Genus Rutilus
Rafinesque, 1820

The roaches are found in Europe and western Asia where there are about 15 species (Bogutskaya and Iliadou, 2006). Two species are found in Iran.

The genus is characterised by having pharyngeal teeth in one row, usually 6-5, more rarely 6-6 or 5-5, with conical crowns on the anterior teeth and posterior teeth slightly hooked and truncated, scales are large to moderate in size, numbering 33-68 and with numerous fine circuli and radii on all fields, few and short gill rakers (17 or less), short gut, usually a light peritoneum, few to moderate numbers of dorsal and anal fin rays (7-13), the dorsal fin commonly having 4-5 unbranched rays, abdomen behind the pelvic fins rounded or with a slight but scaled keel, and various osteological characters (Bogutskaya and Iliadou, 2006).

Rutilus frisii
(Nordmann, 1840)

Common names

ماهي سفيد (= mahi safid or safid mahi, meaning white fish), sifid mahyi and asbalan mahi (in Gilaki), mahi safid daryacheh khazar or mahisephid-e-daryaye khazar (= Caspian Sea white fish), talaji (in Mazanderani).

[kutum, ak-balyk (= white fish) or ziyad in Azerbaijanian; akbalyk or kutum in Turlmenian; kutum in Russian; Southern Caspian roach; pearl roach; Caspian kutum].

Systematics

Leuciscus Frisii was originally described from the market in Odessa and the Danube, Dniester, South Bug, Dnieper and Don rivers, draining to the Black Sea.

Rutilus frisii kutum Kamenskii, 1901 is the generally accepted Caspian Sea basin subspecies although J. Holčík (pers. comm., 1994) considers that this is not a good taxon (see also Holčík and Jedlička, 1994). It was originally described as Leuciscus Frisii var. kutum from the Caspian Sea, essentially in the southern part, spawning in the rivers and streams of Transcaucasia and Persia (Kura, Araxes streams of the Lenkoran district) and in lesser numbers elsewhere. This subspecies is distinguished from the type subspecies from the Black Sea by having fewer lateral line scales (about 55-58 versus about 60-66), shallower body (depth equal to or less than head length versus exceeding head length), anal fin longer than high versus shorter than high, lower lobe of caudal fin usually shorter than head versus longer (young R. f. kutum have this lobe as long as head), and dorsal fin as high as long versus higher than long (young R. f. kutum have dorsal fin higher than long).

Leuciscus frisii caspius Lönnberg, 1900 described "from the Volga delta" has priority over Rutilus frisii kutum which may be a nomen nudum as it is listed in Radde (1899) as "Leuciscus Frisii Nordm. var. Kutum Kam." without a description, the description only appearing in Kamenskii (1901). However, the name caspius has not been used while kutum appears widely in the literature as well as being the common Russian name of the fish. Additionally, as mentioned above, there may be no need of a subspecific name. Kottelat (1997) considers kutum to be a distinct species as do Bogutskaya and Iliadou (2006) and Fricke et al. (2007).

Leuciscus friesii Kessler, 1870 from the Volga delta is presumably a misspelling.

A hybrid of this species and Ctenopharyngodon idella has been bred at the Astaneh Ashrafie Fisheries Research Station (Sefidrud Research Station) and named "Samur" (Iranian Fisheries Research and Training Organization Newsletter, Tehran, 11:6, 1996; 18:6, 1997; Khara et al., 2002; Nouruz Fashkhami et al., 2002). Gynogenesis may have occurred. Artificial hybrids with Rutilus rutilus and Abramis brama have been bred in Iran (Annual Report, 1994-1995, Iranian Fisheries Research and Training Organization, Tehran, p. 39-40, 1996).

Key characters

This species is distinguished from the related Rutilus rutilus by the higher scale count (usually >50) and the posterior part of the swimbladder being elongate and conical or pointed rather than rounded.

Morphology

Dorsal fin with 3 unbranched and 8-10, usually 9, branched rays, anal fin with 3 unbranched and 9-12, usually 10, branched rays, pectoral fin branched rays 16-19 and pelvic branched rays 8-9. Lateral line scales 47-68, mostly 55-58. Scales are regularly arranged over the body. A pelvic axillary scale is present. Scales have numerous circuli, numerous posterior radii, few but distinctively crowded anterior radii (more than in Rutilus rutilus) and an almost central focus which is broken up into a network of lines. The posterior scale margin is crenulate and the anterior margin is wavy to crenulate but scale margins vary greatly between individual scales. Total gill rakers number 7-12 and are very short, hardly reaching the one below when appressed. Pharyngeal teeth usually 6-5, crowns rounded above a slender stalk, posterior teeth with a weakly hooked tip, posteriormost tooth margin may be serrated. The gut is an elongate s-shape. Chromosome number 2n=50 (Annual Report, 1994-1995, Iranian Fisheries Research and Training Organization, Tehran, p. 43, 1996; Noruz Fashkhani and Kosroshahi, 1995; Klinkhardt et al., 1995). Adibmoradi and Sheibani (2002) carried out a histological study of the brain of this species and concluded, in part, that vision was sharp and the fish were good chasers of food.

Meristic values for Iranian specimens are:- dorsal fin branched rays 8(1), 9(58) or 10(1); anal fin branched rays 9(24), 10(35) or 12(1); pectoral fin branched rays 16(15), 17(30), 18(14) or 19(1); pelvic fin branched rays 8(49) or 9(11); lateral line scales 47(1), 49(1), 50(1), 51(1), 52(7), 53(8), 54(6), 55(5), 56(11), 57(7), 58(7) or 59(5); total gill rakers 8(1), 9(21), 10(25) or 11(13); pharyngeal teeth 6-5(23), 5-5(1) or 5-4(1); and total vertebrae 40(1), 41(7), 42(34), 43(3) or 44(2).

