Spawning grounds and nests of Trachidermus fasciatus (Cottidae) in the Kashima and Shiota estuaries system facing Ariake Bay, Japan

 The spawning grounds of a cottid fish, Trachidermus fasciatus, were investigated off the mouth of the Kashima River in Ariake Bay from January to March, 1996–1999. The nests were found about 1.1–2.3 km off the mouth of the river, where the bottom salinity was 8‰–21‰ at low tides. Oyster shells were mainly found to be used as nests. A male, a pair, a male and egg clusters, a pair and egg clusters, or a female were found inside shells. The average number of guarded egg clusters per male in shells was 1.3 ± 0.4. Received: March 2, 2001 / Revised: September 17, 2001 / Accepted: November 13, 2001     

Habitat selection by larvae of a fluvial lamprey, <Emphasis Type="Italic">Lethenteron reissneri</Emphasis>, in a small stream and an experimental aquarium

Habitat selection by fluvial lamprey larvae, Lethenteron reissneri (Petromyzontidae), was studied in a natural stream and an experimental aquarium to clarify microhabitat requirements for future conservation of natural populations. A gross collection survey of lamprey larvae in the Monbetsu River, southeastern Hokkaido, Japan, revealed a remarkable bias toward distribution in sandy-mud beds. An analysis using Jacobs' electivity index showed that the larvae selectively utilized spaces having regard to shallow water, weak current, deep sandy-mud, and fine substrate particles. A comparison of microhabitat use between small- (≤5 cm) and large-sized larvae (>5 cm) indicated that the latter utilized the space with greater ranges in both water depth and substrate particle size than the former. Both the field survey and laboratory experiments on larval selectivity of physical habitat variables clearly demonstrated that substrate particle size was the most important variable for small-sized larvae whereas both water depth and substrate depth were more important for large larvae. These findings should be applicable in directing attempts at fluvial habitat restoration for conservation of this endangered lamprey species. Received: November 1, 2000 / Revised: September 12, 2001 / Accepted: October 18, 2001     

Spawning ecology of the American brook lamprey, <Emphasis Type="Italic">Lampetra appendix</Emphasis>

Synopsis We observed spawning American brook lamprey, Lampetra appendix, in coldwater streams in Minnesota to assess various aspects of their spawning behavior and spawning habitat requirements. Spawning occurred during April and May, at water temperatures ranging from 8.7 to 15.5°C. Average adult lamprey length and mass differed significantly among streams, but there were no significant differences in length or mass between males and females. Overall sex ratio was 1:1, although one stream had significantly more males than females and one stream held significantly more females. Lampreys spawned in groups of 2–14 individuals, averaging 4.2 adults per nest across all streams. Nests were constructed in gravel and cobble substrate just upstream of riffles, spaced at an average density of three nests m⁻². The typical nest was 16 cm in diameter in water 31 cm deep with a bottom current velocity of 14 cm s⁻¹, and was excavated to a depth of 4 cm below the stream bottom; however, some nest characteristics varied significantly in a few streams. Nests were larger in streams with larger spawning groups, deeper water, and slower current velocities. American brook lamprey exhibited spawning behaviors and spawning habitat requirements similar to those of other species of lamprey in North America.     

The physical factor inducing spawning of the Biwa catfish, Silurus biwaensis

The physical factor inducing spawning of the Biwa catfish, Silurus biwaensis (Siluriformes: Siluridae), an endemic species of Lake Biwa, was investigated on the basis of field observations on 150 nights at the spawning site, the rocky shore of the Seta River, the lake's outlet, and environmental data from April to July in 1989–1994. The catfish gathered at the spawning site when the water level of the river rapidly rose and the rocky shoal became submerged, within a wide range of water temperature (15.0°–30.0°C); their spawning was induced in relation to increases in the submerged portion of the rocky shoal. This observation suggests that water level control operations in Lake Biwa might critically affect the spawning of this and other fish species. Received: December 17, 1999 / Revised: September 13, 2000 / Accepted: January 16, 2001     

Reproductive and sexual characteristics in five Lethrinus species in waters off the Ryukyu Islands

