The aim of this study has been to clarify various features of the population of Sillago sihama (FORSSKÅL) in eastern Bingo-nada of the Seto Inland Sea as a fishing resource for the bottom drift-net fishery. It was attempted to approach to this aim by analyzing the interaction between the ecological characteristics of this fish and the catching ability of the bottom drift-net. In doing so, it was necessary to secure information on the fishery biology of this species, on the species- and size-selectivity of the bottom drift-net and on the configuration of the gear while operated underwater. The data upon which this study is based comprise published records on the commercial catch of Sillago sihama, unpublished operation records of individual bottom drift-netting boats, the biological data derived from those sample fish which the author collected according to a special sampling program, the results of experimental operations of the bottom drift-net and the data obtained by surveying the commercial catch made by other fishing methods.
Sillago sihama is a common demersal fish of commercial value in the Seta Inland Sea and its adjacent sea regions (Fig. 8). In the central region of the Sea almost all the commercial catch of this species is made with the bottom drift-net of stretched mesh size of 2.9 or 3.0 cm (Fig. 3). Ever since the commercial bottom drift-netting for this species was begun around 1957 in this region, its annual landing seems to have been fairly stable. While the fishing season of this fishery lasts from May until October, landing is most abundant in May and gradually decreases thereafter with the increase in the cumulative fishing effort (Fig. 6 and 7).
According to the data obtained by the experimental operations of the bottom drift-net, the number of species captured by this gear amounts to 32, and throughout the fishing season Sillago sihama accounts for about 67% of the captured individuals and 58% of the weight of the catch in the average (Table 1 and Fig. 5).
General conclusions obtained in this study are summarized as follows:
1) Much importance has been attached to the morphological characteristics of the scales of the cheek in identifying the species belonging to Family Sillaginidae. In the species dealt with in this study the scales of the cheek are exclusively ctenoid in 7 of the examined 61 individuals, and consist of ctenoid scales mingled with small number of cycloid scales in the other 54 individuals.
In order to make clear the morphological characters of the adult stage of this species, scales on the lateral line, fin rays, and gill-rakers on the first gill-arch were counted (Table 2), and the proportions between various body parts were examined in the sampled individuals (Table 3).
Total length can be converted to standard length by multiplying by 0.86, the factor being derived from the measurements on 321 individuals (Fig. 9). Fork length can be converted to standard length by multiplying by 0.91 (Fig. 10). Weight can be computed from standard length by means of the relationship: W = 9.69 × 10-3 L3.079, where W is estimated round weight in grams and L is standard length in centimetres (Fig. 11).
2) Age and growth of the sampled individuals of Sillago sihama were determined with the aid of the annuli on the scales taken between the pectoral fin and the lateral line. In this species an annulus on a scale is definable as the zone which comprises irregularly arranged ridges in the anterior sector and concurrently intercepts adjacent inner ridges in both lateral sectors (Fig. 12).
It is observed that in the scales from the same individual the relationship between the scale radius (R, the distance from the focus to the anterior margin of a scale) and the radius of a definite annulus (r, the distance from the focus to the anterior margin of the annulus) can be expressed by linear regression (Fig. 13). The relationship between standard length of a fish (L) and the mean radius (R¯) of the scales taken from the above-mentioned body part is linear, and can be expressed as R¯=0.0248 L－0.237 (Fig. 15). The scale radius computed from the standard length with the above empirical equation was termed the "standard scale radius" (R¯). By multiplying this by the ratio of the measured annulus radius (r) to the measured scale radius (R), the "standard annulus radius" (r¯) was calculated. The mode and range of each standard annulus radius were calculated for different age group of the sample fish (Table 5). The W ALFORD's graph drawn with the data shown in Table 5 proved to be linear; this result indicates that the von BERTALANFFY's model should yield reasonably good fit to these growth data (Fig. 16).
The value of the marginal growth index, (R－rn)/(rn－rn-1), changed monthly and attained the minimum value in May (Fig. 17). A new annulus was found along the margin of scales in 35%, 49% and 7% of the fish caught in May, June and July, respectively (Fig. 18). The monthly growth rate in standard length is estimated from the value of the marginal growth index of each month. The season of rapid growth is from April to September, and the growth during this period accounts for 80% of the annual growth (Fig. 19). Frequency distribution of the standard radius of each annulus and that of the estimated standard length at the time of annulus formation are shown in Fig. 21. The growth of this species can be described by PÜTTER-VON BERTALANFFY's equation, Lt=21.4 (1－e-0.369t-0.232), and by BÜCH-ANDRESSEN-FISCHR's equation, Wt=115(1－e-0.369t-232)3.
3) The sex ratio in Sillago sihama is likely to be I: I, but females are slightly more abundant than males in the samples taken during the spawning season (Table 6).
