The biodiversity of the marine realm is unparalleled, and more species are yet to be described. In the light of living in what now has been termed a ‘biodiversity crisis’ or arguably the 7th mass extinction, it is of eminent importance that we rapidly assess the biodiversity still present before more species disappear, including the cryptic and more elusive fauna. In the oceans, biodiversity is traditionally measured by surveying conspicuous and shallow-water species, although the diversity of hidden or cryptic communities can be equal if not higher to these groups. One key example are the cryptic communities that can be found in the deep-sea. Advances in survey techniques and methodologies has led many to showcase this ‘hidden diversity’ – for example, with the utilization of environmental DNA. That said, eDNA has been primarily utilised for detection of conspicuous fauna up until now, and to some degree rare, endangered or invasive species but cryptobenthic species (cryptic species associated with the benthic environment) remain rarely studied. In this thesis, I therefore aimed to assess the feasibility of using eDNA to detect cryptic communities across different ecosystems, from the shallow waters of the tropical coral reef to the deep-sea. I further aimed to assess how contemporary eDNA methods performed against more traditional survey methods such as remote operated vehicles (ROVs) and physical collection methods (net enclosures). Chapter 1 focused on the detection of cryptobenthic diversity and the impact of microhabitat preferences, based on field work on a coral reef in the Maldives archipelago. From here, I wanted to explore the function of these elusive and cryptic communities, instead of simply the detection of their presence. Therefore, in Chapter 2 I investigated how the sand-sifting feeding behaviour of cryptobenthic reef fish impacts the interstitial meiobenthic communities in coral reef sand. Finally, I turned my attention to the deep-sea and aimed to shine a light on the species composition of the often-elusive fish community there. To achieve this, I utilised a novel ‘tool’ to filter the eDNA, namely deep-sea sponges (Chapter 3). In this thesis, I present an introduction to the field and methods used, explain how I tackled these research questions, my scientific contributions, and finally discuss the broader implication of my findings.

Chapter 1
The impact of ocean warming on corals and the larger, pelagic and often charismatic fish species is evident and well-described by many. By contrast, smaller and less distinct species are often overlooked. The diverse and abundant cryptobenthic reef fish and other cryptofauna such as benthic annelids have been understudied, largely due to difficulties associated with surveying, collecting and/or describing these groups. In this study, I was amongst the first to explore how reef degradation or the composition of microhabitats can impact the ‘hidden diversity’ in tropical coral reefs. Here, I showed the huge potential eDNA has in targeting these important basal food web species. Further, I was also able to show distinct correlations between the diversity of cryptobenthic reef fish and certain habitat traits, such as site complexity or proportion of branching corals. Contrasting eDNA sampling methodology, reef water filtration revealed the highest species diversity, whilst sediment-based methods were more variable in the numbers of species detected. Finally, our results suggest that a reduction in reef complexity, as is forewarned to occur due to climate change and other anthropogenic impacts, will likely result in a halving of the overall fish species richness. 
Chapter 2 On oligotrophic coral reefs, where readily available prey is scarce, some cryptobenthic reef fish use unusual and often undocumented feeding strategies. For example, many goby species utilise sandsifting to acquire their prey. Through experimental studies, I aimed to explore how such feeding behaviour can impact the hidden interstitial meiobenthic community of coral sand habitats. I highlighted that cryptobenthic reef fish can effectively feed on this microscopic prey community, significantly reducing their abundance. However, my results also suggested that this feeding activity does not alter the composition of the meiobenthic community. Additionally, I was able to identify key morphological features of fish that were associated with this behaviour. For example, the presence of the epibranchial lobe and protruding taste buds were suggested to be adaptations that are strongly associated with such sand sifting behaviour in fish. The video abstract for this publication can be found at: https://besjournals.onlinelibrary.wiley.com/doi/full/10.1111/1365-2435.14087 
Chapter 3 The deep-sea is the largest biome in the ocean, yet little is known of the residing species present, largely due to the inaccessibility of this ecosystem. In attempts to mitigate sampling difficulties, a few groups have explored the use of eDNA analysis to explore deep-sea diversity, albeit with limited success. Therefore, I tried an alternative approach using a nature-based ‘tool’. I explored whether eDNA extracted from sponge species could better detect the diversity of deep-sea fish communities. These results suggested that this technique was promising in this respect; eDNA from sponge tissues outperformed the traditional water filtration eDNA method and showed a different (yet complementary) species structure detected with visual surveys using remote operated vehicles.