Play in gulls

I usually check for ringed Common Gulls in my local park. Yesterday the flock, about a hundred strong, converged around someone providing food. While I was checking, a boy run towards the gulls flushing them. A few settled on a puddle nearby, and a second winter one was ringed! I approached and took some photos to try and see the code. No luck, the light was wrong and the gull too active. I decided a video could work as I might be able to get a sharp screen-grab. The gull found a pink object (visible on the left side on the photo above) and started to pick it up and drop it repeatedly, the object bobbed up on the water every time and the gull seemed quite interested, until, after a while, it walked away. I had ended up recording a session of gull play behaviour! I walked away and then thought I wanted to check the object, just in case it was some food item, but no, it was a spongy piece of plastic, probably torn from a soft ball. You can watch the clip here:

 I had never thought gulls played until I started following Ralph Hancock's wonderful blog on the birds on his daily walk around Hyde Park and Kensington Gardens. He describes and documents gulls (Herring, Common and Black-headed) playing drop-catch with various objects: sticks, leaves, plane tree seed heads, rubbish, and stones, and his observations made me pay more attention. He also pointed to this photo sequence of a young gull playing drop-catch.
 How can we know that a particular behaviour is play? Usually play is invoked when a behaviour appears to serve no adaptive purpose. Jennifer Gamble and Daniel Cristol, from the College of William and Mary, Virginia, studied drop-catch behaviour in Herring gulls to discount alternative hypothesis to play. Their Herring Gull population often dropped clams they find on the mud flats at low tide over hard substrates - a tarmac road nearby - to break the shells and eat them. Could drop-catch serve be a practical behaviour? They tested two alternative explanations: the 'kleptoparasite hypothesis' where drop-catch behaviour could allow a gull to asses the chances that other gulls that might be nearby, be ready to steal the clam as it hit the ground, or the 'reposition hypothesis' where the gull might drop the clam and then catch it to reposition it before a proper drop. In contrast, if drop-catch is a play behaviour they expected:
1) It would be carried out more frequently by juveniles.
2) Objects other than clams would be used.
3) It would be performed over soft substrates.
Their results rejected the alternative hypothesis and fitted the predictions of the play hypothesis: younger gulls engaged in more drop-catching than older gulls, that drop-catching was more commonly done over soft surfaces (the mud flats, as opposed to the road and gulls were less likely to drop-catch a clam (9%) than a non-clam object (62%). The conditions at which drop-catch happened were also more favourable to play behaviour: Drop-catch happened more often at warmer temperatures (when the gulls are less cold-stressed) and when wind was stronger (and flight is less costly).
 Play has been much less studied in birds than in mammals, and most of the research in birds has been done with corvids. Gulls are quite opportunistic, long-lived species and play could potentially have long-term beneficial consequences: perfecting the development of the coordinated flight behaviour required to succeed in kleptoparasitic attacks (gull often chase other gulls carrying food and if the attacked gull drops the food, individuals able to catch the dropped object in the air might have a higher chance of success of catching it first) or possibly practice the clam dropping behaviour.
 I have kept the bit of pink plastic, but I haven't yet managed to read the code on the ring, it starts by JP...

More information
Gamble, J. R. & Cristol, D. A. Drop-catch behaviour is play in Herring gulls, Larus argentatus. Animal Behaviour 63, 339–345 (2002). Here.