Simon J. Pierce, PhD
Principal Scientist, Marine Megafauna Foundation
I’ve previously co-authored a couple of review articles on photo-identification, the first way back in 2012, and the book chapter that this article is provisionally based on, completed in 2017 and published in Shark Research: Emerging Technologies and Applications for the Field and Laboratory in 2018. I’ve also published quite a few papers in the primary literature based on the photo-identification of whale sharks, manta rays, and also sea turtles. You can see my current peer-reviewed publication list here on the site. Aside from that academic background, we’ve got ongoing photo-ID projects on multiple shark and ray species.
Rather than wait another few years to update things, I thought it might be more useful to create a web resource that can be improved on an ongoing basis. While this may not be formally peer-reviewed, I hope that people will treat this topic as a discussion anyway. To that end, please do feel free to use the comment functionality below. I’d love to get your suggestions and feedback.
I’ll add direct links to articles when I can, and add in a few open-source figures and my own photos to provide more information. Hope you find it useful!
Simon (20th March 2020)
This is a citable document. A suggested reference is as follows:
Pierce, SJ (2020) Photographic identification of sharks and rays: a review and user guide. Nature Tripper (https://naturetripper.com/photographic-identification-of-sharks). Accessed on: (Your current date)
- Advantages of photo-ID
- What are the “rules” for successful photo-ID studies?
- Applications of photo-ID in shark research
- Challenges associated with photo-ID
- Practical considerations for photo-ID
- Collaborative databases and computer-assisted matching systems
- Summary and conclusions
The only previous review of photographic identification of sharks, in 2012 (Marshall and Pierce 2012), considered what was then seen as a relatively new technique that still struggled for acceptance with some peer reviewers and editors. In the years since, the use of photographic identification (henceforth referred to as “photo-ID”) has been widely and rapidly adopted and has become a standard method in studies of elasmobranch population ecology, movement and social behaviors. A small number of large, semi-automated (and, increasingly, fully automated) collaborative online databases are routinely used to facilitate data sharing among research groups. The ubiquity of underwater camera systems has led to a dramatic increase in the volume of visual data posted online. The continuing development of computer vision and machine learning capabilities mean that use of photo-ID will continue to expand, with artificial intelligence systems enhancing, automating, and assuming responsibility for many of the processes and decisions that are currently performed manually.
Here, we define photo-ID as the “recognition of individual fish through their distinctive natural markings, recorded via photographs or video”. Photo-ID has been used in elasmobranch studies since at least the early 1970’s (e.g. Myrberg and Gruber 1974). Pigmentation spots, body markings, scars, and fin morphology have all been used as photo-ID characteristics for a variety of shark and ray species. We are not aware of photo-ID techniques having been used in chimaeras. Many of these species live in deep water, and the more accessible species have not, as yet, been shown to be individually identifiable.
The popularity of photo-ID has been enhanced by the increasing use of waterproofed digital cameras by scientists and marine tourists, such as scuba divers and snorkelers, who are being recruited directly as “citizen scientists” to contribute data to broader research efforts or indirectly through data-mining social media. The non-invasive nature of photo-ID, in that animals do not need to be touched or restrained, and the inbuilt data validation that it offers (i.e. the researcher can directly examine the original photograph and potentially consult independent computer vision algorithms for confirmation; Bonner and Holmberg 2013), allow such efforts to be easily applied to studies of elasmobranch species that are popular focal species in marine tourism, such as manta rays (Mobula spp.), whale sharks (Rhincodon typus), and white sharks (Carcharodon carcharias). The routine sharing of images via online social media websites and apps, and the ability of researchers to solicit data through such platforms, has also played a significant role in expanded participation in these studies (Davies et al. 2012; Robinson et al. 2016).
In this chapter, we present a guide for successful photo-ID studies. We consider the current and potential uses of photo-ID as a study method and examine how contemporary developments in computer science are likely to influence and enhance our use of photographs and videos for science.
DOWNLOAD THE MAGAZINE
Issue #1 is out now! Dive with Galapagos whale sharks, learn about seal superpowers, obsess about wildlife photography, and there’s plenty more to read too. 79 pages of FREE (and wonderful) articles and photos.
Advantages of photo-ID
Recognition of individuals within a study species is a fundamental requirement for population biology and demography research. The use of marker tags has a long and illustrious history in elasmobranch studies, and they continue to provide novel information on life-history variables, stock status, reproductive behavior, migrations and distribution patterns. A review of tagging studies by Kohler and Turner (2001) reported that 64 studies had already used conventional marker tagging in 101 species of shark at that time, and that number has steeply increased since.
However, tagging studies have a number of general limitations and associated practical issues. Most challenges stem from the potential for tags to be shed, removed, damaged or bio-fouled, thereby limiting their effective lifespan in population studies (Dicken et al. 2006; Graham and Roberts 2007; Kohler and Turner 2001; Pierce et al. 2009; Rowat et al. 2009). Reliable re-identification of damaged or heavily fouled tags by being able to view their individual numbers or color codes can also be challenging. Conventional tagging can also be detrimental to individual fitness, may affect natural behavior (Dicken et al. 2006; Feldheim et al. 2002; Fouts and Nelson 1999; Manire and Gruber 1991; Wilson and McMahon 2006), and could lead to increased mortality (Stansbury et al. 2015). The repercussions of these issues depend on the objectives of the research and the focal species. Some studies, such as age validation, require only small numbers of animals to retain tags over time to be successful (Pierce and Bennett 2009; Smith et al. 2003). In other circumstances, a high rate of tag shedding can mean that re-sighting rates have to be treated as unreliable or used with caution (Rowat et al. 2009).
Photo-ID presents an appealing alternative or supplement to conventional tagging techniques in certain situations. The presence of natural identification marks can eliminate the need for physical marker tags, providing a more permanent means of identifying individuals (Dudgeon et al. 2008; Rowat et al. 2009), and marks can be easy to distinguish at a distance. Photo-ID also avoids some of the problems of tags being shed, removed or fouled, and, depending on field practices, can also minimize the risk of inducing stress or behavioral issues. Many shark and ray species are difficult to capture and handle, due to their large size or inaccessibility, and photo-ID offers an alternative approach for identification.
Photo-ID has also become an attractive option for researchers seeking to minimize disturbance on sensitive populations or threatened elasmobranch species, particularly in cases where the interested public may have a negative view of more invasive research methods. Controversy over shark tagging has rarely been discussed in the scientific literature (although see Jewell et al. 2011; Hammerschlag et al. 2014), but we are aware of a number of additional cases where significant opposition has arisen from community groups due to perceived harm, visual disfigurement or perceived behavioral avoidance by the target species. While these perceptions may have a dubious factual basis, a lack of local support for research activities can be detrimental to conservation and outreach efforts. The non-invasive nature of photo-ID means that it can be relatively simple to conduct research off platforms of opportunity, such as tourist vessels, and to maximize data collection and outreach potential through the incorporation of citizen science programs into field studies (see Chapter 15; Davies et al. 2012; Gallagher et al. 2015; Germanov and Marshall 2014).
Although photographic equipment can be an expensive initial outlay, costs tend to remain relatively cheap compared to electronic tags, and ongoing expenses of maintenance or services are minor in comparison to satellite tagging fees or maintaining passive acoustic receiver arrays. Because scientists usually use “off the shelf” consumer products in their work, many researchers, students and volunteers already own suitable cameras and lenses for photo-ID studies. Image capture and processing are also relatively simple and easily trainable, thus broadening the potential pool of study participants.
What are the “rules” for successful photo-ID studies?
Photo-ID is not suitable for all elasmobranch species. The technique has two specific assumptions: a) individuals can be reliably distinguished from one another, and b) individuals can be re-identified over time.
Distinguishing individual sharks
Many elasmobranchs have natural pigmentation patterns on their skin that act as a unique “fingerprint” for each individual (Fig. 1). Examples include the spots on the dorsal surfaces of spotted eagle rays (Aetobatus narinari) (Corcoran and Gruber 1999) and the ventral surfaces of reef manta rays (Mobula alfredi) (Kitchen-Wheeler 2010; Marshall et al. 2011), or on the flanks of whale sharks (Taylor 1994), zebra sharks (Stegostoma fasciatum) (Dudgeon et al. 2008), and sand tiger sharks (Carcharias taurus)(Bansemer and Bennett 2008; Van Tienhoven et al. 2007). Other natural patterns, such as the irregular countershading boundaries on white sharks, are also used (Domeier and Nasby-Lucas 2007). Some species can be distinguished by the color, shape or notches in their dorsal fins, such as white sharks (Anderson et al. 2011, Hewitt et al. 2017), blacktip reef sharks (Carcharhinus melanopterus) (Porcher 2005) and basking sharks (Cetorhinus maximus)(Gore et al. 2016). Scars, bite marks, fin morphology and deformities may be useful in the absence of intrinsic patterns (Anderson et al. 2011; Castro & Rosa 2005; Klimley & Anderson 1996; Sims et al. 2000) and may also help in secondary confirmation where patterns are present (Marshall et al. 2011).
