M was estimated from body weight and habitat
temperature (Ikeda, 1985).We took the temperature value recorded
at the maximum depth of net deployment, which is the temperature
the migrants may find at their upper limit distribution during daytime.
Body weight was estimated from the taxonomic composition samples,
taking fromeach sample an aliquot of 1/2 or 1/4 (depending on the density)
of the original volume. The aliquots were stained for ~24 h in
20 mL of an eosin solution (final concentration 3.75 10−5 mL−1) to enhance
the contrast, and then scanned at a resolution of 600 dpi using a
HP Scanjet 8200 scanner. Gelatinous organisms were mostly
fragmented, so theywere manually removed before scanning.We processed
the imageswith the R-package “EBImage” to identify each animal
and assess its area. Individualweightwas assessed fromthe relationship
between body area and dry mass described by Lehette and Hernández-
León (2009) for crustaceans. To avoid the potential bias due to the large
variety of sizes within each Bongo sample we used the median of the
particle size distribution as representative of the individual body size
in each sample. Larger animals could be underrepresented at diurnal
samples due to DVM (actually, in May the individual body weight for
the Bongo samples was significantly higher at night; Welch's t-test,
p b 0.001). Accordingly, body mass was estimated only from nocturnal
samples. Metabolic rates were converted to carbon by applying the respiratory
quotient of 1.36 reported by Mayzaud et al. (2005) for large
crustaceans, which comprised the bulk of the migrant biomass (see
Results).