If this hypothesis be correct, is it necessary that the sapid substances should actually enter the taste bud, or only affect its peripheral end? Many of the nerve fibrils entering the taste bud pass through its whole length and end quite near the mouth of the gustatory pore. These might be stimulated without the entrance of the stimulus within the taste bud. But there are many of these fibrils which do not reach to the peripheral end of the bud, but stop far short of this point, and then there are others that reach the entrance of the taste pore but turn back and end in the characteristic knob-like formation within the taste bud. In order that these fibrils may be stimulated upon their ends the stimuli would have to enter the taste bud. The answer to the fourth question is indeed the most difficult of all. What is the nature of this stimulation by which a fluid substance shall start an impulse along the nerve paths to the brain, which shall there produce sensation? About this last stage of the process nothing is known either about taste or any of the other senses. But very well-developed theories exist to account for the transformation of the physical stimulus into physiological nerve impulse. For instance, in the case of vision, the stimulus for which consists of ether vibrations, these ether waves cause chemical changes in certain hypothetical substances within the rods and cones of the retina. It is this chemical change which creates the nerve impulse. In the case of hearing, for which the stimulus consists of air vibrations, these waves, being slightly modified by the more superficial portions of the auditory mechanism, finally cause vibrations of the basilar membrane, which, in turn, produces the impulse in the auditory nerve. One of the earliest and simplest conceptions of the nature of the process in the taste organ was a mechanical theory proposed by Boyle, about 1675. He thought that the particles of various sapid substances differed in size and shape and that on account of these differences they produced different effects in their simple contact with the sensory ends of taste. According to Graham, who announced his theory in 1889, sapidity of substances depends on their chemical constitution, colloids being generally insipid and crystalloids being sapid, hence this has been known as a chemical theory. This difference of chemical structure, discussed on page 94, was made to account for the contact, or the lack of it, between the substances and the sensory ends, but does not account for the effect produced upon the sensory ends by the substances reaching them.