The prevention of fouling on mariculture structures is complicated by the choice of net material and the dangers of toxins to cultured species. Multi-filament netting
material is an ideal substrate for fouling: it is non-toxic, contains many crevices that can entrap and protect settling organisms, and has a high surface-area to volume ratio.
Although copper-based antifoulants have proved effective on nets, their use is undesir- able because of environmental effects from broad-spectrum metal-based toxins, together
Ž .
with consumer concerns that can jeopardise market image Lewis, 1994a . Problems associated with mariculture antifouling have recently received publicity.
The Scottish Environment Protection Authority has found sediments underneath fish Ž
. cages to be seriously contaminated with copper Miller, 1998 . Further, the Norwegian
aquaculture industry is working toward a significant reduction in the use of copper-based Ž
antifouling by the year 2010 Norwegian Pollution Control Authority, personal commu- .
nication . There have been incidents where antifouling has adversely affected fish: in the 1980’s, trials with tributyl-tin on cages caused significant effects to farmed salmon
Ž .
Short and Thrower, 1986; Davies and McKie, 1987 and, more recently, boat antifoul- Ž
. ing was implicated in residues within wild fish Kannan et al., 1995a,b .
Internationally, the development of more environmentally acceptable antifoulants is focused on two strategies: the production of ‘‘foul-release’’ surface coatings and
Ž .
coatings that release non-toxic compounds that act as deterrents Clare et al., 1992 . The Ž
former strategy is based on a physical surface effect, where low surface-free-energy low .
surface tension prevents adhesion or reduces adhesion strength of fouling organisms Ž
. Lewis, 1994b . Substantial fouling is prevented because large masses slough from the
surface and are easily removed by hydrodynamic forces, such as across a ship’s hull, or Ž
. by light cleaning Schultz et al., 1999 . The most successful coatings are presently based
on silicone elastomers, and often include oil-based additives that further improve fouling Ž
. resistance Lewis, 1994b; Swain and Schultz, 1996 . Coatings based on other materials,
Ž .
Ž including fluoropolyurethanes
Bultman and Griffith, 1994 , polyurethanes Lewis,
. Ž
. 1994b and perfluorinated polymers Lindner, 1994 , have also displayed excellent
adhesion resistance. Ž
This paper describes the evaluation of a commercial silicone coating Veridian 2000, .
International Coatings as applied to fish-cage netting. The development, composition and adhesion of fouling are compared between white silicone-coated netting, white
uncoated netting and black uncoated netting. The preferential settlement of fouling species in relation to substrate colour and the adaptation of species to low-surface-free
energy are described and discussed. The flexibility and non-toxic properties of silicone coatings make them highly suitable for fish-cage netting. Results show that they can be
used in conjunction with underwater cleaning equipment, alleviating the need to remove netting from the water for cleaning.
2. Materials and methods
2.1. Study site Trials were conducted at the Huon Aquaculture Company at Hideaway Bay,
Ž .
147.0678E, 43.2708S in the Huon River, Tasmania, Australia. The site is fully marine,
except for a 2–5‰ salinity drop within 1 m of the surface after high rainfall in winter. Water temperatures range from 118C in winter to 178C in summer. Water movement is
dominated by tidal flow and current speed varies from 5 to 20 cmrs.
2.2. Panel and raft design Ž
. On September 8, 1995, 21 panels each 1 = 1 m, 96 ply, 20 mm bar netting were
Ž .
immersed to evaluate the silicone-based product, Veridian 2000 International Coatings . Ž .
Seven replicates were used for each of the three netting types: 1 white silicone-coated Ž .
Ž . netting, 2 uncoated white netting and 3 uncoated black netting. Veridian 2000
consists of a white pigmented tie-coat and a clear finish. Nets were coated by repeated immersion in a bucket and the tie coat was dried for 3 h prior to application of the
finish. The finish was dried for 48 h before attachment of netting to panels. A volume of
Ž .
2
3.8 l 1 US gal was sufficient to coat 9 m of net. A white tie-coat was used and
applied to black netting to aid any observation of cracking or coating loss. Mean coating thickness around the twine was 370 mm.
