Larval trawling techniques and analyses (hypothesis 5)
To aid in interpretation of otolith trace element signatures and infer larval behaviour during their pelagic development, larval trawls were conducted between 9–16 Jan 2016, aligning with the Sep –Nov peak spawning period of landlocked G. brevipinnis (McDowall, 1990). This time period ensured that both newly-hatched larvae and older cohorts were present, providing insight into where larvae were located for the duration of their entire pelagic development period. Night time trawls were conducted given that nocturnal migrations of larvae into surface waters commonly occur (Taylor, Graynoth, & James, 2000). Larval habitat use in the lakes was assessed in three habitats: river plumes, nearshore, and offshore. River plume sites were located near river mouths in the turbid water outflowing from four tributaries from each lake, including the two largest tributaries (Fig. 1; Wakatipu: Rees River, Dart River; Wanaka: Makarora River, Matukituki River) and two smaller tributaries (Wakatipu: Greenstone River, Buckler Burn; Wanaka: Albert Burn, Boundary Creek). Nearshore sites were located < 30 m from the shoreline and > 100 m from other tributaries. Offshore sites were > 200 m from the shore. Sampling was conducted in a hierarchical design where each tributary was sampled as a set of plume, nearshore and offshore sites.
Larvae were sampled using an ichthyoplankton tow net (125 x 50 cm, 250 µm mesh) with a flow meter (Ocean Test Equipment, Inc.; www.oceantestequip.com) mounted in the mouth of the trawl. Tows lasted 4 mins at approximately 4 kmh-1. Plumes were often relatively small, so were trawled in a circular motion to remain within them. Trawl contents were immediately placed in 5% formalin. In the laboratory, larvae were separated from organic and inorganic material, and identified to species. To confirm species identification, a subsample of larvae from each trawl was preserved in 90% ethanol in the field, and the identity of 10 randomly selected larvae per trawl was genetically confirmed using BLAST searches (Altschul, Gish, Miller, Myers, & Lipman, 1990) of cytochrome b, amplified following the methods of Waters, Craw, Youngson, and Wallis (2001). Gobiomorphus cotidianus was the only other species likely to be encountered in these lakes, but is morphologically distinct from G. brevipinnis(Jarvis & Closs, 2015). The density of larvae was determined by dividing total number captured by the volume of water filtered (as determined by the net measurements, flow meter, and trawl time), resulting in a standardized estimate of the number of larvae per m3 of water.
A G -test was used to assess differences in larval distribution differences across offshore, near shore and tributary habitats of each tributary. A one-tailed test was employed as larval densities are predicted to be higher in plumes than other sampling locations.