5. BIOLOGY OF SARDINELLA AURITA AND S. MADERENSIS

5.1 LENGTH AT FIRST MATURITY
5.2 LENGTH-WEIGHT RELATIONSHIP
5.3 REPRODUCTION
5.4 GROWTH
5.5 MORTALITY

5.1 LENGTH AT FIRST MATURITY

Information on length at first maturity in the Congo area is given by Fontana and Pianet (1973), based on sampling from 1967 till 1969. They find that all females of S. aurita of over 21 cm and all females of S. maderensis of more than 20 cm have started their sexual cycle (ovary stage III). From their graphs it can be read that 50 percent of the females of S. aurita have started stage III at about 19 cm, and of S. maderensis at slightly over 18 cm.

The authors mention that, at least in S. maderensis, there is evidence of a difference in size at first maturity in different seasons.

The samples only covered one complete year, and no information is available on possible variations in length at first maturity between years.

Off Gabon, north of 2°30'S, the size of S. aurita is much smaller than further south, and 50 percent of the females reach maturity at about 14 cm. This suggests the existence of a small, separate stock of S. aurita in that area.

5.2 LENGTH-WEIGHT RELATIONSHIP

The length-weight relationship has only been determined for the sardinellas in the Congolese area (Fontana and Pianet, 1973). The data refer to the years 1965 and 1967 for S. aurita and the year 1969 for S. maderensis.

The authors found a strong seasonal variation in the exponent (b) of the equation W = a x L^b (in which W represents the weight in grammes, L the fork length in cm, and a and b are constants). However, they state that the length-weight keys calculated with the equations found in the different seasons are all very similar, and therefore they adopted one generalized relationship for each species as follows:

S. maderensis	W = 5.94 x 10-6 x L3.183
S. aurita	W = 4.25 x 10-6 x L3.218

It is probably that there will be a variation in this relationship between years, as is suggested by the observation's of the same authors on the year-to-year variation of the condition factor (the weight, divided by the cube of the length) for each species (see following table).

Seasonal variation in the condition factor ()

Difference (in % of the minimum) between the lowest and the highest monthly value of the condition factor in each year (from Fontana and Pianet, 1973)

5.3 REPRODUCTION

5.3.1 Sardinella aurita
5.3.2 Sardinella maderensis

The reproduction of sardinellas in the Congo area has been studied by various authors, including Fontana (1969), Dessier and Pianet (1971), and Fontana and Pianet (1973). The reproduction of sardinellas in Angola has not been the subject of any publication, although certain pertinent data seem to have been collected in the region. Spawning is very different between the two species.

5.3.1 Sardinella aurita

Fontana and Pianet (1973) conclude that this species spawns only during the cold season, in waters in which the productivity has been increased by hydrographic phenomena such as upwelling. Furthermore, strong variations in the gonadosomatic index of the females (weight of the ovary, divided by the cube of the length) (Table 6a) show important fluctuations in the sexual activity of the stocks exploited by the Pointe Noire seiners (Fontana, 1969):

(a) seasonal fluctuations which indicate the existence of a short spawning season during the “short cold season” in December-January, and the main spawning during the “long cold season” from June to September;

(b) strong year-to-year fluctuations; the sexual activity, as measured by the gonadosomatic index, was strong in 1967, 1968 and 1971, and very weak in 1969 and 1972.

It should be noted that gonadosomatic index is not necessarily an index of the total egg production, which depends both on the spawning activity of the females and on the number of females in the stock. The possibility of the existence of a relation between the gonadosomatic index and recruitment should be studied further.

The uneven distribution of spawning, both in space and in time (seasonally), is confirmed by the observations of Dessier and Pianet (1971) on the distribution of sardinella larvae in the area between Libreville and the Congo River, from May 1968 till July (8 cruises, 80 stations) (Tables 7a and 7b).

The fecundity of S. aurita in Congo has been studied by Fontana and Pianet (1973), who found the following relationship:

F = 436.415 W - 22076.437

where F = fecundity (number of eggs per female), and W = weight in grammes.

No quantitative information is available on the reproduction of S. aurita in the Angola area. The presence of juvenile S. aurita in Angola, and the capture of mature females (Ghéno and Campos Rosado, 1972) indicate however that the species also spawns in the Angola area.

5.3.2 Sardinella maderensis

Fontana and Pianet (1973) concluded - from the analysis of the data on the gonadosomatic index (Table 6b), from the observations on the distribution of the larvae (Table 7a), and from the data on capture of young fish - that spawning in this species takes place throughout the year in the whole coastal area from Gabon to Cabinda, with a maximum during the cold season.

The fecundity of S. maderensis in Congo has be en given by the same authors as:

F = 418 W - 18974

where F = number of eggs per female, W = weight in grammes.

It can be expected that the species also reproduces in Angola, but there are no quantitative observations to support this assumption.

