Cuttlefish hatchlings are on their own right from the start. Emerging unprotected from the egg, the youngsters have to be able to fend for themselves, capturing food, evading predators and merging with the background if they are to survive. ‘The question is, how are they able to perform so many complex behaviours?’ says Ludovic Dickel from the Université de Caen Basse-Normandie, France, adding, ‘Is this repertoire under genetic control or is the animal able to learn before hatching?’ According to Dickel, there is some evidence that the youngsters, which usually prefer to dine on shrimp, can learn to feast

on crab during the first week of life, but how far back does the ability to learn go? Can tiny cuttlefish embryos that are still developing in the egg learn too? Intrigued, Dickel and his colleagues Sébastien Romagny, Anne-Sophie Darmaillacq, Mathieu Guibé and Cécile Bellanger decided to investigate when the developing embryos’ sensory systems begin to function and whether they were capable of learning a simple task (p. 4125).

Collecting eggs on cuttlefish traps, the team waited until the embryos had developed the ability to flex their mantles (stage 23 of development) before testing which of their senses had begun to function. Knowing that cuttlefish hatchlings need to evade European sea bass – they are one of the fish's favourite delicacies – the team tested the embryos' sense of smell by exposing them to the predator's odour, and observed the tiny animals' movements after painstakingly removing the egg's dark outer case. Explaining that the minute cuttlefish pulse their mantles when startled, the team was pleased to see that the developing youngsters flexed their mantles in response to the predators' odour. And when they tested the embryos' sense of touch by gently prodding their mantles with a blunt needle, the tiny cuttlefish also reacted. Even at this early stage of development, the embryos had developed the senses of smell and touch.

Yet, when Dickel and his colleagues tested the embryos' reaction to light, the animals barely stirred: their visual sense had not developed sufficiently. However, when the team repeated the test 1 to 2 weeks later when the visual pigments had developed (stage 25 of development), the cuttlefish pulsed their mantles: the visual system was finally functioning. ‘The visual system is also the last system to mature in vertebrate embryos, so this is an impressive homology between vertebrates and cephalopods as they diverged very early during evolution’, says Dickel.

Having pinpointed when the tiny cuttlefish begin to perceive light, the team tested whether the animals could learn to become desensitised to light. As soon as the embryos' visual system was functional, the team showed the youngsters 150 s long flashes of light interspersed by 30 min intervals of darkness: the youngsters enthusiastically pulsed their mantles whenever the light came on; they were unable to learn to ignore it. However, by stage 30, the embryos picked the idea up quickly. They soon began to lose interest in the light and when the team checked that the older embryos simply weren't tiring of the task – by throwing in a gentle tap from the needle between the bursts of light – the animals regained some of their interest in the light and began pulsing more energetically again. The cuttlefish embryos were not tiring, they were learning. So, cuttlefish embryos are able to collect information while in the egg, and Dickel is keen to find out how much sensory detail the tiny embryos can process before they embark on life in the open.

Feel, smell and see in an egg: emergence of perception and learning in an immature invertebrate, the cuttlefish embryo
J. Exp. Biol.