Now, this idea will, I think, help us to understand how loud and prolonged laughter came to join itself to the combative game of tickling and being tickled. If play—pure, good-natured play—was to be developed out of teasing attacks, it would become a matter of the highest importance that it should be clearly understood to be such. This would mean, first of all, that the assailant made it clear that his aim was not serious attack, but its playful semblance; and secondly, that the attacked party expressed his readiness to accept the assault in good part as sport. It would be of the greatest consequence to the animal that chanced to be in the play-mood and wished to make overtures of friendly combat that he should be sure of an equally gamesome attitude in the recipient of the challenge. One may see this by watching what happens when a dog, unwisely trying to force a frolic on another dog, is met by a growl and possibly by an uncovering of the canine teeth. Now, what better sign of good-temper, of readiness to accept the attack as pure fun, could nature have invented than the laugh? The smile is, no doubt, a pretty good indicator in some circumstances. Yet one must remember that the rudimentary smile of an ape-like ancestor may, now and again, have been misleading, as our own smiles are apt to be. A laugh would presumably be less easy to affect in such circumstances than a smile; and, in any case, it would be far less liable to be overlooked. In saying that the laughter which accompanies tickling and other closely allied forms of play in children owes its value to its being an admirable way of announcing the friendly playful mood, I do not mean that other signs are absent. Dr. L. Robinson reminds us that a tickled child will roll over on his back just like a puppy. The laughter and the rolling over seem to be two congenitally connected {184} modes of abandonment to the playful attack. In the young of other ticklish animals, _e.g._, the puppy, the rolling over may of itself suffice to give the friendly signal. It seems not unlikely that this consideration, the utility of laughter as a guarantee to a playful challenger that his overtures will be received in the proper spirit, applies to the evolution of all laughter which enters into such forms of social play as the pretence to attack, to frighten, and generally what we call good-natured teasing. It has been suggested that teasing might well be taken as the starting-point in the evolution of play.[118] By adopting this idea, and by regarding laughter, in its elementary form, as essentially a feature of social play, we might set out with this consideration of utility in constructing our theory of the evolution of laughter. One is tempted, too, to follow this course by the fact, recognised in common language, that much, at least, of the later and more refined laughter is analogous to the effect of tickling.[119]