AMPULLARY ELECTRORECEPTORS 501 



dorsal canals running posteriorly are the longest, and if ampullary 

 sensitivities are constant, these receptors would be expected to be the most 

 sensitive. In mandibular receptors Murray (1967) observed a sensitivity of 1 

 to 2 juV for changes in impulse frequency detected by ear. In excised 

 receptors from the group with long posteriorly running canals, we have 

 observed a clear change of PSP amplitude produced by single stimuli of 

 about 5 juV (Figure 9). The receptors of different groups therefore may be 

 similar in voltage sensitivity. However, the posteriorly running canals are up 

 to 20 cm long and their receptors would thereby have a tenfold greater 

 sensitivity to voltage gradients than the mandibular receptors with canals 2 

 cm long. Taking the best values, we obtain a maximum sensitivity to voltage 

 gradients of 0.1 juV/cm, which is tenfold less sensitive than the behavioral 

 threshold of 0.01 juV/cm (Kalmijn 1974, 1978). The discrepancy could be in 

 the experimenters' failing to find the most sensitive receptors, as well as in 

 central processing of data from many receptors. The number of receptors 

 with maximum canal length along a given direction is about 10 to 20, and 

 each is innervated by about five nerve fibers. Thus, the total number of 

 afferent inputs is quite small compared to those in visual and auditory 

 systems. 



Although a dramatic difference in threshold of receptors with different 

 canal lengths has not been observed, the long time constants of the canal 

 walls should produce in longer canals greater attenuation of higher frequency 

 signals (Waltman 1966) that are in the range of responsiveness of the re- 

 ceptors (Murray 1967). Shorter canals could therefore not only provide 

 more local information but also give more information about higher fre- 

 quency signals. 



Evolutionary Considerations 



The ampullae of Lorenzini presumably function similarly throughout the 

 elasmobranchs. Their operation must be passive in animals without electric 

 organs. The weakly electric organs of the skates could conceivably be used in 

 active electrolocation because their pulses are of an appropriate amplitude 

 and duration, but because of the small size of the discharge very little is 

 known about normal occurrence (Bennett 1971a). A communication 

 function would also be possible and indeed seems likely, even if active 

 electrolocation is not employed. In the torpedinids the discharge is very 

 powerful, and the question is more one of how they prevent damage to their 

 ampullae. One species of torpedinid, Narcine brasiliensis, is known to have a 

 weakly electric organ that can generate only a low-voltage, low-frequency 

 signal, but its normal activity is unknown (Bennett and Grundfest .1961). 



The freshwater stingray Potamotrygon presents an interesting variant on 

 the general elasmobranch pattern. It has long dwelled in freshwater, and as 

 in the weakly electric teleosts, its skin is of high resistance and its ampullary 

 canals are very short (Szabo et al. 1972, Szamier and Bennett 1971). The 

 body interior is of low resistance compared to the surrounding medium, and 

 the interior is essentially isopotential (Szabo et al. 1972); voltage gradients 