Most studies employing photo-ID methods have reported 100% of individuals to be recognizable, e.g. whale sharks (Meekan et al. 2006), spotted eagle rays (Corcoran and Gruber 1999), adult zebra sharks (Dudgeon et al. 2008), white sharks (Domeier and Nasby-Lucas 2007), whitetip reef sharks (Triaenodon obesus) (Whitney et al. 2011) and reef manta rays (Marshall et al. 2011). Some species, however, exhibit lower percentages of identifiable individuals, such as the 54.8% of nurse sharks (Ginglymostoma cirratum) reported by Castro and Rosa (2005) and 83% of basking sharks in Scotland (Gore et al. 2016). A photo-ID study on lemon sharks (Negaprion acutidens) in French Polynesia determined that the uniform color of the focal species, combined with the poor resilience of small spots or color aberrations, resulted in this species being difficult to identify from natural coloration alone (Buray et al. 2009). Furthermore, the assumption of equal individual identifiability across all life stages has not been addressed for any species, and morphometric changes may alter patterning or the applicability of computer vision to analyze it. While modelling techniques may also be able to compensate, to some extent, for species where a minority of individuals are not identifiable (Hunt et al. 2017), consideration needs to be made as to whether photo-ID is preferable to conventional tagging where the percentage of identifiable individuals is low (Pratt and Carrier 2001).
Even in species with distinct markings, practical considerations may favor tagging over photo-ID. Individuals may be hard to photograph in their natural environment (i.e. elusive, pelagic or deep-water species, or those living in turbid environments), or large population sizes and low re-sighting rates render the collation and management of a photo-ID library logistically difficult. Photo-ID is most useful – and generally applied – in species that have distinctive markings, concentrate reliably in areas accessible to observers, and are reasonably easy to approach and photograph (Marshall et al. 2011).
Re-identifying individuals over time
Individual natural markings must allow for re-identification over time. Animals that have markings caused by fungal infections, or where color is in the mucus layer on the skin, rather than the skin itself, have not yet been shown to be reliably identifiable over sufficient time frames. Natural ventral markings, in the form of spots and shading, are present from before birth in species such as reef manta rays (Marshall et al. 2008) and are not thought to change over the course of an individual’s life span (Couturier et al. 2012). While the long-term stability of individual markings has been documented in multiple elasmobranch species, to over 10 years in blacktip reef sharks (Mourier et al. 2012; Porcher 2005), over 20 years in both white sharks and whale sharks (Anderson et al. 2011, Norman and Morgan 2016), and 30 years in reef manta rays (Couturier et al. 2014), the longevity of markings is species-specific. Therefore, validating their stability across life stages is an important test for photo-ID studies.
Minor pigmentation changes and accumulation of fin damage have been noted in multiple studies, particularly of white sharks, although these were a small minority of individuals in each study population (Domeier and Nasby-Lucas 2007; Robbins and Fox 2013; Towner et al. 2013). Scars, wounds, nicks, and scratches may transform or completely heal over time (Anderson et al. 2011; Castro and Rosa 2005; Domeier and Nasby-Lucas 2007; Marshall and Bennett 2010a; Pratt and Carrier 2001). The intensity of body coloration has also been reported to change over short time periods (Ari 2014). Depending on the severity of the changes to identifiable markings, photo-ID confirmation may be compromised if individuals can no longer be matched with certainty, an issue equivalent to tag loss in a mark-recapture studies using conventional tags. It is important to note, and preferably quantify, the potential for individuals to become less identifiable over time, as this can lead to over-estimation of abundance and underestimation of residency (Gubili et al. 2009; Towner et al. 2013).
The stability of natural marks is easiest to assess through some form of “double tagging”, i.e. using an independent feature to confirm individual identification. For example, conventional tags may be applied to a proportion of observed individuals to validate natural pattern stability over the study period (Dudgeon et al. 2008). Alternatively, positive identification can also be achieved by using more than one identifying feature on species where patterns or scarring are reasonably stable; for instance, using both sides of the animal, or using a combination of natural markings and scarring (Domeier and Nasby-Lucas 2007; Kitchen-Wheeler 2010; Marshall et al. 2011; Meekan et al. 2006; Norman and Morgan 2016), or by adding a completely separate analytical technique, such as individual genotyping (Gubili et al. 2009). Validation can alternatively be achieved in partnership with aquariums that house specimens of a desired species by closely monitoring natural coloration or distinctive markings over time, particularly ontogenetic shifts at a certain age or size class (Bansemer and Bennett 2008). Routine collection of additional metadata, such as sex, maturity status, and size, also provide a useful means of verification.
If ontogenetic shifts in the natural markings of focal species are understood and accounted for, photo-ID may still be successfully employed as long as markings remain stable over the duration of the study (Arzoumanian et al. 2005; Dudgeon et al. 2008). Similarly, identifications are often based largely on scarring patterns in species where distinctive natural markings are absent. Younger animals are unlikely to show as many scars, accumulated bite wounds or reproductive marks, making identification of these age-classes challenging. Removing certain size or age classes from studies, for instance by focusing on individuals of over a minimum size, may be effective in cases where the proportion of unidentifiable individuals in the population can be reliably established (Wilson et al. 1999).
Applications of photo-ID in shark research
While this overview is not intended to consider all potential applications, some of the common uses of photo-ID are presented here.
Residency and movement
Many sharks and rays, even highly mobile species, frequent specific sites (Chapman et al 2015). This site fidelity has long been evident from telemetry studies, and has similarly been documented with photo-ID studies (Anderson and Goldman 1996; Anderson et al. 2011; Bansemer and Bennett 2009; Marshall et al. 2011). Individual sharks may use aggregation areas year-round (Cagua et al. 2015), while other sites are characterized by seasonal or aperiodic visitation (Luiz et al. 2009; Kock et al. 2013, Norman and Morgan 2016). While most elasmobranch population studies using photo-ID have focused on larger, migratory species, others, such as bull sharks (Carcharhinus leucas), zebra sharks and wobbegong sharks (Orectolobus spp.) that aggregate in specific locations or have small home ranges have also been investigated (Brunnschweiler and Baensch 2011; Carraro and Gladstone 2006; Castro and Rosa 2005; Dudgeon et al. 2008; Lee et al. 2014).
Photo-ID can be used to evaluate inter-site movements. For example, reef manta ray movements were established between the Komodo National Park, Nusa Penida and the Gili islands in Indonesia (Germanov and Marshall 2014). Migrations over hundreds of kilometers occur along the eastern coast of Australia in reef manta rays (Couturier et al. 2011; 2014) and sand tiger sharks (Bansemer and Bennett 2009; Barker and Williamson 2010), respectively. Photo-ID, combined with satellite telemetry, even documented a white shark completing a return migration between South Africa and Western Australia (Bonfil et al. 2005).
A small number of international photo-database collaborations have provided insights into the broad-scale movements (or lack thereof) in whale sharks. These have used either the online Wildbook for Whale Sharks photo library (www.whaleshark.org; Arzoumanian et al. 2005) or the downloadable Interactive Individual Identification System (I3S; Speed et al. 2007; Van Tienhoven et al. 2007), both of which semi-automate the photo matching process. Brooks et al. (2010) and Andrzejaczek et al. (2016) found minimal evidence of population-level interchange among known whale shark feeding areas in the Indian Ocean, although later studies have shown routine movement among countries in the Arabian region (Robinson et al. 2016), and between Mozambique and South Africa (Norman et al. In press). Regional movements of whale sharks have also been examined in the Western Atlantic, where regular movements were shown between Belize, Honduras, Mexico and the United States (McKinney et al. 2017). Individual re-sightings occurred between Australia and Indonesia, representing perhaps the longest-distance re-sighting to date via photo-ID [~2700 km; (Norman et al. In press), Mozambique and Tanzania (~1800 km; Norman et al. In press), the Philippines and Taiwan (~1600 km; Araujo et al. 2016)]. Such collaborative efforts require standardized techniques and often data-exchange among research groups. There are now several other global photo-library initiatives underway, such as Manta Matcher (www.mantamatcher.org; Germanov and Marshall 2014; Town et al. 2013), ID the Manta (), and a white shark database (Hughes and Burghardt 2017), so it is likely that more regional- to global-scale studies will be developed in the near future.
Population size and demographics
When photo-ID data are combined with location and date-time metadata, three fundamental data points underpinning population ecology are established: who, when, and where. These data can then be entered into open or closed capture-mark-recapture (CMR) models or — more specifically “sight-resight” models — to estimate a variety of population parameters, including abundance, trajectory, sex and size ratios, survival rate, capture probability, and others (Williams et al. 2002, Hewitt et al. 2017), when the assumptions of the underlying analytical models are met. Photo-ID is well-suited for this purpose because individuals can be monitored and non-intrusively re-identified over short or longer time periods. These identified animals can then be used to investigate seasonal or annual population size within the area of interest (Castro and Rosa 2005; Chapple et al. 2011; Couturier et al. 2014; Deakos et al. 2011; Dudgeon et al. 2008; Holmberg et al. 2009; Marshall et al. 2011, McKinney et al. 2017; Meekan et al. 2006; Rowat et al. 2009). Use of photo-ID may provide an advantage in such studies, since results are less likely to be affected by loss of fitness or mortality resulting from handling, or biased through avoidance behavior by the identified individuals, as could be the case with conventionally-tagged individuals. Photo-ID also provides access to a potentially greater volume of data for intensive modelling if related tourism is present and can be engaged in concurrent data collection. Care must be taken, however, to understand and account for potential biases in externally-sourced data.
Many of the species that have proven suitable for monitoring with photo-ID techniques are also globally threatened species. Tracking changes in abundance and survivorship over time can thus provide valuable information on the decline or recovery of these elasmobranchs (Bradshaw et al. 2007; Holmberg et al. 2008, 2009, Hewitt et al. 2017), and photo-ID can also be used to avoid issues with counting the same individual more than once in sighting-based studies (Rohner et al. 2013). That being said, it is possible to inadvertently bias the results of these models in cases where individual sharks vary in their behavior. This issue is discussed further in section 13.5.