Gonads of five lethrinids, viz., Lethrinus harak, L. miniatus, L. obsoletus, L. ornatus, and L. sp. 2, were collected monthly in waters off the Ryukyu Islands and observed histologically to reveal their spawning periods and size at sexual maturation and sexual transition. The spawning period was from April to November for L. harak, from April to July for L. miniatus, from April to October for L. obsoletus, from May to November for L. ornatus, and from April to October for L. sp. 2. Sexual patterns in the five species were determined by fork length at both sexual maturation and sexual transition. The body size (fork length, FL) and sex ratio (% of female) at 90% maturity in L. harak were 21.1 cm FL and 90%, respectively; those of L. miniatus were 42.2 cm FL and 80%; 25.7 cm FL and 60% for L. obsoletus; about 20 cm FL and 90% for L. ornatus; and about 26 cm FL and 90% for L. sp. 2. Because the sex ratios decreased to 0% at the maximum size classes in L. miniatus, L. ornatus, and L. sp. 2, the sexual patterns in these species were considered to be protogynous hermaphrodite. Although the sex ratio (% of female) once decreased to about 30% at 28 cm FL, rapid increase occurred in the larger size class in L. harak. However, the increase did not result from sexual transition. Hence, the sexual pattern of L. harak was considered to be protogynous hermaphrodite. Because the sex ratio (% of female) at body sizes larger than 23 cm FL was stable at about 60% in L. obsoletus, the sexual pattern was determined to be one of juvenile hermaphrodites.     

Microhabitat use by the larvae of cryptic lamprey species in <Emphasis Type="Italic">Lethenteron reissneri</Emphasis> in a sympatric area

Microhabitat use by the larvae of two cryptic lamprey species in Lethenteron reissneri, Lethenteron sp. N (northern form) and Lethenteron sp. S (southern form) larvae, as well as intraspecific variations on a size-related basis, were examined in a study area (500 × 5 m) located on a stream where they showed sympatric distribution. The total number of individuals collected from each of 31 quadrats (60 × 60 cm) ranged from 0 (0 n/m²) to 13 (36.1 n/m²). The distinct difference in habitat use was not observed between the two species in the present study area. Lamprey density was not significantly independent among the size groups (small and large) for each of Lethenteron sp. N and Lethenteron sp. S. The density of each size group of the two species was significantly correlated with some habitat variables, and stepwise multiple regression analysis revealed a positive association for a substrate with a smaller particle size and deeper depth for large individuals of Lethenteron sp. N and Lethenteron sp. S, respectively, suggesting that these conditions would be suitable for burrowing. On the other hand, small individuals of both species tended to occur in areas where there was a high content of medium sand and a faster current velocity, probably reflecting a preference for respiratory efficiency. Various condition of habitat should be maintained to conserve the cryptic lampreys.     

Spawning season and migration of the myctophid fish <Emphasis Type="Italic">Diaphus theta </Emphasis>in the western North Pacific

 The spawning season, spawning grounds, and migration of the myctophid fish Diaphus theta were studied in the western North Pacific, based on seasonal sampling and estimation of hatching dates. The peak abundance of larvae was observed in July in the transition waters between the Oyashio and Kuroshio fronts. The spawning season ranged from late March to early September, with a peak from May to July. Larvae and juveniles <40 mm in standard length were distributed in the transition waters, whereas larger individuals were collected in the Oyashio and the Western Subarctic waters. These results indicate that this species undergoes a horizontal spawning migration from the Oyashio and Western Subarctic waters into the transition waters crossing the Oyashio front. Received: July 11, 2002 / Revised: October 2, 2002 / Accepted: October 15, 2002 Acknowledgments We are grateful to the captains, officers, and crew of FRV Hokko-Maru, FRV Tankai-Maru, and RV Hakuho-Maru for their assistance at sea. We thank Dr. H. Saito and Mr. H. Kasai for their cooperation in field sampling, and Dr. C. Sassa for his assistance with larval fish taxonomy. Correspondence to:Masatoshi Moku     

Lake water level increase during spring affects the breeding success of bream Abramis brama (L.)

