As response variable to he perceiving of fear in mussels we choose to focus on the formation of byssus treads. Specifically because this inducible defence can be measured within days, while other defences like shell thickness and muscle thickness takes much longer to develop. A toolkit was developed to quickly assess effects of the presence of various cues on the formation and subsequent detachment of byssus treads.
Experimental set-up
A peristaltic water pump was used to pump filtered sea water into 12 containers, 29.0 x 23.5 x 28.5 cm, at a rate of approximately 46.9 mL/min (200.0 rpm). Previously the incoming water was passed through a compartment with carbon chips in order to remove any pre-existing chemical cues in the water. These containers would hold the variables being tested, as listed in Table 1.
Table 1
Mussels were exposed to water with 12 different cues, cues were created in a separate tank as specified in this table.
ID
|
Treatment
|
1
|
Control
|
2
|
Crushed mussels (including meat)
|
3
|
Crushed mussel shells (no meat)
|
4
|
1 Crab
|
5
|
5 Crabs
|
6
|
10 Crabs
|
7
|
1 Crab + 1 Sea Star
|
8
|
1 Crab + 1 Sea Star + Mussels
|
9
|
1 Sea Star
|
10
|
1 Sea Star + Mussels
|
11
|
1 Sea Star + Crushed mussels
|
12
|
Hydrostress
|
An outlet tube on one side of the container allowed water, and any chemical cues released from the variable, to flow into a nearby jig. The jig, held 15 clear, plastic containers of 4.3 x 4.3 x 4.3 cm. Indentations in the edges of the plastic container allows the water to flow through each of the containers. Holes at each end of the jig allows excess water to drain out, so that the water level is constantly maintained at the approximate height of the plastic containers. The plastic containers each housed one mussel for a treatment period of 5 days. The total experimental set up is summarized in Fig. 1. Since the plastic container is clear, byssal threads and/or plaques were easily seen and counted at the end of each treatment period (Fig. 2).
Variables were tested in triplicate in three consecutive runs and the order was randomized over all experimental runs. Throughout the 5 days it was ensured that conditions were constant. If a crab or sea star died during the trial, they were replaced immediately. In order to test if the magnitude of the induced defense is related to the magnitude of the dose, mussels were exposed to a set of treatments in which the number of predators, in this case crabs, increased from 1 to 5 to 10, with the implication (because flow was constant and similar over all treatments) that an increase in predators would increase the amount of chemical cues released in the water. In order to establish if there is a difference in response to the stimuli as perceived via chemical cues, the mussel was exposed to a variety of chemical cues such as predators (ie. crabs and sea stars), dead/injured conspecifics, a combination of those factors, and supplemental variables such as only the shells of mussels.
For the variables in which a predator fed on mussels, the mussel seeds were supplied throughout the entire trial period to ensure that there was always something to eat (approximately 10–15 mussels would be consumed per predator over 5 days). Moreover, the variable (1) Sea Star + Mussels (crushed) were added to the list since it was observed in a pre-trial that the sea stars would not feed on the mussels if they were intact. For the Dead/injured Mussels, 10 mussel seeds were crushed and then added to the container. For the Freshly crushed shells 5 empty, adult valves were crushed and then added to the container. Lastly, for Hydrostress, 15 mussels were placed into the variable container alongside an Aquabee UP 2000/1 pump, which continuously supplied air to agitate the water and imitate hydrodynamically stressful conditions for the mussels present. Thus, it would be determined if the mussels would send out chemical cues as “warning signals” of the environmental disruption for other conspecifics.
2.3 Statistical Analysis
After being exposed to the chemical cues from the variable for 5 days, the clear containers housing the mussels were drained of water and the byssal threads were counted under a Gerber Instruments Colony Counter RIVOY-KT1. The distinction between the number of threads still attached to the mussel, and the number of threads that were detached from the mussel was taken into account. If a container had 2 or 3 mussels due to excess movement between containers results were omitted. However, if the mussel had climbed high enough that the byssal threads attached to the sides of the jig rather then the plastic container threads were included in the final count if visible.
The statistical analysis of the results was performed using a generalized linear model (GLM) for count data with a quasi-Poisson distribution to correct for overdispersion. A pairwise comparison between all treatments was performed with a Tukey test using the package multcomp (Hothorn et al., 2008) a p-value of P < 0.05 was considered significant.