MECHANORECEPTORS AND BEHAVIOR 361 



The corpus is mostly concerned with spinal pathways, whereas the other two 

 centres are specifically related to the acoustico-lateralis system. 



Our knowledge of the projection sites is based on conventional neuro- 

 anatomy (Kappers, Huber, and Crosby, 1936), on some degeneration studies 

 (Campbell and Boord 1971), and on field and unit potential analyses (Ilinsky 

 et al. 1971, Enin and Ilinsky 1972, Enin et al. 1973, Paul and Roberts 

 1976a). 



The fibres of the anterior nerve enter the medulla and pass to a discrete 

 medial lobe which overhangs the fourth ventricle, being separated from the 

 medullary wall by a distinct cleft, and forms the dorsal nucleus or anterior 

 lateral-line lobe. The posterior nerve enters with the fibres of the vagus at the 

 rear of the medulla and ascends laterally almost as far as the auricle to give 

 fibres to the medial nucleus or the posterior lateral-line lobe (Figure 19). 

 Fibres from these lobes then pass to the auricles, while others, which do not 

 enter the lobes, ascend to the lateral cerebellar nucleus. 



The lateral-line projections in the dogfish have been recently described in 

 detail by Paul and Roberts (1977a), who found that the secondary neurons 

 of both lobes were large multipolar cells with spiny dorsal dendrites. These 

 dendrites resemble the dendrites of cerebellar Purkinje cells in extending into 

 the molecular layer, which consists of many unmyelinated parallel fibres and 

 stellate cells. The obvious resemblance of this arrangement to the well- 

 known organization of the cerebellum has lead to the widely accepted idea 

 (Johnston 1902) that the cerebellum developed phylogenetically from the 

 acoustico-lateralis centres. There are, however, important differences in the 

 organization of the two structures. Nevertheless, as is revealed in Figure 19, 

 the lateral-line lobes, the auricles, and the cerebellar corpus are superimposed 

 on, and have neural circuits that are in parallel with, the hind brain centres. 



The electrophysiological studies (Paul and Roberts 1977a) have revealed 

 certain features about the kinds of analysis carried out by the lateral-line 

 lobes. First, they show that the input is excitatory on the large secondary 

 neurons so that even a single input is converted into a multiple discharge. 

 The axons of these cells then project onto the extensive reticular system 

 (Restieaux and Satchell 1956). Second, the failure of the lateral-line cells to 

 follow faithfully any stimulus delivered more frequently than about 100 Hz, 

 because of powerful inhibitory processes, shows that the lobe functions as a 

 low-pass filter, even though the sensory fibres are capable of operating at a 

 higher range. The final feature revealed by the electrophysiology is that the 

 latency variation at the first synapse is large enough to prevent the lobe 

 resolving small time differences between incoming signals. 



The ventral dendrites extend widely and make synaptic connections with 

 afferent fibres in a complex neuropil. The impact of the lateral-line input is 

 therefore widely distributed throughout the lobe, and there is strong 

 smoothing of the input. For example, the pronounced rhythmic activity gen- 

 erated in parts of the lateral line during ventilatory and swimming movement 

 are out of phase and their impact on individual neurons will cancel out. 

 However, natural external disturbances away from the fish cause an almost 

 simultaneous signal to be set up along the transducer array, which will 