In Lake Constance, Eurasian bream Abramis brama (L.) spawn in very shallow littoral areas by the beginning of May. They attach their adhesive eggs to pebble and cobble substratum at <40 cm depth. Increasing water levels before spawning inundate bare substratum to which bream eggs may attach better than to deeper substratum covered by epilithon. Consequently, the water level increase prior to spawning should determine the amount of pristine spawning substratum available to bream and thus influence their breeding success. In order to test this hypothesis, the influence of hydrology and climate on the abundance of age-0 bream was combined with the results from field investigations on the egg survival and abundance of age-0 bream. A strong positive correlation between the mean water level increase during the spawning season of bream (April–May) and the abundance of juvenile bream was found. In contrast, the absolute water level during spawning and during the nursery stage in summer, the cumulative temperature during the egg, larval and juvenile stages and two North Atlantic Oscillation (NAO) indices did not affect the abundance of juvenile bream. The field investigations confirmed that bream eggs attach better to and have higher survival rates on bare substratum than on substratum with epilithon cover. Accordingly, eggs within a spawning habitat of bream were most abundant between 10 and 20 cm depth, where the epilithon cover was lower than at depths exceeding 30 cm. The results of this study confirm an adverse influence of epilithon cover on the attachment and subsequent survival of bream eggs and emphasize the importance of spring inundations for the successful breeding of the bream. Handling editor: J. A. Cambray     

Effect of physical factors on the distribution of filamentous green algae in the Tama River

 The distribution of stream-specific filamentous green algae (SSFG) was investigated in the middle reach of the Tama River, Japan. Cladophora glomerata (L.) Kützing and Stigeoclonium sp. dominated among five taxa of SSFG collected in the riverbed with fast currents. These two species were abundant in shallow (<20 cm) riffles with high current velocity (>30 cm s⁻¹), a habitat characterized by high light intensity and extensive aeration. SSFG biomass fluctuated greatly (range, 1.2–85.3 cm² m⁻²), exhibiting no clear seasonal trend. The tufts of SSFG grown on riffle cobbles decreased rapidly when these cobbles were transferred to habitats with either deep water, low current velocity, or both. Laboratory experiments revealed that aeration retarded the decrease in SSFG in incubation vessels. These results suggest that the turbulent conditions of riffle habitats are important for the growth and survival of SSFG in the Tama River. Received: September 5, 2001 / Accepted: April 25, 2002     

Growth and reproductive cycle of the zebra mussel in the River Rhine as studied in a river bypass

In a Rhine River bypass built at a water pollution control station, the growth and reproductive cycle of the zebra mussel (Dreissena polymorpha) were studied over a period of nearly 3 years. The experimental setup offered the possibility to manipulate the temperature of the continuously passing river water and to calculate influences of temperature and food on the growth of individually marked mussels. Shell growth rates were a function of algal density (measured as chlorophyll a). This function followed saturation curves with temperature-dependent plateau levels, and saturation occurred at values above 40 μg/l chlorophyll a. The relationship between shell growth rate and temperature followed the courses of bell-shaped curves with algal-food-dependent heights. No clear correlations could be found between shell growth and other environmental factors such as dissolved organic matter, total organic content, seston, and turbidity. Two spawning periods per year characterized the reproductive cycle of the zebra mussels at the study site. The first one covered nearly 3 months (April to July) comprising several spawning events. The second shorter one occurred in August. Young mussels of the first spawning period settled at the end of May, and they were able to reproduce a few months later (end of August) when their shell length exceeded 9 mm. This first spawning period of the young-of-the-years coincided with the second spawning period of the older generations. Through the combined analysis of the results of both body weight determinations and shell growth experiments, it was possible to model mussel growth over a period of nearly 3 successive years, and to suggest the presence of an endogenous factor that might be involved in determining the start of the annual shell growth period. Received: 8 April 1997 / Accepted: 3 November 1997     

Spawning Periodicity, Spawning Migration, and Size at Maturity of Green Sturgeon, Acipenser medirostris, in the Rogue River, Oregon

