It’s time for a new guestblog, this one is by the amazing Louw Claassens. Louw is a South-African marine scientist at the Knysna Basin Project and a member of the IUCN Seahorse specialist group. She studies one of the world’s most endangered seahorses, part of her work involves studying their behaviour, which recently resulted in a very interesting publication (go check it out!). In this guestblog she gives you the most important findings of that paper and shares some fantastic video footage. Enjoy!

---

A big part of ecological research is based on observations – where do animals occur, what do they eat, what do they do. Some of these questions can be answered by using standard scientific methods e.g. a population survey can tell you where animals occur (although why is a whole other kettle of fish!). The tricky part sets in when you want to find out what an animal is doing. Conventionally, this entails going to the animal in question and watching it (sounds pretty simple, right?!). But it is here where observational effect (the act of observing has an effect on behaviour) and observational bias (researcher bias as to expected behaviour) creeps in.

One of the 21^st century solutions to these observational problems, is using cameras to study animals, and we are now even able to use cameras to study animals under water (thank goodness for relatively cheap action cameras such as GoPro’s!). Most fish research uses cameras to look at fish diversity, abundance, and habitat use – with limited work on actual fish behaviour. One of the reasons for this is probably owing to the highly mobile nature of most fish species.

So, is there a place for action cameras in fish behavioural research?

We focused on seahorses to answer this question. The conventional way to study seahorse behaviour entails getting in the water and watching the seahorse go about its business. Or, getting some seahorses and conducting observational research in the lab. The first method is problematic owing to two reasons: 1) Observer effect (the seahorse might act differently when you are watching it), and 2) seahorses move quite slowly most of the time, so detecting a behavioral pattern is quite difficult. Not even to mention the costs and time involved in doing this. The latter method might make sense, but it is well known that animal behaviour in captivity is rarely authentic.

Our aim was to test the efficacy of using video cameras to study the natural behaviour of a seahorse, and we had the perfect opportunity to do this! During a recent population survey of the endangered Knysna seahorse (Hippocampus capensis) in the Knysna estuary (South Africa) we found a stable population within a residential marina estate. The seahorses were found to use artificial Reno mattresses (wire cages filled with rocks). We had the seahorses, we had a relatively protected area to deploy cameras, and we had a sturdy structure to attach the cameras to.

In the first instance, we wanted to see if seahorse behaviour changed throughout the day e.g. between the morning, midday and afternoon. To add to this, we had an opportunity to see what happens to seahorse behaviour during the busy December holiday season. To do this, we used boat noise as a potential stressor (as occupancy of the residential marina estate increases from ~30 % to 100 % over the holiday period).

Video: Aggressive behaviour in the Knysna seahorse (Hippocampus capensis) – main action starts at 0:45.

But first we had to see if cameras successfully captured seahorse behaviour and if they could be used in behavioural assessments. We conducted a short trial period to test this, and found that 49 % of footage recorded contained seahorses. Using this data, we created an ethogram (a catalogue or table of all the different kinds of behaviour or activity observed in an animal) for H. capensis:

• Feeding: the seahorse is actively searching for prey animals.
• Irritation: identified by increased clicking and tail adjustments.
• Moving from holdfast to holdfast: seahorse moves around without any feeding behaviour in-between.
• Interaction: interaction behaviour can either be between a male and female as part of courting or between seahorses of the same sex and might entail aggression.
• Stationary: seahorse remains completely still.

Video: Knysna seahorse (Hippocampus capensis) feeding

The next step was to deploy the cameras throughout the day (morning, midday and afternoon) and across the longer time periods (Pre-holiday, Holiday and Post-Holiday). To assess behaviour we used 10-min video sections as a sample and timed all observed behaviour for a single focal animal during the sample.

We recorded hours of footage, of which 57 hours contained suitable footage of seahorse behaviour. Seahorses spent 82 % of their time feeding and we noted courting behaviour exclusively in the morning. This courting behaviour entailed grasping of the female’s tail by the male in an attempt to position himself face to face with the female, followed by swaying movements. We also found that seahorses were more visible and fed more during the morning. There were no differences between the behaviour of males and females.

Graphic footage! Video of a cormorant catching a seahorse (H. capensis)

We observed quite a few cuttlefish, rays and cormorants, but only noted predation by the latter (check out the video above!). Seahorses were also observed happily living side by side with octopus, although octopus are known to eat seahorses in Australia. We also noted some other curious fish, like our temperate butterfly fish (Chaetodon marleyi) (video below) – can you spot the seahorse?

A cold water butterfly fish (Chaetodon marleyi) checking out Louw’s GoPro setup

When we looked at behavioural changes across the longer-term periods, we noted a decrease in visibility and feeding activity of the focal seahorse, with an increase in irritation behaviour, during the holiday period. No courting behaviour was noted during the holiday period – which is a bit concerning, seeing that this species breeding season is from September to March.  Feeding activity and seahorse visibility increased again during the post-holiday period.

So, what does all this tell us? Action cameras are pretty useful in studying natural behaviour of seahorses. Recorded footage can be watched on fast-forward mode which enables a clear view of the behavioural pattern of the animal (something that is quite difficult to see whilst diving, as these guys move so slowly). For H. capensis, it was the first time that natural behaviour was studied, and we gained some valuable information with regards to feeding and interaction behaviour. In addition, it seems that boat noise has a negative effect on the natural behaviour of this species – an aspect which does need further research (preferably, a controlled experimental approach is needed here, to control the vast number of confounding factors that might have played a role!). The use of cameras in natural seagrass habitat also needs to be tested, as visibility might be problematic in dense vegetation.

In the past, the world of underwater research was exclusively meant for the eyes of the researcher/diver. Now, we are able to bring what we experience to the surface and to the lay person. And perhaps the real power of doing this is to create and instill that love and passion for the underwater world that all divers and water lovers have, in all people. I mean, who cannot fall in love with two seahorses doing their morning courting dance?

The secret sex lives of seahorses: mating dance of the Knysna seahorse