for many types of eyes in different conditions. The 

 reviews referred to should be consulted, supplemented 

 with recent work (i 1 1, 114). 



The problems of electroretinography have centered 

 around the following main issues; a) differentiation 

 of rod and cone ERG's, b') analysis of the transition 

 from rod to cone dominance in mixed retinae, c) at- 

 tempts to split the ERG into component responses, 

 d') comparisons between the ERG and the discharge 

 in the optic nerve and e") attempts to localize com- 

 ponents of the ERG to specific structures in the 

 retina. Under this heading also fall the recent experi- 

 ments with penetrating capillary microelectrodes of 

 the Gerard-Ling type. Finally as /) should be men- 

 tioned a steadily expanding literature on electroretin- 

 ography in man from descriptive, theoretical and 

 clinical points of view. All these aspects cannot be 

 discu.sscd with full attention to detail. Leading refer- 

 ences will, however, be given within all of them. 



The pure cone ERG's illustrated in figure 4 (.4 and 

 B) are from the horned toad (33) and a squirrel (8) 

 and are essentially alike. These types were called 

 Trctinograms (69), as long as there were no pure cone 

 mammalian ERG's available. The cone eyes of 

 squirrel species (8, g, 25) have since provided the 

 evidence necessary for identifying the I-type with the 

 cone ERG in mammals as well as other animals. 

 There is no secondary rise or c-wave. Cone ERG's are 

 often negative in between the h- and c/-waves (cf. 

 fig. 4.-I). They also tend to have large a-waves. The 

 rod ERG's (C and Z)) are from guinea pig and cat 

 eyes, the former having a practically pure rod retina, 

 the latter with cones corresponding roughly to the 

 number found in the human peripheral retina. The 

 ERG labeled E is from the pure rod retina of the 

 gecko (47). It diflfers from that of the guinea pig in 

 showing a (/-wave at 'off.' Now guinea pigs afso have 

 off-di.scharges in their optic nerves (a very striking 

 feature of cone eyes) and so the absence of the corre- 

 sponding (-/-wave in their ERG suggests that elements 

 responding at 'off' are fewer in number than in cone 

 eyes and many other types of rod eyes (cf. 69). In the 

 ERG of the cat (C), for instance, the d-wavc is re- 

 duced to a plateau, or a retardation at 'off' in the 

 drop of potential towards the base line. Walls (146) 

 holds the rods of the gecko to be transmuted cones. 

 The a-wave seems to occur in all retinae, provided 

 the light intensity is sufficiently high to elicit it. 



By changing state of adaptation from scotopic to 

 photopic it is also possible to demonstrate in mixed 

 eyes that the on-off differentials, the a- and rf-waves, 

 become faster, the rf-wave in addition Ijecoming 



NEUR.I1L .ACTIVITY IN THE RETINA 699 



Dog Rabbit 



Fio. 5. A comparison of the electroretinogram of the red 

 Irish setter and tlie rabbit. The upper tracing in each record 

 gives the electroretinogram and the lower, the time mark. 

 The Ught is 'on' when the time tracing is displaced upwards. 

 The record shows that the normal positive c-wave of rabbits 

 is replaced by a negative potential in dogs. Time mark: 0.5 sec. 

 Calibration: 50 ^v. [From Parry el al. (118).] 



larger than before (41, 46, 80, 113). This is conven- 

 iently done by using flickering light. Rod ERG's 

 tend to flicker with repeated positi\e i-waves, while 

 in cone ERG's a-waves and rZ-waves (see below) also 

 take part in the response to intermittent illumination. 



By varying stimulus intensity from the threshold 

 upwards it can easily be shown that both the b- and 

 the rf-waves consist of several components of different 

 rates of rise and of different size though complete sep- 

 aration may be diflicult to achieve. Systematic atten- 

 tion to this problem was first given in two papers on 

 the frog eye (78, 84) and, by varying state of adapta- 

 tion and wavelengths, the authors could show that 

 some components belonged to rods (slow ones) and 

 some to cones. Similar differentiations for the human 

 eye were followed by Motokawa & Mita (iio) and 

 Adrian (i, 2). Variations in duration (149) and in 

 rate of rise of the stimulating light (126) have recently 

 been found to be convenient methods of separating 

 scotopic and photopic components. (This matter is 

 further considered below.) 



Deteriorating ERG's tend to become cornea-nega- 

 tive, but in many animals the negative phase is also a 

 normal feature of the response and then is always 

 found to succeed the cornea-positive A-wave (see fig. 

 4.4). In fact, a large negative phase occurs fairly 

 generally in high intensity ERG's and, after light- 

 adaptation, also in rod eyes. Figure 5 illustrates the 

 ERG of a dog (118) which obviously contains a slow 

 negative component not visible in the ERG of the 

 rabbit inserted for comparison. Noell's (114) work is 

 of particular interest from this point of view, as has 

 been discussed by Granit (73). 



The common occurrence of slow or semistationary 

 negative phases during illumination has led most 



