ONTOGENY AND MORPHOLOGY LIMITS OF INTRACOHORT CANNIBALISM DURING THE EARLY-LIFE STAGES OF DOURADO Salminus brasiliensis

Maria Célia Portella*, Flavio F. Ribeiro
 
São Paulo State University (Unesp), Aquaculture Centre of Unesp, Jaboticabal, SP, Brazil, ZIP 14884-900; portella@caunesp.unesp.br
 

Dourado is a native fish species with great potential for aquaculture in South America. Nevertheless, the main bottleneck for its rearing propagation is the high incidence of intracohort cannibalism, hindering the juvenile production. Therefore, the present study detailed the ontogeny and determined the morphology limits of intracohort cannibalism during the early life stage of dourado. The outcomes of this study provide the bases for the development of novel techniques of cannibalism reduction for the species.

Replicated small-scale larvicultures of dourado were performed in 5×30L aquaria in order to assess the ontogeny of the cannibalistic behaviour. A large-scale larviculture was carried-out in parallel, from where periodic fish samples were collected and measured for morphological traits in order to construct predictive models of the limits of cannibalism. Finally, these models were then cross-validated with empirical data collected from a predation experiments were cannibals were challenged with progressively increment of prey sizes.  

Cannibalism was extremely intense in the first days of life, removing 85.7±3.8% of the original population by 05 days after hatching (DAH). The vast majority of predation acts within this period was orientated tail-first. However, its effectiveness reduced down as fish grew older, becoming fully inefficient for fish older than 07 DAH (Figure 1). From this moment onwards, cannibalism turned exclusively head-orientated, but less intense. By 10 DAH, cannibalism removed 97.7±0.5% of the original population. Morphological models predicted that younger cannibals are able to ingest prey larger than their own sizes (Figure 2). However, as cannibals grew larger, this maximum ingestible prey size reduced down to around 80% (Figure 2). The alternative model (TLprey = 0.705TLcannibal + 5.650; where TL is total body length) better matched the empirical results as it considered the inter-individual variability in the morphological traits and therefore better represents the realistic limits of cannibalism for this species.