The panels were attached to a raft tethered within an unstocked salmon cage of 65 m circumference. The raft was constructed of seven 6 m long beams fixed in parallel at
1.5 m intervals. Panels were placed in a one-way block design with each of the three treatments placed randomly within the seven beams of the raft. This design was chosen
to compensate for the north–south current flow and differences in fouling mass typically observed between the northern and southern sides of cages. The panels held the netting
between 0.5 and 1.5 m depth, so that netting was out of the turbulence of the surface waters where fouling development is generally less common. A 25 mm mesh net
Ž
. regularly changed and cleaned was attached to the cage, therefore surrounding the raft
Ž .
and panels, to protect the panels from grazing by fish. Swain et al. 1998 reported that reduced fouling on silicone coatings might be partly attributed to the increased success
of fouling removal by grazing fish. Fouling development was monitored by wide-angle photography using a Nikonos-V
camera with a 20 mm lens. Photographs were taken after 9 days immersion and then at monthly intervals for 5 months.
2.3. Adhesion strength, composition and fouling biomass Ž
. After 140 days immersion, the nine panels i.e., three replicates per net type on the
southernmost three beams were removed to quantify ease of fouling removal and fouling Ž
composition. The central 30 = 30 mesh-hole area was sampled from each panel 56 of .
the total area and divided into four 15 = 15 mesh areas for ease of sampling. The panels were mounted horizontally and each side of each 15 = 15 mesh area cleaned with a
Ž .
water jet at a fixed distance above the net 30 cm . Constant water pressure was Ž
. maintained throughout the trial 115 psi, delivering 0.33 lrs and each area was cleaned
for 10 s. The panel mesh surrounding each sample area was masked by plastic sheeting during cleaning. Dislodged fouling was collected below the panel in a bucket and
strained through 1-mm mesh. Fouling still attached to the net was removed with forceps. Samples were fixed with 10 formalin in seawater and stored at 48C in the dark prior to
analysis.
The total wet weight of remnant and dislodged fouling from each area was measured Ž
. to 0.1 mg after the excess water was drained from each sample through 1-mm mesh .
Each sample was then sorted into species, dried to a constant weight at 608C, and the total mass for each species weighed to 0.1 mg.
2.4. Underwater mechanical cleaning After 163 days immersion, the remaining 12 panels were removed to quantify ease of
Ž fouling removal with a prototype underwater cleaning machine see Hodson et al.,
. 1997 . The cleaning machine was configured with four pairs of 800 mm long contra-
rotating brushes. Pairs were mounted at 1, 2, 3 and 4 m depth with the upper pair used for cleaning panels in the current trial. Brushes were attached to a gearbox mounted on
the handrail of a salmon cage and driven by a petrol-powered hydraulic unit. The brushes rotated at 25 rpm, with the gearbox moving along the handrail at 0.3 mrmin.
Ž .
Two rows on the raft were chosen randomly from the four and the six panels then attached to the side of a salmon cage and cleaned. The netting was then removed from
the six cleaned panels, and from the remaining six non-cleaned panels, and weighed to 0.01 kg. Wide-angle underwater photographs were taken to record the level of fouling
on each panel. The total wet weight of attached fouling was determined by subtracting Ž
2
the weight of clean wet netting silicone coated s 0.75 kgrm ; untreated net s 0.47
2
. kgrm
from the gross wet weight. 2.5. Statistical analysis
Ž .
The biomass per species, total attached fouling algal, invertebrate and combined and Ž
total dislodged material were compared between each netting type black, white and .
Ž .
silicone-coated using one-way analysis of variance ANOVA . Data shown to have significant differences were further analysed using Tukey’s test. Data was analysed by
ANOVA, rather than as a blocked design, because variation between blocks was not significant. Prior to ANOVA, each data set was tested for homogeneity of variance
Ž
. Ž
. Cochran’s test and normality using residuals, Shapiro–Wilk W test . Data that did not
meet these assumptions were transformed to the common log prior to analysis. In cases Ž . Ž
where a zero occurred in a replicate s e.g., where a species was not found on each
. Ž
. Ž .
panel all data were transformed by log Y q 1 Sokal and Rohlf, 1995 .
Results from the trial with in situ mechanical cleaning were analysed by two-way ANOVA, using netting type and cleaning treatment as the two factors, followed by