The Rogue River, Oregon represents one of three important spawning systems for green sturgeon, Acipenser medirostris, in North America. In this paper we describe the spawning migration, spawning periodicity, and size at maturity for green sturgeon caught in the Rogue River during 2000–2004. Green sturgeon were caught by gill net or angling; 103 individuals were tagged with radio or sonic transmitters (externally or internally). Green sturgeon caught by gill net and angling ranged from 145 cm to 225 cm total length. Histological and visual examinations of gonad tissues indicated that most green sturgeon were spawning or post-spawning adults that entered the Rogue River to spawn. Ripe individuals were caught when water temperature was 10–18°C. Specimens carrying transmitters migrated 17–105 km up river; reaches consisting of likely spawning sites were identified based on sturgeon migratory behavior. Most green sturgeon remained in the Rogue River until late fall or early winter when flows increased, after which they returned to the ocean. Eight green sturgeon (males and females) returned to the Rogue River 2–4 years after leaving, entering the river during March, April, and May when water temperatures ranged from 9°C to 16°C. None of the 103-tagged individuals entered the Rogue River during successive years. There appear to be few known natural threats to adult green sturgeon in the Rogue River. However, our data suggest that a high percentage of adults that spawn in the Rogue River (particularly males) were susceptible to harvest by commercial, Tribal, and sport fisheries after leaving the system because they were not adequately protected by maximum size limits during the period of this study. The implications of maximum size limits (or lack of size limits) to green sturgeon are discussed, and recent actions taken by Oregon and Washington Fish and Wildlife Commissions to manage green sturgeon more conservatively are presented.     

Estimation of spawning sites in the spotted flagtail, Kuhlia marginata, based on sperm motility

To estimate the spawning site of the spotted flagtail, Kuhlia marginata, sperm motility was examined under conditions of different salinities. The spermatozoa were immotile in 0 and 5 ppt test solutions and in seminal fluid but lived longer in 20–30 ppt. Sperm were most active at 25–35 ppt. Therefore, we considered that spotted flagtail spawn in 20–35 ppt salinity. Because the mean salinity of the Genka estuary is below the optimum required for spawning, the spawning site in Okinawa Island would be seawater. This evidence suggested that the spotted flagtail is a catadromous species. Received: January 26, 2001 / Revised: July 6, 2001 / Accepted: July 24, 2001     

Displacement,velocity preference,and substrate use of three native California stream fishes in simulated pulsed flows

We studied whether juvenile fishes were able to maintain swimming speed and position during simulated river pulsed flows in a laboratory flume. We used a glass flume (15.24 × 0.6 m) with river-rock substrate to determine the longitudinal displacement, movement distances and frequencies, velocity selection, and substrate use of juvenile (SL range: 6.1 ± 0.2 cm) hardhead Mylopharodon conocephalus (n = 13), rainbow trout Oncorhynchus mykiss (n = 11), and Sacramento sucker Catostomus occidentalis (n = 12) during a 100-min flow pulse, as velocity changed from slow to medium, fast, medium, and slow. Fish were capable of maintaining swimming speed and position up to the maximum flume velocity of 0.46 m·s⁻¹, except for one hardhead that impinged on the rear fish screen. Fish swam faster in the flume during the medium and fast intervals than the slow intervals, but fish speeds were similar among the medium and faster intervals, when some fish took cover behind the rock substrate. In comparison with a Brett-type swim-tunnel, fish showed less increase in mean swimming speed as the flume velocity increased. Fish in the flume were able to use the rock substrate as hydraulic cover, decreasing the encountered water velocity, and, presumably, conserving energy.     

Growth, summer cohort output, and observations on the reproduction of brook silverside, Labidesthes sicculus (Cope) in the Kawartha Lakes, Ontario

The brook silverside, Labidesthes sicculus (Cope), is unique in Canadian waters, as it completes its life cycle in 1 year. Previous studies based on scale ageing had suggested the species was an “annual,” but we confirmed this for the first time by otolith analysis. Growth rates from both back-calculation, and the Gompertz model, indicated an asymptote near the end of the summer, and average summer growth rates of 0.77 and 0.70 mm/day, respectively. The Gompertz model gave the best fit (n = 201, r = 0.744) with an L∞ of 85.4 mm TL, and instantaneous growth rate, g, of 0.0264. Back-counting daily growth increments allowed us to show that broods of young fish were produced throughout the summer, from late May to mid August, with maximum hatch taking place in mid-July. The species is a “batch” (serial) spawner, with only a fraction of the eggs ripening in the ovary and being released at as yet undetermined intervals. Eggs of the larger immature fish in the first summer developed from 0.05 to 0.21 mm in diameter (preserved) by fall, and in mature fish of the following spring and summer, developed to 1.2 mm in diameter (preserved), 1.4 mm fresh, at spawning. Attached to each egg was a filament averaging 2.0 cm in length, adhesive in nature, and presumably for attachment to vegetation. The egg also had microscopic hairs on its surface. No evidence was found to support temperature-dependant sex determination, nor were embryos or sperm found in the ovaries of spawning females, unlike Labidesthes sicculus vanhyningi, (the southern subspecies) which has internal fertilization. The Canadian species possesses a genital papilla through which the eggs were released, and an apparently much smaller male genital papilla than the southern subspecies.     

