COLLETTE and RUSSO: SPANISH MACKERELS 



larly important: Stolephorus in Singapore Straits 

 (Tham 1950, 1953) and Anchoviella in Waltair, 

 India (Rao 1964). 



Interest to fisheries. — There are commercial or 

 artisanal fisheries for S. guttatus in Cambodia 

 (Merceron 1970), Thailand (Tongyai 1971b), Ma- 

 laysia (Pathansali 1968), and India, particularly 

 in the lower Sundarbans, West Bengal (Banerjee 

 and Chakrubarty 1972), the Balashore coast (Roy 

 and Roy 1974), around Madras (Vijayaraghavan 

 1955), the Gulf of Mannar- Palk Bay area (Krish- 

 namoorthi 1957), and Malwan, south of Bombay 

 (Kaikini 1961). It is caught all year round in some 

 areas (Cambodia — Merceron 1970; Ramaswaram 

 I., India — Krishnamoorthi 1957) but there are 

 peaks of abundance that differ from region to 

 region. It is taken in the non-monsoon months 

 (September-May) along the Balashore coast south 

 of Calcutta with the catch increasing from Octo- 

 ber to February (Roy and Roy 1974). Catches peak 

 in September-October, December- February, and 

 May in Waltair near Vishakhapatnam further 

 south along the west coast of the Bay of Bengal 

 (Venkata Subba Rao et al. 1981). The season 

 extends from September- October to March-April 

 in Vizhingam, southern India, (Gopalan Nayar 

 1958) with the peak catches usually in September 

 or October. It is one of the principal species in the 

 drift net seerfish fishery in India, but the catch 

 is not identified to species in the statistics. Indo- 

 nesia reported the only catch identified as S. 

 guttatus (4,254-5,249 t/yr) in 1979-82 (FAO 1984). 

 The primary gear in most areas appears to be the 

 drift gill net which is set overnight, but it is also 

 taken in bamboo stake traps and with hand lines 

 in Thailand (Tongyai 1970) and by trolling or 

 with hook and line in India and Malaysia (Rao 

 1964; Jones 1968; Pathansali 1968). It is utilized 

 fresh or salted in most areas (India — Jones 1967; 

 Cambodia — Merceron 1970; Thailand — Tongyai 

 1971b). It can be stored on ice for 10-13 d (Shenoy 

 and James 1974). Although less abundant than 

 the Indian mackerels {Rastrelliger spp.), it is 

 highly esteemed for food and commands a high- 

 er price in Thailand and India (Tongyai 1966a; 

 Pathansali 1968). 



Distribution. — Indo-West Pacific from Taiwan to 

 the Gulf of Thailand, Java, and Sumatra west 

 around the Bay of Bengal and Arabian Sea into 

 the Persian Gulf (Fig. 55). The northernmost 

 records are from Wakasa Bay, Japan (Nakamura 

 and Nakamura 1982), Taiwan (FMNH 59284), 



Amoy (BMNH 1860.7.20.110), and Swatow, China 

 (Reeves 1927). There are many records and speci- 

 mens from Indochina, the Gulf of Thailand, and 

 the East Indies. There are records and specimens 

 of S. guttatus from Borneo (Bleeker 1860; Har- 

 denberg 1936; Herre and Herald 1951; ANSP 

 72282) and Makassar, Celebes (RMNH 24096). 

 The range extends further out in the East Indies 

 than that of either S. lineolatus or S. koreanus, at 

 least to Bali. The report of S. guttatus from 

 Western Australia (McKay 1970) is based on a 

 specimen (HUMZ F-423) of S. queenslandicus. 

 Earlier reports of S. guttatus from Australia 

 (Macleay 1881; Stead 1906, 1908; Rendahl 1923) 

 are also based on S. queenslandicus (Munro 1943: 

 86). Early reports from New Zealand are based on 

 "a damaged specimen of a Cybium, probably 

 C. guttatum, was obtained at the Chatham 

 Islands...." (Hutton 1895). This report has led to 

 subsequent records (Hutton 1904; Phillipps 1927; 

 Whitley 1968). We concur with Whitley's conclu- 

 sion that this record is "very doubtful". The range 

 extends west into the Persian Gulf (Kuronuma 

 and Abe 1972; ZMK 3-4). 



Geographic variation. — Morphometric data for 

 five populations of S. guttatus were compared 

 with ANCOVA (Table 17): Arabian Sea (n = 7- 

 13), Bay of Bengal (n = 5-9), East Indies U = 14- 

 24), Gulf of Thailand (n = 10-17), and China (n = 

 22-31). Null hypotheses that the 5 sets of regres- 

 sion lines are coincident were accepted for 18 sets, 

 rejected for 8 sets: Sn-ID, Sn-Pi , Head L, maxi- 

 mum depth, maxilla L, orbit (fleshy), interorbit, 

 and 2D-C. The five populations were arranged 

 geographically from west to east as listed above. 

 No significant differences were found between 

 populations in the Arabian Sea and Bay of Ben- 

 gal, but there were significant differences be- 

 tween all other adjacent populations: Bay of Ben- 

 gal vs. East Indies (Sn-Pi ), East Indies vs. Gulf of 

 Thailand (Interorbital), Gulf of Thailand vs. Chi- 

 na (Sn-ID, interorbital, and 2D-C). The Arabian 

 Sea and Bay of Bengal populations were com- 

 bined, the regressions rerun, and compared with 

 the other three populations with ANCOVA. Null 

 hypotheses that the 4 sets of regression lines are 

 coincident were accepted for 15 sets, rejected for 

 11 sets: Sn-ID, Sn-Pi, Head L, maximum body 

 depth, Base ID, Base 2D, Base A, maxilla L, orbit 

 (fleshy), interorbital, and 2D-C. The Newman- 

 Keuls Multiple Range Test was able to distin- 

 guish populations that differed significantly for 7 

 sets of regressions but could not do so for 4: 



635 



