PROFESSOR AT HEIDELBERG 219 



visual field is moving. To obtain a definite answer to this, the 

 further question (of great importance to optics and the theory 

 of knowledge) must be considered : how it comes about that in 

 the movement of the eye, owing to which the light impression 

 varies constantly at each point of the retina, our inference 

 should be that in spite of this variation of all the luminous im- 

 pressions there is no displacement nor alteration of the object, 

 but only a motion of the eye; it is obviously sufficient if this 

 inference obtains for infinitely small displacements of the eye. 

 But in order to convince ourselves that every alteration of the 

 image upon all points of the retina collectively depends on the 

 altered position of the eye alone, and not on any change of 

 the object in the field of vision, we must fulfil the condition in 

 virtue of which the transition of any point of the image from 

 the retinal fovea to a definite point on the retina at infinitesimal 

 distance, can only occur by rotation round a given axis, which 

 is unalterable relatively to the eye. 



It follows from the law of simplest orientation and from the 

 well-known theorem of mechanics, according to which the 

 axial directions of infinitely small revolutions are compounded 

 by the law of the parallelogram of forces, that the movement 

 of the point of fixation to any second point of the visual field 

 at infinitely small distance, must be caused by its revolution 

 round an axis lying in a given plane, which is unalterable in 

 relation to the eye. Now since the axes of rotation for all the 

 movements which occur lie in one plane, no infinitely small 

 rotation of the eye can produce a rotation of the same round 

 the line perpendicular to this axial plane, which Helmholtz terms 

 the atropic line of the eye. But since rotations of finite mag- 

 nitude can obviously not be compounded by this mechanical 

 law, the necessary condition for the maintenance of orientation 

 in the visual field is not altogether fulfilled in the movements of 

 the eye. Hence we must look for a law of eye-movements 

 which makes the sum of all the deviations from this principle 

 a minimum. 



Helmholtz now arrived quite unexpectedly at the law which 

 Listing had already expressed without giving any reason for it. 



If we call that position of the eye, from which all infinitely 

 small movements of the eye occur without rotation round the 

 visual line, the primary position, and all others secondary posi- 