Sexual dimorphism

Females are larger than males. The breeding tubercles are evident in males and may develop as early as the end of summer before the spring spawning season in the following year. Large tubercles are found on the top of the head above the eye level, behind the eye, between the eye and the mouth and on the snout as well as on upper flank scales. Tubercles are white in contrast to the dark head.

Colour

Small specimens up to a year old are silvery on the flanks and belly and the back is steel-grey to pale brown or pale olive. In adults, back scales are circled with black and there is a strong contrast between the back and flank. The anterior part of each flank scale, particularly those of the lateral line, is darkly pigmented. Lateral line scales may have two dots, one above and one below the opening as in Alburnoides bipunctatus but not as pronounced. The sides of the head are silvery with some yellow and darker pigment, the latter particularly in front of the eye. Adults are a bright silvery on the flank. The belly is pearly-white. The iris is silvery, with some spots, and with a marked dark spot above. The dorsal and caudal fins have some grey and a faint orange tint while the pectoral, pelvic and anal fins are colourless to lightly pigmented with black. The pectoral fin may be orange.

Size

Attains 66 cm body length and 4.065 kg in Iran (Farid-Pak, 1968a). Ouseley (1819-1823) ate one almost 3 feet long (ca. 0.9 m). In Iran during the 1950s, catches were 36-67 cm long (Farid-Pak, no date). Reaches 7 kg in weight.

Distribution

Found on the northern shores and drainages of the Black Sea, and the western shores and drainages of the Caspian Sea. Other populations are found in northwestern Turkey, Bulgaria and other parts of Europe, some subspecifically distinct from Rutilus frisii kutum. In Iran, it is found along the whole Caspian coast, entering almost all the rivers to spawn and also the Anzali Mordab and Gorgan Bay (Nedoshivin and Il'in, 1929; Kozhin, 1957; Holčík and Oláh, 1992; Riazi, 1996; Oryan et al., 1998; Abbasi et al., 1999; Kiabi et al., 1999; Abdoli, 2000).

It is also found in Valasht Lake, Mazandaran (landlocked and introduced about 1850 to provide food for the royal family which had a summer residence there) where it possibly hybridises with native Alburnus chalcoides (Armantrout, 1980).

von den Driesch and Dockner (2002) record this species as faunal remains in an excavation at the Great Mosque in medieval Siraf on the Persian Gulf coast. This may well be a misidentification but, if not, indicates that a fondness for this species has a long history considering that a preserved fish or its remains had to be transported all the way from the Caspian Sea.

Zoogeography

A European and western Asian species with its origins in a Danubian or Sarmatian fauna.

Habitat

This species is migratory, spawning in rivers in March-April and returning to the Caspian Sea. Rezavi (1997) found three populations in Iran, one autumn and two spring populations. Il'in (1927a) records schools of fish entering the Anzali Bay as numbering in the several tens of thousands. On one fishing ground next to the Djifrud in February 1914, 12,000 fish were taken simultaneously. However, while the majority of fish migrate into lowland rivers not far from the sea to spawn among bulrushes and cattails, some migrate far upstream into the mountains of Tavalesh (= Talish) and Gilan where spawning conditions are very different. These fish may well overwinter in the river (Derzhavin, 1934). Such fish reached altitudes of about 1000 m before environmental changes inhibited the migration. Young fish migrate downstream, the migration ending in August, and in Azerbaijan enter the sea within 20-50 days. Holčík (1994; 1995) stated that Iranian young from the Anzali Mordab never entered the sea but remained in fresh or brackish water for 1-2 years. Riazi (1996) reports that this species migrates into the Siah-Keshim Protected Region of the Anzali Mordab. Young fish descend the Atrak River and feed in the sea at depths of 4.0-8.5 m. They winter in Iranian waters (Savenkova, 1985). Some fish are reported at depths of 36.6-53.0 m in the Iranian Caspian Sea (Knipovich, 1921).

Age and growth

Azari Takami et al. (1990) described the biology of this species in Iran. Males normally mature between their third and fourth year, sometimes earlier, females during their fourth year. Life span is at least 9 years in Dagestan (Shikhshabekov, 1979) and 8 years in Iran (Holčík and Oláh, 1992). Spawners are 3-8 years old and the principal age groups are 4-5 years for males and 5-6 years for females in the Anzali Mordab (Holčík and Oláh, 1992; Holčík, 1995), older than the 3-4 years of fish reported in 1970-1971. However, recently males are maturing at age 2 and females at age 4 with most spawners at age 3 and 4 years respectively (Holčík and Oláh, 1992). The average size of mature females (700 g) has been decreasing (Bartley and Rana, 1998b). Males grow faster than females until about the third year of life and then more slowly. Males are smaller and have a shorter life span than females (Holčík and Oláh, 1992). An annual increase in length of 17% over the first four years was observed in tagged fish, decreasing to 4.7% annually in subsequent years (for weight the figures are 61.2% and 25.7%). Sex ratios in spring were 3.08 males:1 female and in autumn 0.78 males:1 female (Annual Report, 1995-1996, Iranian Fisheries Research and Training Organization, Tehran, p. 55, 1997).

Fish may spend 1 or 2 years in fresh water after hatching (Holčík and Oláh, 1992) and this affects growth, 1 year fish growing more quickly than those spending 2 years in fresh water, maturing earlier and having a shorter lifespan. Growth in the Anzali Mordab is slower than it was 20 years prior to the report of Holčík and Oláh (1992), probably owing to a higher population density and more competition for food resources. The commercial catch in Iran was 3-7 years old, 39.0-57.1 cm long and weighed 613-2525 g (Razivi et al., 1972).

Growth in Iranian fish farms is up to 1.5 g in 8 weeks after a hatching rate of 75% (Aslaanparviz, 1994).