Spawning population characteristics of pike <Emphasis Type="Italic">Esox lucius</Emphasis> L. in Lake Rubikiai (Lithuania)

The aim of the study was to describe the peculiarities of pike spawning in mesotrophic Lake Rubikiai, to determine the spawning population structure and evaluate the influence of some environmental factors on year-class strength formation. The data were collected in April and May (1994–2011). A total of 1586 individuals were caught. The age of pike ranged from 1 to 12 years; 2–5-yearold males (96.0%) and 3–8-year-old females (89.1%) prevailed. The overall sex ratio of females to males was 1:4.1. One-year-old spawning males (26.5–28.5 cm) and two-year-old spawning females (31.5–35.5 cm) constituted 2.1% and 2.6%, respectively. Water temperature during the spawning period was relatively stable, between 4 and 6°C, during March and increased slightly towards 10°C in the middle of April. No correlation was observed between female length and spawning date. Year-class strength did not correlate with the last day of ice presence and the minimal water level at the end of spawning (WL_min), but statistically significantly negatively correlated with the maximum water level at the beginning of spawning (WL_max) and the difference between WL_max and WL_min. The mean annual survival rate for pike (aged ≥2) was 0.74 and the mean annual instantaneous mortality rate was 0.45.     

Features of the reproductive ecology of the rock catfish Silurus lithophilus

Some aspects of the reproductive ecology of the rock catfish Silurus lithophilus (Siluriformes: Siluridae), a species endemic to the Lake Biwa system, were investigated from April to July in 1989–1994. Field observations were conducted at a rocky shoal, which was also a spawning ground of S. biwaensis, along the shore of the Seta River, the lake's outlet. Spawning of S. lithophilus occurred around midnight at shallow, rocky places on the shoal (5–70 cm in water depth) from early May to mid-July, starting earlier than S. biwaensis (mid-May to mid-July). Spawning tended to occur at low water temperatures irrespective of high-water events, unlike S. biwaensis, and in low densities (fewer than 8 fish per night) compared to S. biwaensis (1–45 fish per night). Although the rock catfish tended to appear and spawn at higher frequencies on nights when S. biwaensis spawned or appeared, it always avoided conflict with the latter species by spawning at sites far from those used by S. biwaensis or after S. biwaensis had left the vicinity. The survival of eggs of S. lithophilus might be favored by a strategy of diverting predator attention from them by synchronizing spawning nights with S. biwaensis. Received: February 27, 2001 / Revised: October 17, 2001 / Accepted: November 8, 2001     

Initiation of sperm motility depending on a change in external osmotic pressure in the noncopulatory marine cottid fish Gymnocanthus herzensteini

 The initiation of sperm motility in a noncopulatory marine cottid fish, Gymnocanthus herzensteini, was examined. The spermatozoa, which were immotile in seminal plasma, initiated motility at osmolalities of more than 500 mOsm kg⁻¹ in NaCl solution and 400 mOsm kg⁻¹ in KCl and mannitol solutions, indicating that the initiation of sperm motility depends on changes in external osmolality, in contrast with that of the sperm of other marine cottid fish, which are motile in seminal plasma. This study revealed that there are plural manner of initiation of sperm motility in marine cottid fish, which are oviparous but include both copulatory and noncopulatory modes. Received: May 24, 2001 / Revised: December 19, 2001 / Accepted: January 8, 2002     

Competition for spawning sites between two gobiid fishes, Bathygobius fuscus and Eviota abax, derived by alteration of mating tactics of the former

Interspecific competition for spawning sites between two gobiid fishes, Bathygobius fuscus and Eviota abax, was studied on a rocky shore. Large males (LM: 55–80mm in standard length: SL) and small males (SM: 34–52mm SL) of B. fuscus acted as nest holders and sneakers, respectively, in the early spawning season (July). The sympatric male E. abax (22–33mm SL) was smaller than SM B. fuscus, and synchronically used rocky holes much smaller than those of LM B. fuscus. In this season, aggressions between the two species were rarely observed. In late season (August), as the number of the LM decreased, the SM converted their tactics to nest holding, occupying nests of a sizes similar to those of E. abax. Eviota abax males were dispossessed by SM and shifted their nest sites to cavities between cobbles and sandy bottom, which might be poorer nest sites than rocky holes.     

