Abstract
Greenshell mussels, Perna canaliculus, are cultured in New Zealand using wild spat sourced from two areas. One of these is the Marlborough Sounds marine farming area where specially designed spat-catching ropes are provided as settlement surfaces for the wild larvae. Catches on these ropes vary spatially and temporally, and post settlement losses are often high. An understanding of the factors influencing that variability is essential if the industry is to reduce the unreliability and uncertainty of its spat supply.
The major focus of this study was an examination of the 'supply-side' hypothesis ie. that variation in recruitment on to the ropes is a function of the supply of larvae and settlers. Temporal and spatial patterns of abundance of larvae, settled plantigrades and recruits were examined during four consecutive years at several sites in the Crail/Beatrix/Manaroa embayment of Pelorus Sound. Larvae were counted in daily pumped water samples taken from four depths. Settlement and recruitment were both measured on commercial spat-catching rope at the same depths. Recruitment was assessed 8 and 12 weeks after settlement. Factors influencing distribution patterns of larvae and settled plantigrades are also examined.
Variability in larval abundance and settlement was high but there was strong evidence to support the hypothesis that the supply of early life stages (larvae and settlers) is the major determinant of the number of recruits 8 weeks after settlement. Field experiments indicated that the role played by processes such as predation and other forms of disturbance increases as the recruits grow older. By 12 weeks, supply of settling larvae is no longer the major determinant of recruit numbers at many sites.
Caging experiments to exclude fish showed that predation by fish on mussel spat varies among sites and, at least on a short term temporal scale, is heavier at 15 m than at 5 m. Other experiments indicated that some common marine farming practices may combine with natural processes to produce recruitment variability also. Many of these practices appear to act indirectly by enhancing the likelihood of processes such as settlement and predation occurring. Practical applications of the data are discussed.