No. 3.] VISUAL CEEES\ IN” VERTEBRATES. 609 
heat, etc., which showed elongation in darkness and contraction 
(as a rule) under the influence of light and heat (Hornbostel, 
78; Angelucci, °85; Engelmann, ’85; Gradenigo, 85; Krause, 
2.7 pe roo Llerzog; 205). 
In an attempt to explain the contraction of the visual cells 
we may employ a photo-contractile hypothesis (Cf. “Photo- 
musculaire theorie” of Dubois et Renaut, 89) and suppose 
that the final nervous impulse results from a mechanical stim- 
ulation (pressure or contact) produced by an intermediate 
process. The intermediate process would be the contraction 
Otaarstructtre (the tree ‘portion of a rod ‘cell) extremely 
sensitive mechanically to differences of light or temperature; 
the contraction in response to heat or light might be conceived 
as comparable to that of a muscle from the effect (heat ?) of 
a motor-nerve discharge. 
If to the photo-contractile hypothesis the objection were 
advanced that the introduction of mechanical stimulation 1s 
an unnecessary complication, it might be answered that it is 
consistent with observed parallel cases,—that the sensitiveness 
of a single cell to light might not be a type of stimulation 
transmissible over the long distances necessary to convey an 
impulse to the central nervous organs, in which case a trans- 
formation by secondary stimulation would be necessary. 
To correlate the above hypothesis with the theory that 
rods are functional only for light of low intensities (see Rivers 
:00), we would suppose that the contraction was not due 
directly to light or heat but indirectly through the action of 
the visual purple, the chemical character of which was changed 
by light action. 
That the lamine of the outer segment in their dimensions 
come with the limits of the wave length of light and show 
a constant thickness has been taken, by some (Schultze, Greef, 
Zenker) to be of considerable significance (see p. 563). 
Schultze states that nocturnal animals and others living in 
diminished light have very long rods, thus increasing the num- 
ber of plates. If stimulation depends on the relation of the 
