It is often a matter of great difficulty to determine which part of the nervous system is originally concerned in these rudimentary hallucinations. It is probable that in normal life they are most frequently due to peripheral disturbance. And it seems reasonable to suppose that where the hallucination remains in this initial stage of a very incompletely interpreted visual or auditory impression, whether in normal or abnormal life, its real physiological source is the periphery. For the automatic excitation of the centres would pretty certainly issue in the semblance of some definite, familiar variety of sense-impression which, moreover, as a part of a complex state known as a percept, would instantly present itself as a completely formed quasi-percept. In truth, we may pretty safely argue that if it is the centre which is directly thrown into a state of activity, it will be thrown into the usual complex, that is to say, _perceptional_, mode of activity. Let us now turn to hallucinations properly so called, that is to say, completely developed quasi-percepts. These commonly assume the form of visual or auditory hallucinations. Like the incomplete hallucinations, they may have their starting-point either in some disturbance in the peripheral regions of the nervous system or in the automatic activity of the central structures: or, to use the language of Baillarger, we may say that they are either "psycho-sensorial" or purely "psychical." A subjective visual sensation, arising from certain conditions in the retina and connected portions of the optic nerve, may by chance resemble a familiar impression, and so be at once interpreted as an effect of a particular external object. More frequently, however, the automatic activity of the centres must be regarded, either in part or altogether, as the physiological cause of the phenomenon. This is clearly the case when, on the subjective side, the hallucination answers to a preceding energetic activity of the imagination, as in the case of the visionary and the monomaniac. Sometimes, however, as we have seen, the hallucinatory percept answers to previous prolonged acts of perception, leaving a kind of reverberation in the structures concerned; and in this case it is obviously impossible to say whether the peripheral or central regions (if either) have most to do with the hallucination.[58]