But we should scarcely speak of "learning" if the only modification consisted in the simple strengthening of native reactions, and at first thought it is difficult to see how the {298} exercise of any reaction could modify it in any other respect. But many reflexes are not perfectly fixed and invariable, but allow of some free play, and then exercise may fix or stabilize them, as is well illustrated in the case of the pecking response of the newly hatched chick. If grains are strewn before a chick one day old, it instinctively strikes at them, seizes them in its bill and swallows them; but, its aim being poor and uncertain, it actually gets, at first, only a fifth of the grains pecked at; by exercise it improves so as to get over half on the next day, over three-fourths after another day or two, and about 86 percent (which seems to be its limit) after about ten days of practice. Exercise has here modified a native reaction in the way of making it more definite and precise, by strengthening the accurate movement as against all the variations of the pecking movement that were made at the start. Where a native response is variable, exercise tends towards constancy, and so towards the _fixation_ of definite habits. A reflex may come to be _attached_ to a new stimulus, that does not naturally arouse it. A child who has accidentally been pricked with a pin, and of course made the flexion reflex in response to this natural stimulus, will make this same reaction to the sight of a pin approaching his skin. The seen pin is a _substitute stimulus_ that calls out the same response as the pin prick. This type of modification gives a measure of control over the reflexes; for when we pull the hand back voluntarily, or wink at will, or breathe deeply at will, we are executing these movements without the natural stimulus being present. Voluntary control includes also the ability to omit a response even if the natural stimulus is present. Holding the breath, keeping the eyes wide open in spite of the tendency to wink, not swallowing though the mouth is full of saliva, holding the hand steady when it is being pricked, and many {299} similar instances of control over reflexes are cases of _detachment_ of a native reaction from its natural stimulus. Not "starting" at a sudden sound to which we have grown used and not turning the eyes to look at a very familiar object, are other instances of this detachment. The _substitute response_ is another modification to be placed alongside of the substitute stimulus. Here a natural stimulus calls out a motor response different from its natural response. The muttered imprecation of the adult takes the place of the child's scream of pain. The loose holding of the pen between the thumb and the first two fingers takes the place of the child's full-fisted grasp.