Food

Zarbalieva (1987) provides data on feeding of this species in the Caspian Sea off Azerbaijan. Fish concentrate on a sandy-shellrock bottom and remain there for most of the year to feed. The crab Rhithropanopeus harrisii dominates in the diet, 67.9-93.7% by weight. Molluscs, mainly Cerastoderma lamarckii, comprise 30% by weight of the food of fish 30-40 cm long. Fish larger than 40 cm seldom take molluscs but occasionally Clupeonella spp. Molluscs used to be the main diet item of this fish. Juveniles in the Anzali Mordab of Iran feed mostly on phytoplankton in contrast to the zooplankton reported for Azerbaijan fish (Holčík, 1995). This is a consequence of the poor productivity of this lagoon. Small Iranian specimens from the Caspian Sea had bivalve shell remains, plant remains and in one case a worm. Oryan et al. (1998) state that Cardium is the main food of this species on the eastern and western coasts of Bandar Anzali. Crabs and Balanus are also important. The hepatosomatic index is highest in February and March, the prespawning period.

The hybrid with Ctenopharyngodon idella fed principally on macrophytes with phytoplankton as a secondary food in pond sin Gilan (Khara et al., 2002).

Reproduction

Fish spawn annually according to an Iranian study by Azari Takami et al. (1990) and probably return to their river of birth. Temperature, and probably river flow, are the factors determining the entrance of fish into the rivers on the spawning migration. Male fish run before females. Adeli Mosabbab and Piri (2005)0 report spawning in Iranian rivers from the end of March to the middle of April, beginning at 10°C. A large female, 60 cm long, 3.24 kg and 7 years old contained 124,712 eggs but some large females may have over 250,000 eggs (see also Farid-Pak (1968a) where fish 38-40 cm long have a mean value of 53,900 eggs, ranging up to 174,400 for fish 62-64 cm long; Abdurakhmanov (1962) gives a fecundity of 290,000 eggs for fish in Azerbaijan). Maximum egg diameter is 2.0 mm (Farid-Pak, 1968a). There are two spawning migrations, winter and spring, and so two forms or stocks of this species (Azari Takami et al., 1990). There may be three stocks based on electrophoretic studies of blood proteins associated with the two spawning migrations (Annual Report, 1994-1995, Iranian Fisheries Research and Training Organization, Tehran, p. 41, 1996). The winter form enters rivers with emergent or submerged aquatic plants at the end of autumn or the beginning of winter. Eggs are deposited on these plants. The principal rivers were the Nahang Roga, Pir Bazar Roga, Sowsar Roga, and Anzali Roga (all draining the Anzali Mordab) but the stock has declined with overfishing and destruction of the habitat. The Safid River is listed as the main spawning ground by the Caspian Sea Biodiversity Database (www.caspianenvironment.org). The spring form enters rivers at the end of winter or the beginning of spring and spawns on gravel or sand. Adults return downstream after spawning, in the early morning. The run occurs both by day and night but increasing water clarity near the end of the spawning season limits runs to the night.

The migration into the "Dinachal" (? = Denya Chal) River occurred on 1 November in 1975, when sexually immature fish were caught. By 11 February, 5 spawners were caught and the peak migration occurred on 26 March when 3,845 fish were caught at a water temperature of 10.5°C. The migration was continuous until 11 May. Water temperatures between 28 March and 23 April in 1976 and 1977 varied between 9.1°C and 13.5°C when peak migrations occurred in the "Havyg" (= ? Haviq) River. Here the spring form of this species started to migrate at 6°C and the optimum water temperature appears to be between 11 and 13°C.

The main migration (99%) in the Anzali Mordab began in February and lasted 3 months at 8-10°C with a much smaller run (only about 1%) in November-December. The Mordab is now overgrown with reeds or has a silty, vegetated bottom and is no longer a spawning site (Holčík and Oláh, 1992), an economic and poetic loss. Holčík and Oláh (1992) translate Il'in (1927a) on Anzali Mordab mahi safid "During this time the reeds rustle as if of wind as the fish rubs pushing their flanks to the reed stalks or pushing forward among reeds standing in rows".

Khaval (1998) reports a spawning migration into the Safid River despite construction, sand removal and pollution. The migration peaks in the second half of March at a water temperature of 11°C.

Each female is flanked by two males at spawning in shallow water. Sometimes the backs of the fish protrude from the water. Females sharply rub their lower abdomen and pectoral area on the gravel bottom, males and females convulse, and eggs are shed and fertilised. Eggs are adhesive and hatch in 20 days at 10-14°C. Embryonic development in glass incubators at 14-16°C is described by Parivar et al. (1993). Cleavage and gastrulation begin within 24-30 hours, hatching at 216 hours (10 days after fertilisation) and resorption of yolk is complete at day 16. Some males and females are injured through contact with the gravel stream bed. Embryonic development takes 10-15 days at 8-16°C and 5-6 days at 20°C (Aslaanparviz, 1994). Young fish go down to the sea in summer according to some works although Holčík and Oláh (1992) believe that they stay in fresh water or brackish river mouths for 1-2 years before entering the sea.

Kousha et al. (2007) examined some aspects of reproductive endocrinology in this species, specifically on sex steroid-binding protein in plasma.

