The Manuscript ID MEE-22-08-634, entitled "Using standardized fish-specific autonomous reef monitoring structures (FARMS) to quantify cryptobenthic fish communities." presents a novel approach to sample cryptobenthic fish, which are often overlooked given the methodological difficulties in appropriately quantifying them (without destructive methods). The authors describe the materials, assemblage, deployment and recovery of these structures (Fish-specific autonomous reef monitoring structures FARMS). The authors identify the strengths of the approach in overcoming constraints such as sampling in particular environments/depths/areas that are out of reach to regular diving and sampling techniques. The authors chose to deploy the FARMS in several different environments and habitats for a variable amount of time. Then the authors compared abundance and diversity measurements across habitats using a Bayesian approach. Within location the authors compared community composition, species richness and or abundance at the species level. The results show (the posterior distribution) variation in abundance and richness across all sampling locations. Table 2 presents the relative abundance of species of FARMS and complementary sampling methods combined OSUs and anaesthetics?)! The authors discuss the ability of the FARMS to accurately sample the local diversity as several of the FARMS sampled mostly one species. At some locations the FARMS sampled several species with habitat associations. Another potential “problem” is if the settlement habitat is really the FARMS or the algae that grow on them. The authors then discuss the effectiveness of the FARMS compared to published lists of cryptobenthic fishes to reveal a tremendous gap (Hawaii and Red Sea) and argue that the time of deployment (maturity) of the FARMS may decrease the differences between FARMS diversity and abundance and the local cryptobenthic fish community.

The approach proposed is attractive for multiple reasons and while specifically targets a specific fringe of the fish communities it will likely attract the interest of marine biologists world wide. My recommendation is to accept after major reviews given that it is vital to compare the efficiency of the ARMS in sampling diversity and abundance in each habitat to published cryptobenthic diversity and abundance. I would have thought that this point alone would have guided the authors choice of sampling sites/locations. There are a few points I believe are important to increase the quality of this MS
• The main comparison here should be between the FARMS and both the “real” cryptobenthic community and alternative sampling methods for this community. This should be at the beginning of the methods, not be left for the discussion!
• The difference between the provided microhabitats in FARMS and the surrounding habitats, particularly sand, seagrasses and mangroves should be taken into consideration. In my view, the benefit of a standardized sampling approach FARMS (for all and any habitat) is outweighed by the large differences between farms and the above-mentioned habitats. This also links to a point made in the introduction that states that cryptobenthic fishes have short dispersal abilities! Dispersal kernels published for a few “cryptic” fishes show very long tails, meaning that there are few larvae reaching far from home locations. My main point here is that by providing a different habitat (FARMS) to the surrounding area/habitat (e.g. sand, seagrasses, etc) one might be sampling a community that is actually NOT present at the site but rather is dispersing from nearby distinct habitats. Did you find any species on sand FARMS that were not associated to sand?
• Earlier work on the use of artificial structures to sample this specific part of the fish community should be mentioned in the introduction and they are really the precursors to the approach (for example: Allen et al 1992. Abundance, Diversity, and Seasonality of Cryptic Fishes and their Contribution to a Temperate Reef Fish Assemblage off Santa Catalina Island, California. Bull. Southern California Acad Sci 91(2) or many others to choose from)
• In the methods there should be a brief justification of the choice of statistical analysis. I wonder why you chose a Bayesian approach if there are no priors set? It could have been interesting adding a variable that translates the mismatch between the sampled habitat and the FARM.
• Some minor points include referencing the techniques you compare the FARMS to Oyster sampling units (Texas), anaesthetic collections (Red Sea).