Photo-ID studies have often focused on areas where sharks aggregate, thereby enabling the study of social interactions between individuals (Jacoby et al. 2012). For example, bonnethead sharks (Sphyrna tiburo), maintained in a captive environment and identified using spot patterns, scars and fin tears, formed size-based dominance hierarchies (Myrberg and Gruber 1974). Fin morphology of blacktip reef sharks allowed Mourier et al. (2012) to show that these sharks formed stable, long-term social bonds in French Polynesia. On the other hand, white sharks identified by their dorsal fins during chumming activities in South Africa co-occurred at random, displaying no preference or avoidance towards particular individuals, although there was a weak tendency for sharks to co-occur with individuals of similar size and the same sex (Findlay et al. 2016).
Support for biology and ecology studies
Photo-ID can be a useful complement to studies of size at maturity (Acuña-Marrero et al. 2014; Deakos 2010; Marshall and Bennett 2010b; Norman and Stevens 2007; Ramírez-Macías et al. 2012; Rohner et al. 2015), gestation period and reproductive periodicity (Bansemer and Bennett 2011; Deakos et al. 2011; Marshall and Bennett 2010b), reproductive behavior (Bansemer and Bennett 2011; Whitney et al. 2004; Yano et al. 1999), survivorship (Bradshaw et al. 2007; Couturier et al. 2014; Smallegange et al. 2016), growth (Norman and Morgan 2016; Sims et al. 2000) and longevity (Anderson et al. 2011; Couturier et al. 2014; Norman and Morgan 2016). One whale shark that has been returning to Ningaloo Reef in Australia from 1995 until at least 2016 (Norman and Morgan 2016), and a male reef manta ray, first sighted when visibly mature in 1982, was re-sighted 30 years later at Lady Elliot Island on the Great Barrier Reef in Australia (Couturier et al. 2014). Where length or other body morphometrics are recorded, photo-ID can be incorporated to investigate individual growth (Graham and Roberts 2007; Rohner et al. 2015).
One of the advantages of photo-ID for such studies is the additional data that can often be collected concurrently to aid in interpretation of results through, for instance, assessing how sexual segregation can influence residency and movement patterns (Bansemer and Bennett 2009; Deakos et al. 2011; Robbins 2007). Incorporating knowledge of breeding status, such as differences in movement patterns between pregnant and non-pregnant females (Bansemer and Bennett 2009), can considerably advance understanding of the ecology and management of these animals.
Photo-ID has also been used to examine predator–prey and competitive interactions in elasmobranchs. Marshall and Bennett (2010a) investigated the frequency and effect of shark predation on a population of reef manta rays in southern Mozambique by examining the size, number and positioning of shark-inflicted bite wounds on individual rays over time. Potential predators were identified through bite mark analysis, and the bite wounds themselves were monitored to track healing rates. Wound healing in whale sharks (Fitzpatrick et al. 2006) and white sharks (Domeier and Nasby-Lucas 2007, Towner et al. 2012) has also been examined in cases of shark bites, and also for human-induced injuries (Riley et al. 2009). A broader comparison between scarring frequency, origins and their influence on survivorship was conducted between whale shark aggregations in Mozambique, the Seychelles and Western Australia (Speed et al. 2008). Photo-ID is also a useful addition to studies investigating and quantifying threats such as fishing-related injuries (Bansemer and Bennett 2008; Riley et al. 2009), net entanglement, and boat strike (Deakos et al. 2011; Speed et al. 2008).
Challenges associated with photo-ID
Like any method, there are drawbacks, limitations and potential sources of error with photo-ID studies that must be considered to limit bias and ensure robust results. Photo-ID requires both the presence of a photographer, either human or automated, and a relatively close approach by the animal. The presence of boats or divers can have a significant effect on shark behavior, either as an attractant (Bruce and Bradford 2013) or repellent (Brunnschweiler and Barnett 2013), particularly where provisioning tourism is a factor (Laroche et al. 2007, Gallagher et al. 2015). Concurrent photo-ID and acoustic telemetry studies have demonstrated that sharks may often be present in the area, but not documented via photo-ID (Brunnschweiler and Barnett 2013; Cagua et al. 2015; Delaney et al. 2012), thereby underestimating residency when using photo-ID alone. This may be overcome in part by using alternative approaches which have less influence on shark behaviour, such as rebreather diving systems rather than open-circuit equipment (Lindfield et al. 2014), or by using remote cameras at defined aggregation sites, such as cleaning stations (O’Shea et al. 2010). Passive acoustic receivers also have a far larger range for detection than visual ID approaches, typically ~500 m, and offer continuous sampling coverage. Their disadvantages are the cost of purchase, installation and maintenance, and the normal factors associated with the use of electronic tags (see section 13.2). Ultimately, the use of complementary methods to assess the bias associated with observer presence can add considerable value to a study, and should be implemented where feasible (Brunnschweiler and Barnett 2013; Cagua et al. 2015; Chapple et al. 2016; Delaney et al. 2012).
Photo-ID is also commonly used in conjunction with mark-recapture models (section 13.4.2). Individual heterogeneity in shark behavior can, however, cause bias in these analyses (Burgess et al. 2014), and this should be explicitly tested for during modelling studies (Burnham et al. 1987; Holmberg et al. 2009). A re-sighting bias can occur for individuals in which multiple sightings have already been obtained (Holmberg et al. 2009; Van Tienhoven et al. 2007). Individual differences in the sightability of sharks can manifest in a variety of circumstances. On a broad scale, heterogeneity in survivorship estimates has been shown in whale sharks off Western Australia due to the presence of large numbers of transient individuals (Holmberg et al. 2008), and inferred in whale shark study populations in Belize (Graham and Roberts 2007) and the Maldives (Riley et al. 2010). Sex- or size-based segregation is also commonly present at aggregation sites (Jacoby et al. 2012). On a finer scale, size-based dominance patterns in white sharks could mean that subordinate sharks are excluded from the area, or from the surface, leading to a lower probability of sightings or successful photo-ID (Burgess et al. 2014). If chum is being used to attract sharks, the sharks may also learn to ignore this stimulus when no food reward is offered, also leading to under-reporting (Laroche et al. 2007). A failure to detect previously-identified individuals that are present would lead to over-estimation of population size, while a failure to detect unmarked individuals (such as by exclusion through dominance behavior) would cause the population size to be under-estimated (Burgess et al. 2014; Irion et al. 2017). Therefore, careful consideration of model assumptions is necessary in mark–recapture studies based on photo-ID data (Holmberg et al. 2008; 2009).
The issue of “tag loss” through changes in natural markings over time has been discussed in section 13.3.2. However, it is important to note that, while regular effort at a study site may detect fine-scale changes in identifiable characteristics as they occur, allowing individual identification to be continuously updated, long breaks between field work may lead to misidentification of previously-identified individuals and consequent over-estimation of population abundance (Towner et al. 2013).
Matching errors between photo-identified individuals, either through incorrect assignment of previously-identified individuals as “new”, or accidental confusion of two different animals with similar markings, can also occur. These can largely be avoided by following the photographic and processing workflow discussed in the following section (13.6). However, neither human, nor algorithmic matching is 100% accurate (Andreotti et al. 2017; Dureuil et al. 2015; Speed et al. 2007; Van Tienhoven et al. 2007). Increased automation, particularly when combined with larger photographic databases, is likely to lead to a larger number of incorrect assignments by matching algorithms (in absolute terms). It is therefore important to quantify and incorporate this error rate into analyses using these datasets. This issue is discussed further in section 13.7.
Practical considerations for photo-ID
Photography (either still photography or videography) is generally agreed to be the best method of recording the appearance of natural markings or scars. Photographs can freeze motion and record extremely detailed information, allowing individuals with similar markings or scars to be reliably separated from one another. Photographs also allow a permanent record to be kept for each encounter that can be examined in detail at a later stage and verified by independent observers. Standardized images can also be used in current or future identification software programs to fully- or partially-automate the image matching process (Arzoumanian et al. 2005; Hughes and Burghardt 2017; Speed et al. 2007; Van Tienhoven et al. 2007).
It is no coincidence that the popularity of photo-ID as a research technique has grown in conjunction with the increasing use of digital photographic equipment (Markowitz et al. 2003). Digital cameras allow confirmation that suitable pictures have been captured in the field, and simplify post-processing workflow, computer-assisted matching, and the storage of images. Decisions on the specific equipment requirements (cameras, lenses) are best made on a case-by-case basis. Digital single lens reflex (SLR) cameras, mirrorless cameras and compact cameras, as well as “action cameras”, such as the GoProTM line and other video cameras, are all in routine use in current studies. Earlier studies have found that video footage was seldom clear enough for successful extraction of photo-IDs (Meekan et al. 2006), but modern video cameras (shooting in High Definition, 4K, or higher resolutions) usually produce acceptable frame grabs in reasonable water visibility (Dureuil et al. 2015). At the time of writing, advanced DSLR cameras still have some autofocus advantages over most mirrorless and compact cameras, which may be particularly useful in dorsal fin photo-ID studies, but the gap is rapidly narrowing. Underwater, advanced compact cameras and mirrorless cameras have advantages in reduced size and system cost, although the cost of a complete system – particularly if one or more flash units are required – is still significant. Almost all modern cameras have sufficient resolution for scientific studies. If color-correction is required for images, which is routinely the case for underwater images, a camera that shoots raw images (.dng or the equivalent proprietary format) is advantageous because these contain significantly more data than images processed in-camera. However, they do require more manual processing time, which may not be a worthwhile trade-off in photo-ID studies (Gore et al. 2016).