By investigating the monthly trend of the gonad index, (gonad weight × 104 )/(cube of standard length), of the sampled individuals, it is inferred that the spawning season lasts from June until August (Fig. 23). The frequency distribution of the diameter of ovarian eggs is shown in Fig. 25 for each individual representing different stages of gonadial maturity (Table 7). It may be seen in this figure that the mode of the diameter shifts according to the gonadial maturity from the immature stage to the spent. In the ovary of mature stage the egg diameter has two modes, namely, at 0.3 and 0.65 mm (Fig. 25). This fact and the fact that remnant mature eggs occur together with the majority of eggs approaching ripeness in a single ovary suggest that this species may spawn more than once in a spawning season. Eggs of the most advanced group (over 0.5 mm in diameter) are likely to be shed at a single spawning; the estimated number of such eggs ranges from 3 to 20 thousand in the 25 females with matured ovaries (Fig. 27). Based on the assumption that all the ovarian eggs over 0.2 mm in diameter be released in the spawning season of the respective year, the fecundity per female per spawning season is estimated 20-30 thousand eggs in the 2-year-old female, 30-50 thousand eggs in the 3-yearold and 50-80 thousand eggs in the 4-year-old (Fig. 27).
All males and females over 2 years of age were observed to be sexually mature. In the 1-year-old female, 8 specimens of the examined 98 were mature and the others were immature. The minimum size of the female with matured ovary is about 10 cm in standard length.
4) Analysis of stomach contents of Sillago sihama reveals that this species is a benthos-feeder and that shrimps and polychaetes form the most important items of the diet in the adult stage. There is marked difference in the composition of stomach contents between the adult and the immature stage: namely, the main items of diet are shrimps and polychaetes in adult fish, whereas they are amphipods and polychaetes in immature fish (Fig. 32).
In this species the ratio of the maximum weight of stomach contents to the body weight declines as the fish grows (Fig. 28). Judging from the monthly changes in the stomach contents weight index, (weight of stomach contents ×102)/(maximum weight of stomach contents for the respective body weight), and in the percentage occurrence of empty stomachs, this species seems to show intense feeding activity during the period from April to September, and its feeding activity appears to take a declining trend after August (Fig. 30). While the main items of the diet are essentially the same in the active feeding season and in the other seasons, it appears that wider variety of food items is accepted in the former season than in the latter (Fig. 31).
Definite seasonal changes were observed in the composition of the stomach contents of immature fish. During May, June and July, am phi pods form the dominant food, but after July they are replaced by polychaetes. In September relative frequencies that principal food items occur in the stomach contents are considerably similar in the adult and the immature fish. (Fig. 33). Available data suggest that in this month immature individuals immigrate into the regions which are inhabited by the adult.
5) Although various species of demersal fishes are captured with the bottom drift-net as mentioned before, this gear shows considerably high ability to catch Sillago sihama. It is presumed that the shape of this net underwater matches the swimming behaviour of this species and therefore results in a high catching ability.
The mesh selectivity of the bottom drift-net for Sillago sihama was estimated by applying KITAHARA's method to the data obtained by comparative fishing experiment. In this experiment 50 nets were used (i.e., 10 nets for each of the 5 different mesh sizes of 2.58, 2.82, 3.00, 3.19 and 3.36 cm in stretched measure) (Fig. 34). Selectivity was analysed as a function of the variable, (Standard length)/(stretched mesh size) (Fig. 43).
The determined selectivity curve has a fairly sharp peak and is slightly skewed to the right (Fig. 45). The bottom drift-net used in the central region of the Seto Inland Sea usually has the stretched mesh size of 2.9 or 3.0 cm; the above mentioned selectivity curve indicates that the standard length corresponding to the peak efficiency of those mesh sizes are 12.8 and 13.3 cm, respectively.
6) The bottom drift-net consists of a long narrow rectangular strip of single-layer netting. In actual operation the net drifts over the sea bed with the tidal current.
Assuming that the various curves of a fishing net in working condition can be approximated by circular arcs of different curvatures, 10 equilibrium equations are theoretically obtained regarding the distributing tensions and the shape of the net.
The shape of the net underwater depends on the difference between the current velocity and the drifting speed of the net (i.e., function V) and can be computed if the current velocity, the drifting speed of the net and the coefficient of kinetic friction between the sinkers and the sea bed are known. The obtained result is shown in Fig. 49.
In order to check the validity of the above theory, the shape of the net was actually observed in water at a river mouth where various current velocities were available. There was satisfactory agreement between the experimental and the theoretical results (Table 9).
7) The swimming behaviour of Sillago sihama in the fishing grounds was discussed by analyzing the portions and sides of the net in which the fish were entangled in relation to the direction of tidal current. From the observations on the entangled individuals, it may generally be said that the majority of the fish keep themselves against the flow while swimming (Table 10). In general, those individuals swimming against the flow are situated farther above the sea bottom than those swimming along the flow. Considering the configuration of the net underwater, it is concluded that 65% of the captured fishes were swimming within 6cm of the sea bed, 20% from 6 to 14cm above the sea bed, and 9% from 14 to 22cm above the sea bed (Fig. 52). It is presumed that the highest swimming level of this species should usually be 40cm above the sea bed. Thus, Sillago sihama can be regarded as a demersal fish from the view point of swimming behaviour.