Parasites and predators

Mokhayer (1976b) records the digenetic trematodes Aspidogaster limacoides and Asymphylodora macracetabulum. Eslami and Kohneshahri (1978) report various helminths from this species in Iranian waters. These are the monogenean Diplozoon paradoxum, the aspidogastrean Aspidogaster limacoides, the digenean Asymphylodora kubanicum and larvae of the nematode Anisakis sp. This fish therefore is potentially a source for human infection with Anisakis. Mokhayer (1989) reports metacercariae of the eye fluke, Diplostomum spathaceum from this species in Iran, which can cause complete blindness and death in commercially important species. Jalali and Molnár (1990a) record the monogeneans Dactylogyrus frisii from this species in the Khosk (= ? Khoshk) River (Caspian basin) and D. rarissimus and two monogenean species, Dactylogyrus spp., from the Safid Rud. Jalali and Molnár (1990b) record the monogeneans Dactylogyrus frisii and D. rarissimus at fish farms in Iran. Molnár and Jalali (1992) report the monogeneans Dactylogyrus suecicus, D. haplogonus and D. turaliensis from this species in the Safid Rud. Gussev et al. (1993b) record the monogenean Dactylogyrus haplogonus from this species in the Safid Rud. Masoumian and Pazooki (1998) surveyed myxosporeans in this species in Gilan and Mazandaran provinces, finding Myxobolus bramae. Safari and Khandagi (1999) record Clostridium botulinum from 2.2% of fresh and smoked samples of this species in Mazandaran Province. Shamsi and Jalali (2001b; 2001c) detail monogenean parasites for Safid Rud and Caspian Sea samples, including 6 species of Dactylogyrus. Jalali et al. (2005) summarise the occurrence of Gyrodactylus species in Iran and record G. prostae and G. sp. in fish from the Safid River.

Amini (2006) records Gyrodactylus sp., nematodes, Diplostomum sp., Trichodina sp., Aeromonas sp. and Hydrophila sp. in fingerlings from hatcheries in Iran. The latter two parasites caused a heavy mortality in the Shahid Rajaii hatchery ponds.

Tavassoli and Moghir (20020 describe a squamous cell carcinoma in the oral cavity of this species, the first record for the Caspian Sea.

Economic importance

Safid mahi is the most popular fish in Iran with the highest economic value (Azari Takami et al., 1990). It may be available out of season as mahi qachaq or bootleg fish, an indication of its popularity. Batmanglij (1999) notes that this fish is baked and served with herbed rice. A stuffing consists of garlic, parsley, tarragon, scallions, coriander, mint, ground walnuts, barberries, raisins, lime juice, salt and pepper, sautéed in butter (he also incorrectly refers to mahi-e safid as striped bass). In 1996 a news report stated that "In the port city of Anzali, one small white fish sells for $10" (http://www.iran-e-azad.org/english/boi/03400201.96). In Rasht, 60% of the inhabitants consume no other fish and "86% of ladies in Rasht can only cook one kind of seafood" (Khairkhah, 1994). Ouseley (1819-1823) reported that this species "seemed most abundant, and was found in all the great rivers of this country near the sea; for several days it had furnished the principal dish of my dinners and often of my breakfasts". Holmes (1845) recorded catches with cast-nets in the Anzali Mordab over 150 years ago which were even then worth £1400-1500. O'Donovan (1882) reported that a small stream in the Atrak River drainage was so crowded with this species that individuals could only move by floundering and jumping over one another, and the horses in crossing the stream, trod them to death by scores. Lönnberg (1900) reported an annual catch of some millions in Persian waters.

It is caught in rivers and lagoons and by large, mechanically hauled beach nets in the Caspian Sea. Sea nets can be 1500 m long and 18 m deep. The fishing season begins in October and reaches a maximum between 20 February and 10 March, ready for the "Now Ruz" or New Year celebrations when many Iranians eat this fish with rice (Emadi, 1979).

The roe of this species is also eaten, salted or unsalted (the fish are called asbalan mahi and the roe shur-e asbal in Gilaki, the local dialect of Gilan). Whole female fish are soaked for one year in a mixture of salt and madder in special clay jars. The jars are traditionally buried in the ground and hermetically sealed but a modern technique has the roe removed from the fish after 20 days. Esmaeilzadeh et al. (2004) studied the nutrient composition and marinade qualities of this fish and compared them to those for grass carp (Ctenopharyngodon idella), the latter being preferable according to the organoleptic properties. The marinades could be stored for 6 months at 10ºC.

The average weight of fish caught in the Anzali Mordab in the early years of this century was 2 kg, the catch in the 1914-1915 season was 47,000 fish and in 1913-1915 123,000 fish. On 26 February 1914 a single haul in the sea near Anzali took 41,045 fish (Il'in, 1927a; Berg, 1948-1949). Il'in (1927a) estimated a yearly catch in Anzali Bay to be 3-4 million fish, a figure conflicting with these other reports.

Nevraev (1929) gives catches for various fishing regions in Iran in the early twentieth century. For the period 1901-1902 to 1913-1914 the catch in the Astara region was 0 to 29,053 individuals, for 1901-1902 to 1917-1918 the catch in the Anzali region was 2565 to 124,195 individuals, in the Safid River region from 1904-1905 to 1917-1918 the catch was 100 to 31,799 individuals, and in Astrabad (= Gorgan) region from 1900-1901 to 1912-1913 the catch was 4000 to 323,500 individuals. The total catch for Iran in the 1914-1915 season was 443,000 fish. The catch in Iran from 1956/1957 to 1961/1962 varied from 197,884 kg to 2,066,580 kg (Vladykov, 1964), from 1965/66 to 1968/69 it varied between 159 and 1252 tonnes (Andersskog, 1970), from 1963/64 to 1968/69 it varied between 121.3 and 1252 t (RaLonde and Walczak, 1970b; 1972), from 1987 to 1991 it varied between 3500 t and 8855 t (Holčík and Oláh, 1992), and between 1989 and 1998 it varied between 11,792 kg and 14,336 kg (Caspian Environmental Programme, 2001a) - catch figures are at variance with each other. Holčík and Oláh (1992) report a catch of 3107 kg in the Anzali Mordab for 1990 and for 1932-1964 a range of 95.1-3488.9 t. They are also caught in rogas and inflowing rivers of the mordab in late winter and early spring. Moghim et al. (1994) estimate that coastal areas of the southern Caspian Sea have a total biomass of 24,000 t with a maximum sustainable yield of 7000 t. In 1993-1994 the total catch of this species, including the illegal catch, was 11,175 t with the total stock estimated at 25,400 t and a maximum sustainable yield of 9300 t. More than 25% of the catch was young fish indicative of non-standard methods being used (Annual Report, 1995-1996, Iranian Fisheries Research and Training Organization, Tehran, p. 19-20, 1997). In 1994-1995, the biomass of this species in Iran was 241,000 t (sic, probably 24,000 t) and the maximum sustainable yield was 9000 t (Annual Report, 1994-1995, Iranian Fisheries Research and Training Organization, Tehran, p. 37, 1996). About 62% of the bony fish catch in the Caspian Sea of Iran in 1993-1994 was this species with the mullet Liza aurata second at 22% (Annual Bulletin 1993-94, Iranian Fisheries Research and Training Organization, Tehran, p. 83, 1995). The catch from beach seine cooperatives along the Iranian coast was 8477 t in in 2004-2005, a 2500 t decrease over the previous year. The biomass of this species in Iranian coastal waters for 2003-2004 was estimated art about 25,000 t (Abdolmalaki, 2006a).