Large sharks or rays, such as whale sharks or manta rays, are best photographed with rectilinear wide-angle or even fisheye lenses (the latter is the personal preference of authors SJP and ADM). These lenses have an ultra-wide field of view that can capture a large portion of the entire animal in a single frame, often capturing a view of additional data such as scars, deformities or sex. However, compensation for distortion may be necessary with this type of lens to avoid misrepresentation of patterns, particularly if measurements are being obtained from photographs (Bansemer and Bennett 2008; Deakos 2010). Where images are significantly distorted, to the extent that computer-based image analysis on natural markings is affected, this could prevent automated individual identification from images. Above the water, a telephoto zoom lens will often be appropriate for photographing dorsal fins as they break the surface.
In some cases, artificial lighting (generally through external flash units or video lights) or color-correction filters are useful to capture detailed natural markings properly in filtered underwater light (Marshall et al. 2011).
A standardized area or areas on the animal’s body should be chosen for each species. A good reference area will be easy to reliably photograph on a free-swimming shark, while also minimizing the influence of the photographer on the animal’s natural behavior. For species that have differing patterns or marks on either side of their body it is generally considered best-practice to photograph the spot patterns or scarring on one pre-determined side of the animal consistently e.g. the left side, to avoid double-counting individuals, as this could lead to over-estimation of population size (Arzoumanian et al. 2005; Dudgeon et al. 2008; Meekan et al. 2006; Van Tienhoven et al. 2007; Whitney et al. 2011), although both sides should be photographed whenever possible (Bonner and Holmberg 2013; Domeier and Nasby-Lucas 2007). Taking photographs of multiple standardized areas (i.e. dorsal and ventral surfaces, or both left and right sides of the body) can make re-sighting identification easier and more accurate by allowing for independent confirmation (Dureuil et al. 2015; Robbins and Fox 2013). Where scars or marks are used as identifying features, assigning a standardized area is also appropriate, e.g. the dorsal fin (Anderson et al. 2011).
Reference points, i.e. body parts that can be used to help scale and rotate the image appropriately, may be required by software-matching systems (Speed et al. 2007; Van Tienhoven et al. 2007). Using a reference point, for instance the area just behind or between the gill slits or pectoral fins, is also a good way to ensure that the photographed area remains consistent. If a photographic study includes historical images or those donated by the public, using areas that appear consistently in non-specialist photographs is also an important consideration.
Minimizing sources of error
Aside from the previously discussed requirements for validation and standardization, a series of other workflow steps should be implemented to ensure accurate matches.
In the field, collection of photo-IDs may be influenced by environmental conditions such as wind, swell, glare, visibility or currents. This variation in detectability is important to account for when standardizing for effort in data analyses, and detailed field logs should be filled out for each photo-ID survey (Evans and Hammond 2004; Rohner et al. 2013).
There are several basic procedures that can be implemented to avoid accidental misidentification of individuals in the field. It is useful to take a photo of a survey sheet or hand signals between each animal, particularly where multiple photos of each individual are taken, so as to associate photos with that individual and distinguish them from any following individuals (Evans and Hammond 2004). This is particularly the case where a manual assessment of sex and size is made. The camera should be set to the local time and date. Careful downloading, storage and labelling of images is also necessary to prevent confusion of when and where photographs were sourced from. Photos should be catalogued carefully e.g. photos from a particular survey could be imported into folders arranged in a hierarchical format, such as year, month and day. Images should be taken perpendicular to the area of interest because the perspective and perception of markings can change with the movement of the animal or the position of the photographer (Bansemer and Bennett 2008; Van Tienhoven et al. 2007), resulting in increased potential for “false negative” matches.
Training programs and reference images are useful for maintaining data quality. This is particularly important when images from citizen scientists are solicited. While such programs can provide an extremely useful boost to the quantity and geographical extent of data collected, it is important to maintain quality and consistency within the dataset that is actually used for matching (Gubili et al. 2009). It can be useful to develop explicit criteria that must be met for photos to be included in the matching dataset. Matching itself can also be sped up, and made more accurate, by categorizing images by metadata such as sex and size (Marshall et al. 2011) or the patterns of coloration or shape that may be relevant to ID classification (Domeier and Nasby-Lucas 2007; Gore et al. 2016). It is important to note that when more than one person is matching images there exists a potential for bias. Thus, implementing a peer-review system for identifications is a useful means of quality control in photographic datasets (Holmberg et al. 2009).
Collaborative databases and computer-assisted matching systems
Collaboration and increased automation are helping photo-ID studies to achieve greater breadth and depth of coverage for sharks. However, these advances also come with new challenges for humans and wildlife.
Collaboration in photo-ID is a clear pathway to overcoming individual project resource constraints by engaging other research efforts, overlapping tourism activities (e.g., diving and snorkeling), and the public in both data collection and potentially curation, analysis, and publication as well. Collaborative photo-ID projects have reported over 10x increases in data collection through collaboration with tourists, enabling more detailed models for population analysis (Holmberg et al. 2008; 2009), and linked shared populations of migratory whale sharks across political, geographic, and individual research project boundaries in the Gulf of Mexico and Caribbean (McKinney et al. 2017). Collaborative platforms and communities provide a powerful foundation for new inquiries among existing researchers, a foundation for the rapid start-up of new field sites, and an open environment for novel, unanticipated, and independent results from disparate participants (Araujo et al. 2016; McKinney et al. 2017; Robinson et al. 2016). Collaboration between scientists and citizen scientists (through data collection and participation in research) can also build relationships between these communities and facilitate buy-in from local stakeholders for the conservation of threatened species.
Critical to the success of collaborative photo-ID efforts is the ability to scale a projects’ data collection and curation in an accessible, standardized, equitable, and secure manner. Web-based software such as the Wildbook® platform (Wildbook 2017), which originated out of collaborative studies of whale sharks (Arzoumanian et al. 2005, Holmberg et al. 2008), can provide URL-based accessibility to securely engage multiple stakeholders, communities, and projects in shared data collection, management and analysis. Fundamentally underpinning collaborative databases is a shared information architecture or schema and an accepted study design, allowing for a common understanding of data definitions and types (e.g., date format, study site boundaries, individual identity, GPS coordinate format, etc.) and establishing common protocols for data capture and management (e.g., optimum angles and equipment for photographing the species). The Darwin Core biodiversity data model (Wieczorek et al. 2012) provides an excellent foundation for a collaborative photo-ID schema, but has required some modification to expressly reflect and store individual identity under photo-ID (Holmberg et al. 2008, Holmberg et al. 2009). One significant advantage of using an existing data standard, such as the Darwin Core, is the ability to more easily exchange data with other facilities, such as pushing species occurrence data to the Global Biodiversity Information Facility (GBIF 2016) and the Ocean Biogeographic Information System (OBIS 2016) for long-term storage and third-party analysis.
While the application of new technology and the a priori creation of a good study design can solve many problems in photo-ID and scalability, the potential for conflict among human participants is present in collaborative projects. Shared user agreements (especially those addressing publication rights) among participants (example: MantaMatcher User Agreement 2017), joint approval of the study design, and non-disclosure agreements can help set expectations early, prevent misunderstands, and define acceptable methods of resolution where unanticipated disagreements arise. An independent managing authority can also assist in conflict resolution and take responsibility for collaborative database advancement, promotion, maintenance and sustainability.
Data security for humans and wildlife
Care must be taken in collaborative photo-ID studies to protect any personal information about human participants (e.g., personal information about contributing members of the public) as well as collected photographs and metadata about individually-identified animals. Improper exposure of personal data can lead to third-party harassment (e.g., email spamming), while improper exposure of wildlife data (e.g., locations and dates) could potentially be used to better target fishing or tourism activity, exposing the study population to greater impact or even mortality. Authentication, authorization (e.g., function-limiting roles for study participants), and accounting (AAA) software security is recommended for collaborative databases, especially as they grow in scope and participation. Wildbook provides open source examples of how this can be flexibly configured and implemented in global-scale research efforts for sharks and rays (Wildbook for Whale Sharks 2017, MantaMatcher 2017, Spotashark 2017). Data security should be addressed at the beginning of photo-ID studies and reflect species- and location-specific threats to personal and wildlife security.
Computer-assisted matching of individuals
Manual (or “by eye”) matching of photos has a declining return on time investment. While dedicated, expert matching can achieve a high proportion of successful matches (Chapple et al. 2009; Gore et al. 2016), manual processing of identification photographs does not scale well. As Duyck et al. (2015) point out, “at 10 seconds per comparison, a 10, 000-sized catalog will take approximately 15 person-years to analyze”. While promising new and more in-depth forms of analysis, growth in data from collaboration in photo-ID can conversely slow data curation and introduce additional human bias in collection and curation.
Computer-assistance in photograph matching of individual sharks and rays has emerged as a scalable and potentially less biased method that also reduces researcher time and effort. Arzoumanian et al. (2005) introduced computer-assisted matching of whale shark photographs based on the natural spots on their flanks, adapting an algorithm originally developed to match celestial star patterns between photographs (Groth 1986). Van Tienhoven et al. (2007) introduced a simple, nearest neighbor-based spot pattern matching algorithm for sand tiger sharks with the I3S software application (I3S 2016). Both algorithms are now available for use on species with spot patterns in the open source Wildbook platform (Wildbook 2017) and specifically implemented for a global shark research community online in www.whaleshark.org (Wildbook for Whale Sharks 2017). Other computer-assisted matching applications for sharks and rays have been developed by Hughes and Burghardt (2016) for white sharks (using the natural shape and notches on the trailing edge of the dorsal fin as a unique fingerprint) and Town et al. (2013) for manta rays based on natural, high contrast markings on their ventral sides. Since multiple areas and forms of individual identification may exist for a single species, advancement of computer assistance has also introduced the need for new research on photographic mark-recapture modelling in the presence of multiple marks (Bonner and Holmberg 2013).