8) In Bingo-nada, central part of the Seto Inland Sea, commercial catch of Sillago sihama is made mainly with the bottom drift-net, although a minor portion of the catch is made with such commercial gears as the "Kogata sokobiki-ami" (a small beam trawl), the "Masu-ami" (a pound net), the "Gochiami" (a boat seine) and the "Ipponzuri" (the angling). Sillago sihama caught with the small beam trawl from May through August are !-year-old; their standard length has a single mode at 7 cm in May and at 9.5cm in August (Fig. 55). The fish caught by the pound net range from 9.5 to 18.5cm in standard length with the fish of 11 to 14cm being dominant (Fig. 56); they are I to 4 years of age, and most of them are 2-year-old (Fig. 56). The fish caught by angling and the boat seining are bigger than those captured with the other fishing gears mentioned above: their standard length ranges from 13 to 18cm and the dominant lengths are between 14 and 16 cm (Fig. 56); while they are from 2 to 4 years of age, the 3-year-old are dominant (Fig. 56). Though the fish occurring in the fishing ground of the bottom drift-netting vary from 10 to 18cm in standard length (Fig. 57), the commercial catch by this fishing method comprises the fish of 11~16cm in standard length, 70% of them measuring 12 to 14cm (Fig. 57); while the catch comprises principally the 2-year-old fish during most part of the fishing season, it comprises the 3-year-old fish at the beginning of the fishing season (May), and the 1-year-old fish toward the end of the fishing season (September and October).
In Fig. 58 the standard lengths of the commercial catch of this species are indicated for each fishing method. At the bottom of this figure are shown the frequency curves of the standard length at the full ages of 1, 2 and 3 years; these curves are based on the results of the scale study. In Fig. 59 are schematically shown the location of fishing ground and the ages of the commercial catch, respectively for each fishing method. The immigration of Sillago sihama from one region to another was discussed by the aid of this figure.
9) An analysis of the operation records of the commercial bottom drift-net fishing indicates the tendency that the catch per boat-day of Sillago sihama decreases gradually with the increase of cumulative fishing effort (Fig. 61). This result suggests that DELURY's model can be employed in studying the abundance, immigration and emigration of this species in a local fishing ground. Assuming that immigration, emigration and change in the length composition are all neglegible in a local stock within a relatively short period of time, DELULY's model was applied to those operation records of the commercial bottom drift-netting in the eastern part of Bingo-nada which consisted in the daily total weight of captured Sillago sihama and the daily number of operating boats during the 1964 fishing season. In this figure the logarithm of the catch per boat-day as plotted against the cumulative fishing effort can be regarded as decreasing fairly regularly; therefore, it does not seem unreasonable to fit a straight line to a group of plots which represents a certain period of time. However, it is noted that the plots lie on the four parallel straight lines (Fig. 61). We may interpret this as that the decrease of catch per boat-day is due to the effect of exploitation and that the jumping up from one straight line to an upper one reflects an increase in the abundance caused by immigration. And this interpretation is supported by the length corn position of the commercial catch at the respective times in the fishing season and by the migratory behaviour of the 1year-old fish as associated with the changes in their feeding habit.
On the basis of the information mentioned above, it may be said that the ratio (a) of the catch per boat-day to the amount of the stock is almost constant over the period of the operation. Therefore, it may be given as a conclusion that 3.8 × 10-4 is obtained as the numerical value of a evaluated from the slope of the lines and that the percentage of the season's total catch to the stock is estimated at 62% for Sillago sihama in this region.
The conclusion as mentioned below on the bottom drift-net fishery has been obtained from the analyses of the interaction between the ecological characteristics of Sillago sihama and the features of the fishing gear, and moreover from the information of the fishing stock. The bottom drift-net has the species- and size-selectivity, and Sillago sihama of 12-14cm in standard length is captured dominantly in the central region of Seto Inland Sea. The catching ability of this gear is very efficient for the fish, and the percentage of the catch to the stock amount is estimated fairly high for the fish. However, the fishing intensity for large (more than 14cm in standard length) and small (less than 12cm) is not so high as medium (12-14cm) in the population. Accordingly, it is recognized that the catch per unit effort of the fish by the bottom drift-netting is fairly stable since from the introduction of the commercial bottom drift-net fishery, and the decrement of the fishing stock of the fish is avoidable at a present situation. The construction and operation of the bottom drift-net are simple and easy, and more it has a high efficiency for catching the fish. Judging from the ecological character of Sillago sihama in the central part of the Seto Inland Sea, the optimum mesh size of the fishing gear operating in this region is 2.8-3.0 cm in stretched measure.