Bartley and Rana (1998a; 1998b) comment that the fishery collapsed in 1980 but has risen from 500 tonnes in 1981 to around 10,000 t in 1996 after restocking from around 400,000 fingerlings/year in 1981 to around 142 million/year in 1997. Rana and Bartley (1998a) note that 7 million 1 g fingerlings were released into the Caspian Sea in 1997 which contradicts their earlier report. There is only a state supported stocking programme but ERM-Lahmeyer International GmbH, DHI Water & Environment and GOPA Consultants (2001a) note that this fish is being successfully managed in Iran while it has been fished almost to extinction in waters off Daghestan and Azerbaijan. Market price is high at about U.S.$5.00/kg, five times the price for silver carp. Export prices are higher, 700 g of smoked "mahi-sefid" cost £21.00 in 2004 (www.superhormuz.com, downloaded 19 January 2004). However this species is difficult to culture beyond the 40-50 g stage and growth is slower than carp.

The conflicting ranges seen in Andersskog and RaLonde and Walczak are typical of the wide variations in reports; figures can only be taken as general guides for many Iranian fisheries. Abdolmalaki (2006c) states that fluctuations are due to destruction of spawning grounds, overfishing and release of fingerlings. The mean catch sizes were 3110 tonnes in 1937-1947, 990 t in 1967-1977 and 8505 t in 1987-1997. A minimum catch of 121 t was taken in 1964 and a maximum of 11,175 t in 1994. Catch-per-unit-effort also showed high variation,. The calculated stock biomass was 1300 t in 1071 and between 18,489 and 25,400 t between 1990 and 2000. The mean biomass in the past 10 years was 22,750 t, a 17-fold increase over the year 1971. The catch in the previous 10 years was 35-46% of the annual stock. A decrease in stock size during 1998-1999 was attributed to an exploitation rate being more than maximum sustainable yield, a decreased mean weight of released fingerlings and a lowered return rate.

Catches in the Bandar-e Anzali area have been as high as 5,480 t in 1939-1940 but fell to 85 t in 1961-1962 (Vladykov, 1964; RaLonde and Walczak, 1972) but have risen again as indicated above. This is attributed to a massive stocking effort with 170 million fingerlings released in Iran in 1991 (cf. Emadi (1993a) below), about a quarter in the Anzali Mordab (Holčík and Oláh, 1992). The Sari hatchery produced 400 million "white fish" over the previous 10 years (Tehran Times, 30 May 1998). Fingerling production in 1996 was 142.1 million (Bartley and Sana, 1998a). Sea ranching increased the yearly catch to 8500 t in 1991, the highest recorded catch in the past being 5850 t in 1940 (Emadi, 1993a). Natural stocks in the past were very high. Migrating fish were so dense at the Mordab mouth that they were caught in buckets and jumping fish literally fell into boats.

Savenkova (1990) comments on the marketing of this species in southwestern Turkmenistan. Salehi (2003) analysed the economics of production of finglerings in Iran. The cost of labour was 50%, feed and fertiliser 20%, maintenance 10%, and harvesting, handling and releasing 6%. The average cost per fingerlings was 37 rials in 2001. As the average rate of fingerling return was 8.3% and the average weight and age of commercially caught fish was 815 g and 3.7 years, it was expected that 19,257,494 individuals weighing 16,000 t would be harvested over the Iranian year 2004-2005. The value would be an estimated 345 billion rials at the 2001 price (8050 rials to a U.S. dollar in 2001; $42.86 million).

Experimental culture of triploid mahi safid has been carried out at the Gilan Fisheries Research Centre (Iranian Fisheries Research and Training Organization Newsletter, 2:2, 1993) to increase fish weight for exploitation but apparently was unsuccessful although the techniques worked with grass carp, Ctenopharyngodon idella (Annual Bulletin 1993-94, Iranian Fisheries Research and Training Organization, Tehran, p. 70-71, 1995). In addition, monoculture and polyculture of this species has been investigated, the latter with grass carp (Ctenopharyngodon idella) and silver carp (Hypophthalmichthys molitrix) (Iranian Fisheries Research and Training Organization Annual Report, 1992-93; Danesh Khoshashi, 1997). From an average initial weight of 7 g, monocultured fish weighed 158 or 177 g on average at the end of each of two one-year periods, polycultured fish weighed 158 or 168 g, and maximum weight attained was 250 or 300 g (Annual Report, 1994-1995, Iranian Fisheries Research and Training Organization, Tehran, p. 38, 1996). Danesh Khoshasi (1997) gives figures of 177 g for mono-cultured fish after 6 months in ponds, poly-cultured fish weighed 185 g (with a maximum of 300 g) while stocking with 7 g fish gave 168 g and 192 g for mono- and poly-culture fish respectively (maximum weight 250 g)(these two reports on the same experiment have confusing figures). The fish were fed on pellets especially made for this species, nitrate and phosphate fertiliser and cow and chicken manure were added to the ponds, and water temperature was 15.0-27.64°C.