Growth and tissue mass cycles in the infaunal bivalve Yoldia eightsi at Signy Island, Antarctica

Shell growth in Yoldia eightsi was measured over an austral summer and winter in 1992. In specimens < 12 mm length, growth was not significantly different between summer and winter periods, and the fastest recorded rate, 6.3 μm day⁻¹ was for 5-mm individuals during the winter. In summer, specimens of all lengths grew significantly, but in winter bivalves > 27 mm length did not increase in length. Tissue dry and ash-free dry mass (AFDM) cycles were assessed at monthly intervals between December 1988 and January 1991. ANCOVA indicated significant interannual and seasonal effects on this cycle. Tissue mass increased in the summer, coinciding with the phytoplankton bloom and the period of maximum sedimentation of organic material from the water column. A standard 20-mm-length animal reached a maximum AFDM of 114 mg in February 1990. The minimum value (68 mg AFDM) throughout the 2 years of measurements was in early December 1988, at the end of the austral winter. Periods of tissue mass increase were, therefore, decoupled from shell growth, at least in juveniles. Tissue mass was significantly higher in 1990 than 1989, which was mainly due to high organic contents in the summer (January to May). This was not consistent with the pattern of organic content in the sediments at the study site, but was in phase with the cycle in sediment chlorophyll a content. Tissue mass increase depended on major resource input during the summer, but Y. eightsi was capable of maintaining winter condition from stocks of benthic microalgae in years of poor ice cover. Tissue mass declined between April and July each year. This was accompanied by large falls in tissue ash content, and coincided with the spawning period in early June. These are the first monthly tissue mass data collected over a 2-year period for an Antarctic mollusc. They are the first such data indicating seasonal variation in tissue mass and showing a decoupling of shell and tissue growth in a polar bivalve. The P/B ratio calculated from these data was 0.106, which is slightly lower than previous values found for this species, but is in line with general values for Antarctic marine benthos. Accepted: 6 December 1999     

Seasonal variations in biomass, abundance and composition of zooplankton in the subarctic waters north of Iceland

The seasonal variations in biomass, abundance and species composition of zooplankton in relation to hydrography and chlorophyll a were studied in the subarctic waters north of Iceland. The sampling was carried out at approximately monthly intervals from February 1993 to February 1994 at eight stations arranged along a transect extending from 66°16′N–18°50′W to 68°00′N–18°50′W. The mean temperature at 50 m depth showed a clear seasonal pattern, with lowest water temperatures in February (∼1.1°C) and the highest in July (∼5.4°C). The spring growth of the phytoplankton began in late March and culminated during mid-April (∼7.0 mg Chl a m⁻³). Both the biomass and the abundance of total zooplankton were low during the winter and peaked once during the summer in late May (∼4 g m⁻² and ∼38,000 individuals m⁻²). A total of 42 species and taxonomic groups were identified in the samples. Eight taxa contributed ∼90% of the total zooplankton number. Of these Calanus finmarchicus was by far the most abundant species (∼60% of the total zooplankton). Less important groups were ophiuroid larvae (∼9%), Pseudocalanus spp. (∼8%), Metridia longa (∼4%), C. hyperboreus (∼3%), Acartia longiremis (∼2%), chaetognaths (∼2%) and euphausiid larvae (∼2%). The dominant copepods showed two main patterns in seasonal abundance: C. finmarchicus, C. hyperboreus and C. glacialis had one annual peak in numbers in late May, while Pseudocalanus spp., M. longa and A. longiremis showed two maxima during the summer (July) and autumn (October/November). Ophiuroid larvae and chaetognaths (mainly Sagitta elegans) peaked during the middle of July, while the number of euphausiid eggs and larvae was greatest from May to July. The succession in population structure of C. finmarchicus indicated its main spawning to be in April and May, coincident with the phytoplankton spring bloom. A minor spawning was also observed sometime between August and October. However, the offspring from this second spawning contributed only insignificantly to the overwintering stock of C. finmarchicus. Received: 12 September 1997 / Accepted: 1 March 1998