in Seasonal Dimorphism. 399 
of tightly closing tins, which were half filled with moist 
earth or sand, this being kept sufficiently wet to ensure the 
proper saturation of the air. In all cases the pupz were not 
removed until the pigments could be clearly seen under the 
wing-covers. Although the tins were opened several times a 
day for the purpose of ventilation, it was impossible to 
entirely prevent the growth of mould, which did a considerable 
amount of damage, and the saturated atmosphere was evi- 
dently prejudicial to the health of the insects. It was notice- 
able that the mould did not affect healthy pups which were 
normally suspended ; but in many cases the fastenings of the 
larva were dissolved by the moisture, and it would fall on 
the wet earth just before or after pupation. The slight 
injuries thus sustained proved fatal in every instance. 
The full results obtained may best be shown in tabular 
form (pp. 400-401). 
From these tables it will be seen that out of thirty- 
three imagos no less than thirty remained unaltered by their 
abnormal treatment and developed the ordinary coloration 
characteristic of the dry season. With regard to the three 
exceptions, it cannot be supposed that their divergence was 
due to the direct action of the moisture, owing to the over- 
whelming counter evidence, and they must be regarded 
simply as instances of that overlapping tendency which is so 
abundantly evident under natural conditions and of which it 
is difficult at present to offer any adequate explanation. In 
the case of the two Tertas Desjardinsi in the second series, [ 
made a check observation in the field at the time of their 
emergence by examining fifty specimens of Terzas captured 
at random, only taking into consideration those which were 
in good condition and had probably lately emerged. Of these 
I found six showing marked variations in the direction of 
their respective summer forms, or a total of 12 per cent. In 
my experiments, out of twenty-four Terias only two showed 
such variation, giving a percentage of 8-3. We may thus 
broadly consider that the artificial conditions to which they 
were subjected had no effect whatever upon these insects and 
that their behaviour was practically identical with what it 
would have been in their natural mode of life. This result 
is certainly at variance with my own anticipations (at least so 
far as the genus Terzas is concerned) and is in direct opposition 
to Doherty’s conclusions. 
It cannot, however, be legitimately concluded from these 
experiments that humidity plays no part in the causation of 
seasonal dimorphism, for field observations have convinced me 
that the lateness of the summer rains will certainly retard the 
