Really, there was no other way?

Well,

we actually tried various approaches at the same time.
In addition to the NGS on spraints and jellies (described in the previous page) we attempted to collect samples by using hair-traps and to catch hints on otters' marking behaviour by camera-traps.
Let's start from the latter.

Otters are invisible (TO IR)

Known also as trail cameras, camera-traps are photo/videocameras automatically activated by the animal transit. They are largely used in naturalistic research to record behavioural and activity patterns of animals. Most of the camera-traps are triggered by a PIR (passive infrared) sensor able to distinguish the animal from the background (preventing activations caused by trees or leaves in motion) while minimising the drain on the battery. In other words, a PIR sensor detects the difference in the heat emitted by the animal and the background.

For 150 days we had a photo-trap pointing on a rock an otter used to mark: no photos.
Actually, no photos of otters, because the field of view extended up to the other river bank, and the camera recorded wild boars, wolves, deers, plus many species of micro-mammals and birds.
For all that, we often did find fresh spraints on the rock: the problem was specifically linked to otters.

We suspected the cause of otters' invisibility to PIR sensors to be found in the two thick layers of fur which cover the animal. We speculated the thermal insulation to be so efficient that, in the few seconds the animal spends out of water to mark, it wouldn't radiate enough heat to separate from the background.

To test this hypothesis we tweaked a photo-trap blinding the PIR sensor and equipping a pressure trigger placed on the marking rock.
After only two nights we recorded the following video, the nearby "standard" photo-trap didn't notice a thing.

Otter at the marking site

To confirm the hypothesis we needed a counterexample. We set a photo-trap along a riverbank (it's very difficult to spot places on land frequented by otters, that's why it's so important to record them on the marking sites). The camera was pointed to a small stretch of pebbles and stones with shallow, nearly still water: if an otter were to pass here would have had the time to heat up enough to trigger the PIR sensor.
And so it was: after few days the camera recorded the following video.

An otter on the Aventino River bank

The prototype we used to record our first video was created with some wire and a couple of wooden blades (actually two medical tongue depressors). Once we proved the effectiveness of the technique, we switched to more durable materials. We got in contact with Ettore Centofanti (Fototrappolaggio.it) who helped us to design a new version of the pressure trigger. The improved one was given a plastic disc to both expand the activation area, and to protect the trigger from the weather (one of the troubles we bumped into was the rain closing the small gap between the contacts continuously activating the camera, depleting the batteries, and filling the memory cards).

proof of concept
improved trigger

Thanks to the new version we were able to asses the cohabitation of otters and coypus (Myocastor coypus, an invasive alien species from South America). The new triggers still have some troubles though, one of which is related to the frequent overflows that the dams cause to the river.
Many times, even with good weather and the new plastic shield, we found batteries depleted and memory cards full. We later learned that, at irregular intervals, the water level downstream the Ateleta Dam increases and decreases by dozen of centimeters, in some cases completely submerging the marking rocks and our triggers with them.
Ultimately, the technique is far from perfect and it has much room for improvements, but it has already shown good prospects.

Myocastor_coypus, invasive species

The hair-traps

Knowing how low-quality would be the genetic material provided by marking signs, we attempted other ways to obtain DNA samples. A DNA source widely used in mammal studies is the fur (more precisely the hair follicle). There are different models of hair-traps varying from velcro strips to special glues, to barbed wire strands (used primarily to sample ursids as they use it to rub their back).

We were looking for a method to sample one individual at the time while all the previous methods have no ways to tell apart different sources. We then built a small wooden corridor with some stretched coil springs inside: the passing animal would release the springs which would trap some hairs inside the coils, and, when sprung, no other animal would be able to "contaminate" the sample.

size of the wooden box
a view of the inner springs

This kind of device clearly needs some time in the wild to lose the "just-built-by-humans" smell. The box was left to be weather-beaten for a while and then placed in the field. And then we waited: it could take some time for an animal to grow confident enough to enter.

After five weeks we finally found one of the four hair-traps sprung and some hair trapped in to the coils.
In a preliminary microscope analysis the hair seemed to be compatible with three mustelids species: Lutra lutra (the otter), Neovison vison (the american vison) and Martes foina (the stone marten).
The subsequent genetic analysis attributed the sample to the american vison, the second alien species overlapping the otter's range.

Neovison vison, invasive species

This method seemed to work fairly well but turned out to be very time-consuming: in addition to the actual hair-traps construction, we had to take into account the weeks before obtaining a sample and the daily checks. For these reasons this approach was put aside, waiting for the next opportunity to attempt again.

We
and you