J. W. Jbnkinson 
161 
Symmetry and the lines ruled across the slides, in groups of 15°. The ranges, for 
the reason already given, extend on each side from 0° to 180°, the angle being 
measured between the middle point of the grey crescent and the zero point. In 
Fig. 2 are given the corresponding frequency polygons, but the groups are here 
just twice as large, 30°. 
I. Let us consider those eggs in which the axis is horizontal, while the eggs 
are closely pressed and kept in darkness (I. B). 
There is a striking tendency for the Plane of Symmetry to lie in the gravita- 
tion symmetry plane; and for the grey crescent to appear on that side on which 
the white pole was turned up, namely 0°, hardly ever on the opposite. When, how- 
ever, eggs under the same conditions of pressure and gravity are exposed to 
day-light coming from + 90° we find (I. A), in addition to the high frequency at 0°, 
another about — 90^ and a minor rise in the polygon at + 90°. The rise at 0° is 
mainly accounted for by gravity, but the large rise at — 90°, and the lesser rise at 
+ 90° can only be explained, as far as I am able to see, by supposing that the Plane 
of Symmetry has a tendency to be developed in the direction of the incident light, 
the grey crescent appearing preferably, if I may so put it, on that side which is 
turned away from the light. 
II. This explanation is strengthened by the position of the grey crescent when 
the eggs are close, and the axes vertical (II. A). Here thei'e is the same rise at 
— 90°, the same drop at -f 90°, as in the case just considered, but when the light is 
made to fall from 0° (II. B) the drop at + 90° is replaced by a rise which is as 
great as that on the opposite side. This effect is not caused by the pressure, for it 
reappears when the eggs are spaced and their axes have been vertical from the 
beginning (IV.), in which the direction of light is again from -1- 90°. In all these 
cases (II. A and B, and IV.) there are also high frequencies at 0° and 180°. In 
II. B these are in the direction of the light, but those in II. A and IV. and the 
corresponding ones at —90° and +90° in II. B are at right angles to the light. 
An anomaly which I cannot, I fear, explain is the excessive frequency (in II. A 
and, though to a less extent, in IV.) about 180° as compared with 0°. One might, 
possibly, suppose that the light was not exactly from one end, but oblique, but I 
doubt if this would do*. 
In II. B where the light comes from 0° it is interesting to observe that this 
anomaly does not occur. The frequency about -j- 90° is practically as great as that 
about —90°; the frequency however at 180° — the side turned from the light — is 
decidedly greater than that at 0°, which is in consonance with the other results. 
III. The results recorded in column III. A and B, and in the corresponding 
polygons are not so unmistakable. There is, indeed, in III. B, the same evident 
drop at -1- 90°, the side from which tlie light comes, and there is as great a frequency 
at —90°, the side turned from the light, as at 0° and at 180°, but the rise at 0° which 
* For the probable explanation of this, see p. 168. 
Biometrika vii 
21 
