Abstract:Based on the fishing records of bull trawlers operated in the Gulf of North Part(Tonkiu Gulf) by the Nanhai Fisheries Company during the years 1957-1960,seasonal and annual changes of species composition in the catches for differentregions of the gulf were examined in present studies. Thirteen subregions are adoptedin this paper corresponding to the customary boundaries of principle fishing groundsgenerally used. Their numbers and extents are tabulated in a table accompanied with achart. The percentages of species are derived from the average catch per haul ineach subregion or whole region over the whole period of 1957-1960. Rank correalationcoefficients, known as Spearmen's ρ, are calculated for the pairs of subregions, seasonsand years of species compositions, with which the status of changes is discussed. Quantitative data for twenty-two species or groups of fishes and one group ofsquid are available in fishing records. The calculation of the percentages in speciescomposition for the whole region after the following figures: Nemipterus spp. 18. 6,Lujianus erythropterus 10. 7, Upeneus spp. 9. 6, Priacanthus spp. 9.0, Saurida spp.8.1, Caranx malabaricus 4. 0, Arius sp. 3. 4, Squid 2.4, Parargyrops edita 2. 4,Pentaprion Longimanus 1. 5, Therapon theraps 1. 3, Pomadasys hasta 1. 1, Wak spp.0. 9, dogfishes 0.8, Psenes indicus 0. 6, Taius tumifrons 0. 4, Epinephelus spp.0. 4, Trichiurus haumala 0. 2, Pagrosomus major 0. 1, Muraenesox cinerus 0. 1, andother species including unclassified fishes 22. 8. In gerenal, the percentages of speciescomposition for subregions are not similar to those of the gulf as a whole, butmost of them have considerable identifical trends. They are shown in the histograms(Fig. 2(1) (2)]. The rank correlation coefficients of species composition between subregions over four years have been calculated. They show that positive correlations are presentedbetween most subregions. Most significant-levels of correlation coefficient ρ≥0. 66are found between nine subregions with either subregion Ⅳ or Ⅸ, while subregion Ⅴwith only one subregion bears such level of significance correalation.There is no sig-nificant correlaion found between subregion Ⅱ-Ⅶ, Ⅱ-Ⅷ, Ⅱ-Ⅻ, Ⅵ-2-Ⅴ orⅥ-2-Ⅶ, their coefficient ρ'is 0.40. The dissimilarities of the seasonal and annual changes in relative indices of abun-dance of species composition for different subregion are noticeable. They are illus-trated by diagrams. with the help of a number of text-figures to show the curves ofrank correlation coefficients plotted by coefficient values between seasons as wellas between years reciprocally, the variations are found to be more conspicuous whenthe seasons turning from March to April and from September to October, and to beless conspicous when the season turning from December to January and from June toJuly. According to the figure of curves mentioned above, four types of trends inseasonal variations of species composition for different subregion and the gulf asa whole have been recognized. Subregiou Ⅷ, Ⅸ and Ⅺ and whole region are denotedas type 1, subregion Ⅱ, Ⅲ, Ⅴand Ⅵ-1 as type 2, subregion Ⅶ,Ⅹ and Ⅻ as type3, and subregion Ⅳ as type 4. their characteristics are described. The classificationof seasonal variation trends mentioned above is found to be applicable to those ofannual variation with minor differences, In contradiction to subregion Ⅷ belongto type 2, subregion Ⅹ belong to type 3 in seasonal variation, they are consideredas type 4 and type 2 of annual vatiation respectively. It seems probable that relationship of geographical positions of the subregionis one of the essential factor which makes the seasonal and annual variations ofspecies composition much alike among some subregion or differing among the others.In general, the neighboring regions usually maintein their trends is a similar fashionin these variations than the remote ones. The index of diversity after Magalef(1960) has been applied to demonstrate the status of the demersal fish communityof the regions investigated.