5.4 GROWTH

5.4.1 Sardinella aurita
5.4.2 Sardinella maderensis

5.4.1 Sardinella aurita

The growth of S. aurita has been analysed by Ghéno (1968, 1975) from samples obtained from the Pointe Noire (Congo) fishery. The growth studies have been based on readings of the year rings in scales together with other characteristics, such as:

(a) spawning seasons, which have been determined from the state of development of the gonads. As described in Section 4.5, the spawning seasons are usually clearly distinguishable in S. aurita

(b) length frequency analysis

(c) determination of number of vertebrae which allows to identify groups of individuals born at the same sea temperature

“birthday”	1 January	1 May
L¥	25.97	26.00
K	1.023	1.210
to	-0.08	-0.02

It appears that the growth of S. aurita is very rapid during the first years of life followed by a slow growth of the adults, resulting in a high K. Thus, the growth of S. aurita seems to be well established for the area around Pointe Noire (Congo). It may, however, not be correct to accept these data for the whole population of S. aurita in the Gabon-Angola area, because the growth in different parts of the region may be influenced by local hydro-graphic conditions or by the existence of local races with different growth patterns. This is demonstrated by the existence, north of Cape Lopez, of S. aurita with biological characteristics which are very different from those of the same species exploited in the Congo area.

5.4.2 Sardinella maderensis

Ghéno and Le Guen (1968) have studied the growth of S. maderensis caught in the Congo area in 1966 and 1967.

The growth has been determined from scale readings and the identification of groups of fish born under similar hydrographic conditions, as indicated by their average number of vertebrae. This analysis was complicated by the absence of specific spawning seasons in this species.

The von Bertalanffy growth equation fitted to the data gave the following growth parameters:

L_¥ = 24.92
K = 0.984
t_o = 0.287

Thus, the growth of S. maderensis shows the same characteristic as that of S. aurita, a very rapid growth during the younger ages and a slow growth thereafter. As in S. aurita, it is possible that the growth differs between areas and between years. It would appear that the growth studies could be improved by the analysis of all data collected in Pointe Noire since 1968, when Ghéno and Le Guen carried out their analysis.

It should be noted that the large difference between the lengths of S. maderensis in the catches in the Congo area, and those reported for Angola in 1974 (Annex 3), may have been at least partly due to differences in the standard measure (fork length versus total length) and cannot, therefore, be taken as an indication of difference in size of the fish in different areas. It is still possible, however, that the growth of S. maderensis in Angola is different from that in Congo, due to differences in hydrographic conditions.

5.5 MORTALITY

5.5.1 Sardinella aurita
5.5.2 Sardinella maderensis

5.5.1 Sardinella aurita

The apparent total mortality rate (Z) of S. aurita in the Congo area has been estimated by Ghéno (1972) . The author made use of the fact that the yearclass 1964 of this species was very poor, which facilitated the distinction between the younger fish and those born in 1963 or earlier. Thus, Z was determined for the decline in the catch per unit of effort of the fishes of the yearclass 1963 and the older yearclasses together, in the period 1964 to 1969. A similar procedure was followed for the yearclasses 1968 and older caught in 1970 and 1971.

The two analyses gave very similar estimates of the apparent total mortality rate, 0.76 and 0.79 respectively. The method may, however, have given rise to biased results, for the following reasons:

(a) the data used in the studies were obtained from the only purse seiner fishing in the area in that period, and the catches of this seiner may not have given a proper representation of the sardinella stock in the area;

(b) no account has been taken of possible changes in the fishing power of the vessel used in the studies;

(c) there may have been year-to-year fluctuations in the vulnerability of the stocks or of certain ages of fish, in relation with hydrographic conditions;

(d) there are still uncertainties about the age of the fish, in particular in the older ones.

It may be noted that according to Beverton and Holt (1959) and Cushing (1968) in the relation M/K has for clupeids usually a value of close to l.

This would mean that, with estimates of K close to l as given above, M, the natural mortality rate, would also be around 1.

Whereas it is difficult to draw any definite conclusions from the available information, the data would seem to suggest that the fishing mortality rate of S. aurita in the area was low in the periods for which mortality estimates have been obtained.

5.5.2 Sardinella maderensis

The apparent total mortality rate (Z) of S. maderensis has been estimated by Ghéno in the same study as that of S. aurita (1972). Two different methods have been applied:

(a) the approximation given by Beverton and Holt (1956)

where Z = apparent total mortality rate
L_¥ and K = parameters of von Bertalanffy's growth equation
l_c = average length at recruitment
l = average length of the fish in the catch

Ghéno used the following values: L_¥ ⁼ 19.50 cm, K = 0.98 cm, l = 24.93 cm and l_c = 21.24 cm, which gave an estimate of Z = 2.12. The Ad hoc working group noted that the value of Z obtained is very sensitive to differences in the value of ; consideration of the length composition of the catches suggested that l_c could be closer to 16 cm than to the value used of 21.24 cm. Applying the value of lc of 16 cm in the equation would, for example, lead to an estimate of Z of 0.70.

(b) the conversion of the length composition of the catches in 1964-69 into age-composition on the basis of the von Bertalanffy growth equation established for the species. The average decline in the catches by age group then represents the apparent total mortality rate in the stock which has been estimated by this method at 1.93.

There are several factors which may have introduced a bias in this estimate:

(i) the conversion of a length composition into age-composition on the basis of data on the average growth seems very difficult in view of the very slow growth of fish over 20 cm long, and of the variability in length of fish of the same cohort. This variability, as well as the variability in growth between cohorts, is probably much greater than the theoretical annual growth of the older fish. This also causes problems in allocating fish of a length greater than L_¥ to the various yearclasses;

(ii) the estimate is likely to be sensitive to the value of the growth parameters used in the conversion from length to age, and it would seem that the growth parameters determined for S. maderensis cannot be considered as very precise;

(iii) the length composition of the catches may represent a biased sample of the length composition in the stocks in the Pointe Noire area.

Therefore, these estimates of Z should be taken as very provisional and need to be considered with caution.