Successful implementation of one or more computer-assisted algorithms offers a powerful incentive for collaboration, providing a demonstrable savings of time and effort in exchange for collaborative access to data. Such web-based implementations and collaborations (MantaMatcher 2017, Wildbook for Whale Sharks 2017) have led to new insights into whale shark abundance (Holmberg et al. 2008, 2009) and movement across borders and research catalogs (McKinney et al. 2017). Implementation online through web-browser access further reduces barriers of accessibility and usability across borders and studies, shifting the computationally-intensive matching operations away from disparate desktop systems and resource-constrained users and into more scalable cloud-computing environments. For example, n-number of virtual computers and CPUs are flexibly and scalably engaged in Amazon Web Services (AWS 2017) to quickly match 28, 000+ left-side whale shark spot patterns in parallel for www.whaleshark.org (Wildbook for Whale Sharks 2017), allowing for global access to rapid matching (often completed in less than four minutes) through powerful and increasingly inexpensive grid computing.
One important note about current computer-assisted matching systems for sharks and rays: all existing systems either require some amount of human intervention (e.g., manually mapping spots onto whale shark photos before computer-assisted matching) (Arzoumanian et al. 2005; Van Tienhoven et al. 2007) or can be significantly optimized by an optional manual step, such as cropping images down to a pre-defined area of the body (Town et al. 2013) or selecting reference points in the image (Hughes and Burghardt 2016). The required operations are generally less than two minutes per photo and are rewarded with significant time savings overall, but none of the systems completely remove the burden of human photographic curation or human analysis of the resulting list of potential matches.
In the near future, evolution of computer-assisted photo-ID for wildlife will significantly engage artificial intelligence (e.g., computer vision trained by deep convolutional neural networks) and remove systematic human involvement in photo analysis (Menon et al. 2017; Parham et al. 2017). This will then enable the extraction of photographic data from disparate data sources, including social media (Menon et al. 2017), and datamining of video archives such as YouTube.com (Wildbook for Whale Sharks 2017). This will reduce or remove the role of human input in answering fundamental questions, such as whether the photo-ID’s are of new or previously identified individuals, or determining how many individuals are present in the study population.
Critical to a fully automated future in photo-ID are a number of required research efforts, including a benchmark of population estimates, biases, and errors from human-curated photo-ID versus fully-automated computer estimates of the same dataset using only a cloud of photographs and related metadata (e.g., location and date) as inputs. Current population models require fixed duration “capture” sessions with longer time periods between captures, effectively leaving out data that can be continuously obtained from collaborative activities, such as diving and snorkeling tourism. Research into continuous-time population models that allow for higher volumes of data to be collected at daily intervals could allow for more accurate parameter estimates and detailed ecological insights. More information is also needed on the relative biases involved where studies include multiple modes of data collection, such as data collected from trained researchers, lightly or untrained tourists, and from social media sources (e.g., YouTube), or which are collected at different spatial and temporal scales.
Leading the way forward
Photo-ID studies for sharks and rays have significantly led broader efforts for computer-assisted research on wildlife populations. The data sets acquired and carefully curated over the past two decades (Wildbook for Whale Sharks 2017, MantaMatcher 2017, Spotashark 2017) are likely to provide the foundation for the development of new techniques in computer vision and population analysis for both marine and terrestrial species. Artificial intelligence and computer vision are already increasing data volume (Wildbook for Whale Sharks 2017) and reducing the required effort for analysis. This trend is likely to push researchers into new interactions with machines (computers, drones, etc.) and changing roles and responsibilities within research projects, with an increasing focus on understanding and successfully implementing technology. By integrating the cameras of tourists and citizen scientists into research work, and augmenting researchers with computer vision and artificial intelligence, we can plausibly imagine a wildlife research and conservation community that is continuously informed about animal population sizes and their individual interactions, movements, and behaviors.
Summary and conclusions
Photo-ID is a relatively simple research technique, usually requiring only off-the-shelf components and a basic level of training. Its non-invasive nature lends itself to use on opportunistic platforms, such as tourist vessels, and the extension of data collection via citizen science programs, enhancing outreach and public engagement potential. However, the simplicity of photo-ID should not be confused with a lack of power.
Use of photo-ID in shark research continues to increase, and camera sensors and battery life continue to improve. Cameras are getting smaller, sensors are increasing in their resolving power and low-light capabilities, and the number of potential platforms they can be affixed to is rapidly expanding. It is now possible to extract photo-IDs from autonomous underwater and aerial “drones” and dedicated research platforms such as animal-mounted tags, baited remote underwater video survey (BRUV) systems and remotely operated vehicles (ROVs). It is likely that photo-ID will be increasingly used in conjunction with remote cameras placed at areas like cleaning stations and areas of feeding and reproductive importance, providing improved sampling coverage and standardization over time (Bicknell et al. 2016; Oliver et al. 2011; O’Shea et al. 2010). Taken together, this expansion of use-case scenarios will allow for photo-ID studies of species that may not be adequately surveyed by scientific or recreational divers, or live below normal diving depths.
As well as being an important study methodology in its own right, photo-ID can facilitate or extend studies using complementary techniques, such as telemetry studies, either by bolstering sample size (Guttridge et al. 2017) or by allowing continued long-term monitoring of individuals following tag loss. In the case of white sharks, a combination of satellite tagging and photo-identification allowed return migration from Australian to South African waters to be established (Bonfil et al. 2005), while a whale shark was tracked from the Gulf of Mexico to the mid-Atlantic off Brazil, and back (Hueter et al. 2013). Tagging and sighting data can, in fact, be combined within mark-recapture models to improve the precision of results, and to help mitigate the lower detectability of photo-ID only studies (Chapple et al. 2016; Dudgeon et al. 2015, Lee et al. 2014). Incorporating individual IDs into other study methods that may include predictable biases, such as inflation of shark counts in underwater visual census (Ward-Paige et al. 2010) and the under-estimation of abundance from BRUV systems (Willis et al. 2000), can also enhance the results obtained from these techniques.
The resolution of modern cameras allows the extraction of considerable information from photographs. This could include details such as parasite loading (Mucientes et al. 2008) and other individual health and fitness information, such as infection or body proportions. While it may be a reputational risk to write this in a book about sharks, our marine mammal research colleagues are currently well ahead in this area, and it is worth perusing the literature in that field to assess the possibilities (e. g. Hunt et al. 2013; 2015). Elasmobranch photo-ID studies will be enhanced with increasing use of photogrammetric techniques, which can concurrently evaluate the length, body condition and mass of individuals (Shortis et al. 2009; Waite et al. 2007).
A primary benefit of using photo-ID is the ability to expand data collection, through integration with citizen science initiatives, and enhanced collaboration opportunities due to the ease of matching standardized photos between research groups. Use of photos means that it is easy to verify the accuracy of public submissions, and could be particularly useful for population studies of elasmobranchs that appear to be at lower than optimal densities for cost-effective dedicated surveys, or cryptic species. Many of the larger photo-identifiable species, such as white sharks and basking sharks, routinely traverse political boundaries (Bonfil et al. 2005; 2010; Gore et al. 2008; Skomal et al. 2009). Photo-ID provides a cost-effective means of assessing population-level interchange between discrete areas, the products of which can improve population estimates and stock delineation. To fulfil this potential, there is a need for standardization of species-specific techniques between research groups, and increased movement towards routine data-sharing. Data collection, processing and sharing will all be facilitated by computer-assisted datamining and identification, enabling a “big data” approach to shark science.
Photo-ID studies are steadily expanding to new species and sites, and asking more ambitious questions. Photo-ID offers a useful alternative or adjunct to conventional tagging where its assumptions and practical constraints are met, and the widespread adoption of this research technique through the scientific community is enhancing opportunities for the public to become directly involved in projects. This can benefit researchers while offering an educational experience for interested participants. As emerging technologies increasingly allow the diverse ecology and behaviors of sharks to be observed first-hand, we hope that more and more scientists will bring their cameras along for the journey.
I want to thank my regular collaborators and co-authors, particularly Andrea Marshall, Chris Rohner and Jason Holmberg, for helping inform and shape my thoughts on photo-identification. Colin Simpfendorfer and Jeff Carrier provided useful feedback on the book chapter that this article was originally based on. While no specific funding was applied to produce this resource, I greatly appreciate the support from our Patreon community and other MMF funders.
