PELVIC GIRDLE POLYMORPHISM AND REPRODUCTIVE BARRIERS IN THE NINESPINE PUNGITIUS (L.) FROM NORTHWEST RUSSIA

by

VALERY V. ZIUGANOV1) and ALEXEY A. ZOTIN2) (Institute of Developmental Biology, Russian Academy of Sciences, Vavilov Str. 26, Moscow 117808, Russia)

(With2 Figures) (Acc.10-IV-1995)

Summary Populations of the ninespine stickleback with the complete loss of the pelvic girdle are described from Russia. In some areas in northern Karelia, these populations are reproduc- tively isolated from parapatric populations of the 'wild type' with pelvic girdle, while in adjacent areas, populations are polymorphic and have intermediate girdle phenotypes. This seems to reflect the initial stage of divergence. Experiments in large ponds show a selective advantage of the loss of the pelvic girdle under pressure from insects predators and a selective disadvantage under predatory . Mate choice tests indicate reproductive isolation between pelvic phenotypes.

Introduction

Some groups of fish fail to develop the pelvic complex (NELSON, 1971). Several , however, are known which have populations in which some individuals develop the pelvic complex while other individuals do not develop it. Most of these species belong to the family Gasterosteidae (ZIUGANOV, 1991). Reduction and loss of the pelvic girdle have been reported for threespine stickleback Gasterosteus aculeatus, Culaea inconstans, and ninespine stickleback Pungitius pungitius (NELSON, 1971; MOODIE & REIMCHEN, 1976; GILES, 1983; REIMCHEN, 1984; WOOT- TON, 1984; CAMPBELL, 1985; BLOUW & BOYD, 1992; BELL & ORTi, 1994). In the ninespine stickleback reduction of the pelvic girdle has been found in three regions - Canada, Ireland and Greece (NELSON, 1971;

1)E-mail address: [email protected] 2) We thank Drs T.C.M. BAKKER,M.A. BELL,and T.E. REIMCHENfor fruitful discussion of manuscript. 1096

STEPHANIDIS, 1971; BLOUW & BOYD, 1992). We describe here for the first time pelvic spine polymorphisms from a fourth region - the northern Karelia, Russia and examine reproductive barriers and influence of pre- dation. REIMCHEN ( 1980) proposed that stickleback without spines were at an advantage during predation by aquatic insects such as odonate naiads. Our experiments with predatory fish and insect predators showed an advantage of the complete girdle morph under pressure of fish predation, whereas without a girdle had better survival chances under insect predation.

Materials and methods Populations. Investigations were carried out from June to September 1986. Five freshwater isolated lakes, one creek, and two bays of the White Sea were studied in the region of Chkalov village, Loukhi district, Karelia (66018' N, 33025' E) (Fig. 1). Nineteen hundred specimens of P. pungitius(L.) were sampled by deep net. The total length of fish was 37-60 mm. Pelvic structure was examined in all specimens. We observed the following pelvic phenotypes in the ninespine stickleback P. pungilius(Fig. 2): i) phenotype 'C' - specimens with complete pelvic complex; ii) phenotype'!' - specimens lacking part of the pelvic complex; iii) phenotype 'A' - specimens lacking all of the pelvic complex.

Mate choicetests. Fish for mate choice tests were sampled in June 1986. Females were kept in tanks 30 cm wide, 30 cm high, and 100 cm long for 1-2 days under natural 24 hours sunlight at 18,C. Each tank contained five females and these were fed live worms and Daphnia spp. daily. Tanks were constructed with wooden walls and a wooden bottom with longitudinal aper- tures 3 mm in width. These were placed in the littoral zone of a natural pond so that three- quarters of each tank were immersed. Males were placed in experimental tanks as soon as they were captured. Mate choice tests were carried out with different pelvic phenotypes using methods from FOSTER(1977). Males with different pelvic phenotype were simultaneously introduced into opposite corners of the experimental tank (30 cm wide, 30 cm high, and 100 cm long). An opaque partition divided the tank into two equal parts. Males constructed their nests usually within 6-24 hours. Three rocks, one clump of filamentous green algae, and one clump of (three strands, 50 cm in length) were placed on the sandy bottom at both ends of this tank. On nest completion, the partition was removed and a gravid female was introduced into the tank. Mate selection was determined by which nest the female entered and deposited her eggs. If courtship was not successful in the first twenty minutes the test was stopped. 'C' and 'A' males of similar body sizes were selected for the mate choice tests (total length of 45-55 mm; t-tests, p > 0.05). The size of the females used of both phenotypes was similar (t-tests, p > 0.05). Only gravid females that demonstrated the 'head-up' posture and males that had nuptial coloration and performed the zig-zag dance were used in the experiments. Experiments were begun immediately after the discovery of receptive fish. Each male and female were used once. For choice tests, 22 males from each of two study sites (Levin