This species has been used in Iran for experimental studies, e.g. on the acute LC₅₀ and bioconcentration of mercuric chloride (Gharaei et al., 2006); on the toxicity and LC₅₀ of phenol and 1-naphthol (Shariati et al., 2004); the effects of anionic detergents and diazinon on the LC₅₀, both separately and in combination (Tehranifard et al., 2002; Tehranifard et al., 2007); etc.

The hybrid with grass carp, Ctenopharyngodon idella, reached an average weight of 100 g and a length of 22 cm after 5 months in the first report cited above and 6.7 g and 9.17 cm after 4 months in the second report. Larvae were fed live food twice a day and hybrids larger than fingerling size ate grass. The hybrid phenotype resembled the safid mahi but was a herbivore. It was expected that this hybrid would be used as a new culturable "species".

Robins et al. (1991) list this species as important to North Americans. Importance is based on its use in aquaculture and as food.

Conservation

Azari Takami et al. (1990) list the following reasons for a decline in the commercial catch of this species:- a) regression of the Caspian Sea which decreased the surface area of the Anzali Lagoon and increased the growth rate of aquatic vegetation, b) mechanisation of farming and a consequent increased demand for irrigation water leading to reduced river flows during the spawning migration, c) use of fertilisers and pesticides in rivers draining into the Anzali Mordab (and pesticides and herbicides such as Diazinon, Malathion, Machete and Saturn have a highly toxic effect on fingerlings of this species (Piri et al., 1999)), d) pumping of river water for irrigation causing mass mortalities of fry, e) industrial development increasing the pollution load, and f) excessive catches of adults to the extent that all spawners in a river were taken. Emadi (1979) added such factors as road construction and the removal of sand from banks and river beds, erosion caused by felling trees and shrubs along river banks and in the mountains, construction of bridges and raising of their substructures which formed barriers to migration, and climatic changes. Illegal fishing and non-standard nets threaten the stocks while fingerling release (120-140 million) and improvement of natural spawning areas through rises in water level have contributed to stock increases (Annual Report, 1995-1996, Iranian Fisheries Research and Training Organization, Tehran, p. 19-20, 1997). Three million fingerlings weighing 3-5 g were released into the Anzali Lagoon (Iranian Fisheries Research Organization Newsletter, 49:4, 2006).). Poaching is rampant, even on the spawning grounds, and is also a significant factor in decline of this species (RaLonde and Walczak, 1972).

In the 1970s, rivers were rented to fishermen to exploit to an unlimited degree such that all spawning fish were taken and there was no recruitment for 4 years (Carl Bond Archives, Oregon State University, Corvallis).

The catch declined from 5854 t in 1918 to 172 t in 1937. Actual catches are about 20-30% larger because of local sale and consumption (Emadi, 1979). RaLonde and Walczak (1970b) note a decline from 1556 t in 1957 to 162.1 t in 1967 so the catch does not decline evenly. Emadi (1979) also pointed out that intensive sea fishing and fishing in rivers hindered spawning of the winter form, and led to a situation where only the spring form spawned. Leasing of rivers to fishermen was discontinued in an attempt to alleviate the decline in this species. The migration distance up rivers has decreased from about 25 km to about 8 km in recent years because of water abstraction and dams for agriculture (Bartley and Rana, 1998b).

A further problem is the restocking programme only takes spring-run fish; the fall-run stock may no longer exist (but see below). In addition, stocks from various rivers are mixed and may result in outbreeding depression, the loss of adaptations to specific rivers in terms of migration patterns, spawning time, behaviour and other factors (Bartley and Rana, 1998b).

Measures to combat loss of this valued fish included a ban on fishing in the Anzali Mordab and its tributary rivers (catches here were up to 1000 t per year (Emadi, 1979)), effective control of illegal fishing and artificial spawning experiments (although the latter was insufficient to replace stocks). For several years starting in 1925, artificial breeding raised larvae for release in the ten most important rivers. In 1976-1977, fingerlings were raised to increase the stock in the Caspian Sea. Over 5 million fry were produced by the "Havyg" (? = Haviq) Hatchery alone in 1977.

Emadi (1979) recorded releases of 28-44 million fingerlings and 250 million larvae "in recent years". A farm in the Siah Kal region of Gilan Province near Rasht was expected to produce 40 million white fish "roe", presumably fry, in 1985 (Kayhan International, 20 May 1984). The Dr. Beheshti Hatchery near Rasht expected to release more than 60 million fingerlings in the Iranian year 1993-1994 (Abzeeyan, Tehran, 4(5):VI, 1993). This hatchery had a peak production for the period 1973-1993 of over 140 million fingerlings in 1989, rising from the low period of 1973-1982 with less than 10 million fingerlings, and with subsequent decreases to about 50 million fingerlings in 1993 (Abzeeyan, Tehran, 5(3 & 4):IX-X, 1994). The number of fingerlings released from 1986 to 1991 climbed from 38 million to 170 million (Holčík and Oláh, 1992). Krasznai (1987) also refers to propagation of this species at "Sad-e Sangar" (Dr. Beheshti) and Siah Kal Fish Farms near Rasht in Gilan. Emadi (1993a) gives fingerling production at government hatcheries as follows: 25.3 million in 1983, 28.3 million in 1984, 38.0 million 1985, 51.7 million in 1986, 72.0 million in 1987, 84.3 million in 1988, 140.2 million in 1989, 156.3 million in 1990, 110.0 million in 1991, and 145.0 million in 1992. The Shahid Rajaee Hatchery in Sari released 358 million fingerlings, possibly in a single year (Abzeeyan, Tehran, 4(7):VII, 1993), although 70 million are reported as released annually to Mazandaran rivers in 1995 (Abzeeyan, Tehran, 6(8):III, 1995) and production of 70 million kutum fingerlings annually is reported in 2001 (Iranian Fisheries Research Organization Newsletter, 28:3, 2001). In 1997, 142 million fingerlings were produced for restocking (Bartley and Rana, 1998b). In 1999-2000, 150 million juveniles were released into the Caspian Sea (Iranian Fisheries Research Organization Newsletter, 23:4, 2000). From October to March 2000, 80 million juveniles raised in the Shahid Ansari aquaculture and breeding centre in Gilan were released into the Caspian Sea and neighbouring water bodies (Iranian Fisheries Research Organization Newsletter, 26:2, 2001). Billard and Cosson (2002) cite a mean of 100 million alevins released per year, reaching 140 million in some years, and give a brief overview of production facilities. Mosabbab and Piri (2005) record the release of fry into the Gorgan River of the southeast Caspian Sea as 16,663 million in 2000, 19,119 in 2001, 12,263 million in 2002, 11,931 in 2003 and 10,413 in 2004, the decrease being due to a fall in capture of brood fish. Amini (2006) reports release of fingerlings into the Larim, Goharbaran, Shirood, Tonekabon, Sardabrud, Mirud, Babol, Asbuchin and Sorkhrud rivers in 2000-2001 numbering 53.7 million and 62.5 million. Fork length was 36.7 mm and 43.3 mm and condition factor 1.13 and 1.2 0 respectively by year. In 2006, a report has more than 150 million juveniles being released into the Caspian Sea every year (www.iranfisheries.net, downloaded 28 July 2006). Sources obviously conflict on exact numbers, nevertheless marked variations in production are evident over short periods. The South Caspian Fisheries of Azerbaijan also released larvae in Soviet waters, more than 150 million in 1983 for example (Zarbalieva, 1987).

