Growth patterns of Mirogrex terraesanctae
Sections:
Development of the growth hypothesis
The first publication dealing with growth in the Kinneret sardine (named Acanthobrama terraesanctae at that time) was Steinitz’s 1959 paper in which he sums up Dr. Lissner’s data on this fish, after Lissner’s death. When I began to study lavnun growth in the 1970’s, I followed Lissner in using scale ‘readings’ for age determination and obtained similar results (Gophen & Landau, 1977). Further study revealed various ‘readings’ from different scales of the same fish.
In addition it became apparent that lavnun growth is slower than that deduced from any of the scale ‘readings'. Fishermen reported ‘clouds’ of tiny sardines in September, five months after the end of the spawning season, and specimens as small as 3 cm total length were taken by small-mesh beach-seine in the late summer.
Opercula of the lavnun are more suited for growth studies than its scales, which are only a few mm in diameter and probably shed during the fish's lifetime. However, the myriad of growth rings on opercula cause difficulty in identification of annuli (annual rings).
Studies on larval and juvenile growth, in collaboration with Dr. Paul Walline, solved the problem of annulus identification (Landau et al, 1988). Conveniently for research, the lavnun attaches its spawn to stones in shallow water, January to March. Brought to the laboratory for hatching, this spawn provided larvae of known age, so that Dr. Walline could identify the daily rings on larval otoliths. Lavnun larvae taken by neuston net in the lake were found to be slow-growing relative to larvae of most other fish species: in 6 – 15 mm specimens, increments averaged only 0.167 mm/day (5 mm/month). Analysis of cast-net and beach-seine samples taken March to August indicated a similar growth rate in 18 – 32 mm lavnun.
Between April and June, concentrations of the Kinneret sardine 3 – 6 cm total length are observed by fishermen and have been taken by beach-seine and cast-net (Landau et al, 1988; Landau, 1991). Since spawning extends from November to March, it seems reasonable to assume that these are young of the year. However, operculum observations and growth data for larvae and small specimens dictate otherwise: sardines of ~4.5 cm mean length are yearlings.
In examining opercula, I sought structural differentiation in the area formed when the fish was 3 to 6 cm in length. A wavy ridge could be felt with a fine needle and seen with the aid of polarized light, extending from margin to margin of the operculum. These criteria were used to define annuli, which I counted and recounted until consistent ‘age readings’ were obtained.
From age-at-length data, the von Bertalanffy parameters L (infinity) and K (growth co-efficient) were calculated: 16.4 cm and 0.25 for males, 19.7 and 0.l4 for females (see figure entitled 'Mirogrex growth curves' and (Landau, 1991). At the size of full recruitment to the commercial fishery, 13 cm TL, lavnun are 5 to 7 years of age. Brood biomass reaches a maximum at ~8 years, if exploitation is not excessive.
Female lavnun as large as 25 cm have been observed; however, according to records going back as far as 1942 (Ricardo-Bertram, 1947; Landau, 1991), few females are over 20 cm, and male sardines seldom exceed 16 cm. This suggests constancy in growth pattern. Whereas the maximum age of lavnun in the lake may be higher than 25 years, few survive beyond 17 years, either male or female.
Validation and application of the growth hypothesis
'The proof of the pudding is in the eating'. If the growth hypothesis is correct, then changes in lavnun biomass are predictable several years in advance, by analysis of size distribution. And, with the identity and relative strength of year-classes established, it should be possible to relate changes in lavnun biomass to conditions in the lake and its catchment.
Size distribution data collected in the 1950's foretold the dramatic rise in lavnun abundance in the early 1960’s which was described by Reich (1978), and the timing of this rise links it to drainage of Lake Hula (Landau, 1991). The slower rise that followed may also be attributed to conditions in the catchment area, as fish pond acreage increased during this period.
While peak catches were being taken in the 1986/87 season, I predicted a stock decline due to recruitment failure, as shown by the scarcity of small specimens (Landau et al, 1988; Landau, 1991). Catch-per-unit-effort in the commercial fishery decreased after 1987 (Landau, 1996) resulting in the near-total collapse of the fishery in 1993 (see Mirogrex biomass;Hambright & Shapiro, 1997). Inspection of size-frequency data showed a decline in brood strength beginning in 1979, and becoming steeper in 1982. Thus the decline in sardine abundance was linked to reduction in fish pond drainage into L. Kinneret, and to the first operation of settling ponds in the Hula valley.
