150 EDWARD B. POULTON. 
made up of short sections corresponding in length to the thick- 
ness of the adjacent imbricated cells. From this structure I 
formed the conclusion that the filaments are probably built up 
of units contributed by the cells which lie outside them. Wilson 
and Martin describe them as made up of bead-like varicosities, 
each of which is placed between a pair of cells (simple or com- 
pound) in vertical succession (I. c., pl. xxvi, fig. 15, for a diagram- 
matic representation). At the meeting of the Physiological 
Society (1884) I suggested that the filaments might be nerve 
terminal organs, and that the imbricated arrangement of cells 
around them might result in pressure whenever the free sur- 
face of the rod was in contact with external objects, thus 
effecting the stimulation of the nerve. Professor Schafer pointed 
out, in the discussion, that the highly refringent character of 
the filaments is opposed to the view that they are terminal 
' nerve-fibrils. I have often tried to find a connection with the 
abundant nerves below the base of the push-rod, but such an 
investigation needs the fresh tissues. Wilson and Martin now 
describe the filaments as naked axis-cylinders prolonged from 
the nerve-fibres below the rod, which suddenly lose their me- 
dullary sheaths soon after entering the epithelium. Although 
their drawings from gold-stained preparations (pl. xxvi, fig. 21) 
appear to leave no doubt that the filaments are connected with 
nerves and form some kind of terminal organ, it is obviously 
erroneous to speak of them as “naked axis-cylinders” (p. 196) or 
“ nerve-fibrils” (p. 197). They are highly refringent and have 
none of the characters of these nerve structures. It is note- 
worthy, too, that although the filaments are represented in 
pl. xxvi, fig. 22, as black varicose threads, the authors accurately 
state in the description (p. 200) that the fibrils “ are uot black, 
but only highly refracting.” Furthermore, if the photomicro- 
graphs establish nothing else, they certainly prove that this 
conclusion as to the nature of the filaments is erroneous ; 
neither naked axis-cylinders nor nerve-fibrils could have caused 
the appearances seen in transverse section in pl. xxv, fig. 13, 
and in longitudinal section in fig. 8. 
These authors state (p. 196) that I failed to recognise the 