Acuña-Marrero D, Jiménez J, Smith F, Doherty Jr PF, Hearn A, Green JR, Paredes-Jarrín J, Salinas-de-León P (2014) Whale shark (Rhincodon typus) seasonal presence, residence time and habitat use at Darwin Island, Galapagos Marine Reserve. PLoS One 9:e115946
Anderson SD, Chapple TK, Jorgensen, SJ, Klimley AP, Block BA (2011) Long-term individual identification and site fidelity of white sharks, Carcharodon carcharias, off California using dorsal fins. Mar Biol 158:1233-1237
Anderson SD, Goldman KJ (1996) Photographic evidence of white shark movements in California waters. Calif Fish Game 82:182–186
Andrzejaczek S, Meeuwig J, Rowat D, Pierce S, Davies T, Fisher R, Meekan M (2016) The ecological connectivity of whale shark aggregations in the Indian Ocean: A photo-identification approach. Royal Soc Open Sci 3:160455
Araujo G, Snow S, So CL, Labaja J, Murray R, Colucci A, Ponzo A (2016) Population structure, residency patterns and movements of whale sharks in Southern Leyte, Philippines: results from dedicated photo‐ID and citizen science. Aquat Conserv 27:237-252
Ari C (2014) Rapid colouration changes of manta rays (Mobulidae). Biol J Linnean Soc 113:180-193
Arzoumanian Z, Holmberg J, Norman B (2005) An astronomical pattern-matching algorithm for computer-aided identification of whale sharks, Rhincodon typus. J Appl Ecol 42:999–1011
Andreotti S, Holtzhausen P, Rutzen M, Meÿer M, van der Walt S, Herbst B, Matthee CA (2017) Semi-automated software for dorsal fin photographic identification of marine species: application to Carcharodon carcharias. Mar Biodiv 1-6
Bansemer CS, Bennett MB (2008) Multi-year validation of photographic identification of grey nurse sharks, Carcharias taurus, and applications for non-invasive conservation research. Mar Freshwater Res 59:322-331
Bansemer CS, Bennett MB (2009) Reproductive periodicity, localised movements and behavioural segregation of pregnant Carcharias taurus at Wolf Rock, southeast Queensland, Australia. Mar Ecol Prog Ser 374:215–227
Bansemer CS, Bennett MB (2011) Sex- and maturity-based differences in movements and migration patterns of grey nurse sharks, Carcharias taurus, along the eastern coast of Australia. Mar Freshwater Res 62:596–606
Barker SM, Williamson JE (2010) Collaborative photo-identification and monitoring of grey nurse sharks (Carcharias taurus) at key aggregation sites along the eastern coast of Australia. Mar Freshwater Res 61:971–979
Bicknell AWJ, Godley BJ, Sheehan EV, Votier SC, Witt MJ (2016) Camera technology for monitoring marine biodiversity and human impact. Front Ecol Environ 14:424-432
Bolger DT, Morrison TA, Vance B, Lee D, Farid H (2012) A computer‐assisted system for photographic mark-recapture analysis. Methods Ecol Evol 3:813-822
Bonfil R, Meyer M, Scholl MC, Johnson R, O’Brian S, Oosthuizen H, Swanson S, Kotze D, Paterson M (2005) Transoceanic migration, spatial dynamics and population linkages of white sharks. Science 310:100–103
Bonfil R, Francis MP, Duffy C, Manning MJ, O’Brian SO (2010) Large-scale tropical movements and diving behaviour of white sharks Carcharodon carcharias tagged off New Zealand. Aquatic Biol 8: 115–123
Bonner SJ, Holmberg J (2013) Mark-recapture with multiple, non-invasive marks. Biometrics 69:766-775
Bradshaw CJA, Mollet HF, Meekan MG (2007) Inferring population trends for the world’s largest fish from mark–recapture estimates of survival. J Anim Ecol 76:480–489
Brooks K, Rowat D, Pierce SJ, Jouannet D, Vely M (2010) Seeing spots: photo- identification as a regional tool for whale shark identification. WIO J Mar Sci 9:185–194
Bruce BD, Bradford RW (2013) The effects of shark cage-diving operations on the behaviour and movements of white sharks, Carcharodon carcharias, at the Neptune Islands, South Australia. Mar Biol 160:889–907
Brunnschweiler JM, Baensch H (2011) Seasonal and long-term changes in relative abundance of bull sharks from a tourist feeding site in Fiji. PLoS One 6:e16597
Brunnschweiler JM, Barnett A (2013) Opportunistic visitors: long-term behavioural response of bull sharks to food provisioning in Fiji. PLoS ONE 8:e58522
Buray N, Mourier J, Planes S, Clua E (2009) Underwater photo-identification of sicklefin lemon sharks, Negaprion acutidens, at Moorea (French Polynesia). Cybium 33:21-27
Burgess GH, Bruce BD, Cailliet GM, Goldman KJ, Grubbs RD, Lowe CG, MacNeil MA, Mollet HF, Weng KC, O’Sullivan JB (2014) A re-evaluation of the size of the white shark (Carcharodon carcharias) population off California, USA. PLoS One 9:e98078
Burnham KP, Anderson DR, White GC, Brownie C, Pollock KH (1987) Design and analysis methods for fish survival experiments based on release-recapture. Am Fish Soc Monogr 5:1-437
Cagua EF, Cochran JEM, Rohner CA, Prebble CEM, Sinclair-Taylor TH, Pierce SJ, Berumen ML (2015) Acoustic telemetry reveals cryptic residency of whale sharks. Biol Lett 11:20150092
Carraro R, Gladstone W (2006) Habitat preferences and site fidelity of the ornate wobbegong shark (Orectolobus ornatus) on rocky reefs of New South Wales. Pac Sci 60:207–223
Castro ALF, Rosa RS (2005) Use of natural marks on population estimates of the nurse shark, Ginglymostoma cirratum, at Atol das Rocas Biological Reserve, Brazil. Environ Biol Fish 72:213–221
Chapman DD, Feldheim KA, Papastamatiou YP, Hueter RE (2015) There and back again: review of residency and return migrations in sharks, with implications for population structure and management. Annu Rev Mar Sci 7:547-570
Chapple TK, Jorgensen SJ, Anderson SD, Kanive PE, Klimley AP, Botsford LW, Block BA (2011) A first estimate of white shark, Carcharodon carcharias, abundance off Central California. Biol Lett 7:581–583
Chapple TK, Chambert T, Kanive PE, Jorgensen SJ, Rotella JJ, Anderson SD, Carlisle AB, Block BA (2016) A novel application of multi-event modeling to estimate class segregation in a highly migratory oceanic vertebrate. Ecology 97:3494-3502
Corcoran MJ, Gruber SH (1999) The use of photo-identification to study the social organization of the spotted eagle ray, Aetobatus narinari. Bahamas J Sci 11:21–27
Couturier LIE, Marshall AD, Jaine FRA, Kashiwagi T, Pierce SJ, Townsend KA, Weeks SJ, Bennett MB, Richardson AJ (2012) Biology, ecology and conservation of the Mobulidae. J Fish Biol 80:1075–1119
Couturier LIE, Dudgeon CL, Pollock KH, Jaine FRA, Bennett MB, Townsend KA, Weeks SJ, Richardson AJ (2014) Population dynamics of the reef manta ray Manta alfredi in eastern Australia. Coral Reefs 33:329-342
Davies TK, Stevens G, Meekan MG, Struve J, Rowcliffe JM (2012) Can citizen science monitor whale-shark aggregations? Investigating bias in mark-recapture modelling using identification photographs sourced from the public. Wildlife Res 39:696-704
Deakos MH (2010) Paired-laser photogrammetry as a simple and accurate system for measuring the body size of free-ranging manta rays Manta alfredi. Aquatic Biol 10:1–10
Deakos MH, Baker JD, Bejder L (2011) Characteristics of a manta ray Manta alfredi population off Maui, Hawaii, and implications for management. Mar Ecol Prog Ser 429:245–260
Delaney DG, Johnson R, Bester MN, Gennari E (2012) Accuracy of using visual identification of white sharks to estimate residency patterns. PLoS One 7:e34753
Dicken ML, Booth AJ, Smale MJ (2006) Preliminary observations of tag shedding, tag reporting, tag wounds, and tag biofouling for raggedtooth sharks (Carcharias taurus) tagged off the east coast of South Africa. ICES J Mar Sci 63:1640 – 1648
Domeier ML, Nasby-Lucas N (2007) Annual re-sightings of photographically identified white sharks (Carcharodon carcharias) at an eastern Pacific aggregation site (Guadalupe Island, Mexico). Mar Biol 150:977-984
Dudgeon CL, Noad MJ, Lanyon JM (2008) Abundance and demography of a seasonal aggregation of zebra sharks Stegostoma fasciatum. Mar Ecol Prog Ser 368:269-281
Dudgeon CL, Pollock KH, Braccini JM, Semmens JM, Barnett A (2015) Integrating acoustic telemetry into mark-recapture models to improve the precision of apparent survival and abundance estimates. Oecologia 178:761-772
Dureuil M, Towner AV, Ciolfi LG, Beck LA (2015) A computer-aided framework for subsurface identification of white shark pigment patterns. Afr J Mar Sci 37:363-371
Duyck J, Finn C, Hutcheon A, Vera P, Salas J, Ravela S (2015) Sloop: A pattern retrieval engine for individual animal identification. Pattern Recognition 48:1059-1073
Evans PGH, Hammond PS (2004) Monitoring cetaceans in European waters. Mammal Rev 34:131–156
Feldheim KA, Gruber SH, Ashley MV (2002) The breeding biology of lemon sharks at a tropical nursery lagoon. Proc R Soc Lond (Biol) 269:1655 – 1661