In 1992 fishermen observed unusually great concentrations of tiny sardines. As the winter of 1991/92 was especially rainy, an increase in abundance due to the influx of organic matter could be expected. On the basis of the growth hypothesis and the assumption that the sardine concentrations were mainly comprised of the 1991/92 brood, I predicted (correctly) that in 1997 the commercial stock would be bolstered by the recruitment of this strong brood (Landau, 1993).
After sampling the 'dilul' catch and inspecting their opercula, I identified two exceptionally strong broods, the 1990/91 and 1991/92 year-classes (Landau, 1996; 1997). The slow-growth hypothesis is validated by the gradual size-distribution changes in the 'dilul' fishery between December 1993 and Jan - March 1998 (Landau, 1997; 1998). The Figure 'Size and age distribution of lavnun in the 'dilul' fishery' presents data for males and females combined, and mean size-at-age calculated from the von Bertalanffy growth equation (given above) is super-imposed on the size frequency graphs.
Countering criticism of the growth hypothesis
In spite of the strong evidence in its favor, the slow-growth hypothesis described above has not been adopted by Kinneret researchers and administrators. Likewise, the peak estimates of 15,000 - 20,000 T for the mid-1980's (Landau, 1991) and 30,000 for 1997(Landau, 1998) are not given consideration. All 'acceptable' sardine stock estimates vary between 1000 and 5000 T. Yet in order to support the 'dilul' project, this small sardine stock must be shown to have an impact on water quality through its predation on zooplankton.
Ostrovsky & Walline (1999) achieve this goal by imposing high growth rates on the lavnun, from which they derive high rates of food consumption relative to fish biomass. Ignoring the years of dwindling abundance after 1987, they claim that the 'stability' of the lavnun stock was upset by an especially strong 1992 cohort, so that food requirement, calculated at 70 times fish biomass per year, was not met by zooplankton productivity in 1993. Then starvation caused the collapse of the commercial fishery in 1993. To restore stability, removal of undersized sardines is recommended.
Choosing scales for aging the lavnun, Ostrovsky & Walline (1999) propose to overcome "the failure of previous studies" by combining methods, including "observations of the size of the smallest lavnun at different times of the year". But they do not mention that 3 cm lavnun have been found in August. Nor do their ‘reliable basic data’ include the recurring concentrations of ~4.5-cm lavnun, April-June, mentioned by Landau, Walline and Gophen (1988).
Noting that a bimodal group of lavnun, 5-10 cm total length, taken by close-mesh purse seine in May 1993, was isolated from larger-sized lavnun ( >14 cm), Ostrovsky & Walline (1999) claim a yearly recurrence of 5 - 10 cm lavnun, supporting their hypothesis that this group consists of yearlings.
In the extensive records of lavnun size-distribution, there is no evidence of such a recurring, isolated grouping. On the contrary, in previous years there was a continuum in size distribution data, typical of slow-growing fish. This is especially clear in the 1942 data for commercial cast-net catches, due to the low minimum size prevailing at that time (Ricardo-Bertram, 1947; Landau, 1991).
The size distribution gap observed in 1993 resulted from recruitment failure followed by super-abundance. Observations on the 'dilul' fishery illustrate the gradual bridging of this gap, i.e. slow growth (Landau, 1997; 1998 and Figure entitled 'Size and age distribution of lavnun in the 'dilul' fishery'). Ostrovsky & Walline (1999) suppressed this information.
A size of 5 - 10 cm in yearlings cannot reconciled with observations on early growth by Dr. Walline (Landau et al, 1988) as this assumes a continuous increment of 5 – 6 mm/month well into the second year of the fish’s life. It is well known that fish growth slows down before the second year and may cease completely in winter.
Applications of the lavnun slow-growth hypothesis to the assessment of stock, food consumption and parasite-induced mortality can be found in the Mirogrex biomass section . Use of the growth hypothesis to identify ecological factors is discussed in the preceding sub-section, as well as in Landau (199l) and other sources.