinstein & Austin (1991) noted "that many genera are poorly defined and that their monophyly, and that of some subfamilies, is questionable."
Since 1907 the numbers of Trigonalidae collected, mainly through
Malaise traps and rearing, has greatly increased. Because Schulz's
genera were poorly defined and his generic keys confusing, different
workers have placed species in unrelated genera, or have created
new genera synonymous with existing ones. The purpose of this
study is to revise generic concepts, to resolve some of the taxonomic
confusion which resulted from a lack of understanding of the variability
of genera and species, and to propose monophyletic higher taxa
based on shared derived characters.
Monophyly and distinguishing characters of the Trigonalidae
Despite the diversity of form within the family, the Trigonalidae
is clearly monophyletic, and is well-defined by a suite of derived
morphological and behavioral characters. These characters clearly
distinguish and separate the family from other groups, so that
despite superficial mimicry there are no extant taxa that can
be confused with Trigonalidae. One feature unique to all members
of the family is the presence of sparse white scales or setae
on the outside of the middle flagellomeres of the female antenna
(Fig. 13). At magnifications of X 30-50 they usually appear as
an oval patch of sparse white spots. On dirty, wet, or greasy
specimens they may not be visible. Additional features, apparently
unique to trigonalids but secondarily lost in Nomadina
Westwood, Bakeronymus Rohwer, and Pseudonomadina
Yamane & Kojima are: mandibles asymmetrical (Fig. 19); hind
trochantellus diagonally divided and appearing 2-segmented (=hind
trochanters apparently 3-segmented); and propodeal spiracle covered
by a prominent flap. Bareogonalos, as well as Nomadina,
Bakeronymus, and Pseudonomadina, always
have the hind trochantellus undivided, and some species of Bareogonalos
have symmetrical mandibles. Trigonalids are also characterised
by a supra-antennal elevation or raised area above and mesad of
the torulus that may be reduced to absent in some Bareogonalos
and related genera (Figs 14-20). Other, possibly plesiomorphic,
diagnostic characters are: maxillary palps longer than mandibles
and 6-segmented, with first segment very small (maxillary palps
as long as mandible and 4-segmented in Bakeronymus,
rudimentary and 4 or fewer segments in Nomadina
and Pseudonomadina); tarsi with plantar lobes (Fig.
21); tarsal claws cleft (Fig. 21); forewing with pronounced costal
cell and complete wing venation (generally 10 closed cells), hindwing
with two closed cells; propodeal foramen open ventrally; and metasomal
terga with short dorsolateral sutures (visible on tergum II and
often covered on the following terga). Most trigonalids have 18-28
antennal segments (16-26 flagellomeres), although some have as
few as 13 segments and one specimen has been found with 32 segments.
Tyloids (raised areas without pubescence)(Figs 10-12) on the outside
of the middle antennal segments of males of many species are not
part of the groundplan but are derived within the family. Females
of many species have 'metasomal armature' (Figs 1-3), i.e., projections
or raised ledges, posteriorly on sternum II or III or both, a
feature also derived within the family and not part of the groundplan.
Distribution
Trigonalidae have a cosmopolitan distribution, with the exception of arctic and alpine regions. The greatest abundance of taxa occur in the tropics and only one species is found in Europe. Most genera are widely distributed. For example, Orthogonalys Schulz is known from Japan, Taiwan, north-western Vietnam, north-eastern India, Madagascar, Tanzania, eastern South America, and eastern North America. Taeniogonalos Schulz is even more widespread, and is found in these areas as well as in Australia, Papua New Guinea, the Solomon Islands, Indonesia, Sri Lanka, and throughout India. Some genera are more restricted, e.g., Mimelogonalos Schulz from Australia (with an undescribed close relative in Papua New Guinea) and Seminota Spinola from the neotropics. Nomadina is also known only from the neotropics, but there are two closely related Asian genera, Bakeronymus, with a single species known only from the Philippines and Taiwan, and Pseudonomadina, known only from the Philippines. Bareogonalos is found around the perimeter of the Pacific Ocean, in western Mexico, north-western North America, eastern Siberia, Japan, Taiwan, Java, and Sumatra. The same species of Bareogonalos is found in Siberia, Japan, and Java. Other trigonalid species found in Indonesia may be distributed as far away as India [Taeniogonalos thwaitesii (Westwood)] and China [Lycogaster celebesiensis (Szepligeti) and Taeniogonalos fasciata (Strand)].
Distributions of trigonalid genera linking widely separated regions are supporting evidence for the ancient origins of the group and appear to be relictual. However, none of the known hosts, except for sawflies, can be considered ancient. The most widely distributed trigonalid, Taeniogonalos, is a relatively derived genus known to directly parasitise sawflies as well as indirectly tachinids and ichneumonids (Weinstein & Austin, 1991). The least derived trigonalid, Orthogonalys, is widely distributed, absent only from Europe and Australia. Host information for the genus is limited to a few rearings from Lepidoptera pupae which are believed to be intermediate hosts and a tachinid parasitoid of Lepidoptera (see generic discussion).
Systematics
What little systematic stability this family has enjoyed historically has been an artefact due to its obscurity and the difficulty in developing adequate generic concepts. Schulz (1907a) and those who followed him placed the most distinct and autapomorphic genera, e.g., Bareogonalos, Seminota (with Xanthogonalos Schulz), and Nomadina (with Bakeronymus) in their own subfamilies. The remaining genera were either united in one subfamily (Bischoff, 1938), or divided into two subfamilies, based on the presence or absence of female metasomal armature (Schulz, 1907a; Benoit, 1951; Weinstein & Austin, 1991). Although the armature is an obvious character unique to the trigonalids, it has been lost independently in some species of the genera Taeniogonalos and Trigonalys. In addition, armature presence or absence varies geographically in the species Taeniogonalos thwaitesii, and Riek (1962a) used a series of figures to show that the armature of Taeniogonalos venatoria Riek also varies from prominent to reduced. Therefore, it is clear that other characters, in addition to armature, must be considered in assigning species to subfamilies, tribes, and genera.