The Inland Water Aquaculture Research centre in Anzali has artificially propagated the autumn or fall-run stock (www.iranfisheries.net, downloaded 28 July 2006). Three million fish were released in July 2006, probably at 3-4 g with more to be released at a heavier weight. Paykhan Heyrati and Dorafshan (2007) compared the effectiveness of two kinds of dopamine antagonist combined with a gonadotropin releasing hormone analogue on ovulation and fertilisation success under hatchery conditions. Spawning induction techniques using hormones provide high quality gametes in quantity for aquaculture programmes.

Azari Takami et al. (1990), Woynarovich (1985) and Bartley and Rana (1998b) detail the technique for artificial spawning, incubation and raising of fingerling mahi safid. Adults are caught as they enter rivers on the spawning migration by blocking the river with wooden tripods interlaced with twigs as screens and using a meshed net. The Shaeed Ansari Fish Farm (of Shilat, the Iranian Fisheries Company) takes fish from 5 main rivers, forming a brood stock of 10,000 females and 20,000 males (Bartley and Rana, 1998b). Almost 100% of migrating fish are caught. Ripe fish are stripped and the eggs fertilised with sperm. Eggs from 2 females are pooled and mixed with milt from 2 males. The adhesive layer is washed away with water and continuous stirring. Once the eggs are completely separated from each other they are placed in incubator jars. Eggs may be placed in net-bottomed trays in the river for 2 days before being sent to a hatchery. Egg development takes a minimum of 7 days but larval development only takes 1-2 days. Larvae are fed with a mixture of milk and eggs until they are 5 days old and then put in the earthern rearing ponds. In the Shaeed Ansari Fish Farm, fingerlings are released in a river mouth at 1 g and most enter the Caspian Sea within 3 days (Bartley and Rana, 1998b). Fingerlings are grown in ponds to a weight of 2-3 g and then released in the Safid River which carries them down to the sea (Petr, 1987). Production of 1-2 g fingerlings attains 3 tonnes per hectare in fish farms (Emadi, 1993a). Farabi et al. (2007) studied brood stocks and fingerlings in the Shirod, Tonekabon, Tajan and Goharbaran rivers from March 2004 to March 2005. The mean length, weight and condition factor for broodstock females and males were 43.75 and 36.5 cm, 1189.5 and 678.13 g and 1.42 and 1.38 respectively. Eggs collected over a 62 day period weighed 4931 kg, mostly at the end of March and beginning of April. Percentage of survival of eggs in the four rivers listed above was 94.5, 95.1, 87.7 and 96.9 respectively. Newly hatched larvae averaged a total length of 6±2 mm and a weight of 2±0.2 g. The number of fish produced at less than 1g was 16,942,454 representing 19.9% of the total released in Mazandaran in 2004. Fish in the 1.0-1.5 g weight class, suitable for release, numbered 62,905,247. Parasites of fry were Diplostomum, Dactylogyrus, Butriocephalus, nematodes, Trichodina and Epistelis.

Studies on the best methods of rearing fingerlings and fry have been carried out in Iranian hatcheries. The use of rotifers (Brachionis plicatilis) as food is more effective in larval growth than concentrated food preparations; after 40 days the rotifer-fed fish reached 440 mg and 31.36 mm on average and those fed concentrated food reached 17.2 mg and 14.75 mm (Fallahi et al., 2006; Kapourchali, 2006). Neverian et al. (2006) studied variation in protein levels fed to advanced fry at an isocaloric digestable energy of 3400 kcal/kg. The highest treatment at protein level 35% showed increased body protein, decreased fat and improved flesh quality in the carcass. Trichlorfon, an organophosphate pesticide is used to protect zooplankton such as rotiferans, food for larval safid mahi, against predators. However it does slow growth of the fish and the larvae must only be added to treated water 72 hours after the last application of the chemical (N. Chookbar, www.netiran.com, downloaded 18 April 2005; Chobkar et al., 2005). The sex steroid binding protein has been investigated as a biomarker for sexual behaviour in this species (Kousha et al., 2006). Spawning induction studies have been carried out on broodfish in Iran using carp pituitary extract and GnRH analogue, the latter alone or combined with metoclopramide. The latter treatment gave 90% ovulated females, 71.3% ovulation index and 68.4% fertilisation success, significantly higher than values for carp pituitary extract. However, the latency period was longer. Paykhan Heyrati et al. (2006) and Paykhan Heyrati and Dorafshan (2007) induced spawning with a gonadotropin releasing hormone analogue, both alone and in combination with the dopamine antagonists domperidone and metoclopramide, the combination with both antagonists being equally successful. Noverian et al. (2008) studied the effect of dietary digestible energy level on growth indices of advanced fry. Body protein and fat improved when the dietary energy was raised from 2500 to 2800 kcal kg^-1.

