From each subsample, three replicate measurements were analysed by each analytic 
method/device. Subsample bottles were inverted five times before withdrawing each replicate 
to ensure subsamples were well mixed, and all steps of each analysis were performed 
independently on each replicate. Due to time constraints and operational considerations (e.g. 
limited materials), some analytic methods could not be conducted for all trials and in some 
cases, the number of replicates was also limited (e.g. flow cytometry). The total number of 
replicates analyzed for a given comparison is shown in the text above each plot. 
Note that flow cytometric analysis was performed on subsamples preserved with 
formaldehyde (final concentration 2.3 %) and stored at -20°C until analysis; therefore the 
analysis of phytoplankton did not discriminate between viable and non-viable cells but reflects 
the total number of phytoplankton present. 
2.4 Statistical Analysis 
All analyses were performed in the R statistical programming environment (R Core Team 
2015). Since there could be variability between samples collected at the same time using 
different sample collection devices in the ship's engine room, analytic results are only 
compared amongst measurements within the same sample (i.e. collected using the same 
sample collection device from the same sampling point in the same trial). The use of multiple 
sample collection devices is a potential advantage of our experimental design, since it reduces 
any bias that might be introduced by the choice of sample collection device (i.e. potential that 
health and/or diversity of sample may differ between collection methods, which may affect 
performance of analytic methods).