Previously, venational characters were given considerable weight in defining genera (Cameron, 1899) and subfamilies (Schulz, 1905). Some of these characters are now known to vary even between the left and right side of individual specimens. Indeed, Benoit (1951) argued against using the form and size of the submarginal cells as an important systematic character. However, if many specimens are available, some generalisations may be informative. For example, while there are exceptions, most Orthogonalys have a petiolate submarginal cell II, as do most Nomadina that we examined. Mimelogonalos, Lycogaster, and Bareogonalos consistently have submarginal cell II broadly attached, while many Taeniogonalos and Trigonalys are intermediate, so that Rs meets 1m-cu, with enough variation so that the submarginal cell is sometimes petiolate or broadly attached.
Because the ultimate host is not selected at oviposition, the size of the host can vary considerably, and thus the size of individuals within a trigonalid species can also vary greatly. The extent of markings can also vary. Tsuneki (1991) illustrated and discussed the range of variation possible in the size and markings of Taeniogonalos sauteri Bischoff. Series of specimens collected at the same location and time often have individuals that depart from the norm for what is considered a definitive character. For example, a specimen of Seminota marginata (Westwood) has symmetrical mandibles [BMNH] or one Bareogonalos jezoensis (Uchida) has a flat dorsellum [CARM]. This variation and anomaly makes it difficult to establish valid differences among species, especially when few specimens are available.
Finally, most earlier workers incorrectly believed that the metasomal armature was found on males only (Banks, 1908; Buysson & Marshall, 1892; Cameron, 1897, 1899; Cresson, 1865; Harrington, 1898; Magretti, 1897; Schrottky, 1906; Smith, 1851; Westwood, 1841, 1843) and thus identified females with metasomal armature as males. Even Schulz (1905) initially made this error, although his later works were accurate.
Taxa Covered: (Links not working)
Afrigonalys, 45
albomaculata, 40
alishana, 69
alticola, 69
boliviana, 40
bouvieri, 77
canadensis, 50
celebesiensis, 52
centrimaculta, 40
chadwicki, 68
championi, 73
cisandina, 54
claripennis, 69
costalis, 69
crassiceps, 73
debilis, 40
depressa, 58
Discenea, 70
dubia, 62
elongata, 40
enderleini, 68
erythromelaina, 46
fasciata, 68
fasciatipennis, 68
fasciatus, 73
flavescens, 73
flavicincta, 68
flavocincta, 68
flavonotata, 50
flavoscutellata, 69
formosana, 40, 68
fukuiensis, 40
fulvoscutellata, 68
gigantea, 40
gracilicornis, 68
gundlachii, 69
hagoromonis, 40
hahnii, 80
harmandi, 80
heinrichi, 52
henicospili, 69
hirasana, 40
hova, 40
huisuni, 50
inquirenda, 58
intermedia, 69
Ischnogonalos, 61
javana, 69
jezoensis, 50
Jezonogonalos, 75
jucunda, 69
kerala, 69
Labidogonalos, 62
lachrymosa, 73
laeviceps, 58, 69
leprieurii, 58
Liaba, 53
lugubris, 69
Lycogaster, 50
Lycogastroides, 62
Lycogastrula, 70
Lycogonalos, 62
maculata, 69
maculifrons, 73
maga, 69
magnifica, 68
marginata, 58
marujamae, 75
maschuna, 69
maynei, 69
melanoleuca, 73
mexicana, 58
micanticeps, 73
Mimelogonalos, 75
minima, 69
minuta, 77
Nanogonalos, 63
nasuta, 54
natalensis, 73
nigricauda, 77
nigrithorax, 77
Nippogonalos, 48
nipponica, 81
Nomadina, 53
Nomadinini, 44
ornata, 69
ornatissima, 46
Orthogonalinae, 36
Orthogonalys, 36
partiglabra, 77
phylogenetica, 54
pictifrons, 69
pictipennis, 69
Platygonalys, 53
Poecilogonalos, 62
Pseudogonalos, 78
Pseudonomadina, 55
pulchella, 40
pullata, 53
punctulata, 77
raymenti, 69
robertibuyssoni, 60
rufiventris, 73
rufofasciata, 69
sanctaecatharinae, 73
Satogonalos, 36
satoi, 69
sauteri, 69
schulzi, 70
scutellaris, 50
semibrunnea, 70
Seminota, 56
semirubra, 46
severini, 73
seyrigi, 41
similis, 69
smithii, 54
Stygnogonaloides, 70
Stygnogonalos, 70
Taeniogonalos, 62
taihorina, 70
Taiwanogonalos, 63
Tapinogonalos, 36
taschenbergi, 58
tenebrosa, 70
Teranishia, 80
thwaitesii, 70
tricolor, 69, 70
Trigonalinae, 42
Trigonalini, 61
Trigonalis, 70
Trigonalos, 70
Trigonalys, 70
typicus, 48
venatoria, 70
violaceipennis, 53
Xanthogonalos, 58
yuasai, 69
zairensis, 70
zimmeri, 70
This page is maintained by Dave Carmean with an eye towards speed and clarity, and last modified 15 April 1997. Comments or suggestions are welcomed!