Ethological and morphological adaptations of Psophus stridulus LINNAEUS 1758 to habitat islands ( Caelifera : Acrididae )

This article proposes the hypothesis that the Orthopteren species Psophus stridulus restricted itself to habitat islands during its range expansion in the Holocene, and that it is specifically adapted to these. Literature analysis shows, that P. stridulus never colonises large areas, but is always concentrated in small isolated patches. Observations in Poland and Western Siberia corroborate this findings. The hypothesis is sup­ ported by the specific autecology and ethology of the species. To verify the hypothesis, a number of ethological, morphological and genetic features of this species can be understood as selective adaptations. In the context of nature conservation, these features have been discussed previously as existence-threaten­ ing trends.


Introduction
Psophus stridulus is considered to be a well studied species among European grasshoppers (DETZEL1998). The species has been described to colonise meso-to xerophilous habitats with plant communities characterising late successional stages which stay constant for longer time periods (VAlsANEN etal. 1991, H e B & RlTSCHEL-KANDEL 1992,BUCHWEITZ 1993, K o l b & F i s c h e r 1994, V o s s e n & P i p e r 1996, V a r g a 1997, J a n Be n & R e ic h 1998, BONSEL & RUNZE 2000). P. stridulus lays its eggs in the soil, but the exact require ments on soil conditions are not sufficiently known. BÖNSEL & RUNZE (2000) pointed out for Central Europe, that calcareous soils and constant water supply are common features of many habitats of P. stridulus. An habitat of P. stridulus, isolated for centuries, can be found in north-east Poland. Its conditions resemble those of the species' original habitats. The existence of these habitat islands, as well as the long-term stability and rarity of all known Central European habitats led to the hypothesis, that the species restricted itself to specific habitat islands, and that it is adapted to these conditions. This hypothesis can explain historic as well as recent records of P. stridulus in Central Europe in isolated habitats (HARZ 1960, HEROLD 1916, HOLST 1986, LUNAU 1940, OSCHMANN 1969, STERNAD 1998. Also in its region of origin, the vast Siberian forest steppe (H a r z 1960, INGRISCH & KÖHLER 1998, P. stridulus colonises only habitat islands with very rare sets of ecological conditions (BÖNSEL 2003). This leads to the theory of a gradual selection towards colonisation of habitat islands in the current area of distribution. Following this thought, the high number of ovarioles of the species (see RUBTZOV1934 in INGRISCH andKÖHLER 1998) can be interpreted as a result of saved energy, which are not allocated to flight muscles, but to increased ovarioles production. The same phenomenon can also be found in other species of grasshoppers, inhabiting similar specific habitats (see INGRISCH and KÖHLER 1998). The species does not form metapopulations (BÖNSEL & RUNZE 2000, jA N ßE N & REICH 1998, STERNAD 1998. In contrast, other species occur ring in insular habitats like ombrotrophic mires are adapted to isolation of their habitat by forming metapopulations (STERNBERG 1995), thereby preserving their genetic variability. In P stridulus, rather than metapopulation structures, other mechanisms seem to have been selected for. For one, the sex-ratio is shifted towards a higher proportion of males. This phenomenon was interpreted as a sampling error by STERNAD (1998), whereas BÖNSEL & RUNZE (2000), based on their own and similar observations by other authors (BUCH-''X'EITZ 1993, KOLB & F i s c h e r 1994), exclude sampling error and interpret this uneven distribution of sexes as a special adaptation to the existence on habitat islands. P. stridulus appears to be adapted to climatic irregularities by prolongation of diapause or by hatching in intervals. This idea is supported by observations of larvae in unusual times of the year (HEMP & Z e h m 1997, BÖNSEL & RLTNZE 2000). Genetic analysis of nine records of P stridulus in the northern Frankenalp showed clear differences among them (STERNAD 1998). STERNAD (1998) attributes this to a bottleneck-effect and interprets the isolation of the occurrences as a threat to their existence. In the light of the discussed species' restric tion to specific habitat islands and the adaptive consequences thereof (e.g. the females' inability to fly), genetic differentiation would be a consequence. All this supports the hypothesis presented in this paper. Here data on further ethological and morphological adaptations of Psophus stridulus to habitat islands will be presented, also supporting the hypothesis.

Study area
The study area is located in the Biebrza valley in North-eastern Poland (53°19'N; 22°34'E), in one of the last intact percolation mires of Europe (SUCCOW & JESCHKE 1990) with a total area of 80.000 ha. Here a population of A stridulus inhabits a mineral outcrop covered with sand, which is isolated by the surrounding bog (BÖNSEL & RUNZE 2000). The study area covers 1.5 ha, 25% of it shaded by Quercuspetraea and Tilia cordata overstory. 5% of the area are open sandy patches. 70% are covered by a herb layer which can be subdivided in various sub-layers. Soil type and plant community can be classified as limestone dry grassland. Following the law of similar habitat preference (WALTER & BRECKLE 1991), habitat conditions are comparable to those of other Central-European populations of P stridulus (BÖNSEL & RUNZE 2000).

