It is well known that day/light cycles regulate the circadian rhythms of marine fishes. Evidence for diversication corresponding to a “temporal niche” is readily apparent in the eyes of marine fishes, where adaptations to dim-light conditions have constrained optical and trophic diversity. True nocturnality has evolved independently within dozens of marine fish families, making fishes an excellent system for assessing how human-driven disturbances, such as light pollution or contaminants, affect circadian rhythms in wild species. While the vast species-richness of marine ray finned fishes presents a potential boon to research, it also presents the inordinate challenge of first quantifying activity patterns across a quarter of all living vertebrates.
I recently had the opportunity to attend a special conference hosted by the Bulletin of Marine Science in Miami this November that focused on “fish at night”. Our work for this conference focused on overviewing just what we know about which fishes actually are nocturnal. It turns out there are some pretty striking gaps in our knowledge. Light Pollution in Curaçao. Changes in light effect not only the behavior of humans and other animals, but have also been linked to long-term health problems. For most species of fish, we have no idea how artificial light pollution impacts there lives .
We just published a study that investigated where our knowledge gaps in circadian rhythms are. It turns out that there are some readily apparent biases in where we collect data on marine fish activity patterns.
By surveying the literature and coding every study on fish activity we could find we compiled a database of over 800 marine species that have been studied. This is pretty amazing in itself. The ability to track species and study marine species is incredibly difficult. Given the difficulties of working underwater, it is probably not surprising that most behavioral studies come from coral reefs. In contrast, we know far less about the fishes off the coast in New England.
I actually have spent quite a bit of time in the waters of Connecticut and Rhode Island and can tell you first hand that designing a transect study to visually monitor or film activity would be very difficult. I have had numerous dives where visibility was measured in inches, not feet. That said, good days can produce some really excellent dives such as the one we filmed below.
In contrast, a good day of tropical diving can often have visibility beyond 100 feet, and most days will have excellent visibility.
However, even in tropical areas, we are overlooking large groups of taxa. If we map what species have been studied unto a phylogeny of all fishes, it is clear that some groups are well represented, other groups (like blennies and gobies) are barely represented at all.
(A) Significant clustering of sampled (dark) and unsampled (light) in taxon sampling in the Rabosky et al. (2013) topology pruned to only reef associated species relative to a diversity tree containing all known reef associated fish species. (B) Significant clustering of sampled (dark) and unsampled (light) in activity data relative to a diversity tree containing all known reef associated fish species. Black lines in the phylogeny represent no deviation in under or over dispersion of taxon sampling. Names represent signicantly under-sampled named clades. Cartoon illustration of a butterfly fish (Chaetodontidae) represents one of the reef fish clades highlighted as highly sampled in our analyses.
More globally, this group specific pattern persists.
(A) Patterns of over (dark) and under-dispersion (light) in taxon sampling in the Rabosky et al. (2013) topology pruned to only marine species relative to a diversity tree containing all known marine fish species. (B) Patterns of over (dark) and under-dispersion (light) in activity data relative to a diversity tree containing all known marine fish species. Black lines in the phylogeny represent no deviation in under or over dispersion of taxon sampling. Names represent significantly over-dispersed named clades. Illustration of a sturgeon (Acipenseridae) represents the earliest diverging lineage in our analyses.
This mapping exercise is illustrative a major problem in science that matters quite a bit. We live in the age of big data and compiling lots of historic information is easier than ever. However, independent studies are not conducted with the intent of future meta-study. As such, it is up to the investigator if their synthesis of the literature is well-designed or not. As it stands right now, making sweeping statements about the ecology and evolution of circadian rhythms in marine fishes would be largely driven by data from a few groups of tropical reef species.
In the theme of the Fish at Night symposium, shining a light on fish at night is in some ways analogous to exploring biases lurking beneath the surface of an existing data structure. Focusing research efforts on one underrepresented component of marine vertebrates holds great potential for scientific discovery. Likewise, addressing emerging challenges associated with the accumulation of large data sets, such as sampling bias, also holds great promise.
Assessing sampling bias patterns can increase the efficiency of experimental design, saving both research time and costs. Scrutinizing data for bias patterns can also spur the development of new methodologies while empowering new discoveries. As global biodiversity patterns continue to change in response to new pressures associated with the Anthropocene, such scrutiny of ecological or behavioral data will be essen- tial if we are to accurately forecast and manage the future of the planet’s biodiversity.