Savenkova (1990) lists several reasons for reductions in catches in the Atrak River stock of Turkmenistan on the border with Iran. These include ineffective fish passes, poor marketing strategies, and the low quality of fish reaching the spawning grounds.

A fine of 1,500 rials is imposed specifically for illegal angling of this species (Anonymous, 1977-1978).

Kiabi et al. (1999) consider this species to be of least concern in the south Caspian Sea basin according to IUCN criteria. Criteria include commercial fishing, sport fishing, abundant in numbers, habitat destruction, widespread range (75% of water bodies), absent in other water bodies in Iran, and present outside the Caspian Sea basin. The 2000 IUCN Red List lists this species as DD (Data Deficient). Coad (2000a), using 18 criteria, found this species to be one of the top 4 threatened species of freshwater fishes in Iran. Extinct in Turkey (Fricke et al., 2007).

Further work

The heavy demand for this fish as a traditional food warrants extensive management and conservation of stocks.

Sources

Iranian material: CMNFI 1970-0513, 9, 51.1-66.1 mm standard length, Gilan, Shafa River estuary (37º35'N, 49º09'E); CMNFI 1970-0516, 27, 42.3-61.0 mm standard length, Gilan, Lomir River (38º14'N, 48º52'30"); CMNFI 1970-0518, 9, 43.4-60.7 mm standard length, Gilan, Haviq River estuary (38º10'N, 48º54'E); CMNFI 1970-0520, 19, 46.7-73.9 mm standard length, Gilan, Astara River (ca. 38º25'N, ca. 48º52'E); CMNFI 1970-0563, 24, 40.8-58.1 mm standard length, Gilan Caspian Sea, Kazian Beach (ca. 37º29'N, ca. 49º29'E): CMNFI 1979-0088, 1, 104.3 mm standard length, Gilan, Safid River (no other locality data); CMNFI 1979-0471, 1, 106.0 mm standard length, Mazandaran, Caspian Sea west of Alamdeh (36º35'N, 51º48'E); CMNFI 1979-0686, 30, 42.8-52.2 mm standard length, Gilan, Safid River above ferry (37º24'N, 49º58'E); CMNFI 1980-0159, 1, 117.0 mm standard length, Gilan, Caspian Sea at Kazian Bridge (37º28'30"N, 49º28'E); CMNFI 1980-0160, 1, 88.5 mm standard length, Iran, Caspian Sea basin (no other locality data); CMNFI 1993-0140, 2, 64.5-71.5 mm standard length, Mazandaran, Tirom River, Ramsar (36º51'48"N, 50º48'E).

Comparative material: BM(NH) 1879.11.14:31-32, 2, 410.0-430.0 mm standard length, Russia, Astrakhan (ca. 46º24'N, ca. 48º05'E).

Rutilus rutilus
(Linnaeus, 1758)

Common names

كلمه (= kolme, kollme, kolmeh or koolmeh, meaning unknown), mahi kolme, kolme Gorgan, telagi or talaji (in Gilaki and Mazandarani, meaning unknown, latter three used for R. r. caspius natio knipowitschi), mahi cheshm qermez (= redeye fish), kolme Kura or kolme Anzali for R. r. caspius natio kurensis, kolme Gorgan or kolme Turmenistan for R. r. caspius natio knipowitschi

[kulma, kyumen, xazar kulmasi and Kur kulmasi in Azerbaijan; kasli akcapagy in Turkmenian; Armyanskaya plotva or Armenian roach, karmraki and last for schelkovnikovi, all in Armenia; vobla, severo-kaspiiskaya vobla or North Caspian vobla, Astrakhanskaya vobla or Astrakhan vobla, Kurinskaya vobla or Kura vobla, Astrabadskaya vobla or Astrabad vobla, Turkmenskaya vobla or Turkmenian vobla, all in Russian; roach being the English equivalent of vobla].

Systematics

Cyprinus Rutilus was originally described from lakes in Europe, the type being from Sweden.

Holčík and Skořepa (1971) revised the roach and found no reason to maintain subspecies, of which they list 12. Rutilus rutilus caspicus (Jakovlev, 1870) (North Caspian or Astrakhan vobla) was the subspecies found in the Caspian Sea basin with natio knipowitschi Pravdin, 1927 (Astrabad or Turkmenian vobla) in Gorgan (= Astrabad) and Gasan-kuli bays and the Gorgan, Atrak and Qareh Su rivers and natio kurensis Berg, 1932 (Kura vobla) in Kyzylagach Bay, the Kura River, Astara and rarely the Anzali Mordab. The Kura vobla was said to differ from the Astrabad vobla by a greater body depth, smaller eye and more rapid growth. R. r. caspicus was distinguished from other subspecies by having modally 9 branched dorsal fin rays, darker fins, more inferior mouth, higher dorsal and anal fins, longer pectoral and pelvic fins, a deeper head, and larger eyes (see Berg (1948-1949) for more details). R. r. caspicus was originally described as Leuciscus rutilus var. caspicus from the Volga River delta, Russia. The types are unknown (Eschmeyer et al., 1996). Eschmeyer et al. (1996) also list Leuciscus rutilus var. wobla Grimm, 1896 described from the North Caspian Sea and mouths of rivers (Volga, Ural, Emba, Terek, Kura, Astara) entering the Caspian Sea. This is presumably a synonym of Rutilus rutilus judging from the subspecific name which is the Russian word for this species (wobla = vobla or roach). Kottelat (1997) lists it as a nomen nudum.