Salps were collected using double oblique Bongo tows, with 0.7m diameter frames equipped with 202 µm nets, General Oceanics flow meters, and an RBR temperature depth recorder. Salp specimens (typically 10) from each tow had their guts excised, and chl a and phaeopigments gut contents were measured. A power function was used to fit the size-specific Gpig (chl a + phaeo) contents for each tow, allowing the estimation of Gpig for each size bin per tow, and this was multiplied by the abundance in each size bin. Gut passage time (GPT) was calculated using a modified equation, based on (von Harbou, Dubischar et al. 2011 doi:10.1007/s00227-011-1709-4) where GPT(h) = 2.607*ln(OAL, mm) - 2.6. Grazing was estimated as: G (h-1) = Gpig /GPT, and scaled using a Q10=2. Daily salp grazing rates were obtained by assuming 14 h of day and 10 h of night, coincident with the times and latitudes at which we sampled these communities during the Salp Particle expOrt and Ocean Production (SalpPOOP) campaign. Cycle estimates were calculated by first averaging all day and all night tows separately, and then adding the two estimates. Fecal pellet production was calculated by assuming an egestion efficiency of 0.36 (Huntley, Sykes et al. 1989 doi:10.1007/BF00238291, Pakhomov 2004 doi:10.1016/j.dsr2.2001.03.001, Pakhomov and Froneman 2004 10.1016/j.dsr2.2000.11.002) and converting to carbon values using C:Chl ratios from the phytoplankton growth and grazing experiments combined with NPP, and reported in mg C m-2 d-1. Data is reported by size after binning in 5mm size bins (ranging 1-135mm), and for oozooids and blastozooids separately.

This study was funded by the Ministry for Business, Innovation and Employment (MBIE) of New Zealand, NIWA core programmes Coast and Oceans Food Webs (COES- COES1901) and Ocean Flows (COOF-COOF1902), the Royal Society of New Zealand Marsden Fast-track award to Moira Décima, and NSF award #OCE-1756610 to Michael R. Stukel and Karen E. Selph.