Findlay R, Gennari E, Cantor M, Tittensor, DP (2016) How solitary are white sharks: social interactions or just spatial proximity? Behav Ecol Sociobiol 70:1735-1744
Fitzpatrick B, Meekan M, Richards A (2006) Shark attacks on a whale shark (Rhincodon typus) at Ningaloo Reef, Western Australia. Bull Mar Sci 78:397-402
Fouts WR, Nelson DR (1999) Prey capture by the Pacific angel shark, Squatina californica: visually mediated strikes and ambush-site characteristics. Copeia 1999:304 – 312
Gallagher AJ, Vianna GMS, Papastamatiou YP, Macdonald C, Guttridge TL, Hammerschlag N (2015) Biological effects, conservation potential, and research priorities of shark diving tourism. Biol Cons 184:365-379
GBIF (2016) What is GBIF? The Global Biodiversity Information Facility. https://www.gbif.org/what-is-gbif. Accessed 3 November 2016
Germanov ES, Marshall AD (2014) Running the gauntlet: regional movement patterns of Manta alfredi through a complex of parks and fisheries. PLoS One 9:e110071
Gore MA, Rowat D, Hall J, Gell FR, Ormond RF (2008) Transatlantic migration and deep mid-ocean diving by basking shark. Biol Lett 4:395–398
Gore MA, Frey PH, Ormond RF, Allan H, Gilkes G (2016) Use of photo-identification and mark-recapture methodology to assess basking shark (Cetorhinus maximus) populations. PLoS One 11:e0150160
Graham R, Roberts CM (2007) Assessing the size, growth and structure of a seasonal population of whale sharks (Rhincodon typus Smith 1828) using conventional tagging and photo identification. Fish Res 84:71–80
Groth EJ (1986) A pattern-matching algorithm for two-dimensional coordinate lists. Astron J 91:1244–1248
Gubili C, Johnson R, Gennari E, Oosthuizen WH, Kotze D, Meÿer M, Sims DW, Jones CS, Noble LR (2009) Concordance of genetic and fin photo identification in the great white shark, Carcharodon carcharias, off Mossel Bay, South Africa. Mar Biol 156:2199-2207
Guttridge TL, Van Zinnicq Bergmann MPM, Bolte C, Howey LA, Finger JS, Kessel ST, Brooks JL, Winram W, Bond ME, Jordan LKB, Cashman RC, Tolentino ER, Grubbs RD, Gruber SH (2017) Philopatry and regional connectivity of the great hammerhead shark, Sphyrna mokarran in the U.S. and Bahamas. Front Mar Sci 4:3
Hammerschlag N, Cooke SJ, Gallagher AJ, Godley BJ (2014) Considering the fate of electronic tags: interactions with stakeholders and user responsibility when encountering tagged aquatic animals. Methods Ecol Evol 5:1147-1153
Hewitt, AM, Kock AA, Booth AJ, Griffiths CL (2017) Trends in sightings and population structure of white sharks, Carcharodon carcharias, at Seal Island, False Bay, South Africa, and the emigration of subadult female sharks approaching maturity. Environ Biol Fish
Holmberg J, Norman B, Arzoumanian Z (2008) Robust, comparable population metrics through collaborative photo-monitoring of whale sharks Rhincodon typus. Ecol Appl 18:222–233
Holmberg J, Norman B, Arzoumanian Z (2009) Estimating population size, structure, and residency time for whale sharks Rhincodon typus through collaborative photo identification. Endanger Species Res 7:39–53
Hueter RE, Tyminski JP, de la Parra R (2013) Horizontal movements, migration patterns, and population structure of whale sharks in the Gulf of Mexico and northwestern Caribbean Sea. PLoS One 8:e71883
Hughes B, Burghardt T (2017) Automated visual fin identification of individual great white sharks. Int J Comput Vis 122:542-557
Hunt KE, Moore MJ, Rolland RM, Kellar NM, Hall AJ, Kershaw J, Raverty SA, Davis CE, Yeates LC, Fauquier DA, Rowles TK (2013) Overcoming the challenges of studying conservation physiology in large whales: a review of available methods. Conserv Physiol 1:cot006
Hunt KE, Rolland RM, Kraus SD (2015) Conservation physiology of an uncatchable animal: the North Atlantic right whale (Eubalaena glacialis). Integr Comp Biol 55:577-86
Hunt TN, Bejder L, Allen SJ, Rankin RW, Hanf DM, Parra GJ (2017) Demographic characteristics of Australian humpback dolphins reveal important habitat toward the southwestern limit of their range. Endanger Species Res 32:71-88
I3S (2016) I3S: Interactive Individual Identification System. https://www.reijns.com/i3s. Accessed 1 January 2016
Irion DT, Noble LR, Kock AA, Gennari E, Dicken ML, Hewitt AM, Towner AV, Booth AJ, Smale MJ, Cliff G (2017) Pessimistic assessment of white shark population status in South Africa: Comment on Andreotti et al. (2016). Mar Ecol Prog Ser 577:251-255
Jacoby DMP, Croft DP, Sims DW (2012) Social behaviour in sharks and rays: analysis, patterns and implications for conservation. Fish and Fisheries 13:399-417
Jewell OJD, Wcisel MA, Gennari E, Towner AV, Bester MN, Johnson RL, Singh S (2011) Effects of smart position only (SPOT) tag deployment on white sharks Carcharodon carcharias in South Africa. PLoS One 6:e27242
Kitchen-Wheeler AM (2010) Visual identification of individual manta ray (Manta alfredi) in the Maldives Islands, Western Indian Ocean. Mar Biol Res 6:351-63
Klimley AP, Anderson SD (1996) Residency patterns of white sharks at the South Farallon Islands, California. In Great White Sharks: The Biology of Carcharodon carcharias (Klimley, A. P. and Ainley, D. G., eds), pp 309–316. San Diego, CA: Academic Press Inc.
Kock A, O’Riain MJ, Mauff K, Meÿer M, Kotze D, Griffiths C (2013) Residency, habitat use and sexual segregation of white sharks, Carcharodon carcharias in False Bay, South Africa. PLoS ONE 8:e55048
Kohler NE, Turner PA (2001) Shark tagging: a review of conventional methods and studies. Environ Biol Fish 60:191–223
Laroche RK, Kock AA, Dill LM, & Oosthuizen WH (2007) Effects of provisioning ecotourism activity on the behaviour of white sharks Carcharodon carcharias. Mar Ecol Prog Ser 338:199–209
Lee KA, Huveneers C, Gimenez O, Peddemors V, Harcourt RG (2014) To catch or to sight? A comparison of demographic parameter estimates obtained from mark-recapture and mark-resight models. Biodivers Conserv 23:2781-2800
Lindfield SJ, Harvey ES, Mcllwain JL, Halford AR (2014) Silent fish surveys: bubble-free diving highlights inaccuracies associated with SCUBA-based surveys in heavily fished areas. Methods Ecol Evol 5:1061-1069
Luiz OJ, Balboni AP, Kodja G, Andrade M, Marum H (2009) Seasonal occurrences of Manta birostris (Chondrichthyes: Mobulidae) in southeastern Brazil. Ichthyol Res 56:96-99
Manire CA, Gruber SH (1991) Effect of M-type dart tags on the field growth of juvenile lemon sharks. Trans Am Fish Soc 120:776–780
MantaMatcher (2017) MantaMatcher: The Wildbook for Giant Mantas. Marine Megafauna Foundation and Wild Me. https://www.mantamatcher.org. Accessed 1 October 2017
MantaMatcher User Agreement (2017) MantaMatcher: The Wildbook for Giant Mantas. Marine Megafauna Foundation and Wild Me. https://www.mantamatcher.org/userAgreement.jsp. Accessed 1 October 2017
Markowitz TA, Harlin AD, Wursig B (2003) Digital photography improves efficiency of individual dolphin identification. Mar Mamm Sci 19:217–223
Marshall AD, Bennett MB (2010a) The frequency and effect of shark-inflicted bite injuries to the reef manta ray (Manta alfredi). Afr J Mar Sci 32:573–580
Marshall AD, Bennett MB (2010b) Reproductive ecology of the reef manta ray (Manta alfredi) in southern Mozambique. J Fish Biol 77:169–190
Marshall AD, Dudgeon CL, Bennett MB (2011) Size and structure of a photographically identified population of manta rays Manta alfredi in southern Mozambique. Mar Biol 158:1111-1124
Marshall AD, Pierce SJ, Bennett MB (2008) Morphological measurements of manta rays (Manta birostris) with a description of a foetus from the east coast of Southern Africa. Zootaxa 1717:24–30
Marshall AD, Pierce SJ (2012) The use and abuse of photographic identification in sharks and rays. J Fish Biol 80:1361-1379
McKinney JA, Hoffmayer ER, Holmberg J, Graham RT, Driggers WB III, de la Parra-Venegas R, Galván-Pastoriza BE, Fox S, Pierce SJ, Dove ADM (2017) Long-term assessment of whale shark population demography and connectivity using photo-identification in the Western Atlantic Ocean. PLoS One 12: e0180495
Meekan MG, Bradshaw CJA, Press M, McLean C, Richards A, Quasnichka S, Taylor JG (2006) Population size and structure of whale sharks (Rhincodon typus) at Ningaloo Reef, Western Australia. Mar Ecol Prog Ser 319:275-285
Menon S, Berger-Wolf TY, Kiciman E, Joppa L, Stewart CV, Parham J, Crall J, Holmberg J, Van Oast J (2017) Animal population estimation using Flickr images. 2nd International Workshop on the Social Web for Environmental and Ecological Monitoring (SWEEM 2017). Troy, New York