Registration of males' and females' behaviour
The capture-recapture experiments took place in the beginning of August 2000 and 2001 and lasted for 8-10 days in each season (BÖNSEL & RUNZE 2000). Within the habitat, a grid of 20x20m cells was staked out using marked pegs thus facilitating mapping of catches. All males and females of the Polish habitat were caught and marked by writing individual numbers on the males ' wings and the females ' thoraces. After 2 days, almost all individuals of the habitat were caught once. A maximum of 18 individuals was marked only in the course of the following recapture experiment. In order to monitor night movement pat tern, in 2001 altogether 20 females and 30 males were marked additionally with reflecting foil (7610 HIGH GAIN SCOTCHLITE) by wrapping a piece of the self-adhesive foil around the tibia of the hind leg and sticking it together (see OPITZ & KÖHLER 1997). Individuals marked like this could be detected in the light of a flashlight within distances up to 50 m even in total darkness. Because of their immobility, the females could be found even after a couple of days still near the place of last capture. Accordingly, they could easily be refound by using torches at night. Using this method, 70% of the recaptures of females were made at night. In contrast, the males were easy to detect during daytime by their movements, but only accidental recaptures were made at night. Therefore males were caught especially at daytime. Acoustic signals were detected with a bat-detector.

Processing of data and statistical analysis
The informations of the daily grid maps were transfered to a geographical information system (GIS; ArcView), and the distances covered by males and females were calculated. The home ranges of both sexes were determined by drawing the smallest convex polygon which contains all recapture points (JENNRICH and TURNER 1969) and calculating its area. The maximum distances covered were sorted by quantity of recapture. t-Tests were used to compare the means of groups with the same quantity of recaptures between males and females. t-Tests were also used to compare the abundances of the different types of sensilla on the antennae between males and females. Separate t-test were conducted for the ven tral and the dorsal position.

Electron microscopy
A total of 14 antennae of male and female individuals have been scanned with an electron microscope: 2 males and 2 females from North-eastern Poland, 2 males and 2 females from Central Siberia, 2 males and 2 females from South-western Germany and 1 male and 1 female from South-eastern Germany. The individuals were killed by ether and pre served in a 4% solution of glutar aldehyde. An artefact free critical spot drying of the prepared antennae was carried out using K 850 EMITECH. The conductivity of the preparation was achieved by evaporating with gold in the vacuum (Sputter-Coaters SCD 004). SEM observations were made with a Carl Zeiss DSM 960 A. The preparations were standing vertically on the slide. Thus a dorsal as well as ventral observation of the antennal structure was possible. For living imagines dorsal refers to pointing backwards and ventral to pointing forward with erected antennae.

Spatial distribution and movement pattern of males and females
In 2000, altogether 187 males and 60 females were recorded, in 2002 209 males and 61 females. The rates of recapture of males were approximately equal in both years, whereas the rate of recapture of females increased in 2001 by additional marking with reflecting foil (table 1).     The north-western part of the study area was less frequented by the imagines in both years ( fig. 1,2). This is attributed to shading by groups of Quercuspetraea and Tilia cordata. Any preferences for other vegetation structures were not found. The males proved to be the more active gender in all observations: In both years, the maximum distances covered by the males were significandy larger than those of the females (p < 0.1) ( fig. 1/2/3; tab. 1). Due to the high activity of the males, their trajectories crossed frequently. Territories with defined borders did not exist. The females hardly moved at all, but stayed in a narrow corridor during the whole time of the studies, partly in dense clusters ( fig. 1). Highest activity of males was observed near clusters of females ( fig. 2). Males moved in about 85% of the observations by crawling and in 15% by flying. Flights are rarely accompanied by rattling. Rattling did not influence the females' behaviour, they stayed passive and even hid when high numbers of males were present. In 2000, 4, and in 2001, 2 matings were observed. The male approached the female silently and mated with it. Mating was observed even in light rain. In rainy weather no acoustic signals were heard at all. No acoustic signals from females were recorded at any time of the study.

