Abstract.-The spawning seasonality, fecundity, and daily Reproductive biology and egg production of three species of short-lived c1upeids, the sardine egg production of three species of Amblygaster sirm, the herring Herklotsichthys quadrimaculatus, Clupeidae from Kiribati, and the sprat Spratelloides delicatulus were examined in Kiribati to assess whether vari­ tropical central Pacific able recruitment was related to egg production. All species were David A. Milton multiple spawners, reproducing throughout the year. Periods of Stephen J. M. Blaber increased spawning activity were Nicholas J. F. Rawlinson not related to seasonal changes in (SIRO Division of Fisheries, Marine Laboratories, the physical environment. Spawn­ ing activity and fish fecundity P.O. Box J20. Cleveland, Queensland 4 J63. Australia were related to available energy reserves and, hence, food supply. The batch fecundity ofA. sirm and S. delicatulus also varied inversely with hydrated oocyte weight. The maximum reproductive life The sprat Spratelloides delicatulus, Changes in abundance may be span of each species was less than the herring Herklotsichthys quadri­ related to variable or irregular re­ nine months and averaged two to maculatus, and the sardine Ambly­ cruitment, because many clupeoids three months. Each species had a gaster sirm are the dominant tuna (especially clupeids and engraulids) similar spawning frequency of three to five days, but this varied baitfish species in the Republic of have little capacity to compensate more in A. sirm and S. delica­ Kiribati (Rawlinson et aI., 1992). for environmental variation during tulus. Amblygaster sirm had the All three species inhabit coral reef the period ofpeak spawning and egg highest fecundity and potential lagoons and adjacent waters. production (Cushing, 1967, 1971). lifetime egg production, but the Sprats school in shallow water Most clupeids, including some number of eggs produced per ki­ logram of fish was highest in the around reefs and adjacent seagrass tropical species, are multiple small sprat S. delicatulus. during the day. Herring also form spawners (Alheit, 1989). Multiple Monthly estimates of the daily dense schools in shallow water spawning should be advantageous egg production of each species along the shoreline and among for short-lived species because it varied with the proportion of the reefs during the day (Williams and enables them to maintain rela­ population that was spawning. Estimates of egg production Clarke, 1983). Unlike the other tively stable population sizes in showed little similarity to the fre­ species, sardines school near the unpredictable environments quency distribution of birthdates bottom ofthe lagoon during the day (Armstrong and Shelton, 1990). back-calculated from length-fre­ (Conand, 1988). All species disperse Multiple spawning has been estab­ quency samples. The distribution into the mid and upper waters of lished for few tropical clupeids of back-calculated birthdates con­ firmed that fish spawned in all the lagoon during the night to feed (e.g., Sardinella brasiliensis; Isaac­ months, but the proportion born and become available to the com­ Nahum et aI., 1988). Of the three each month varied widely from mercial fishery. major baitfish species in Kiribati, species to species and year to year. A major source of lost fishing only S. delicatulus has been shown The reproductive strategy of these time by pole-and-line vessels in to be a multiple-spawner (Milton species ensures that successful spawning is likely, and so the Kiribati has been irregular baitfish and Blaber, 1991). All three species level of recruitment is more de­ catches (MacInnes, 1990). These are subject to high natural mortal­ pendent on post-hatching survival important tuna baitfish species ity in Kiribati (Rawlinson et aI., rates than on egg production. have shown large seasonal and 1992), thus lifetime egg production interannual fluctuations in abun­ dance since they were first re­ corded during the 1940's (McCar­ 1 McCarthy, D. 1985. Fishery dynamics and 1 biology of the major wild baitfish species thy, 1985 ; Rawlinson et aI., 1992). particulary Spratelloides delicatulus, from Both A. sirm and H. quadrima­ Tarawa, Kiribati. Kiribati Fisheries Div., culatus disappear from baitfish Tarawa, Kiribati, 53 p. catches for variable periods and 2 Kiribati Fisheries Division. 