Mourier J, Vercelloni J, Planes S (2012) Evidence of social communities in a spatially structured network of a free-ranging shark species. Anim Behav 83:389-401
Mucientes GR, Queiroz N, Pierce SJ, Sazima I, Brunnschweiler JM (2008) Is host ectoparasite load related to echeneid fish presence? Res Lett Ecol 2008:107576
Myrberg Jr AA, Gruber SH (1974) The behavior of the bonnethead shark, Sphyrna tiburo. Copeia 1974(2):358-374
Norman BM, Stevens JD (2007) Size and maturity status of the whale shark (Rhincodon typus) at Ningaloo Reef in Western Australia. Fish Res 84:81–86
Norman BM, Morgan DL (2016) The return of “Stumpy” the whale shark: two decades and counting. Front Ecol Environ 14:449-450
Norman BM, Holmberg JA, Arzoumanian Z et al. (In press) Undersea constellations: ‘citizen scientists’ elucidate the global biology of a threatened marine mega-vertebrate. Bioscience
OBIS (2016) Ocean Biogeographic Information System. Intergovernmental Oceanographic Commission of UNESCO. www.iobis.org. Accessed 1 January 2016
Oliver SP, Hussey NE, Turner JR, Beckett AJ (2011) Oceanic sharks clean at coastal seamount. PLoS One 6:e14755
O’Shea OR, Kingsford MJ, Seymour J (2010) Tide-related periodicity of manta rays and sharks to cleaning stations on a coral reef. Mar Freshwater Res 61:65–73
Parham J, Crall J, Stewart C, Berger-Wolf T, Rubenstein D (2017) Animal population censusing at scale with citizen science and photographic identification. AAAI 2017 Spring Symposium. Palo Alto, California
Pierce SJ, Bennett MB (2009b) Validated annual band pair periodicity and growth parameters of blue-spotted maskray (Neotrygon kuhlii) from southeast Queensland, Australia. J Fish Biol 75:2490–2508
Pierce SJ, Pardo SA, Bennett MB (2009a) Reproduction of the blue-spotted maskray, Neotrygon kuhlii (Myliobatoidei: Dasyatidae) in south-east Queensland, Australia. J Fish Biol 74:1291–1308
Porcher IF (2005) On the gestation period of the blackfin reef shark, Carcharhinus melanopterus, in waters off Moorea, French Polynesia. Mar Biol 146:1207–1211
Pratt HL Jr., Carrier JC (2001) A review of elasmobranch reproductive behaviour with a case study on the nurse shark, Ginglymostoma cirratum. Environ Biol Fish 60:157–188
Ramírez-Macías D, Vázquez-Haikin A, Vázquez-Juárez R (2012) Whale shark Rhincodon typus populations along the west coast of the Gulf of California and implications for management. Endanger Species Res 18:115-128
Ramírez-Macías D, Queiroz N, Pierce SJ, Humphries NE, Sims DW, Brunnschweiler JM (2017) Oceanic adults, coastal juveniles: tracking the habitat use of whale sharks off the Pacific coast of Mexico. PeerJ 5:e3271
Riley MJ, Harman A, Rees RG (2009) Evidence of continued hunting of whale sharks Rhincodon typus in the Maldives. Environ Biol Fish 86:371–374
Riley MJ, Hale MS, Harman A, Rees RG (2010) Analysis of whale shark Rhincodon typus aggregations near South Ari Atoll, Maldives Archipelago. Aquatic Biol 8:145–150
Robbins RL (2007) Environmental variables affecting the sexual segregation of great white sharks Carcharodon carcharias at the Neptune Islands South Australia. J Fish
Robbins R, Fox A (2013) Further evidence of pigmentation change in white sharks, Carcharodon carcharias. Mar Freshwater Res 63:1215-1217
Robinson DP, Jaidah MY, Bach S, Lee K, Jabado RW, Rohner CA, March A, Caprodossi S, Henderson AC, Mair JM, Ormond R, Pierce SJ (2016) Population structure, abundance and movement of whale sharks in the Arabian Gulf and Gulf of Oman. PLoS One 11:e0158593
Rohner CA, SJ Pierce, AD Marshall, SJ Weeks, MB Bennett, Richardson AJ (2013) Trends in sightings and environmental influences on a coastal aggregation of manta rays and whale sharks. Mar Ecol Prog Ser 482:153-168
Rohner CA, Richardson AJ, Prebble CEM, Marshall AD, Bennett MB, Weeks SJ, Cliff G, Wintner SP, Pierce SJ (2015) Laser photogrammetry improves size and demographic estimates for whale sharks. PeerJ 3:e886
Rowat D, Speed CW, Meekan MG, Gore M (2009) Population abundance and apparent survival of the Vulnerable whale shark, Rhincodon typus, in the Seychelles aggregation. Oryx 43:591–598
Rowat D, Brooks K, March A, McCarten C, Jouannet D, Riley L, Jeffreys G, Perri M, Vely M, Pardigon B (2011) Long-term membership of whale sharks (Rhincodon typus) in coastal aggregations in Seychelles and Djibouti. Mar Freshwater Res 62:621–627
Shortis M, Harvey E, Abdo D (2009) A review of underwater stereo-image measurement for marine biology and ecology applications. Oceanogr Mar Biol 47:257–292
Sims DW, Speedy CD, Fox AM (2000) Movements and growth of a female basking shark re-sighted after a three-year period. J Mar Biol Assoc UK 80:1141–1142
Skomal GB, Zeeman SI, Chrisholm JH, Summers EL, Walsh HJ, McMahon KW, Thorrold SR (2009) Transequatorial migrations by basking sharks in the western Atlantic Ocean. Curr Biol 19:1019–1022
Smallegange IM, van der Ouderaa IBC, Tibiriçá Y (2016) Effects of yearling, juvenile and adult survival on reef manta ray (Manta alfredi) demography. PeerJ 4:e2370
Smith SE, Mitchell RA, Fuller D (2003) Age-validation of a leopard shark (Triakis semifasciata) recaptured after 20 years. Fish Bull 101:194–198
Speed CW, Meekan MG, Bradshaw JA (2007) Spot the match – wildlife photo- identification using information theory. Front Zool 4:1–11
Speed CW, Meekan MG, Rowat D, Pierce SJ, Marshall AD, Bradshaw CJA (2008) Scarring patterns and relative mortality rates of Indian Ocean whale shark. J Fish Biol 72:1488–1503
Spotashark (2017) Spotashark and the North Carolina Aquariums. https://www.spotashark.com. Accessed 1 October 2017
Stansbury AL, Götz T, Deecke VB, Janik VM (2015) Grey seals use anthropogenic signals from acoustic tags to locate fish: evidence from a simulated foraging task. Proc R Soc Lond (Biol) 282:20141595
Taylor G (1994) Whale Sharks, the Giants of Ningaloo Reef. Angus & Robertson, Sydney
Town C, Marshall A, Sethasathien N (2013) Manta Matcher: Automated photographic identification of manta rays using keypoint features. Ecol Evol 3:1902–1914
Towner AV, Wcisel MA, Reisinger RR, Edwards D, Jewell OJ (2013) Gauging the threat: the first population estimate for white sharks in South Africa using photo identification and automated software. PloS One 8:e66035
Towner A, Smale MJ, Jewell O (2012) Boat strike wound healing in Carcharodon carcharias. Global Perspectives on the Biology and Life History of the White Shark. CRC Press, Boca Raton:77-84
Van Tienhoven AM, Den Hartog JE, Reijns RA, Peddemors VM (2007) A computer-aided program for pattern-matching of natural marks on the spotted ragged-tooth shark Carcharias taurus. J Appl Ecol 44:273–280
Waite JN, Schrader WJ, Mellish JE, Horning M (2007) Three-dimensional photogrammetry as a tool for estimating morphometrics and body mass of Stellar sea lions (Eumetopias jubatus). Can J Fish Aquat Sci 64:296 – 303
Ward-Paige C, Mills Flemming J, Lotze HK (2010) Overestimating fish counts by non-instantaneous visual censuses: consequences for population and community descriptions. PLoS One 5:e11722
Whitney NM, Pratt HL, Carrier JC (2004) Group courtship, mating behaviour, and siphon sac function in the whitetip reef shark, Triaenodon obesus. Anim Behav 68:1435–1442
Whitney NM, Pyle RL, Holland KN, Barcz JT (2011) Movements, reproductive seasonality, and fisheries interactions in the whitetip reef shark (Triaenodon obesus) from community-contributed photographs. Environ Biol Fish 93:121–136
Wieczorek J, Bloom D, Guralnick R, Blum S, Doring M, et al. (2012) Darwin Core: An Evolving Community-Developed Biodiversity Data Standard. PLoS One 7:e29715
Wildbook (2017) Wildbook: Software to Combat Extinction. Wild Me. www.wildbook.org. Accessed 1 October 2017
Wildbook for Whale Sharks (2017) Wildbook for Whale Sharks. Wild Me. https://www.whaleshark.org. Accessed 1 October 2017
Williams BK, Nichols JD, Conroy MJ (2002) Analysis and management of animal populations. Academic Press, San Diego, CA
Willis TJ, Millar RB, Babcock RC (2000) Detection of spatial variability in relative density of fishes: comparison of visual census, angling, and baited underwater video. Mar Ecol Prog Ser 198:249–260
Wilson B, Hammond PS, Thompson PM (1999) Estimating size and assessing trends in a coastal bottlenose dolphin population. Ecol Appl 9:288–300
Wilson RP, McMahon CR (2006) Measuring devices on wild animals: what constitutes acceptable practice? Front Ecol Environ 4:147–154
Yano K, Sato F, Takahashi T (1999) Observations of the mating behavior of the manta ray, Manta birostris, at the Ogasawara Islands, Japan. Ichthyol Res 46:289–296