MORPHOLOGICAL AND ALLOZYME VARIATION IN THE COMPLEX (, TETTIGONIOIDEA)

by

L. OUDMAN, M. DUIJM1 & W. LANDMAN2 (Departmentsof Geneticsand Zoology,University of Groningen,P. O. Box 14, 9750 AA Haren, The Netherlands)

SUMMARY

A number of (sub)species of Ephippigerephippiger are analysed for geographical varia- tion in morphological characters and allozyme frequencies. About 700 specimens from 33 sites, mainly from the southern part of , are studied. Genetic distances are very low. All characters show clinal variation, but most clines differ from each other. It is concluded that within the material studied no species nor subspecies can be distinguished: there are only a great number of local forms without taxonomic status. The taxa E. cruciger,E. cunii and E. ephippigermoralesagacinoi have to be sup- pressed. Instead there is only one very polymorphic species. Traditionally this taxon should be called E. ephippigervitium complex, but according to KRUSEMAN(1988) the name E. ephippigerdiurnus Dufour 1841 should be used. KEY WORDS:Ephippiger ephippiger, biosystematics, enzyme electrophoresis, male appendages, principal component analysis, clines.

INTRODUCTION

A number of species belonging to the Ephippiger occur in France and the adjacent parts of and . Dujm & OUDMAN (1983) dis- tinguished 3 groups, between which no mating was possible: 1. E. provincialis (Yersin, 1854); 2. E. terrestris (Yersin, 1854) with-in accordance with NADIR (1980)-the subspecies E. t. terrestris, E. t. bormansi and E. t. caprai; 3. the group E. ephippiger vitium (Serville, 1831 ), E. e. vicheti Harz 1966, E. cruciger (Fieber, 1853) and E. cunii (Bolivar, 1877). This distinction of three groups is confirmed by analysis of allozyme frequencies (OUDMAN et al., 1989). Dugs & OUDMAN (1983) regarded group 3 as a superspecies, con- sisting of a number of semispecies with partial mating incompatibility. HARTLEY & WARNE (1984), however, described this group as a species 455 complex with two subspecies: E. ephippiger vitium (also comprising E. cruciger and E. e. moralesagacinoz) and E. ephippiger cunii. More information is needed for both the elucidation of evolutionary relationships within this group and to assess the resulting taxonomic status of the various taxa. In particular, the exact geographical distribution of the characters used for discrimination has to be map- ped, and the systematic relevance of these characters determined. In this paper the taxa E. e. vitium, E. e. moralesagacinoi, E. cruciger and E. cunii are studied. Geographic patterns are given for morphological characters which are commonly used in . In addition allozyme frequencies were estimated for a number of populations. These frequencies can reveal evolutionary relationships between these populations, independent of morphological characters (e. g. BERLOCHER, 1984). Overall patterns of morphological characters and allozyme frequencies are compared by means of principal component analysis. In the allozyme analysis the group of taxa studied here is compared to Ephippiger t. terrestris.

MATERIALS AND METHODS Collectionsites About 700 specimens from 34 sites (table I, fig. 1) were analysed. Only samples of more than 5 specimens were used for metric analysis. A minimum of 20 individuals was aimed at to estimate allozyme frequencies. Every sample was collected in as small an area as possible. Within 24 hours after being captured the were measured, photographed and frozen in solid carbon dioxide (-79°C) for later storage at -28°C. Specimens from four sites were borrowed from the Institute for Taxonomical Zoology (Zoological Museum), Amsterdam. In table I these sites are marked with (ITZ). Not all characters could be measured from these dried animals. Samples of collected specimens from all sites will be deposited in the entomological collection of the Institute for Taxonomical Zoology (Zoological Museum), Am- sterdam.

Morphology Five characters of body and extremities were measured in the field: body weight (BW), body length (LB), length of pronotum (LP), length of hindfemur (LF) and length of protibia (LT). Weight was determined by means of a Pesola pocket spring balance (0 to 5 g in 0.1g; for exceptionally heavy specimens 0 to 10 g in 0.2g). Length was measured by means of sliding calipers (0.1mm). Colour of pronotum and abdomen was recorded in the field. Cerci and epiproct of freshly killed specimens were photographed with a Medical Nikkor (2:1). From these photographs two measurements of the cerci were taken (fig. 2a): the length of the basal part (CB) from the basal edge to a projection of the inner tooth, and the length of the apical part of the cercus (CA) from the top to the projec- tion of the inner tooth. The total length of the cercus (CL) then is CL = CA + CB. Three measurements of the epiproct were taken (Fig. 2b): the maximum width at the