1989. Fisher­ ies Division 1989 Annual Rep., Ministry Manuscript accepted 24 September 1993 can be absent for months or years of Natural Resources Development, Fishery Bulletin 92:102-121 (1994) (Kiribati Fisheries Division, 19892). Tarawa, Kiribati, 38 p. 102 Milton et al.: Reproductive biology and egg production of three species of Clupeidae 103 may be increased ifthey spawned multiple batches In this study, we examined the variability in re­ of eggs. productive biology of the three major baitfishes in Egg production of multiple spawning species de­ Kiribati to determine the influence of adult repro­ pends on reproductive life span, the time between ductive variability on subsequent recruitment. Our spawnings, and the age structure ofthe population objective was to test the hypothesis that reproduc­ (Parrish et al., 1986). Batch fecundity ofS. delica­ tive biology of short-lived clupeids is adapted to tulus varies widely between sites, both within and maintaining relatively stable population sizes. We between countries (Milton et aI., 1990). In a short­ determined potential life-time egg production and lived species such as S. delicatulus «5 months; whether estimated egg production is related to the Milton et aI., 1991), reproductive life span may have frequency distribution ofback-calculated birthdates. an important influence on potential lifetime egg production. Batch fecundity of H. quadrimaculatus does not Methods and materials appear to vary throughout its distribution, and ranges from 4,000 to 10,000 eggs (Marichamy, 1971; Study areas Hida and Uchiyama, 1977; Williams and Clarke, The Republic of Kiribati covers an area of 3 x 106 1983; Moussac and Poupon, 1986; Conand, 1988). km2 in the central Pacific ocean and comprises three Fish mature at about 90 mm in length at six months main island groups (Gilbert, Phoenix, and Line Is­ of age (Williams and Clarke, 1983), and they sur­ lands) (see Inset Fig. 1). The Gilbert Island group vive for at least one year (Milton et aI., 1993). Little is the most populated, consisting of 16 coral reef is known offecundity and egg production ofA. sirm. islands. All islands in the group have a typical ocean Fecundity of the species is related to length and platform coral reefstructure and have been built up weight, with a mean of 20,000 eggs per batch, and by sc1eractinian corals and coralline algae on a sub­ individuals probably spawn more than one batch of merged mountain (Gilmour and Colman, 19903). eggs (Conand, 1988). Most atolls consist ofsmall islets lying on the east­ Temperate clupeids vary widely in life-history ern side of a lagoon with an open western side due parameters (e.g., Clupea spp., Jennings and to the prevailing easterly winds. Most typically have Beverton, 1991). Food availability and environmen­ passages between the islets through which water is tal conditions affect the size and number of eggs of exchanged. Pacific herring (Clupea pallasi) (Hay and Brett, The four study sites (Abaiang, Butaritari, Tarawa, 1988). Results of studies of temperate clupeoids and Abemema) were typical ofislands in the Gilbert suggest that they do not spawn during periods of Island group; all had narrow islets on their south­ high food abundance, but store energy as fat for ern and eastern sides, except Abaiang (Fig. 1). La­ later reproductive activity (Hunter and Leong, 1981; goons were mainly shallow (20-30 m deep), often lIes, 1984). There are no similar studies of tropical with large areas of intertidal seagrass or sand on clupeids. Encrasicholina heterolobus, a tropical their eastern sides. Bottom topography ofthe deeper engraulid, does not deplete energy reserves in the parts ofthe lagoon was generally smooth, with some liver or soma during spawning (Wright, 1990). Fish coral outcrops. Our study sites were similar to those with higher condition factor (K) also had higher described by Hobson and Chess (1978) in the fecundity. Marshall Islands. Stored energy or fish condition that may influence both spawning frequency and batch fecundity have Environmental parameters a marked influence on egg production and, hence, affect subsequent recruitment (Ricker, 1954; On each sampling occasion, we measured the time Beverton and Holt, 1957). Adult reproductive varia­ of collection, sea surface temperature (OC), cloud tion should strongly influence recruitment in short­ cover (okters), wind direction and speed. and moon lived tropical species that have short larval phases phase because these factors may be related to and rapid growth. An example is S. delicatulus spawning or recruitment (Dalzell, 1985, 1987; which, in the Solomon Islands, live a maximum of Peterman and Bradford, 1987; Milton and Blaber, five months and mature at about two months ofage 1991). For each site, monthly rainfall data for 1989 (Milton and Blaber, 1991; Milton et aI., 1991). Amblygaster sirm and H. quadrimaculatus live less than two years (Milton et aI., 1993) and mature in 3 Gilmour, A. J., and R. Colman. 1990. Report on a consultancy on a pilot environmental study ofthe outer island development 6-12 months (Williams and Clarke, 1983; Conand, program, Republic of Kiribati. Graduate School ofthe Environ­ 1988).