SEM structure of the antennae
The four typical forms of sensilla of the Orthoptera, Sensillum chaeticum, S. trichodeum, S. basiconicum and S. coeloconicum, were found on all 22 segments of the antennae (see OCHIENG et al. 1998). Highest abundances of all types of sensilla were found on the ventral sides of the antennae. Sensillum chaeticum and S. trichodeum showed an almost regular distribution over all segments of the antennae in both sexes ( fig. 4), without signifi cant differences in their abundances between males and females (p > 0.01). S. basiconicum and S. coeloconicum were the most frequent sensilla, their highest densities could be found on the ventral side of the 10th to 22nd segment ( fig. 4, table 2). Higher densities of Sensillum basiconicum were found in males than in females, with a significant difference on the ventral side of the antenna (p < 0.01). Sensilla coeloconica were found on the 6th to 22nd segment. Here the males showed sig nificantly higher densities than the females on the dorsal as well as on the ventral side of the antennae (p < 0.01) ( fig. 5/6).

D iscussion
According to the movement patterns of males and females (see also BUCHWEITZ 1993, K o l b & F i s c h e r 1994, J a n Be n & R e i c h 1998), individuals neither occupy nor defend defined territories. This is indicated by frequent crossing of the males' trajectories.
Rattling is a rare event. JACOBS (1953) interprets rattling as part of courtship. If so, the females would have to approach a possible winner in contest, or at least the loser among the males would have to retreat. However, this was not observed in both years of the study. Therefore rattling is not considered to be part of courtship or competition.
The matings observed in rain support the theory that finding of females is not due to acoustic signals. The scarcity of the females and their passivity are considered to be the major reason for the high activity of the males, forcing them to scan the habitat repeatedly for females. The rattle may stimulate the females to excrete pheromones, thereby attracting the males. This stimulation could be the reason, why from time to time some males ratde instinctively, thereby causing the females to reveal their hidden position by excretion of pheromones. This mechanism of finding females by their pheromones is known from few other species of grasshoppers (see BLANEY & S i m m o n d s 1990, S i d d iq i & K h a n 1981, W h i t m a n 1982 and 1990). The high abundance of Sensilla coeloconica in male P. stridulus ( fig. 4), which serve for detection of chemical stimuli (OcHIENG et al. 1998), leads to the hypothesis that their function is perception of females ready to mate. A similar dimorphism of sensilla is also known from another species of grasshoppers (BLAND 1982) and numerous species of butterflies (KEIL et al. 2001, STEINBRECHT 1999. However, further investigations on the functions of the different types of sensilla are necessary, since it is possible that there are differences in function between species and also between the sexes (KEIL et al. 2001, OBENG-OFORI et al. 1994, OCHIENG et al. 1998, STEINBRECHT 1998, STEINBRECHT et al. 1996. The genetic cohesion, which is selected for the specific biotope, has to be protected from destruction e.g. by gene drift (MAYR 1967). This is possible only if as much genetic mate rial as possible is taken along to the next generation. In P. stridulus, the females seem to be responsible for passing on as many genes as possible in multiple ways. Short-time excretions of pheromones in small doses reveal their hidden positions to the males, but the time limitation of the signal prevents attraction of too many males at a time. A male being by chance close to a female excreting pheromones can find it and mate with it. Afterwards it will leave, unaware of the other females usually hidden close by ( fig. 1), and continue scan ning the habitat for other females. The female can attract another male by further excre tion of pheromones after some time. The clusters of females increase the chance, that more than one male perceives the signal and that the group of females is found by at least one of the males attracted. Further it is known, that P. stridulus is one of the grasshopper species able to store sperm (Re i n h a r d t & J e n t z s c h 1999, R e in h a r d t 2000). Considering the high probability of matings with different males, this mechanism increases the number of possible gene com binations in next generation. This phenomenon was described for numerous species in cluding humans (COOK & GAGE 1995, BAKER & BELLIS 1993, BlRKHEAD & MOLLER 1998, G a g e & B a k e r 1991, G a g e & B e r n a r d 1996, P a r k e r 1970, P a r k e r et al. 1990al. , SIMMONS 1987. The behaviour of the females thereby serves for increasing and main taining the existing genetic variability in the following generations. In this context also investment of females in higher proportions of male offspring has to be discussed. The high mobility of the males increases the chance of mating with different females, thereby passing on many genes to the following generations. Sensu DAWKINS (1976), male P stridulus are typical players, setting high stakes and taking high risks for the survival of their genes. In contrast, the females are careful investors, trying to minimise their risks and maximising the chance to pass on as many genes as possible to the next generation. These strategies can be the result of habitat isolation, or, on the other hand, can have made the colonisation of habitat islands possible at all (see KANESHIRO 1976KANESHIRO /1980. Precision of adaptations went so far that the mayor part of flight muscles were sacrified in favour of increased ovariole production. Further studies should explore whether the females in fact announce their readiness for mating by pheromones, and whether this is the reason for the sex dimorphism of the antennal structures. Also, the physiological plasticity of the females, enabling directed investment in higher proportions of male offspring (see TELFORD & DANGERFIELD 1990), need further examination.