Estimating microcarnivore diel consumption pattern
Fish diel consumption of food can be estimated from samples at the midpoint of feeding activity by sampling gut contents at the middle and end of the normal feeding window (Arnal and Côté 2000). Our estimate for nocturnal microcarnivores was based on previously unpublished gut-content data collected in May of 1977 by one of the authors (Grippo). Diel pattern was assessed by sampling gut content of three Myripristis jacobus at midnight and at dawn. Fish were collected from a shallow 3 m reef in Taque Bay, St. Croix, U.S. Virgin Islands. Fish were speared and iced, and their gut contents removed by dissection. Gut contents were twice rinsed in fresh water and centrifuged to remove salts, then oven dried. Each sample was then weighed. The ratio of late-night sample to diel feeding was estimated by summing the mean mass of midnight (late night) gut content with the mass of dawn gut content and dividing this sum by the mass of midnight gut content. Standard error of these measurements was estimated using 10,000 bootstrap samples with replacement.
Calculation of derived traits
Estimates of fish biomass
For nocturnal microcarnivorous fish, biomass of fish was approximated using mass-from-length estimates (Marks and Klomp 2003) based on published parameters summarized on the Fishbase website (www.fishbase.org). Parameter estimates based on multiple sources provide less variability. In cases where only one or a few parameter-estimate sources were available, estimates derived from a broader taxonomic level were used—for example, estimates from all congenerics pooled and used in place of species-level parameters. The seven sources of parameter estimates available on Fishbase for Haemulon flavolineatum were pooled. Only two sources of parameter estimates were available for M. jacobus so the 26 sources available for Myripristis on Fishbase were pooled and used. Fish were identified using a common field fish identification reference (Humann and DeLoach 2002), and as this source predates taxonomic revisions moving various Apogon sp. into other genera, we pooled all 37 parameter-estimate sources for the family Apogonidae for Apogon sp. estimates. Squirrelfish parameters were derived by pooling parameter estimates from all 13 sources for the genus Holocentrus and all 23 sources for the genus Sargocentron. The mean standard length (SL) of fish caught in each taxon was used to compute the mean estimated mass of individuals caught in that taxon. This estimated mean fish mass was then combined with the mean number of gnathiids consumed to assess the efficacy of each fish taxon at gnathiid removal.
Estimates of gnathiids removed per reef area
In each bootstrap iteration, gnathiid consumption at a single point in the diel cycle was estimated by resampling gnathiid counts from gut-content data. The multiplier used to estimate daily gnathiid consumption by microcarnivores was similarly resampled each bootstrap iteration. The multiplier for cleaning gobies was randomly varied in each bootstrap sample by randomly varying the morning and afternoon sampling estimates using the standard errors reported by Arnal and Côté (2000). In each bootstrap iteration, the gnathiid consumption per fish predator biomass was estimated by combining the product of the resampled estimate of gnathiid gut-content counts for that fish taxon and the appropriate diel-period multiplier. This value was then divided by the estimated biomass of the average-sized fish of that taxon based on SL and determined as described above. The gnathiid consumption by each fish taxon per surface area of reef plot for each of the three plots sampled was similarly determined by combining the product of the appropriate resampled gut-content gnathiid count and diel-period multiplier with the fish count for that taxon on that reef plot.
References
Arnal C, Côté IM (2000) Diet of broadstripe cleaning gobies on a Barbadian reef. J Fish Biol 57:1075–1082
Humann P, DeLoach N (2002) Reef Fish Identification — Florida, Caribbean, Bahamas. New World Publications, Jacksonville, FL
Marks KW, Klomp KD (2003) Appendix two: fish biomass conversion equations. Atoll Res Bull 496:625–628